Neville F. Hacker
Epithelial ovarian cancer has the highest fatality-to-case ratio of all the gynecologic malignancies because more than two-thirds of patients have advanced disease at diagnosis (1). It presents a major surgical challenge, requires intensive and often complex therapies, and is extremely demanding of the patient's psychological and physical energy. Serous carcinomas, which are the most common, are now believed to be related etiologically to fallopian tube and peritoneal cancer (2,3,4,5,6,7).
There are more than 21,550 new cases of ovarian cancer annually in the United States, and more than 14,600 women can be expected to succumb to their illness (1). It is the fifth most common cancer in women in the United States after cancers of the lung, breast, colon, and uterus. It accounts for 4% of all female cancers and 31% of cancers of the female genital tract. Ovarian cancer is the fourth most common cause of death from malignancy in women.
A woman's risk at birth of having ovarian cancer some time in her lifetime is nearly 1.5%, and the risk of dying from ovarian cancer is almost 1% (2). Five-year survival from ovarian cancer of all stages and histologies has improved significantly in the United States from 37% in 1975-77 to 45% in 1996-2003 (p <.05) (1). The death rate per 100,000 has decreased by 8% from 9.51 in 1991 to 8.75 in 2004.
Approximately 90% of ovarian cancers are derived from cells of the coelomic epithelium or modified mesothelium(2). The cells are a product of the primitive mesoderm, which can undergo metaplasia. Neoplastic transformation can occur when the cells are genetically predisposed to oncogenesis or exposed to an oncogenic agent.
Epithelial ovarian cancers are thought to arise from a single layer of cells that covers the ovary or that lines cysts immediately beneath the ovarian surface. These cells are generally quiescent but proliferate following ovulation to repair the defect created by rupture of a follicle. Our understanding of the molecular pathogenesis of ovarian cancer is rapidly increasing. It is now recognized that there are at least two different molecular pathways that lead to the development of ovarian cancers and that these result in tumors that have quite distinct biological behaviors and probably different cells of origin (3). There are those tumors (so-called type I tumors) that arise from ovarian surface epithelium and müllerian inclusions, either from endosalpingiosis or invagination of the ovarian surface epithelium during repair of ovulation or implantation of cells from endometrium. This process typically involves a relatively slow and multistep pathway and accounts for many early stage cancers such as endometrioid, clear cell, mucinous, and low-grade serous cancers. In contrast, the more common high-grade serous cancers (type II) have a phenotype that resembles the fallopian tube mucosa, and they commonly have p53 mutations. These tumors appear to develop rapidly and are almost always at an advanced stage at presentation (3).
Approximately 75% to 80% of epithelial cancers are of the serous histologic type. Less common types are mucinous (10%), endometrioid (10%), clear cell, Brenner, and undifferentiated carcinomas, each of the latter three representing less than 1% of epithelial lesions (2). Each tumor type has a histologic pattern that reproduces the epithelial features of a section of the lower genital tract. For example, the serous or papillary pattern has an appearance similar to that of the glandular epithelium lining the fallopian tube, and it has been proposed that many serous epithelial cancers may originate in the distal fallopian tubal epithelium. Thus, these lesions involve both the ovaries and tubes, making it difficult to ascertain the site of origin (4,5,6,7) (see discussion below). Mucinous tumors contain cells that resemble the endocervical glands, and the endometrioid tumors resemble the endometrium. More specific details of the histology are discussed in Chapter 5.
An important group of tumors to distinguish is the tumor of low malignant potential, also called the borderline tumor (8,9,10,11). Borderline tumors tend to remain confined to the ovary for long periods of time, occur predominantly in premenopausal women, and are associated with a very good prognosis. They are encountered most frequently between the ages of 30 and 50 years, whereas invasive carcinomas are found more commonly between the ages of 50 and 70 years (2).
Although uncommon, metastatic implants may occur with borderline tumors. Such implants have been divided into noninvasive and invasive forms. The latter group has a higher likelihood of developing progressive, proliferative disease in the peritoneal cavity, which can lead to intestinal obstruction and death (9,10,11).
The primary malignant transformation of the peritoneum has been called peritoneal carcinoma or primary peritoneal papillary serous carcinoma. This has the appearance of a “müllerian” carcinoma and simulates ovarian cancer clinically. In such cases, there may be microscopic or small macroscopic cancer on the surface of the ovary and extensive disease in the upper abdomen, particularly in the omentum. Peritoneal carcinoma also explains how “ovarian cancer” can arise in a patient whose ovaries were surgically removed many years earlier (12,13,14).
The peak incidence of invasive epithelial ovarian cancer is 56 to 60 years (2,15). The agespecific incidence of this disease rises precipitously from 20 to 80 years of age and subsequently declines (16). The average patient age of those with borderline tumors is approximately 46 years (2,8). Eighty percent to 90% of ovarian cancers, including borderline forms, occur after the age of 40 years, whereas 30% to 40% of malignancies occur after the age of 65.
The chance that a primary epithelial tumor will be of borderline or invasive malignancy in a patient younger than age 40 years is approximately 1 in 10, but after that age it rises to 1 in 3 (2). Less than 1% of epithelial ovarian cancers occur before the age of 20, twothirds of ovarian malignancies in such patients being germ cell tumors (2,16). Approximately 30% of ovarian neoplasms in postmenopausal women are malignant, whereas only about 7% of ovarian epithelial tumors in premenopausal patients are frankly malignant (2).
Ovarian cancer has been associated with low parity and infertility (17). Although there has been a variety of epidemiologic variables correlated with ovarian cancer—such as increased risk with talc use, galactose consumption, and decreased risk with tubal ligation (see Chapter 7)—none has been so strongly correlated as prior reproductive history and duration of the reproductive career (17,18). Early menarche and late menopause increase the risk of ovarian cancer (18). These factors and the relationship of parity and infertility to the risk of ovarian cancer have led to the hypothesis that suppression of ovulation may be an important factor. Theoretically, the surface epithelium undergoes repetitive disruption and repair. It is thought that this process might lead to a higher probability of spontaneous mutations that can unmask germ-line mutations or otherwise lead to the oncogenic phenotype (see Chapter 1).
Some serous ovarian cancers may actually arise in the fallopian tube (4,5,6,7), a situation that may have been obscured by an overly rigid definition of fallopian tube cancer.Molecular and genetic evidence supports this etiology, as well as the finding of very early histologic alterations in the tubal epithelium that appear to be precursor lesions in serous carcinomas (3).
In a cohort study of more than 1.1 million Norwegian women, a positive association was found between body mass index (BMI), height, and risk of ovarian cancer, particularly of the endometrioid type in women younger than 60 years (19). Women who had a very high BMI and were clinically obese in adolescence and childhood had a relative risk of 1.56 of developing ovarian cancer compared with women with a medium BMI.
There has been considerable controversy as to whether fertility-enhancing drugs increase the risk of ovarian cancer. In a metaanalysis of eight case-control studies of fertility drugs and ovarian cancer (20), there were 5,207 women with cancer compared with 7,705 controls. The relative risk (RR) of fertility drug exposure for ovarian cancer was 0.97—that is, the use of the drugs was not associated with an increased risk. However, in the same cohort, nulliparity (compared with multiparity >4) carried a RR of 2.42, and infertility per se for 5years or longer (compared with <1 year) carried a RR of 2.7. These results support the hypothesis that the higher risk in these women is related to infertility, independent of fertility drug use.
Most case-control and cohort studies have failed to link hormone-replacement therapy to an increased risk of epithelial ovarian cancer (21). A large cohort study has reopened controversy regarding this issue (22). Among 44,241 postmenopausal women in the Breast Cancer Detection Demonstration Project, 329 developed ovarian cancer. Women who had received estrogen-replacement therapy only for more than 10 years without progestin were at increased risk of developing ovarian cancer. By 20 years, the relative risk was 3.2-fold.
The incidence of ovarian cancer varies in different geographic locations throughout the world. Western countries, including the United States and the United Kingdom, have an incidence of ovarian cancer that is three to seven times greater than in Japan, where epithelial ovarian tumors are considered rare (2). In Asia, the incidence of germ cell tumors of the ovary appears to be somewhat higher than in the West. Japanese immigrants to the United States exhibit a significant increase in the incidence of epithelial ovarian cancer, the rate eventually approaching that of white U.S. women. The incidence of epithelial tumors is about 1.5 times greater in whites than in blacks (1).
As parity is inversely related to the risk of ovarian cancer, having at least one child is protective of the disease, with a risk reduction of 0.3 to 0.4. The oral contraceptive reduces the risk of epithelial ovarian cancer (17). Women who use the oral contraceptive for 5 or more years reduce their relative risk to 0.5—that is, there is a 50% reduction in the likelihood of developing ovarian cancer. Women who have had two children and have used the oral contraceptive for 5 or more years have a relative risk of ovarian cancer as low as 0.3, or a 70% reduction (23). Therefore, the oral contraceptive pill is the only documented method of chemoprevention for ovarian cancer, and it should be recommended to women for this purpose. When counseling patients regarding birth control options, this important benefit of the oral contraceptive should be emphasized. This is also important for women with a strong family history of ovarian cancer.
Fenretinide, a retinoid, was thought to be a chemoprophylactic agent for ovarian cancer (24). However, in a prospective, randomized, placebo-controlled trial, there was no difference in the incidence or survival from ovarian cancer after 5 years (25). The Gynecology Oncology Group (GOG) initiated a confirmatory trial, but it was closed because of poor accrual.
Because some serous epithelial tumors might arise in the fallopian tube and because there is a higher rate of tubal carcinoma in women with BRCA1 and BRCA2 mutations, it is essential that risk-reducing prophylactic surgery include the removal of both ovaries and both fallopian tubes. The performance of a prophylactic salpingo-oophorectomy will signifi-cantly reduce, but not eliminate, the risk of ovarian and fallopian tube cancer (13,14) because the entire peritoneum is potentially at risk. Peritoneal carcinomas can occur even after prophylactic oophorectomy. Because the ovaries provide protection from cardiovascular and orthopedic diseases, prophylactic oophorectomy should not be routinely performed in premenopausal women at low risk for ovarian cancer.
There is no proven effective method of screening for ovarian cancer. Given the false positive results for both CA125 and transvaginal ultrasonography, particularly in premenopausal women, these tests are not cost-effective and should not be used routinely to screen for ovarian cancer. In the future, new markers or technologies may improve the specificity of screening, but proof of this will require large prospective studies (27,28,29,30,31,32,33,34,35,36,37,38,39,40,41). Screening in women who have a familial risk may have a better yield, but additional studies are necessary.
Routine annual pelvic examinations are disappointing for the early detection of ovarian cancer (26). Screening with transabdominal and transvaginal ultrasonography has been encouraging (27,28,29), but specificity has been limited. However, recent advances in transvaginal ultrasonography (30,31,32,33) have allowed a very high (>95%) sensitivity for the detection of early stage ovarian cancer, although this test alone might require as many as 10-15 laparotomies per ovarian cancer detected (34). Transvaginal color-flow Doppler to assess the vascularity of the ovarian vessels has been shown to be a useful adjunct to ultrasonography (31,32,33), but it has not been shown to be useful in screening.
CA125 has been shown to contribute to the early diagnosis of epithelial ovarian cancer (34,35,36,37,38,39,40,41). CA125 has been cloned, and although the function of the molecule is unclear, the elucidation of the MUC16 gene and its control may enhance the understanding of this important marker in ovarian cancer (42,43,44). Binding of MUC16 to mesothelin appears to mediate cell adhesion and facilitates peritoneal metastasis in animal models (45,46,47). Regarding the sensitivity of the test, CA125 can detect 50% of patients with stage I disease and 60% with stage II (28). Data suggest that the specificity of CA125 is improved when the test is combined with transvaginal ultrasonography (35) or when the CA125 levels are followed over time (41,42,43,44,45,46,47,48,49,50,51). In this manner, the risk of ovarian cancer (ROC) algorithm might help to improve the efficacy of screening (51).
Altered levels of more than 50 biomarkers have been reported in the serum or urine of ovarian cancer patients. A number of novel markers for ovarian cancer have been identified in recent years, including mesothelin, a 110-kd fragment of EGFR (sEGFR), lysophosphatidic acid, HE4, prostasin, osteopontin, and human kallikreins 6 and 10 (52,53,54,55,56,57,58).Multiplex assays can measure more than 50 biomarkers with a few hundred microliters of serum (59). Using the multiplex technology, a combination of CA125, HE4, sEGFR, and soluble vCAM-1 has distinguished 90% of stage I ovarian cancer patients from 98% of healthy controls.
Use of surface-enhanced laser desorption and ionization (SELDI) with subsequent resolution by mass spectroscopy has demonstrated a pattern of low molecular weight moieties that has been reported to distinguish sera from ovarian cancer patients from those of healthy individuals with 100% sensitivity and 95% specificity (60). Prospective replication of these results and determination of sensitivity for stage I disease should be available in the near future. Methodological issues have been raised regarding these data (61). SELDI may also identify a limited number of protein peaks that could be assayed by more conventional techniques (62).
Another approach is the measurement of plasma DNA levels and allelic imbalance by a technique known as digital single-nucleotide polymorphism analysis. In a study by Chang and colleagues (63), this analysis had an 87% positive correlation (13 of 15) in stages I and II and a 95% correlation (37 of 39) in patients with stages III and IV disease.
A complete discussion of screening is presented in Chapter 7.
Genetic Risk for Epithelial Ovarian Cancer
Ovarian cancers appear to arise from a single clone—that is, they are monoclonal—and thus are initiated from a single mutation (see Chapter 1) (64). Conversely, there is evidence that peritoneal carcinomas may have a multifocal origin (65).
Hereditary Ovarian Cancer
The risk of ovarian cancer is higher than that of the general population in women with certain family histories (66,67,68,69,70,71,72,73,74,75,76,77,78,79,80). Although most epithelial ovarian cancer is sporadic, as many as 10% to 14% of women with epithelial ovarian cancer have a germ-line mutation in BRCA1 or BRCA2 (67,79,80). Further discussion of germ-line mutations and their biology is presented in Chapter 1.
BRCA1 and BRCA2
Most hereditary ovarian cancer results from mutations in the BRCA1 gene, which is located on chromosome 17 (66), with a small proportion associated with mutations in BRCA2, which is located on chromosome 13 (41). Although these appear to be responsible for most hereditary ovarian cancers, it is likely that there are other, as yet undiscovered, genes that also predispose to ovarian or breast cancer or both (75).
In the past, it had been thought that there were two distinct syndromes associated with a genetic risk: site-specific hereditary ovarian cancer and hereditary breast or ovarian cancer syndrome. However, it is now accepted that these groups essentially represent a continuum of mutations of BRCA1 and BRCA2 with different degrees of penetrance within a given family (43,51). There are other less common genetic causes of ovarian cancer, and there is a higher risk of ovarian and endometrial cancer in women with the Lynch II syndrome, which is also known as the hereditary nonpolyposis colorectal cancer syndrome (HNPCC syndrome) (76).
The mutations are autosomal dominant, and thus a complete family history and pedigree analysis, including both maternal and paternal sides of the family, must be carefully evaluated (70). There are numerous distinct mutations that have been identified on each of these genes, and the mutations have different degrees of penetrance, which may account for the preponderance of either breast cancer, ovarian cancer, or both, in any given family. Based on analysis of women who have a mutation in the BRCA1 gene and are from high-risk families, the lifetime risk of ovarian cancer may be as high as 28% to 44%, and the risk has been calculated to be as high as 27% for those women with a BRCA2mutation (67,68,74). In women with a BRCA1 or BRCA2 mutation, the risk of ovarian and breast cancer may be as high as 54% and 82%, respectively (78).
Hereditary ovarian cancers generally occur in women approximately 10 years younger than those with nonhereditary tumors (67,78). As the median age of epithelial ovarian cancer is in the mid- to late 50s, a woman with a first- or second-degree relative who had premenopausal ovarian cancer may have a higher probability of carrying an affected gene.
Breast and ovarian cancer may exist in a family in which there is a combination of epithelial ovarian and breast cancers, affecting a mixture of first- and second-degree relatives.Women with this syndrome tend to have their breast cancers at a very young age, and the breast cancers may be bilateral. If two first-degree relatives are affected, this pedigree is consistent with an autosomal dominant mode of inheritance (66,71).
There is a higher carrier rate of BRCA1 and BRCA2 mutations in women of Ashkenazi Jewish descent, Icelandic women, and in other ethnic groups (72,73,75). There have been three specific mutations that are carried by the Ashkenazi population: 185delAG and 5382insC on BRCA1, and 6174delT on BRCA2. The total carrier rate of at least one of these mutations for a patient of Ashkenazi Jewish descent is 1 in 40 or 2.5%, which is considerably higher than the general white population. The increased risk is a result of the founder effect—that is, a higher rate of mutations that have occurred within a defined geographic area.
The risk of ovarian cancer depends on the number of first- or second-degree relatives with a history of epithelial ovarian carcinoma or breast cancer, and on the age of onset. The degree of risk is difficult to determine precisely unless a full pedigree analysis is performed, and all patients should be referred to a familial cancer service for genetic counseling.
Risch et al. (79) found that the hereditary proportion of invasive ovarian tumors was approximately 13% and was as high as 18% in the large subgroup of serous ovarian cancers. This was independent of family history. Similar findings have been reported in a smaller study from Poland (80). If confirmed in other population-based studies, this will have major implications on genetic testing and argues for consideration of genetic testing in most patients with high-grade serous cancers irrespective of the pedigree and family history.
Lynch II Syndrome or HNPCC Syndrome
HNPCC syndrome, which includes multiple adenocarcinomas, involves a combination of familial colon cancer (known as the Lynch I syndrome); a high rate of ovarian, endometrial, and breast cancers; and other malignancies of the gastrointestinal and genitourinary systems (76). The mutations that have been associated with this syndrome are MSH2, MLH1, PMS1, and PMS2. The risk that a woman who is a member of one of these families will develop epithelial ovarian cancer depends on the frequency of this disease in first- and second-degree relatives, although these women appear to have at least three times the relative risk of the general population. A full pedigree analysis of such families should be performed by a geneticist to more accurately determine the risk.
Management of Women at High Risk for Ovarian Cancer
The management of a woman with a strong family history of epithelial ovarian cancer must be individualized and depends on her age, her reproductive plans, and the level of risk. A thorough pedigree analysis is important. A geneticist should evaluate the family pedigree for at least three generations. Decisions about management are best made after careful study of the pedigree and, whenever possible, verification of the histologic diagnosis of the family members' ovarian cancer as well as the age of onset and other tumors in the family.
The value of testing for BRCA1 and BRCA2 has been clearly established, and guidelines for testing now exist (70,77,78). The importance of genetic counseling cannot be overemphasized because the decision is complex. The American Society of Clinical Oncologists has offered guidelines that emphasize careful evaluation by geneticists, careful maintenance of medical records, and a clear understanding in a genetic screening clinic of how to counsel and manage these patients. There remain concerns of how the information will be used, the impact on insurability, how the results will be interpreted, and how the information will be used within a specific family—for example, to counsel children.
Although there are some conflicting data, the behavior of breast cancers arising in women with germ-line mutations in BRCA1 or BRCA2 appears to be comparable to that of sporadic tumors (69). Women with breast cancer who carry these mutations, however, are at a greatly increased risk of ovarian cancer as well as of a second breast cancer. The lifetime risk of ovarian cancer is 54% for women who have a BRCA1 mutation and 23% for those with a BRCA2 mutation; for the two groups together, there is an 82% lifetime risk of breast cancer(78).
Although recommended by the National Institutes of Health Consensus Conference on Ovarian Cancer (81), the value of screening with transvaginal ultrasonography, CA125 levels, or other procedures has not been clearly established in women at high risk. Bourne et al. (49) have shown that this approach can detect tumors about 10 times more often than in the general population, and thus they recommend screening in high-risk women. However, the findings of two prospective studies of annual transvaginal ultrasonography and CA125 screening in 888 BRCA1 and BRCA2 mutation carriers in the Netherlands and 279 mutation carriers in the United Kingdom are not encouraging and suggest a very limited benefit, if any, of screening even in high-risk women (82,83).
Despite annual gynecological screening, Hermsen et al. (82) reported that a high proportion of ovarian cancers in BRCA1 and BRCA2 carriers were interval cancers, and the majority of all cancers diagnosed were at an advanced stage. Similar findings were reported by Woodward (83). Therefore, it is unlikely that annual screening will reduce mortality from ovarian cancer in BRCA1 and BRCA2 mutation carriers (82,84).
This important question has also been addressed in GOG 199, a study of screening with annual transvaginal ultrasonography and CA125 ROCA compared to prophylactic bilateral salpingooophorectomy. Study accrual is complete as of November 2006 with 2,605 participants enrolled: 1,030 (40%) in the surgical cohort and 1,575 (60%) in the screening cohort. Five years of prospective follow-up ends in November 2011 (85).
Data derived from a multiinstitutional consortium of genetic screening centers indicate that the use of the oral contraceptive pill is associated with a lower risk of development of ovarian cancer in women who have a mutation in either BRCA1 or BRCA2 (86). In women who had taken the oral contraceptive pill for 5 or more years, the relative risk of ovarian cancer was 0.4, or a 60% reduction in the incidence of the disease. Another study, however, failed to confirm this finding (87). Tubal ligation may also decrease the risk of ovarian cancer in patients with BRCA1 (but not BRCA2) mutations, but the protective effect is not nearly as strong as oophorectomy (88).
The value of prophylactic oophorectomy in these patients has been documented (89,90,91,92,93,94). Women at high risk for ovarian cancer who undergo prophylactic oophorectomy have a risk of harboring occult neoplasia: In one series of 98 such operations, three patients (3.1%) had a lowstage ovarian malignancy (90). The protection against ovarian cancer is very high: the performance of a prophylactic salpingo-oophorectomy reduces the risk of BRCA-related gynecologic cancer by 96% (92). Although the risk of ovarian cancer is diminished, there remains a small risk of subsequently developing a peritoneal carcinoma, a tumor that may also have a higher predisposition in women who have mutations in theBRCA1 and BRCA2 genes. In these series, the risk of development of peritoneal carcinoma was 0.8% and 1%, respectively (90,91). Prophylactic salpingo-oophorectomy in premenopausal women reduced the risk of developing subsequent breast cancer by 50% to 80% (90,91).
The role of hysterectomy is more controversial. Although most studies show no increase in the rate of uterine and cervical tumors, there are some reports of an increased risk of papillary serous tumors of the endometrium (95). Women on tamoxifen are at higher risk for benign endometrial lesions (e.g., polyps) and endometrial cancer. Therefore, it is reasonable to consider the performance of a prophylactic hysterectomy in conjunction with salpingooophorectomy, and this decision should be individualized.
Grann and associates reported the application of Markov modeling—that is, quality-adjusted survival estimate analysis—in a simulated cohort of 30-year-old women who tested positive for BRCA1 or BR CA2 mutations (96). The analysis predicted that a 30 year-old woman could prolong her survival beyond that associated with surveillance alone by 1.8 years with tamoxifen, 2.6 years with prophylactic salpingo-oophorectomy, 4.6 years with both tamoxifen and prophylactic salpingo-oophorectomy, 3.5 years with prophylactic mastectomy, and 4.9 years with both prophylactic surgeries. Quality-adjusted survival was estimated to be prolonged by 2.8 years for tamoxifen, 4.4 years with prophylactic salpingo-oophorectomy, 6.3 years for tamoxifen and prophylactic salpingo-oophorectomy, 2.6 years with mastectomy, and 2.6 years with both operations.
The survival of women who have a BRCA1 or BRCA2 mutation and develop ovarian cancer is longer than that for those who do not have a mutation. In one study, the median survival for mutation carriers was 53.4 months compared with 37.8 months for those with sporadic ovarian cancer from the same institution (97). These findings have recently been confirmed in a population-based-study from Israel in which Chetrit et al. reported that, among Ashkenazi women with ovarian cancer, those with BRCA1 and BRCA2 mutations had an improved long-term survival (38% vs. 24% at 5 years). This may result from distinct clinical behavior or from a better response to chemotherapy (98).
Current recommendations for management of women with high risk for ovarian cancers are summarized below (77,78,81,86,87,88,89,90,91,92,93,94,95,96):
The majority of women with epithelial ovarian cancer have vague and nonspecific pelvic, abdominal, and menstrual symptoms (99,100,101,102,103). Goff et al. (104) recently developed an ovarian cancer symptom index and reported that symptoms associated with ovarian cancer were pelvic or abdominal pain, urinary frequency or urgency, increased abdominal size or bloating, and difficulty eating or feeling full. These symptoms were particularly suspicious when they were present for less than 1 year and lasted longer than 12 days a month. The index had a sensitivity of 56.7% for early ovarian cancer and 79.5% for advanced stage disease.
A study from the Royal Hospital for Women in Sydney compared 100 patients with early stage epithelial ovarian cancer with 100 patients with advanced stage disease. Ninety percent of women with early disease and 100% with advanced disease reported at least one symptom. With early disease, abdominal pain was reported by 51% and abdominal swelling by 32%; with advanced disease, abdominal swelling was reported by 62% and abdominal pain by 40%. Seventy percent of patients with early disease and 69% of those with advanced disease reported symptoms of less than 3 months duration. Patients with tumors less than 5 cm in diameter were three times more likely to have advanced disease. Patients with grade I tumors were 40 times more likely to have early stage disease when compared to patients with grade 3 tumors (105). These findings were confirmed in a population-based study from Australia in which there did not appear to be a significant difference in the duration of symptoms or the nature of symptoms in patients with early as opposed to advanced stage disease (106).
These two studies reinforce the concept that early and late-stage ovarian cancer are biologically different entities and argue against the widely held misconception that ovarian cancer is diagnosed at an early stage because the symptoms are recognized earlier than in patients with more advanced disease (84,105,106,107).
The most important sign is the presence of a pelvic mass on physical examination. A solid, irregular, fixed pelvic mass is highly suggestive of an ovarian malignancy. If, in addition, an upper abdominal mass or ascites is present, then the diagnosis of ovarian cancer is almost certain. Because the patient usually reports abdominal symptoms, she may not be subjected to a pelvic examination, and the presence of a tumor may be missed. Pleural effusions commonly occur in association with ascites and very occasionally in the absence of ascites in patients with advanced disease.
The diagnosis of an ovarian cancer requires an exploratory laparotomy. The preoperative evaluation of the patient with an adnexal mass is outlined in Fig. 11.1.
Ultrasonographic signs of malignancy include an adnexal pelvic mass with areas of complexity such as irregular borders; multiple echogenic patterns within the mass; and dense, multiple, irregular septae. Bilateral tumors are more likely to be malignant, although the individual characteristics of the lesions are of greater significance. Transvaginal ultrasonography may have a somewhat better resolution than transabdominal ultrasonography for adnexal neoplasms (27,28,29,30). Doppler color-flow imaging may enhance the specificity of ultrasonography for demonstrating findings consistent with malignancy (31,32,33).
Figure 11.1 Preoperative evaluation of the patient with an adnexal mass.
The size of the lesion is of importance. If a complex cystic mass is more than 8 to 10 cm in diameter, then the probability is high that the lesion is neoplastic, unless the patient has been taking clomiphene citrate or other agents to induce ovulation (100). In the premenopausal patient, a period of observation is reasonable, provided the adnexal mass is not clinically suspicious (i.e., it is mobile, mostly cystic, unilateral, and of regular contour). Generally, an interval of no more than 2 months is allowed for observation. If the lesion is not neoplastic, it should remain stable or regress, as measured by pelvic examination and pelvic ultrasonography. If a mass increases in size or complexity, then it must be presumed to be neoplastic and removed surgically.
In postmenopausal women with unilocular cysts measuring 8 to 10 cm or less and normal serial CA125 levels, expectant management is acceptable, and this approach may decrease the number of surgical interventions (108,109,110).
Premenopausal patients whose lesions are clinically suspicious (i.e., large, predominantly solid, relatively fixed, or irregularly shaped) should undergo laparotomy, as should postmenopausal patients with complex adnexal masses of any size.
Before the planned exploration, the patient should undergo routine hematologic and biochemical assessments. A preoperative evaluation in a patient older than 40 years undergoing laparotomy should include a radiograph of the chest. An abdominal and pelvic computed tomographic (CT) or magnetic resonance imaging (MRI) scan is of no value in patients with a definite pelvic mass (111,112,113,114,115). Patients with ascites and no pelvic mass should have a CT or MRI scan to look particularly for liver or pancreatic tumors (112). The value of positron-emission tomography (PET) scans is being evaluated but may contribute to the specificity of the CT scan findings (114).
The preoperative evaluation should exclude other primary cancers metastatic to the ovary. A colonoscopy is indicated in selected patients with symptoms and signs suspicious for colon cancer. This would include any patient who has evidence of frank or occult blood in the stool or gives a recent history of diarrhea or constipation. A gastroscopy is indicated if there are upper gastrointestinal symptoms such as nausea, vomiting, or hematemesis (116). Bilateral mammography is indicated if there is any breast mass, because occasionally breast cancer metastatic to the ovaries can simulate primary ovarian cancer. Patients who have irregular menses or postmenopausal bleeding should have an endometrial biopsy and an endocervical curettage to exclude the presence of uterine or endocervical cancer metastatic to the ovary.
Differential Diagnosis Ovarian epithelial cancers must be differentiated from benign neoplasms and functional cysts of the ovaries (117,118,119). A variety of benign conditions of the reproductive tract—such as pelvic inflammatory disease, endometriosis, and pedunculated uterine leiomyomata—can simulate ovarian cancer. Nongynecologic causes of a pelvic tumor, such as an inflammatory or neoplastic colonic mass, must be excluded (100). A pelvic kidney can simulate ovarian cancer.
The risk of malignancy index (RMI), first described by Jacobs in 1990, is one method of differentiating between benign and malignant masses (117). Utilization of this index may facilitate better triage of suspicious pelvic masses to gynecologic oncologists. The RMI incorporates the menopausal status, an ultrasonic score, and the serum CA125 level.
In an analysis of 204 consecutive patients with an ovarian mass at the Royal Hospital for Women in Sydney, an RMI of <200 correctly identified 83 of 108 (77%) benign ovarian masses (118). An RMI of >200 correctly identified 11 of 19 (58%) borderline ovarian tumors and 70 of 77 (91%) invasive ovarian cancers. An RMI of >200 had a sensitivity of 84%, specificity of 77%, positive predictive value of 76%, and a negative predictive value of 85% for the detection of both borderline and invasive ovarian tumors (118).
Patterns of Spread
Ovarian epithelial cancers spread primarily by exfoliation of cells into the peritoneal cavity, by lymphatic dissemination, and by hematogenous spread.
Transcoelomic The most common and earliest mode of dissemination is by exfoliation of cells that implant along the surfaces of the peritoneal cavity. The cells tend to follow the circulatory path of the peritoneal fluid. The fluid tends to move with the forces of respiration from the pelvis, up the paracolic gutters, especially on the right, along the intestinal mesenteries, to the right hemidiaphragm. Therefore, metastases are typically seen on the posterior cul-de-sac, paracolic gutters, right hemidiaphragm, liver capsule, the peritoneal surfaces of the intestines and their mesenteries, and the omentum. The disease seldom invades the intestinal lumen but progressively agglutinates loops of bowel, leading to a functional intestinal obstruction. This condition is known as carcinomatous ileus.
Lymphatic Lymphatic dissemination to the pelvic and paraaortic lymph nodes is common, particularly in advanced-stage disease (120,121,122,123). Spreading through the lymphatic channels of the diaphragm and through the retroperitoneal lymph nodes can lead to dissemination above the diaphragm, especially to the supraclavicular lymph nodes (120).
Burghardt et al. performed systematic pelvic and paraaortic lymphadenectomy on 123 patients (122) and reported that 78% of patients with stage III disease have metastases to the pelvic lymph nodes. In another series (123), the rate of positive paraaortic lymph nodes was 18% in stage I, 20% in stage II, 42% in stage III, and 67% in stage IV.
Hematogenous Hematogenous dissemination at the time of diagnosis is uncommon, with spread to vital organ parenchyma, such as the lungs and liver, in only some 2% to 3% of patients. Most patients with disease above the diaphragm at the time of presentation have a right pleural effusion. Systemic metastases are seen more frequently in patients who have survived for some years. Dauplat et al. from UCLA (124) reported that distant metastasis consistent with stage IV disease ultimately occurred in 38% of the patients whose disease was originally intraperitoneal. Sites of hematogenous spread and their median survivals were as follows:
Significant risk factors for distant metastases were malignant ascites, peritoneal carcinomatosis, large metastatic disease within the abdomen, and retroperitoneal lymph node involvement at the time of initial surgery.
The outcome of patients after treatment can be evaluated in the context of prognostic factors, which can be grouped into pathologic, biologic, and clinical factors.
The morphology and histologic pattern, including the architecture and grade of the lesion, are important prognostic variables (125,126,127,128,129,130). In general, stage for stage, histologic type is not of prognostic significance, with the exception of clear cell carcinomas, which are associated with a worse prognosis than the other histologic types (128,129).
Histologic grade, as determined either by the pattern of differentiation or by the extent of cellular anaplasia and the proportion of undifferentiated cells, seems to be of prognostic significance (129,131). However, studies of the reproducibility of grading ovarian cancers have shown a high degree of intraobserver and interobserver variation (130). Because there is significant heterogeneity of tumors and observational bias, the value of histologic grade as an independent prognostic factor has not been clearly established. Baak et al. (131) have presented a standard grading system based on morphometric analysis, and the system appears to correlate with prognosis, especially in its ability to distinguish low-grade or borderline patterns from other tumors.
Several biologic factors have been correlated with prognosis in epithelial ovarian cancer (132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161,162,163,164). Using flow cytometry, Friedlander et al. (133) showed that ovarian cancers were commonly aneuploid. Furthermore, they and others showed that there was a high correlation between FIGO* stage and ploidy; that is, low-stage cancers tend to be diploid and high-stage tumors tend to be aneuploid (132,133,134,135,136,137,138,139) (Fig. 11.2). Patients with diploid tumors have a significantly longer median survival than those with aneuploid tumors: 5 years versus 1 year, respectively (134). Multivariate analyses have demonstrated that ploidy is an independent prognostic variable and one of the most significant predictors of survival (135).
Figure 11.2 Survival of patients with stage I epithelial ovarian cancer based on ploidy evaluation. (From Tropé C, Kaern J, Vergote I. Adjuvant therapy for early-stage epithelial ovarian cancer. In: Gershenson DM, McGuire WP, eds. Ovarian cancer: controversies in management. New York: Churchill Livingstone, 1998:41-63, with permission.)
More than 100 protooncogenes have been identified, and studies have focused on the amplification or expression of these genetic loci and their relationship to the development and progression of ovarian cancer (140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155). For example, Slamon et al. (142) reported that 30% of epithelial ovarian tumors expressed HER2/neu oncogene and that this group had a poorer prognosis, especially those patients with more than five copies of the gene. Berchuck et al. (143) reported a similar incidence (32%) of HER2/neu expression. In their series, patients whose tumors expressed the gene had a poorer median survival (15.7 months versus 32.8 months). Others have not substantiated this finding (144), and a review of the literature by Leary et al. (145) revealed an overall incidence of HER2/neu expression of only 11%.
Additional prognostic variables include p53, bcl-2, K-ras, Ki67, interleukin 6, PTEN, lysophospholipids, and platelet-derived growth factor (150,151,152,153,154,155,156,157,158,159,160). Further discussion of these molecular variables is presented in Chapter 1.
The in vitro clonogenic assay has been studied in ovarian cancer. A significant inverse correlation has been reported between clonogenic growth in vitro and survival (161,162,163,164). Multivariate analysis has found that clonogenic growth in a semisolid culture medium is a significant independent variable (163). The use of an “extreme drug resistance assay” has been suggested as a possible means of directing therapy by defining platinum-sensitive and resistant tumors in vitro but, the assay does not independently predict or alter the outcome in either a primary or recurrent disease setting (164).
In addition to stage, the extent of residual disease after primary surgery, the volume of ascites, patient age, and performance status are all independent prognostic variables(165,166,167,168,169,170,171,172,173,174). Among patients with stage I disease, Dembo et al. (165) showed, in a multivariate analysis, that tumor grade and “dense adherence” to the pelvic peritoneum had a significant adverse impact on prognosis, whereas intraoperative tumor spillage or rupture did not. A subsequent study by Sjövall et al. confirmed these findings (166). A multivariate analysis of these and several other studies was performed by Vergote et al. (168), who reported that poor prognostic variables for early stage disease were the tumor grade, capsular penetrance, surfaces excrescences, and malignant ascites, but not iatrogenic rupture.
Initial Surgery for Ovarian Cancer
Ovarian epithelial malignancies are staged according to the International Federation of Gynecology and Obstetrics (FIGO) system, and the staging system of 1988 is presented inTable 11.1 and in Fig. 11.3. The TNM staging is correlated with the FIGO stage in Fig 11.3. The FIGO staging is based on findings at surgical exploration. A preoperative evaluation should exclude the presence of extraperitoneal metastases. The 2008 FIGO staging for Ovarian epithelial malignancies remains unchanged.
A thorough surgical staging should be performed because subsequent treatment will be determined by the stage of disease. In patients in whom exploratory laparotomy does not reveal any macroscopic evidence of disease on inspection and palpation of the entire intraabdominal space, a careful search for microscopic spread must be undertaken.
In earlier series in which patients did not undergo careful surgical staging, the overall 5-year survival for patients with apparent stage I epithelial ovarian cancer was only approximately 60% (175). Since then, survival rates of 90% to 100% have been reported for patients who have been properly staged and found to have stage IA or IB disease (175,176,177,178,179).
Table 11.1 FIGO Staging for Primary Carcinoma of the Ovary
Figure 11.3 Staging ovarian cancer (FIGO and TNM). (From Heintz APM, Odicino F, Maisonneuve P, Quinn MA, Benedet JL, Creasman WT, et al. Carcinoma of the ovary. In Pecorelli S, ed. Twenty-Sixth Annual Report on the Results of Treatment in Gynaecological Cancer. Int J Gynecol Oncol 2006;95(suppl 1):S161-S192, with permission.)
Technique for Surgical Staging
In patients whose preoperative evaluation suggests a probable ovarian malignancy, a midline or paramedian abdominal incision is recommended to allow adequate access to the upper abdomen. When a malignancy is unexpectedly discovered in a patient who has a lower transverse incision, the rectus muscles can be either divided or detached from the symphysis pubis to allow better access to the upper abdomen (see Chapter 20). If this is not sufficient, the incision can be extended on one side to create a “J” incision.
The ovarian tumor should be removed intact, if possible, and a frozen histologic section obtained. If ovarian malignancy is present and the tumor is apparently confined to the ovaries or the pelvis, then thorough surgical staging should be carried out. This involves the following steps:
As many as three in ten patients whose tumor appears confined to the pelvis have occult metastatic disease in the upper abdomen or the retroperitoneal lymph nodes. At surgical staging, metastases in apparent low-stage epithelial ovarian cancer are found in the diaphragm in approximately 7% of patients, in paraaortic lymph nodes in 15%, pelvic nodes in 6%, omentum in 9%, and peritoneal cytolology in 26% (121,175,176,177,178).
The importance of careful initial surgical staging is emphasized by the findings of a cooperative national study (175) in which 100 patients with apparent stage I and II disease who were referred for subsequent therapy underwent additional surgical staging. In this series, 28% of the patients initially thought to have stage I disease were “upstaged,” as were 43% of those thought to have stage II disease. A total of 31% of the patients were upstaged as a result of additional surgery, and 77% were reclassified as having stage III disease.Histologic grade was a significant predictor of occult metastasis; that is, 16% of the patients with grade 1 lesions were upstaged, compared with 34% with grade 2 and 46% with grade 3 disease.
After a comprehensive staging laparotomy, only a minority of women will have local disease (FIGO stage I). Of the 21,650 women diagnosed yearly with epithelial ovarian cancer in the United States, nearly 4,000 have disease confined to the ovaries (1,182). The prognosis for these patients depends on the clinical-pathologic features, as outlined below. Because of this emphasis on the importance of surgical staging, the rate of lymph node sampling has increased in the United States, with a study showing that for women with stage I and II disease, the percentage having lymph nodes sampled increased from 38% to 59% from 1991 to 1996 (183).
Early Stage Ovarian Cancer
The primary treatment for stage I epithelial ovarian cancer is surgical—that is, a total abdominal hysterectomy, bilateral salpingo-oophorectomy, and surgical staging (175,184). In certain circumstances, a unilateral oophorectomy may be permitted, as discussed below. Based on the prognostic variables outlined above (126,135,165,166,167,168,169,170,172,173,184), early stage epithelial ovarian cancer can be subdivided into low-risk and high-risk disease (Table 11.2).
The principal treatment of borderline ovarian tumors is surgical resection of the primary tumor (3,185,186,187,188,189,190,191,192). There are no data to suggest that either adjuvant chemotherapy or radiation therapy improves survival (193,194,195). After a frozen section has determined that the histology is borderline, premenopausal patients who desire preservation of ovarian function may be managed with a “conservative” operation—that is, a unilateral oophorectomy (3,186,188). In a study of patients who underwent unilateral ovarian cystectomy only for apparent stage I borderline serous tumors, Lim-Tan et al. (187) found that this conservative operation was also safe; only 8% of the patients having recurrences 2 to 18 years later, all with curable disease confined to the ovaries. Recurrence was associated with “positive margins” of the removed ovarian cyst.
In a retrospective series of 339 patients by Zaetta et al. (191), seven (2%) progressed to invasive carcinoma, five serous and two mucinous. Although the recurrence rate after fertilitysparing surgery was 18.5% versus 4.6% after nonfertility-sparing surgery, all but one woman with recurrence of borderline tumor or progression to carcinoma was cured. The disease-free survival was 99.6% for stage I, 95.8% for stage II, and 89% for stage III. Thus, hormonal function and fertility can be maintained in the majority of patients with borderline tumors. In patients in whom an oophorectomy or cystectomy has been performed and a borderline tumor is later documented in the permanent pathology, no additional staging surgery is necessary, but the patient should be monitored with transvaginal ultrasonography.
Table 11.2 Prognostic Variables in Early Stage Epithelial Ovarian Cancer
Fertility Preservation in Early Stage Ovarian Cancer
In patients who have undergone a thorough staging laparotomy and in whom there is no evidence of spread beyond the ovary, the uterus and contralateral ovary can be retained in women who wish to preserve fertility (196,197). In a study by Park et al. (197), 59 women with stages IA-IC underwent fertility-sparing surgery, and there were no recurrences in women whose disease was grade 1 or 2. Women with grade 3 or higher stage disease had a significantly higher recurrence rate and lower survival. Women who undergo fertility-sparing surgery for low-stage, low-grade epithelial ovarian cancer should be followed carefully with routine transvaginal ultrasonography and determination of serum CA125 levels. Generally, the other ovary and the uterus should be removed at the completion of childbearing (see treatment section below).
Advanced-Stage Ovarian Cancer
The surgical management of all patients with advanced-stage disease is approached in a similar manner, with modifications made for the overall status and general health of the patient, as well as the extent of residual disease present at the time treatment is initiated. A treatment scheme is outlined in Fig. 11.4.
If the patient is medically stable, she should undergo an initial exploratory procedure with removal of as much disease as possible(199,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216). The operation to remove the primary tumor as well as the associated metastatic disease is referred to asdebulking or cytoreductive surgery. Most patients subsequently receive combination intravenous chemotherapy with an empiric number of cycles, six to eight. In some patients with completely resected disease, intraperitoneal chemotherapy may be considered. In selected patients who are not candidates for initial cytoreductive surgery, neoadjuvant chemotherapy may be given for a few cycles before surgery, as discussed below. Second-look laparotomy has not been shown to improve outcomes. It is best limited to investigational protocols (198).
The preoperative assessment of resectability is limited. Using a cutoff of 500 IU, CA125 levels have been suggested as a means of predicting the probability of an optimal resection (211,214), but others have shown that these determinations have low predictive value (215). CT , MR, and CT-PET scans have been used to try to predict suboptimal resection(216,217,218,219). In a series by Dowdy et al. (216), the presence of diffuse peritoneal thickening and ascites on CT scan was associated with a 32% optimal debulking, as opposed 71% in the group that did not have these findings, with a positive predictive value of 68%. However, in a larger multiinstititonal validation study, the accuracy of CT (218) in predicting suboptimal cytoreduction dropped to as low as 34% in some cohorts. CT-PET was also found to have limited positive predictive value (219).
Vergote and colleagues from Belgium reported the use of open laparoscopy in 173 patients with a pelvic mass, an omental “cake,” or large volume ascites to exclude other primary tumors and to determine resectability. Seventy-one of the patients (41%) developed a port site metastasis (220).
In our experience, it is virtually always possible to remove the primary tumor (if necessary, with an en bloc resection of the rectosigmoid colon) and the omental cake, so the major reason for recommending neoadjuvant chemotherapy is to improve the medical fitness of the patient.
Patients with advanced-stage epithelial ovarian cancer documented at initial exploratory laparotomy should undergo cytoreductive surgery(199,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216). The operation typically includes the performance of a total abdominal hysterectomy and bilateral salpingo-oophorectomy, along with a complete omentectomy and resection of any metastatic lesions from the peritoneal surfaces or from the intestines. The pelvic tumor often directly involves the rectosigmoid colon, the terminal ileum, and the cecum (Fig. 11.5). In a minority of patients, most or all of the disease is confined to the pelvic viscera and the omentum so that removal of these organs will result in extirpation of all gross tumor, a situation that is associated with a reasonable chance of prolonged progression-free survival.
The rationale for cytoreductive surgery relates to general theoretic considerations (207,221,222): (i) the physiologic benefits of tumor excision and (ii) the improved tumor perfusion and increased growth fraction, both of which increase the likelihood of response to chemotherapy or radiation therapy.
Figure 11.4 Treatment scheme for patients with advanced-stage epithelial ovarian cancer. PFI, progression-free interval
Figure 11.5 Extensive ovarian carcinoma involving the bladder, rectosigmoid, and ileocecal area. (From Heintz APM, Berek JS. Cytoreductive surgery for ovarian carcinoma. In: Piver MS, ed. Ovarian malignancies. Edinburgh: Churchill Livingstone, 1987:134, with permission.)
Physiologic Benefits Ascites may be sometimes reasonably well controlled after removal of the primary tumor and a large omental cake. Also, removal of the omental cake often alleviates the nausea and early satiety that many patients experience. Removal of intestinal metastases may restore adequate intestinal function and lead to an improvement in the overall nutritional status of the patient, thereby facilitating the patient's ability to tolerate subsequent chemotherapy.
Tumor Perfusion and Cellular Kinetics A large, bulky tumor may contain areas that are poorly vascularized, and such areas will be exposed to suboptimal concentrations of chemotherapeutic agents. Similarly, these areas are poorly oxygenated, so radiation therapy, which requires adequate oxygenation to achieve maximal cell kill, will be less effective. Thus, surgical removal of these bulky tumors may eliminate areas that are most likely to be relatively resistant to treatment.
In addition, larger tumor masses tend to be composed of a higher proportion of cells that are either nondividing or in the “resting” phase (i.e., Go cells, which are essentially resistant to the therapy). A low-growth fraction is characteristic of bulky tumor masses, and cytoreductive surgery can result in smaller residual masses with a relatively higher growth fraction.
The fractional cell kill hypothesis of Skipper (221) postulates that a constant proportion of the tumor cells are destroyed with each treatment. This theory suggests that a given dose of a drug will kill a constant fraction of cells as long as the growth fraction and phenotype are the same. Therefore, a treatment that reduces a population of tumor cells from 109 to 104 cells also would reduce a population of 105 cells to a single cell. If the absolute number of tumor cells is lower at the initiation of treatment, then fewer cycles of therapy should be necessary to eradicate the cancer, provided that the cells are not inherently resistant to the therapy.
The larger the initial tumor burden, the longer the necessary exposure to the drug and, therefore, the greater the chance of developing acquired drug resistance. However, because the spontaneous mutation rate of tumors is an inherent property of the malignancy, the likelihood of developing phenotypic drug resistance also increases as the size of the tumor increases. The chance of developing a clone of cells resistant to a specific agent is related to both the tumor size and its mutation frequency (221,222). This is one of the inherent problems with cytoreductive surgery for large tumor masses: Phenotypic drug resistance may have already developed before any surgical intervention.
Goals of Cytoreductive Surgery
The principal goal of cytoreductive surgery is the removal of all of the primary cancer and, if possible, all metastatic disease. If resection of all metastases is not feasible, then the goal is to reduce the tumor burden by resection of all individual tumors to an “optimal” status. Griffiths (199) initially proposed that all metastatic nodules should be reduced to≤1.5 cm in maximum diameter and showed that survival was significantly longer in such patients.
Subsequently, Hacker and Berek (201,204,205,206,207) showed that patients whose largest residual lesions were ≤5 mm had a superior survival, and this was substantiated by Hoskins et al. presenting the data of the Gynecologic Oncology Group (203). The median survival of patients in this category was 40 months, compared with 18 months for patients whose lesions were ≤1.5 cm and 6 months for patients with nodules >1.5 cm (Fig. 11.6). Clearly, those patients whose disease has been completely resected have the best prognosis, and approximately 60% of patients in this category will be free of disease at 5 years (Fig. 11.7).
The resectability of the metastatic tumor is usually determined by the location of the disease. Optimal cytoreduction is difficult to achieve in the presence of extensive disease on the diaphragm, in the parenchyma of the liver, along the base of the small-bowel mesentery, in the lesser omentum, or in the porta hepatis (212).
The ability of cytoreductive surgery to influence survival is limited by the extent of metastases before cytoreduction, possibly because of the higher likelihood of phenotypically resistant clones of cells in large metastatic masses (201,204). Patients whose metastatic tumor is very large (i.e., >10 cm before cytoreductive surgery) have a shorter survival than those with smaller areas of disease (134) (Fig. 11.8). Extensive carcinomatosis, the presence of ascites, and poor tumor grade, even with lesions that measure <5 mm, are also poor prognostic factors (204,205,206,207,212).
Figure 11.6 Survival versus diameter of largest residual disease. (From Hacker NF, Berek JS, Lagasse LD, Nieberg RK, Elashoff RM. Primary cytoreductive surgery for epithelial ovarian cancer. Obstet Gynecol 1983;61:413-420, with permission from the American College of Obstetricians and Gynecologists.)
Figure 11.7 Survival of patients with stage IIIC epithelial ovarian cancer based on the maximum size of the residual tumor after exploratory laparotomy and tumor resection. (From Heintz APM, Odicino F, Maisonneuve P, Quinn MA, Benedet JL, Creasman WT, et al. Carcinoma of the ovary. In: Pecorelli S, ed. Twenty-Sixth Annual Report on the Results of Treatment in Gynaecological Cancer. Int J Gynecol Oncol 2006;95(suppl 1):S161-S192, with permission.)
Figure 11.8 Survival versus diameter of largest metastatic disease before cytoreduction. (From Hacker NF, Berek JS, Lagasse LD, Nieberg RK, Elashoff RM. Primary cytoreductive surgery for epithelial ovarian cancer. Obstet Gynecol 1983;61:413-420, with permission from the American College of Obstetricians and Gynecologists.)
The supine position on the operating table may be sufficient for most patients. However, for those with extensive pelvic disease for whom a low resection of the colon may be necessary, the low lithotomy position should be used. Debulking operations should be performed through a vertical incision in order to gain adequate access to the upper abdomen as well as to the pelvis.
After the peritoneal cavity is opened, ascitic fluid, if present, should be evacuated. In some centers, fluid is submitted routinely for appropriate in vitro studies, particularly the clonogenic assay. In cases of massive ascites, careful attention must be given to hemodynamic monitoring, especially in patients with borderline cardiovascular function.
A thorough inspection and palpation of the peritoneal cavity and retroperitoneum are carried out to assess the extent of the primary tumor and the metastatic disease. All abdominal viscera must be palpated to exclude the possibility that the ovarian disease is metastatic, particularly from the stomach, colon, or pancreas. If optimal status is not considered achievable, extensive bowel and urologic resections are not indicated except to overcome a bowel obstruction. However, removal of the primary tumor and omental cake is usually both feasible and desirable.
Pelvic Tumor Resection
The essential principle of removal of the pelvic tumor is to use the retroperitoneal approach (206,207). To accomplish this, the retroperitoneum is entered laterally, along the surface of the psoas muscles, which avoids the iliac vessels and the ureters. The procedure is initiated by division of the round ligaments bilaterally if the uterus is present. The peritoneal incision is extended cephalad, lateral to the ovarian vessels within the infundibulopelvic ligament and caudally toward the bladder. With careful dissection, the retroperitoneal space is explored, and the ureter and pelvic vessels are identified. The pararectal and paravesicle spaces are identified and developed as described in Chapter 9.
The peritoneum overlying the bladder is dissected to connect the peritoneal incisions anteriorly. The vesicouterine plane is identified, and, with careful sharp dissection, the bladder is mobilized from the anterior surface of the cervix. The ovarian vessels are isolated, doubly ligated, and divided.
The hysterectomy, which is often not a “simple” operation, is performed. The ureters need to be carefully displayed in order to avoid injury. During this procedure, the uterine vessels can be identified. The hysterectomy and resection of the contiguous tumor are completed by ligation of the uterine vessels and the remainder of the tissues within the cardinal ligaments.
Because epithelial ovarian cancers tend not to invade the lumina of the colon or bladder, it is usually feasible to resect pelvic tumors without having to resect portions of the lower colon or the urinary tract (223,224,225,226). However, if the disease surrounds the rectosigmoid colon and its mesentery, it may be necessary to remove that portion of the colon to clear the pelvic disease (Fig. 11.9) (223,224). This is justified if the patient will be left with “optimal” disease at the end of the cytoreduction. After the pararectal space is identified in such patients, the proximal site of colonic involvement is identified, the colon and its mesentery are divided, and the rectosigmoid is removed along with the uterusen bloc. A reanastomosis of the colon is performed, as described in Chapter 20.
It is rarely necessary to resect portions of the lower urinary tract (225). Occasionally, resection of a small portion of the bladder may be required. If so, a cystotomy should be performed to assist in resection of the disease. Rarely, partial ureteric resection may be necessary, followed by primary reanastomosis (ureteroureterostomy), ureteroneocystostomy, or transureteroureterostomy, as described in Chapter 20.
Advanced epithelial ovarian cancer often completely replaces the omentum, forming an omental cake. This disease may be adherent to the parietal peritoneum of the anterior abdominal wall, making entry into the abdominal cavity difficult. After freeing the omentum from any adhesions to parietal peritoneum, adherent loops of small intestine are freed by sharp dissection. The omentum is then lifted and pulled gently in the cranial direction, exposing the attachment of the infracolic omentum to the transverse colon. The peritoneum is incised to open the appropriate plane, which is developed by sharp dissection along the serosa of the transverse colon. Small vessels are ligated with hemoclips. The omentum is then separated from the greater curvature of the stomach by ligation of the right and left gastroepiploic arteries and ligation of the short gastric arteries (Fig. 11.10).
Figure 11.9 Resection of the pelvic tumor may include removal of the uterus, tubes, and ovaries, as well as portions of the lower intestinal tract. The arrows represent the plane of resection.
The disease in the gastrocolic ligament can extend to the hilus of the spleen and splenic flexure of the colon on the left and to the capsule of the liver and the hepatic flexure of the colon on the right. Usually, the disease does not invade the parenchyma of the liver or spleen, and a plane can be found between the tumor and these organs. However, it will occasionally be necessary to perform splenectomy to remove all the omental disease (213,226,227) (Fig. 11.11). Diaphragm stripping and diaphragm resection have been used to optimally resect upper abdominal disease in selected cases (213).
The disease may involve focal areas of the small or large intestine, and resection should be performed if it would permit the removal of all or most of the abdominal metastases.Apart from the rectosigmoid colon, the most frequent sites of intestinal metastasis are the terminal ileum, the cecum, and the transverse colon. Resection of one or more of these segments of bowel may be necessary (223,224,225,226).
Resection of Other Metastases
Other large masses of tumor that are located on the parietal peritoneum should be removed, particularly if they are isolated masses, and their removal will permit optimal cytoreduction. Resection of extensive disease from the surfaces of the diaphragm is generally neither practical nor feasible, although solitary metastases may be resected, the diaphragm sutured, and a chest tube placed if necessary for a few days (228). The use of the Cavitron Ultrasonic Surgical Aspirator (CUSA), the argon beam laser, and the loop electrosurgical device may help facilitate resection of small tumor nodules, especially those on flat surfaces (229,230,231).
Figure 11.10 Separation of the omentum from stomach and transverse colon. (From Heintz APM, Berek JS. Cytoreductive surgery for ovarian carcinoma. In: Piver MS, ed. Ovarian malignancies. Edinburgh: Churchill Livingstone, 1987:134, with permission.)
Figure 11.11 Omental “cake” densely adherent to the spleen.
Resection of Pelvic and Paraaortic Lymph Nodes
The performance of a pelvic and paraaortic lymphadenectomy in patients with stage IIIC-IV disease has been reported to prolong survival (123), but an international randomized study failed to confirm this (232). Patients who were optimally cytoreduced in the peritoneal cavity were randomized between systematic pelvic and paraaortic lymphadenectomy versus resection of bulky nodes only. There were 216 evaluable patients in the systematic lymphadenectomy arm and 211 in the nodal debulking arm of the study. Patients in each group were well matched for clinical characteristics such as stage of disease, grade of tumor, and residual disease status. Although there was a 6-month improvement in progression-free survival, there was no difference in 5-year overall survival [48.5% versus 47% respectively; 95% confidence interval (CI)—8.4% to 10.6%].
Du Bois and colleagues analyzed three phase III German trials to try to retrospectively determine the role of lymphadenectomy (233). There were a total of 3,336 patients, of whom 1,059 (32%) had no macroscopic residual intraperitoneal disease. Retroperitoneal lymphadenectomy was performed in 757 of the 1,059 patients (72%), and there was a significant survival advantage for the group having the lymphadenectomy (66% vs. 55%; p = 0.003). No advantage could be demonstrated if there was any macroscopic residual disease. The authors concluded that a randomized controlled trial was warranted to address the question of systematic pelvic and paraaortic lymph node dissection versus nodal debulking in patients with no intraperitoneal residual disease.
Feasibility and Outcome
An analysis of the retrospective data available suggests that these operations are feasible in 70% to 90% of patients when performed by gynecologic oncologists (205,207,210,213). Major morbidity is in the range of 5% and operative mortality in the range of 1% (182,187). Intestinal resection in these patients does not appear to increase the overall morbidity of the operation (205,207).
Some have questioned the ability of cytoreductive surgery to improve the overall outcome of patients with ovarian cancer (234). Concern has been expressed that these operations are excessively morbid and that modern chemotherapies are sufficient. No randomized prospective study has ever been performed to define the value of primary cytoreductive surgery. However, all retrospective studies indicate that the diameter of the largest residual tumor nodule before the initiation of chemotherapy is significantly related to progression-free survival in patients with advanced ovarian cancer. In addition, quality of life is likely to be significantly enhanced by removal of bulky tumor masses from the pelvis and upper abdomen.
In a metaanalysis of 81 studies of women who underwent cytoreductive surgery for advanced ovarian cancer, Bristow and colleagues (235) documented that the extent of debulking correlated with incremental benefits in survival-that is, the greater the percentage of tumor reduction, the longer the survival. Each 10% increase in cytoreduction equated to a 5.5% increase in median survival. Women whose cytoreduction was greater than 75% of their tumor burden had a median survival of 33.9 months compared with 22.7 months for women whose tumors were cytoreduced to less than 75%. (p <0.001).
Patients with stage IV disease can benefit from optimal cytoreduction (208,236). In selected cases, resection of isolated hepatic metastasis to <1 cm results in an optimal resection of disease (236).
A prospective randomized study of “interval” cytoreductive surgery was carried out by the European Organization for the Research and Treatment of Cancer (EORTC). Interval surgery was performed after three cycles of platinum-combination chemotherapy in patients whose primary attempt at cytoreduction was suboptimal. Patients in the surgical arm of the study demonstrated a survival benefit when compared with those who did not undergo interval debulking (237). Most of these patients had not had an aggressive attempt to debulk their tumor at their initial surgery. In a 10-year follow-up analysis, the risk of mortality was reduced by more than 40% in the group that was randomized to the debulking arm of the study (238). Therefore, the performance of a debulking operation as early as possible in the course of the patient's treatment should be considered the standard of care(239).
A prospective phase III study of interval cytoreductive surgery was conducted by the GOG (240), but the study design was different because the patients entered on the trial had already undergone a maximal attempt at tumor resection at their initial surgery. The randomized findings showed no difference between the group of patients who had an additional attempt at debulking after three cycles of chemotherapy compared with those who did not. The median survival of the 216 women who underwent interval cytoreduction was 32 months compared with 33 months for the 209 women who did not undergo surgical cytoreduction.
There is evidence that the survival of women with advanced ovarian cancer is improved when the surgeon is specifically trained to perform cytoreductive surgery (241) and when there is centralization of care (242). Therefore, whenever feasible, patients with advanced ovarian malignancy should be referred to a subspecialty unit for primary surgery, and every effort should be made to attain as complete a cytoreduction as possible.
Treatment with Chemotherapy and Radiation
Early Stage Low-Risk Ovarian Cancer
Guthrie et al. (184) studied the outcome of 656 patients with early stage epithelial ovarian cancer. No untreated patients who had stage IA, grade 1 cancer died of their disease; thus, adjuvant radiation and chemotherapy were unnecessary. Furthermore, the Gynecologic Oncology Group carried out a prospective, randomized trial of observation versusmelphalan for patients with stages IA and IB, grades 1 and 2 disease. Five-year survival for each group was 94% and 96%, respectively, confirming that no further adjuvant treatment is needed for such patients.
Early Stage High-Risk Ovarian Cancer
In patients whose disease is high risk—for example, more poorly differentiated or in whom there are malignant cells either in ascitic fluid or in peritoneal washings—additional therapy is indicated. Treatment options include chemotherapy or whole-abdominal radiation (186,187,188,189,190,191,192,193,194,195,196,197,198,199,200,201,202,203,204,205,206). Some comparisons of these modalities have been made and are summarized below.
Chemotherapy for patients with early stage high-risk epithelial ovarian cancer can be either single agent or multiagent (243,244,245,246,247,248,249). Some researchers have questioned the wisdom of overly aggressive chemotherapy in women with early stage disease, suggesting that the evidence for a durable impact on survival is marginal (250,251). Furthermore, the possibility of leukemia with alkylating agents and platinum make the administration of adjuvant therapy risky unless there is a significant benefit (244,245).
Cisplatin, carboplatin, cyclophosphamide, and paclitaxel are all active as single agents against epithelial ovarian cancer and have been administered either alone or in various combinations in the adjuvant setting. There were some early series in which cisplatin or cyclophosphamide or both (PC) were used to treat patients with stage I disease (252,253,254,255,256,257,258,259,260). In a GOG trial of three cycles of cisplatin and cyclophosphamide versus intraperitoneal 32P in patients with stages IB and IC disease, the progression-free survival of women receiving the platinum-based chemotherapy was 31% higher than that of those receiving the radiocolloid (254). Similar results were reported from a multicenter trial performed in Italy by the Gruppo Italiano Collaborativo Oncologica Ginecologica for progression-free survival, although there was no overall survival advantage (259). Carboplatin can be substituted for cisplatin in the therapy of these patients (260) because it is much better tolerated, has fewer side effects, and appears to have similar efficacy.
Two large, parallel, randomized phase III clinical trials have been recently reported on women with early stage disease: the International Collaborative Ovarian Neoplasm Trial 1 (ICON1) and the Adjuvant Chemotherapy Trial in Ovarian Neoplasia (ACTION) (260,261).
In the ICON1 trial, 477 patients from 84 centers in Europe were entered. Patients of all stages were eligible for the trial if, in the opinion of the investigator, it was unclear whether adjuvant therapy would be of benefit. Most patients were said to have stages I and IIA disease, but optimal surgical staging was not required, so it is likely an unquantified number of these women had stage III disease. Adjuvant platinum-based chemotherapy was given to 241 patients, and no adjuvant chemotherapy was given to 236 patients. The 5-year survival was 73% in the group that received adjuvant chemotherapy compared with 62% in the control group [hazard ratio (HR) = 0.65, p = 0.01] (261).
In the ACTION trial, 440 patients from 40 European centers were randomized; 224 patients received adjuvant platinum-based chemotherapy, and 224 patients did not (260). Patients with stages I and IIA, grades 2 and 3 were eligible. Only one-third of the total group was optimally staged (151 patients). In the observation arm, optimal staging was associated with a better survival (HR = 2.31, p = 0.03); in the nonoptimally staged patients, adjuvant chemotherapy was associated with an improvement in survival (HR = 1.78, p = 0.009). In optimally staged patients, no benefit of adjuvant chemotherapy was seen. Therefore, in the ACTION trial, the benefit from adjuvant chemotherapy was limited to the patients with nonoptimal staging, suggesting that patients might only benefit if they had a higher likelihood of occult microscopic dissemination.
Figure 11.12 Overall survival in patients with early-stage ovarian carcinoma (The ICON1/ACTION Trials). Adjuvant cisplatin- or carboplatin-based single agent or combination chemotherapy (n = 465 patients) (blue line) versus no adjuvant chemotherapy (n = 460 patients) (purple line) until clinical progression. The hazard ratio is 0.67 (95% CI = 0.50 to 0.90, p = 0.008 using log-rank test) in favor of chemotherapy. Five-year survivals were 82% for the adjuvant chemotherapy group versus 74% for those who do not receive adjuvant chemotherapy (221). These data should be interpreted with caution because most of the patients were not completely staged, and the benefit of treatment appears to be only in patients who did not have a complete staging laparotomy. (From Trimbos JB, Parmar M, Guthrie D, Swart AM, Vergote I, Bolls G, et al. International Collaborative Ovarian Neoplasm Trial 1 and Adjuvant Chemotherapy in Ovarian Neoplasm Trial: two parallel randomized phase III trials of adjuvant chemotherapy in patients with early-stage ovarian carcinoma. J Natl Cancer Inst 2003;95:105-112, with permission.)
When the data from the two trials were combined and analyzed (262), a total of 465 patients were randomized to receive platinum-based adjuvant chemotherapy and 460 to observation until disease progression (Fig. 11.12). After a median follow-up of more than 4 years, the overall survival was 82% in the chemotherapy arm and 74% in the observation arm (HR = 0.67, p = 0.001). Recurrence-free survival was also better in the chemotherapy arm: 76% versus 65% (HR = 0.64, p = 0.001). The results of this analysis must be interpreted with caution because most of the patients did not undergo thorough surgical staging, but the findings suggest that platinum-based chemotherapy should be given to patients who have not been optimally staged.
The GOG reported the results of a randomized study of three cycles versus six cycles of carboplatin and paclitaxel in 457 patients with early stage ovarian cancer (231). An unexpectedly large number of patients (126, or approximately 28%) had incomplete or inadequately documented surgical staging. The recurrence rate was 24% lower (HR = 0.76, CI = 0.5-1.13, p = 0.18) for six versus three cycles, but this was not statistically significant. The estimated probability of recurrence at 5 years was 20.1% for six cycles and 25.4% for three cycles. They concluded that three cycles of adjuvant carboplatin and paclitaxel was a reasonable option for women with high-risk early stage ovarian cancer.
Table 11.3 Randomized Trials in Stage I Epithelial Ovarian Cancer (Since 1995)
The current GOG trial includes patients with high-risk stage I and II disease and offers three cycles of carboplatin and paclitaxel followed by a randomization to either observation versus 26 weeks of weekly low-dose (40 mg/m2) paclitaxel. High-risk stage I is defined as stages IA or IB, grade 3, stage IC, or clear cell carcinomas.
A summary of randomized phase III trials reported since 1995 for the treatment of patients with low-stage disease is presented in Table 11.3 (247,248,252,260,262).
There are two general approaches to the treatment of low-stage epithelial ovarian cancer with radiation: intraperitoneal radiocolloids and whole-abdominal radiation therapy. In one retrospective study of 32P, the 5-year survival was 85% (215). In a series of patients with stage I disease treated with whole-abdominal radiation (246), the 5-year relapse-free survival rate was 78%, but many of these patients had high-risk variables (e.g., poor histologic grade).
A prospective trial was conducted by the GOG of patients with stage IB, grade 3, stage IC, or stage II with no residual disease. Twelve cycles of melphalan were compared with intraperitoneal 32P, and there was no difference in survival (248). However, in a multicenter Italian trial (182), a randomized comparison of six cycles of cisplatin as a single agent versus 32P showed an 84% disease-free survival with cisplatin and 61% with 32P(p <0.01). Furthermore, the GOG protocol that randomized cisplatin and cyclophosphamide versus 32P showed that the platinumbased chemotherapy was superior (254). Therefore, although 32P produces results similar to single-agent melphalan, platinum-based chemotherapy is preferable (Table 11.3). Pelvic radiation alone is not as effective as melphalan in these patients and should not be used in ovarian cancer (243).
Recommendation for Adjuvant Treatment of Early Stage Ovarian Cancer
Low-Risk Early Stage Disease
No adjuvant chemotherapy is recommended for these patients.
High-Risk Early Stage Disease
Advanced-Stage Ovarian Cancer
Systemic chemotherapy is the standard treatment for metastatic epithelial ovarian cancer (264,265,266,267,268,269,270,271,272,273,274,275,276,277,278,279,280,281,282,283,284,285,286,287,288). After the introduction of cisplatin in the latter half of the 1970s, platinum-based combination chemotherapy has become the most frequently used treatment regimen in the United States (240). Paclitaxel became available in the 1980s, and this drug was incorporated into ovarian combination chemotherapy in the 1990s (264,265,266,267,268,269,270,271). Comparative trials of paclitaxel, cisplatin, and carboplatin are summarized below (Table 11.4).
Figure 11.13 Survival of patients with advanced-stage ovarian cancer: a metaanalysis of multiple trials comparing cisplatin-containing combination chemotherapy with regimens without cisplatin. (From Advanced Ovarian CancerTrialists Group. Chemotherapy in advanced ovarian cancer: an overview of randomized clinical trials. BMJ1991;303:884, with permission.)
Table 11.4 Randomized Trials Involving Platinum and Taxanes in Patients with Advanced-Stage Epithelial Ovarian Cancer
Cisplatin Combination Chemotherapy
Combination chemotherapy has been shown to be superior to single-agent therapy in most studies of initial chemotherapy in patients with advanced epithelial ovarian cancer(272,273,274,275,276,277,278,279,280,281,282,283,284,285). After cisplatin became available for the treatment of ovarian cancer, a prospective study conducted in England showed that cisplatin was better than an alkylating agent, cyclophosphamide, as a single agent (277). Concurrently, cisplatin was tested in a variety of different combinations, and the platinum-containing regimens were superior (278). A metaanalysis compared outcomes for patients given cisplatin-containing combination chemotherapy, with those for patients not receiving cisplatin (275). The cisplatin group had a slight survival advantage from 2 to 5 years, but this difference disappeared by 8 years (Fig 11.13).
Most studies using the PC (cisplatin and cyclophosphamide) or PAC (cisplatin, doxorubicin, and cyclophosphamide) regimen have reported similar survival rates(279,280,281,282,283,284). The GOG's randomized prospective comparison of equitoxic doses of PAC versus PC showed no benefit to the inclusion of doxorubicin in the combination (280). Although a metaanalysis of the combined data from four trials showed a 7% survival advantage at 6 years for those patients treated with the doxorubicin-containing regimen (284), the survival curves converged at 8 years.
The next major advance in the treatment of advanced-stage disease was the incorporation of paclitaxel into the chemotherapeutic regimens. A series of randomized, prospective clinical trials with paclitaxel-containing arms have defined the current recommended treatment protocol in advanced epithelial ovarian cancer (272,273,274,288). These studies are listed in Table 11.4.
Paclitaxel was shown to be a very active agent in ovarian cancer (264,265,266,267,268,269,270). The overall response rate of paclitaxel in phase II trials was 36% in previously treated patients (268), which is a higher rate than was seen for cisplatin when it was first tested.
Reporting the Gynecologic Oncology Group data (Protocol 111), McGuire et al. showed that the combination of cisplatin (75 mg/m2) and paclitaxel (135 mg/m2) was superior tocisplatin (75 mg/m2) and cyclophosphamide (600 mg/m2), each given for six cycles (272). In suboptimally resected patients, the paclitaxel-containing arm produced a 36% reduction in mortality (Fig 11.14). These data were verified in a trial conducted jointly by the European Organization for the Research and Treatment of Cancer, the Nordic Ovarian Cancer Study Group (NOCOVA), and the National Cancer Institute of Canada (NCIC), in which patients with both optimal and suboptimal disease were treated (273). In this study, thepaclitaxel-containing arm produced a significant improvement in both progression-free interval and overall survival in both optimal and suboptimal groups (Fig 11.15). Based on these two studies, paclitaxel should be included in the primary treatment of all women with advanced-stage epithelial ovarian cancer unless precluded by toxicity.
A three-arm comparison of paclitaxel (T) versus cisplatin (P) versus PT in suboptimal stage III and IV patients (protocol 132) showed equivalency in the three groups, but crossover from one drug to the other was permitted (274). The study essentially showed that the combination regimen was better tolerated than the sequential administration of the agents in suboptimally resected patients.
The second-generation platinum analogue, carboplatin, was introduced and developed to have less toxicity than its parent compound, cisplatin. Fewer gastrointestinal side effects, especially nausea and vomiting, are observed with carboplatin, and there is less nephrotoxicity, neurotoxicity, and ototoxicity (285,286). Carboplatin is associated with a higher degree of myelosuppression than cisplatin.
The initial studies showed that carboplatin and cisplatin had approximately a 4:1 equivalency ratio. Thus, a standard single-agent dose of approximately 400 mg/m2 carboplatin was used in most phase II trials. The optimal way to dose carboplatin is by using the area under the curve (AUC) and the glomerular filtration rate according to the Calvert formula (287), as discussed in Chapter 4. The target AUC is 5 to 7 for previously untreated patients with ovarian cancer. A platelet nadir of approximately 50,000/dL is a suitable target (287).
Carboplatin and Paclitaxel
Two randomized, prospective clinical studies have compared the combination of paclitaxel and carboplatin with paclitaxel and cisplatin (288,289) (Table 11.4). In both studies, the efficacy and survivals were similar, but the toxicity was more acceptable for the carboplatin-containing regimen.
In the first trial, GOG Protocol 158, the randomization was carboplatin AUC 7.5 and paclitaxel 175 mg/m2 over 3 hours versus cisplatin 75 mg/m2 and paclitaxel 135 mg/m2 over 24 hours (288). The progression-free survival of the carboplatin-containing arm was 20.7 months versus 19.4 months for the control arm (Fig. 11.16). The overall survival was 57.4 months for the carboplatin arm versus 48.7 months for control arm. The relative risk of progression for the carboplatin plus paclitaxel group was 0.88, and the RR of death was 0.84.The gastrointestinal and neurotoxicity of the carboplatin arm were appreciably lower than in the cisplatin arm.
A similar result was obtained in a large randomized trial in Germany (289) in which the dose of carboplatin was AUC = 6, and paclitaxel was 185 mg/m2 over 3 hours compared with the same dose of paclitaxel and cisplatin 75 mg/m2. The overall survival was 44.1 months for the carboplatincontaining arm versus 43.3 months for the control arm. Thus, the preferred regimen in patients with advanced stage disease is the paclitaxel plus carboplatin combination.
Figure 11.14 Survival of patients with suboptimal stage III and IV epithelial ovarian cancer treated with paclitaxel and cisplatin versus cyclophosphamide and cisplatin: a Gynecologic Oncology Group study (Protocol 111). (From McGuire WP, Hoskins W, Brady MF, Kucera PR, Partridge EE, Look KY, et al. Cyclophosphamide and cisplatincompared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med 1996;334:1-6, with permission.)
Figure 11.15 Survival of patients with stage III and IV epithelial ovarian cancer treated with paclitaxel and cisplatin or cyclophosphamide and cisplatin: results of a European cooperative group trials study. Survival by treatment, and survival by treatment group (optimal vs. suboptimal). [From Stuart G, Bertelsen K, Mangioni C, Tropé C, James K, Kaye S, et al. Updated analysis shows a highly significant improved survival for cisplatin-paclitaxel as first line treatment of advanced stage epithelial ovarian cancer: mature results of the EORTC-GCCG, NOCOVA, NCIC CTG and Scottish Intergroup Trial. Proc Am Soc Clin Oncol 1998;34: 1394(abst), with permission.]
Figure 11.16 Survival of patients with stage III epithelial ovarian cancer treated with carboplatin and paclitaxel versus cisplatin and paclitaxel: a Gynecologic Oncology Group study (Protocol 158) Top: Survival by treatment. Bottom: Survival by treatment group (microscopic vs. macroscopic) [From Ozols RF, Bundy BN, Greer BE, Fowler JM, Clarke-Pearson D, Burger RA, et al. Phase III trial of carboplatin and paclitaxel compared with cisplatin and paclitaxel in patients with optimally resected stage III ovarian cancer: a Gynecologic Oncology Group study. J Clin Oncol 2003;21:3194-3200, with permission.]
The International Collaborative Ovarian Neoplasm 3 (ICON3) trial was a study of 2,074 women with all stages of ovarian cancer, including 20% who had stage I or II disease (290). The combination of carboplatin plus paclitaxel was compared with two nonpaclitaxel regimens, carboplatin (70%) or cyclophosphamide, doxorubicin, and cisplatin (CAP) (30%).
The regimens were chosen before randomization and were based on the clinical preference of the treating physician. One-third of patients who received carboplatin or CAP subsequently received second-line paclitaxel, and this additional chemotherapy was often given before clinical progression. With a median follow-up of 51 months, the carboplatinplus paclitaxel and the control groups had a similar progression-free survival (0.93) and overall survival (0.98). The median survival for the paclitaxel plus carboplatin and control groups was 36.1 and 35.4 months, respectively. The median duration of progression-free survival was 17.3 and 16.1 months, respectively. The researchers concluded that single-agent carboplatin and CAP were as effective as paclitaxel and carboplatin for first-line chemotherapy. Because carboplatin as a single agent had a lower toxicity than the other regimens and the median survival was similar in a prior trial that had compared carboplatin and CAP as first-line treatment (291), the researchers suggested that carboplatin alone was the preferred therapy.
The design of this study limits drawing any definitive conclusions. Patients with FIGO stages I to IV ovarian cancer were included, the extent of primary surgery was variable, and the study was not audited by an independent data-monitoring committee. Furthermore, the majority (85%) of patients who relapsed after single-agent carboplatin subsequently receivedpaclitaxel. The investigators suggested that one explanation for the favorable outcome with carboplatin alone was that 30% of patients had dose escalation based on their nadir counts not falling significantly, although this was not protocol driven. They have commenced a study comparing standard AUC dosing of carboplatin with dose escalation based on nadir counts, and this will help address the question of whether “optimal” dosing of carboplatin makes a difference. The findings of the ICON study were not conclusive and have not altered the standard of care.
Carboplatin and Docetaxel
Docetaxel has a different toxicity profile to paclitaxel. The Scottish Gynaecological Cancer Trials Group (SCOT-ROC) study randomly assigned 1,077 women with stages IC to IV epithelial ovarian cancer to carboplatin with either paclitaxel or docetaxel (292). The efficacy of docetaxel appeared to be similar to paclitaxel: the median progression-free survival was 15.1 months versus 15.4 months, and the docetaxel group had less extremity weakness, neurologic effects, arthralgias, and myalgias than the paclitaxel group. However,the docetaxel plus carboplatin regimen was associated with significantly more myelosuppression and its consequences: serious infections and prolonged severe neutropenia. Therefore, additional study will be necessary to determine whether docetaxel should supplant paclitaxel in the primary treatment of epithelial ovarian cancer.
A large intergroup, international trial (GOG 182/SWOG 182/ICON5/ANZGOG) compared the standard combination of carboplatin and paclitaxel with these drugs in combination withgemcitabine, topotecan, or liposomal doxorubicin in sequential doublets or triplets (292,293).This was the largest randomized trial ever carried out in women with advanced ovarian cancer and recruited more than 4000 patients. The results have been presented, and there was no apparent difference between any of the arms in terms of progression-free survival or median survival although there were differences in the side effects experienced in the different arms. The conclusion was that the combination of carboplatin and paclitaxelshould remain the standard of care (294). A summary of these trials is presented in Table 11.4.
Intravenous Chemotherapy The issue of dose intensification of cisplatin was examined in a prospective trial conducted by the GOG (295). In this study, 243 patients with suboptimal ovarian cancer were randomized to receive 50 mg/m2 or 100 mg/m2 cisplatin plus 500 mg/m2 cyclophosphamide. There was no difference in response rates in those patients with measurable disease, and the overall survival times were identical. There was greater toxicity associated with the high-dose regimen. A Scottish group reported that patients who received 100 mg/m2 cisplatin plus 750 mg/m2 cyclophosphamide had a significantly longer median survival compared with those receiving 50 mg/m2 cisplatin plus the same dose ofcyclophosphamide (296). The overall median survival time was 114 weeks in the high-dose group and 69 weeks in the low-dose group (p = 0.0008), but this difference disappeared with longer follow-up (297). Therefore, the doubling of the dose of cisplatin does not improve the survival of these patients.
Dose escalations of paclitaxel and carboplatin require granulocyte colony-stimulating factor (G-CSF) because of the combined myelosuppressive effects, but there is no evidence to support a role for a more intensive course of either agent (270,298).
Intraperitoneal Chemotherapy A randomized, prospective trial in 546 evaluable patients of intraperitoneal cisplatin versus intravenous cisplatin (100 mg/m2), each given with 750 mg/m2 cyclophosphamide, was performed jointly by the Southwest Oncology Group (SWOG) and the GOG in patients with advanced ovarian cancer following optimal cytoreduction (residual nodules <2 cm diameter) (299). The intraperitoneal cisplatin arm had a somewhat longer overall median survival than the intravenous arm, 49 versus 41 months (p = 0.03).In the patients with minimal residual disease (<0.5 cm maximum residual diameter), who would be expected to derive the most benefit, there was no difference between the two treatments, 51 versus 46 months (p = 0.08) (Table 11.5).
In a follow-up trial of 532 patients (GOG Protocol 114), the dose-intense arm was initiated by giving a moderately high dose of carboplatin (dose AUC = 9) for two induction cycles followed by intraperitoneal cisplatin 100 mg/m2 and intravenous paclitaxel 135 mg/m2 over 24 hours versus intravenous cisplatin 75 mg/m2 and intravenous paclitaxel 135 mg/m2(300). In the dose-intense arm, progression-free median survival was 27.6 months compared with 22.5 months for the control arm (p = 0.02). However, there was no difference in overall survival (52.9 months versus 47.6 months, p = 0.056). Thus, it is unclear if dose intensification with intraperitoneal cisplatin has a long-term impact on the survival of these patients. A phase II trial of intravenous paclitaxel plus intraperitoneal cisplatin and paclitaxel was well tolerated and associated with a 2-year survival of 91% (302).
A randomized prospective GOG study compared intraperitoneal cisplatin and paclitaxel with intravenous cisplatin and paclitaxel (301). Four hundred and twenty-nine patients were randomly assigned, and 415 were eligible. The median progression-free survival was 23.8 months in the IP arm versus 18.3 months in the IV arm (p = 0.05). The median overall survival was 65.6 months in the IP group and 49.7 months in the IV group (p = 0.03) (Fig. 11.17). Ninety percent of patients in the IV arm received the six planned cycles of therapy, whereas only 42% of patients received the assigned 6 cycles of IP therapy with the remainder switching to IV therapy. The reasons for discontinuing were primarily for catheter-related problems, but there were also significantly more side effects in the IP group, with more patients experiencing severe fatigue, pain, hematological toxicity, nausea, and vomiting as well as metabolic and neurotoxicity. It is likely that with more training, appropriate dose modifications, and better antiemetics, the toxicity can be reduced.
The results of this study, together with the previous studies, led to an NCI clinical announcement recommending that women with optimally debulked stage III ovarian cancer be considered for IP chemotherapy. There has been a Cochrane Review as well as a separate metaanalysis that concluded that IP chemotherapy was associated with better outcomes than intravenous chemotherapy (304,305). The metaanalysis included six randomized trials with a total of 1,716 ovarian cancer patients. The pooled HR for progression-free survival of IP cisplatin as compared to IV treatment regimens was 0.792 (95% CI: 0.688-0.912, p = 0.001), and the pooled hazard ratio for overall survival was 0.799 (95% CI: 0.702-0.910, p = 0.0007). The authors concluded that these findings strongly supported the incorporation of an IP cisplatin regimen in the first-line treatment of stage III optimally debulked ovarian cancer (304).
Similar conclusions were reached in the Cochrane review. The reviewers concluded that their analysis established that IP chemotherapy was associated with an increased overall survival and progression-free survival in patients with optimally debulked stage III ovarian cancer. However, they also commented on the potential for catheter-related complications and increased toxicity with IP therapy and concluded that the optimal dose, timing, and mechanism of administration should be addressed in the next phase of clinical trials (305).
The role of IP chemotherapy is still contentious, with some researchers arguing that the trials to date were not pure tests of IP therapy and were flawed. In addition, concerns have been raised about the technical difficulties and increased toxicity of IP therapy (306).
Some authors have suggested that for patients with advanced disease, chemotherapy may be given before cytoreductive surgery. A series performed at Yale by Schwartz et al. (307) suggested that the survival of patients treated with neoadjuvant chemotherapy was comparable to that of patients historically treated with primary cytoreductive surgery in the same institution.
Table 11.5 Randomized Trials of IV Versus IP Chemotherapy in Patients with Advanced Epithelial Ovarian Cancer
Neoadjuvant chemotherapy might be appropriate in selected patients who are at high risk for operative morbidity or mortality [e.g., those with significant cardiac disease (308,309) or those with large pleural effusions], but primary cytoreductive surgery should be considered the standard of care for most patients. Bristow et al. (310) recently reported the results of a systemic overview of neoadjuvant chemotherapy and concluded that neoadjuvant chemotherapy represented a viable alternative strategy for the limited number of patients felt to be optimally unresectable by an experienced ovarian cancer surgical team; however, currently available data suggest that the survival outcome achievable with initial chemotherapy may be inferior to successful primary cytoreductive surgery.
Figure 11.17 Overall survival after intraperitoneal vs. intravenous cisplatin and paclitaxel chemotherapy. (From Armstrong DK, Bundy B, Wenzel L, Huang HQ, Baergen R, Lele S, et al. Intraperitoneal cisplatin and paclitaxel in ovarian cancer. N Eng J Med 2006:354:34-43, with permission.)
Table 11.6 Combination Chemotherapy for Advanced Epithelial Ovarian Cancer: Recommended Regimens
The EORTC has completed a large randomized trial of surgery followed by six cycles of carboplatin and paclitaxel versus three cycles of chemotherapy followed by surgical debulking and a further three cycles of chemotherapy. As these data become available, the results will help address the potential role of neoadjuvant chemotherapy.
Chemotherapy and Molecular Targeted Therapies
Inhibition of angiogenesis with drugs such as bevacizumab has demonstrated activity and benefit in women with recurrent ovarian cancer; in view of this, there are two large randomized trials investigating the impact, if any, of the addition of bevacizumab to standard carboplatin and paclitaxel in patients with advanced ovarian cancer. There is evidence in other tumor types such as breast cancer, colon cancer, and lung cancer that the addition of bevacizumab to chemotherapy increases response rates and progression-free survival and also survival in some studies (311,312,313).
GOG 218 is a phase III, three-arm randomized, double-blind, placebo-controlled trial. Patients in arm 1 will receive six cycles of carboplatin and paclitaxel and placebo starting with the second cycle and continuing for ten additional cycles after the completion of chemotherapy. Patients in arm 2 will receive six cycles of carboplatin and paclitaxel plusbevacizumab starting with cycle 2 and administered with chemotherapy followed by ten cycles of placebo, and in arm 3 patients will receive bevacizumab to start with cycle 2 ofcarboplatin and paclitaxel and then ten additional cycles of bevacizumzab after the completion of chemotherapy. This study is designed to investigate the benefit of bevacizumabin combination with chemotherapy as well as for maintenance. The bevacizumab is administered at a dose of 15mg/kg and starts with the second cycle of chemotherapy to decrease the risk of gastrointestinal perforation, which has been a rare complication of this agent in the setting of its use in colorectal cancer. The ICON7 study is similar but is a two-arm study of carboplatin and paclitaxel plus or minus bevacizumab 7.5 mg/kg administered every 3 weeks with chemotherapy and then as maintenance therapy. The results of these studies will be available in several years.
Chemotherapeutic Recommendations for Patients with Advanced Ovarian Cancer
For first-line chemotherapy of advanced epithelial ovarian cancer, we recommend the following (Table 11.6).
Consolidation and Maintenance of Complete Clinical Response to First-Line Chemotherapy
Because as many as 80% of women with advanced-stage disease who completely respond to their first-line chemotherapy will ultimately relapse, several trials have been conducted that administer a drug to these patients immediately following their primary treatment in an effort to decrease the relapse rate.
In a study conducted by the GOG and SWOG, 277 women with advanced ovarian cancer who had a complete clinical response to first-line chemotherapy were randomized to receive 3 or 12 cycles of additional single-agent paclitaxel (175 or 135 mg/m2 every 28 days) (314). Patients were excluded if they had developed grade 2 or 3 neurotoxicity during their initial chemotherapy. Because of cumulative toxicity, the mean number of actual cycles of paclitaxel received by the group assigned to receive 12 cycles was 9. The treatment-related grade 2 to 3 neuropathy was more common with longer treatment, 24% versus 14% of patients, respectively. The study was closed after a median follow-up of only 8.5 months, and an interim analysis showed a significant 7-month prolongation in median progression-free survival (28 versus 21 months) with 9 versus 3 months of consolidationpaclitaxel. However, there was no difference in median overall survival. The rate of disease progression increased significantly after maintenance therapy was discontinued, which suggested that long-term survival would not be likely to be improved. Furthermore, it is unlikely that a survival benefit will be seen with longer follow-up because patients assigned to three cycles were given the option of receiving an additional nine courses of paclitaxel after the study was discontinued (315).
Four additional courses of topotecan were administered to patients following six cycles of carboplatin and paclitaxel in two randomized trials, one conducted in Italy (316) and the other in Germany (317). In the larger trial conducted in Germany, 1,059 evaluable patients were randomly assigned to six cycles of paclitaxel (175 mg/m2 over 3 hours) andcarboplatin (AUC 5) with (537 patients) or without (522 patients) four additional cycles of topotecan (1.25 mg/m2 IV days 1 to 5 every 3 weeks) (316). In the Italian trial, 273 women were randomly assigned to receive four additional cycles (137 patients) of topotecan at a dose of 1 mg/m2 on days 1 to 5 every 3 weeks or no further chemotherapy (136 patients) (316). There were no significant differences in either progression-free or overall survival in patients who received four to six cycles of maintenance topotecan.
In a randomized clinical trial of intraperitoneal cisplatin for consolidation versus observation, there was no difference in survival between the treatment arms (316).
The clinical benefit of consolidation and maintenance chemotherapy seems doubtful. Patients and their physicians may consider prolonged single-agent paclitaxel an option, but it should not be considered the standard of care.
Administration of Chemotherapy and Amelioration of Toxicity
Paclitaxel The principal concern of combining paclitaxel and carboplatin is the potential for enhanced bone marrow toxicity. In general, shorter infusions of paclitaxel (e.g., 3 hours) tend to reduce the likelihood of bone marrow depression when combining the drug with carboplatin (270), although in practice this is not usually a problem. Conversely, when paclitaxel is combined with cisplatin, the principal concern is the potentiation of neurotoxicity. This toxicity can be minimized by using a lower dose of paclitaxeladministered over a longer period of time (e.g., 135 mg/m2 over 24 hours).
Carboplatin The renal and gastrointestinal toxicities of carboplatin are modest compared with cisplatin; thus, patients do not require prehydration, and outpatient administration is standard practice. Carboplatin does tend to have more bone marrow toxicity than cisplatin. Growth factors such as G-CSF have facilitated the administration of drug combinations that have neutropenia as a dose-limiting toxicity, although they are not commonly required with carboplatin and paclitaxel. The use of growth factors is discussed more fully inChapter 2.
There are some data to suggest that amifostine can reduce carboplatin and paclitaxel-induced neurotoxicity (319). In a phase III randomized trial of 187 women, the incidence of grade 3-4 neutropenia was lower in the arm with amifostine (31.3% vs. 37.9%; p = 0.03), as was the incidence of severe mucositis (4.7% vs. 15.4%, respectively; p <0.0001).Amifostine appeared to be protective against neurotoxicity (grade 3-4 neurotoxicity, 3.7% vs. 7.2%; p = 0.02). In a similar study, De Vos et al. (320) found that amifostine had only minor albeit statistically significant benefit in decreasing neurotoxicity, without preventing paclitaxel plus carboplatininduced bone marrow toxicity. The drug has not found widespread use.
Cisplatin Cisplatin combination chemotherapy is given every 3 to 4 weeks by intravenous infusion over 1 to 1.5 hours. Cisplatin requires appropriate hydration and can be administered on either an inpatient or outpatient basis. Hydration is administered with intravenous one-half normal saline at a rate of 300 to 500 dL per hour for 2 to 4 hours until the urinary output is greater than 100 dl per hour. When the urinary output is satisfactory, the cisplatin is infused in normal saline and the intravenous fluid rate is decreased to 150 to 200 dl per hour for 6 hours and then discontinued if the patient is stable.
The principal toxicities of this regimen are renal, gastrointestinal, hematologic, and neurologic. The renal and neurologic toxicities generally limit the duration of treatment to six cycles.
The acute and delayed gastrointestinal toxicity of cisplatin (i.e., nausea and vomiting) can be minimized with appropriate antiemetics, including a 5HT3 antagonist together withdexamethasone and aprepitant, an NK1 receptor blocker. All three published guidelines recommend this approach to prevent acute nausea and vomiting, and they also recommenddexamethasone and aprepitant to reduce delayed nausea and vomiting (321,322,323). For example, the ASCO guidelines recommend a 5HT3 antagonist (e.g., ondansetron 8 mg IV,palonsetron 0.25 mg, or one other 5HT3 antagonist) as well as dexamethasone 12 mg orally and aprepitant 125 mg orally prior to cisplatin followed by dexamethasone 8 mg daily from days 2 to 4 and aprepitant 80 mg orally on days 2 and 3 as appropriate antiemetic cover (323).
An alternative to first-line combination chemotherapy for selected patients with metastatic ovarian cancer is the use of whole-abdominal radiation therapy. This approach is not used in the United States, but it was standard treatment in some institutions in Canada for patients with no residual macroscopic tumor in the upper abdomen (246). It has been compared with oral chlorambucil and appears to be superior (246), but it has not been tested against combination chemotherapy. A trial of three cycles of high-dose cisplatin andcyclophosphamide “induction” chemotherapy followed by whole-abdominal radiation therapy to “consolidate” the initial response has been reported (324). No apparent benefit could be shown by adding wholeabdominal radiation after chemotherapy in patients with optimal disease.
There is no evidence that hormonal therapy alone is appropriate primary therapy for advanced ovarian cancer (325).
There is some interest in the use of biologic response modifiers in ovarian cancer, and in a trial of gamma interferon (γ-interferon) with cisplatin and cyclophosphamide, there appeared to be a benefit to the addition of the interferon (326). A trial of carboplatin and paclitaxel with or without γ-interferon concluded that there were more adverse events with γ-interferon and no survival benefits in women with advanced ovarian cancer (327).
Trials of monoclonal antibodies directed toward ovarian cancer-associated antigens have been conducted (328,329,330,331,332,333,334). Women who were in clinical remission following platinum and taxane chemotherapy were studied in a randomized, prospective trial of maintenance OvaRex, a monoclonal antibody directed toward CA125, and no progression-free survival benefit was seen (330,331). Studies with monoclonal antibodies directed at human milk fat globulin (HMFG) tumor-associated antigens for consolidation have shown no survival benefit (331); however, there was an improved control of intraperitoneal disease, which was offset by increased extraperitoneal disease (331). Herceptin, an humanized antibody directed toward the extracellular protein produced when the HER2/neu oncogene is overexpressed, has been used extensively in breast cancer, where it has been shown to improve the response rate to chemotherapy in selected patients with HER2 amplified breast cancer. A trial of herceptin antibody in HER-2/neu overexpressing ovarian cancers has been conducted by the GOG, and the response rate was low: 9.7% (334). The rationale for the use of these agents in ovarian cancer is discussed in Chapter 2.
Many patients who undergo optimal cytoreductive surgery and subsequent chemotherapy for epithelial ovarian cancer will have no evidence of disease at the completion of treatment. The second-look operation was previously considered to be standard practice, but is now seldom performed (335,336,337,338,339,340,341,342,343).
Elevated CA125 levels are useful in predicting the presence of disease, but normal levels are an insensitive determinant of the absence of disease. In a prospective study (347), the positive predictive value was shown to be 100%—that is, if the level of CA125 was >35 U/dL, disease was always detectable in patients at second-look laparotomy. The predictive value of a negative test was only 56%—that is, if the level was <35 U/dL, disease was present in 44% of the patients. A review of the literature suggests that an elevated CA125 level predicts persistent disease at second-look in 97% of the cases (36).
Serum CA125 levels can be used during chemotherapy to follow those patients whose level was elevated at the initiation of treatment (36,347). The change in level generally correlates with response. Those patients with persistently elevated levels after three cycles of treatment most likely have resistant clones. Rising levels on treatment almost invariably indicate treatment failure. A retrospective study determined that a doubling of the CA125 level from its nadir in those patients with a persistently elevated level accurately predicts disease progression (348).
The Gynecologic Cancer Intergroup (GCIG) developed a standard definition for CA125 level progression, which is now widely used in clinical trials. Patients with elevated CA125 levels pretreatment who normalize their levels must demonstrate CA125 levels greater than or equal to twice the upper limit of normal on two occasions at least 1 week apart. Patients with elevated CA125 levels pretreatment, which never normalize, must show evidence of CA125 levels greater than or equal to twice the nadir value on two occasions at least 1 week apart (349). However, a trial of 527 patients, randomized to early treatment of relapse based on CA 125 level alone versus delay treatment commencing when clinical symptomatic recurrence appear, showed no survival benefits from early treatment based on raised serum marker levels alone (350).
In patients with stage I to III epithelial ovarian cancer, radiologic tests have generally been of limited value in determining the presence of subclinical disease. Ascites can be readily detected, but even quite large omental metastases can be missed on CT scan (112,351,352). If liver enzymes are abnormal, then the liver can be evaluated with a CT scan or ultrasonography. CT-scan-directed fine-needle aspiration cytology will indicate tumor persistence if positive, but a negative result is not definitive. The false negative rate of a CT scan is approximately 45% (351).
Positron-emission tomography with or without CT imaging may help in the detection of relapse, but there appears to be a higher false positive rate with PET compared with CT (113,114,115). A review concluded that PET had a sensitivity of 90% and a specificity of 85% for the detection of recurrent ovarian cancer and that it appeared to be particularly useful for the diagnosis of recurrent disease when CA125 levels were rising and conventional imaging was inconclusive or negative (113). Technologic advances have led to combined 18fluorodeoxyglucose (18FDG)-PET-computed tomographic that provide contemporaneous 18FDG-PET and CT images. The role of 18FDG-PET-CT for the detection of recurrent ovarian cancer is promising, and this technique may be especially useful for the selection of patients with late recurrent disease who may benefit from secondary cytoreductive surgery (353,354,355). MRI can be used as an alternative to CT in patients with allergies to the contrast medium (115).
A second-look operation is one performed on a patient who has no clinical evidence of disease after a prescribed course of chemotherapy in order to determine the response to therapy.
The technique for a second-look laparotomy is essentially identical to that for a staging laparotomy. Areas of previously documented tumor are the most important areas to biopsybecause they are most likely to give a positive result. Any adhesions or surface irregularities should be sampled. In addition, biopsy specimens should be taken from the pelvic sidewalls, the pelvic cul-de-sac, the bladder, the paracolic gutters, the residual omentum, and the diaphragm. At least a pelvic and paraaortic lymph node sampling should be performed in those patients whose nodal tissues have not been previously removed.
Approximately 30% of patients with no evidence of macroscopic disease will have microscopic metastases (335). Also, in many patients with microscopic disease, macroscopic disease will be detected in only the occasional biopsy or cytologic specimen. Therefore, a large number of specimens (20,21,22,23,24,25,26,27,28,29,30) should be obtained to minimize the false negative rate of the operation. In selected patients in whom gross residual tumor is discovered at second-look surgery, resection of isolated masses may be performed. The removal of all macroscopic areas of disease might facilitate response to salvage therapies (356,357), as well as permitting the collection of tissue for in vitro analyses.
Second-look laparotomy has not been shown to influence patient survival, although the information obtained at second-look correlates with subsequent outcome and survival(336,337,338,339,342,343). Patients who have no histologic evidence of disease have a significantly longer survival than those in whom microscopic or macroscopic disease is documented. The operation should be performed selectively—for example, in patients receiving therapy in a setting where second-line therapies are undergoing clinical trials.
The likelihood that a patient will have a recurrence after a negative second-look laparotomy ranges from 30% to 60% at 5 years (339,340,341,342). The majority of recurrences after a negative second-look laparotomy are in patients with poorly differentiated cancers (342). Patients whose tumors are initially stage I and II have negative second-look laparotomy rates of 85% to 95% and 70% to 80%, respectively. In patients with optimally resected stage III disease treated with the platinum and paclitaxel regimen, the negative second-look rate is approximately 45% to 50% (256).
Second-look laparoscopy may be useful for patients on experimental treatment protocols. The advantage of laparoscopy is that it is a less invasive operation; the disadvantage is that visibility may be limited by the frequent presence of intraperitoneal adhesions (344,345,346). The development of newer techniques for retroperitoneal lymph node dissection has potentially increased the utility of the endoscopic approach to second-look.
Secondary cytoreduction maybe defined as an attempt at cytoreductive surgery at some stage following completion of first-line chemotherapy (356). Patients with progressive disease on chemotherapy are not suitable candidates for secondary cytoreduction, but patients who are clinically free of disease and undergo second-look laparotomy may benefit if all macroscopic residual disease can be resected (357). Patients with recurrent disease are sometimes candidates for surgical excision of their disease. Tumor resection under these circumstances should be restricted to those who have a disease-free interval of at least 12, but preferably 24 months, or those in whom all macroscopic disease can be resected,regardless of the disease-free interval (358,360,361).
Chemotherapy for Recurrent Ovarian Cancer
The majority of women who relapse will be offered further chemotherapy with the likelihood of benefit related in part to the initial response and the duration of response. The goals of treatment include improving disease-related symptoms, maintaining or improving quality of life, delaying time to progression, and possibly prolonging survival, particularly in women with platinum-sensitive recurrences.
Many active chemotherapeutic agents (platinum, paclitaxel, topotecan, liposomal doxorubicin, docetaxel, gemcitabine, and etoposide) as well as targeted agents (bevacizumab) are available, and the choice of treatment is based on many factors including likelihood of benefit, potential toxicity, and patient convenience (367,368). Women who relapse later than 6 months after primary chemotherapy are classified as platinum-sensitive and usually receive further platinum-based chemotherapy, with response rates ranging from 27% to 65% and a median survival of 12 to 24 months (369,370). Patients who relapse within 6 months of completing first-line chemotherapy are classified as platinum-resistant and have a median survival of 6 to 9 months and a 10% to 30% likelihood of responding to chemotherapy. Patients who progress while on treatment are classified as having platinum-refractory disease. Objective response rates to chemotherapy in patients with platinum-refractory ovarian cancer are less than 20% (368).
Table 11.7 Second-Line Chemotherapy in Recurrent or Persistent Epithelial Ovarian Cancer
Patients with platinum-refractory and platinum-resistant ovarian cancer are commonly treated with chemotherapy and may have a number of lines of therapy depending on response, performance status, patient request, and doctor recommendations. A study comparing topotecan with liposomal doxorubicin demonstrated the low response rates and poor prognosis for women with platinum-resistant ovarian cancer (371). In a subset analysis of platinum-resistant patients, the median time to progression ranged from 9.1 to 13.6 weeks for topotecan and liposomal doxorubicin, respectively. The median survival was 35.6 weeks for pegylated liposomal doxorubicin (PLD) and 41.3 weeks for topotecan (p = 0.455). Objective response rates were recorded in 6.5% of patients who received topotecan and in 12.3% of those who received PLD (p = 0.118). It is not known whether the treatment improved symptom control or quality of life because this was not specifically addressed.
The potential adverse effects associated with chemotherapy should not be underestimated and have been well documented in trials involving women with recurrent ovarian cancer.The three most commonly used drugs are paclitaxel, topotecan, and liposomal doxorubicin. The reported adverse effects associated with paclitaxel included alopecia in 62% to 100% of patients, neurotoxicity (any grade) in 5% to 42%, and severe leukopenia in 4% to 24%. Topotecan is associated with myelosuppression in 49% to 76% of patients, which is significantly greater than with liposomal doxorubicin or paclitaxel. Liposomal doxorubicin is associated with palmar-plantar erythrodysesthesia of any grade in more than 50% of patients, and it is severe in 23%. In addition, severe stomatitis has been reported in as many as 10% of patients (368). A large international study that is evaluating the impact of chemotherapy on quality of life and symptom improvement in patients with platinum-resistant or platinumrefractory ovarian cancer is currently underway.
The response rates for second-line chemotherapies are in the range of 10% to >40% for most drugs tested by the oral or intravenous route and depend on many factors. The most important predictors of response are response to first-line chemotherapy and the treatment-free interval (370,371,372,373,374,375,376,377,378,379,380,381,382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397,398,399,400,401,402,403,404,405,406,407,408,409,410,411,412,413,414,415,416,417,418,419,420,421,422,423,424,425,426,427,428,429,430,431,432,433,434) (Table 11.7).
Some studies suggest that a combination of platinum plus paclitaxel may be better in some patients for second-line therapy than platinum alone (376,377). In a study of 25 women who relapsed 6 months or longer after first-line carboplatin and paclitaxel, retreatment with the same combination had a response rate of 91% and a median progression-free survival of more than 9 months (377). Other studies suggest that single-agent therapy (cisplatin or carboplatin) should be considered the standard of care for platinum-sensitive disease (378,379).
In most studies, there is a lack of survival advantage and greater toxicity with multiagent compared with single-agent regimens. However, the combination of carboplatin andpaclitaxel is relatively well tolerated and has been shown to have a higher response rate and to be associated with a longer survival than carboplatin alone (379,380).
The use of combination platinum plus paclitaxel chemotherapy versus a single-agent platinum has been tested in two multinational randomized phase III trials (381) and a randomized phase II study (382). In a report (381) of the ICON4 and AGO-OVAR-2.2 trials, 802 women with platinum-sensitive ovarian cancer who relapsed after being treatment free for at least 6 to 12 months were randomized to platinum-based chemotherapy (72% carboplatin, or cisplatin alone; 17% CAP; 4% carboplatin plus cisplatin; and 3% cisplatin plusdoxorubicin) or paclitaxel plus platinum-based chemotherapy (80% paclitaxel plus carboplatin; 10% paclitaxel plus cisplatin; 5% paclitaxel plus both carboplatin and cisplatin; and 4%paclitaxel alone). The AGO-OVAR-2.2 trial did not accrue its planned number of patients. In both trials, a significant proportion of the patients had not received paclitaxel as part of their initial chemotherapeutic regimen. Combining the trials for analysis, there was a significant survival advantage for the paclitaxel-containing therapy (HR = 0.82), with a median follow-up of 42 months. The absolute 2-year survival advantage was 7% (57% vs. 50%), and there was a 5-month improvement in median survival (29 vs. 24 months). Progression-free survival was better with the paclitaxel regimen (HR = 0.76); there was a 10% difference in 1-year progression-free survival (50% vs. 40%) and a 3-month prolongation in median progression-free survival (13 vs. 10 months). The toxicities were comparable, except there was a significantly higher incidence of neurologic toxicity and alopecia in the paclitaxelgroup, whereas myelosuppression was significantly greater with the non-paclitaxel-containing regimens. These data support the slight advantage of a second-line regimen containing both paclitaxel and a platinum agent in patients who have not received paclitaxel in their primary chemotherapeutic regimen.
Two randomized trials have compared carboplatin alone to carboplatin and gemcitabine or liposomal doxorubicin (435,436). There was a higher response rate with the combination therapy and a longer progression-free survival, but the studies were not powered to look at overall survival. In the GCIG study comparing carboplatin and gemcitabine withcarboplatin alone, the response rate was 47.2% for the combination and 30.9% for carboplatin, with the progression-free survival being 8.6 months and 5.8 months, respectively(435). A SWOG study of carboplatin versus carboplatin and liposomal doxorubicin was closed early because of poor accrual, but with 61 patients recruited the response rate was 67% for the combination and 32% for carboplatin. The progression-free survival was 12 months versus 8 months; intriguingly, the overall survival was 26 months compared to 18 months (p= 0.02) (436). A phase II study from France confirmed the high response rate of 67% with carboplatin and liposomal doxorubicin in patients with platinum-sensitive recurrent ovarian cancer. A large GCIG study (CALYPSO) comparing carboplatin and liposomal doxorubicin with carboplatin and paclitaxel has completed recruitment, and the results are expected in the next 18 months (437).
Platinum-Resistant and Refractory Disease
In cisplatin-refractory patients—that is, those progressing on treatment—response rates to second-line carboplatin are less than 10% (375). The management of women who are platinum-resistant (i.e., progressing within 6 months of completion of chemotherapy) is difficult, and noncrossresistant agents are usually selected, but there does not appear to be one best treatment. Single-agent therapy is typically used because combination regimens are associated with more toxicity without any apparent additional benefit. Response rates of 48% to 64% have been reported with dose-dense weekly carboplatin (AUC = 4) and paclitaxel (90 mg/m2), and this deserves further study (438). There are a variety of active drugs; the most frequently used are paclitaxel, docetaxel, topotecan, liposomal doxorubicin, gemcitabine, oral etoposide, tamoxifen, and bevacizumab. Other active agents include vinorelbineand ifosfamide and newer drugs such as trabectidin.
The poor results achieved with chemotherapy in this population of patients are well demonstrated by a recent trial in which 195 patients with platinum-resistant ovarian cancer were randomized to receive either liposomal doxorubicin or gemcitabine. In the gemcitabine and PLD groups, median progression-free survival was 3.6 versus 3.1 months, median overall survival was 12.7 versus 13.5 months, and overall response rate was 6.1% versus 8.3%. In the subset of patients with measurable disease, overall response rate was 9.2% versus 11.7%, respectively. None of the efficacy end points showed a statistically significant difference between treatment groups. The liposomal doxorubicin group experienced significantly more hand-foot syndrome and mucositis, whereas the gemcitabine group experienced significantly more constipation, nausea and vomiting, fatigue, and neutropenia (439).
Some researchers have attempted to treat patients with non-platinum drugs to prolong the platinum-free interval, hoping that their use will allow the tumor to become platinum-sensitive during the interval use of non-cross-resistant agents. The rationale for this approach was that the platinum-free interval was equivalent to the treatment-free interval; before the availability of other active drugs, these two terms were synonymous. However, there are no data to support the hypothesis that the interposition of another drug can produce an increased platinum sensitivity as a result of a longer interval since the last platinum treatment.
Single-agent paclitaxel shows objective responses in 20% to 30% of patients in phase II trials of women with platinum-resistant ovarian cancer (380,383,384,385,386,387,388). The main toxicities are asthenia and peripheral neuropathy. A dose of 135 to 175 mg/m2 every 3 weeks, either as a 3-hour or 24-hour infusion, may be considered standard because a randomized trial showed similar response rates for both regimes (384). Three-hour infusions produce more neurotoxicity but less myelosuppression. Higher doses of paclitaxel (250 mg/m2 vs. 175 mg/m2 per dose) using hematopoietic growth factor support can result in higher response rates but with significantly more toxicity and no survival benefit (386).
Weekly paclitaxel is active, and the toxicity, especially myelosuppression, is less than with the 3-week regimens. In a study of 53 women with platinum-resistant ovarian cancer, weekly paclitaxel (80 mg/m2 over 1 hour) had an objective response of 25% in patients with measurable disease, and 27% of patients without measurable disease had a 75% decline in serum CA125 levels (386).
Docetaxel also has some activity in these patients (389,390,391). The GOG studied 60 women with platinum-resistant ovarian or primary peritoneal cancer (391). Although there was a 22% objective response rate, the median response duration was only 2.5 months, and therapy was complicated by severe neutropenia in three-quarters of the patients.
Topotecan is an active second-line treatment for patients with platinum-sensitive and platinum-resistant disease (392,393,394,395,396,397,398,399,400,401,402,403,404,405,406,407). In a study of 139 women receiving topotecan 1.5 mg/m2 daily for 5 days, response rates were 19% and 13% in patients with platinum-sensitive and platinum-resistant disease, respectively (392). The predominant toxicity of topotecan is hematologic, especially neutropenia. With the 5-day dosing schedule, 70% to 80% of patients have severe neutropenia, and 25% have febrile neutropenia with or without infection (359,360). In some studies, regimens of 5 days produce better response rates than regimens of shorter duration (392,393,394,395,396,397,398,399,400,401,402), but in others, reducing the dose to 1.0 mg/m2/day for 3 days is associated with similar response rates but lower toxicity (403,404). In a study of 31 patients, one-half of whom were platinum refractory (405), topotecan 2 mg/m2/day for 3 days every 21 days had a 32% response rate. Continuous infusion topotecan(0.4 mg/m2/day for 14-21 days) had a 27% to 35% objective response rate in platinum-refractory patients (398,399). Weekly topotecan administered at a dose of 4 mg/m2/week for 3 weeks with a week off every month produced a response rate similar to the 5-day regimen with considerably less toxicity. Therefore, this is now considered the regimen of choice for this agent (407).
Oral topotecan, not currently available in the United States, results in similar response rates with less hematologic toxicity (401). The intravenous and oral formulations of topotecanwere compared in a randomized trial of 266 women as a third-line regimen after an initial platinumbased regimen (406). Compared with intravenous topotecan (1.5 mg/m2 daily for 5 days every 3 weeks), oral topotecan (2.3 mg/m2/day for 5 days every 3 weeks) produced a similar response rate (13% vs. 20%), less severe myelosuppression, and only a slightly shorter median survival (51 vs. 58 weeks).
Liposomal doxorubicin (Doxil in the United States and Caelyx in Europe) has activity in platinum- and taxane-refractory disease (371,408,409). The predominant severe toxicity ofliposomal doxorubicin is the hand-foot syndrome, also known as palmar-plantar erythrodysesthesia or acral erythema. This morbidity is observed in 20% of patients who receive 50 mg/m2 every 4 weeks (408). Liposomal doxorubicin tends to produce a low rate of both neurologic toxicity and alopecia. In a study of 89 patients with platinum-refractory disease, including 82 paclitaxel-resistant patients, liposomal doxorubicin (50 mg/m2 every 3 weeks) produced a response in 17% (one complete and 14 partial responses) (409). In another study, an objective response of 26% was reported, although there were no responses in women who progressed during first-line therapy (408).
There have been two randomized trials comparing liposomal doxorubicin with either topotecan or paclitaxel. In a study of 237 women who relapsed after receiving one platinumcontaining regimen, 117 of whom (49.4%) had platinum-refractory disease (371), liposomal doxorubicin 50 mg/m2 over 1 hour every 4 weeks was compared with topotecan1.5 mg/m2/day for 5 days every 3 weeks. The two treatments had a similar overall response rate (20% vs. 17%), time to progression (22 vs. 20 weeks), and median overall survival (66 vs. 56 weeks). The myelotoxicity was significantly lower in the liposomal doxorubicin-treated patients than with those receiving topotecan. In a second study comparing liposomal doxorubicin with single-agent paclitaxel in 214 platinum-treated patients who had not received prior taxanes (410), the overall response rates for liposomal doxorubicin andpaclitaxel were 18% versus 22%, respectively, and median survival durations were 46 and 56 weeks, respectively. Neither was significantly different. In practice, most patients are treated with a starting dose of 40mg/m2 of liposomal doxorubicin every 4 weeks because of the toxicity associated with the higher dose and the need to commonly dose reduce when 50mg/m2 is used.
Gemcitabine has been associated with response rates of 20% to 50% with, 6% in patients who are platinum-resistant (413,414,415,416,417). The principal toxicities are myelosuppression and gastrointestinal. The drug has been used in doublet combinations with cisplatin or carboplatin with acceptable responses and toxicities and in the triplet combination with carboplatin and paclitaxel (415).
The most common toxicities with oral etoposide are myelosuppression and gastrointestinal: Grade 4 neutropenia is observed in approximately one-fourth of patients, and 10% to 15% have severe nausea and vomiting (418,419,420). Although an initial study of intravenous etoposide reported an objective response rate of only 8% among 24 patients (418), a subsequent study of oral etoposide given for a prolonged treatment (50 mg/m2 daily for 21 days every 4 weeks) had a 27% response rate in 41 women with platinum-resistant disease, three of whom had durable complete responses (419). In 25 patients with platinum- and taxane-resistant disease, eight objective responses (32%) were reported. Oral etoposideshould be considered one of the principal drugs to be used in patients with paclitaxel- and platinum-resistant disease.
Other Chemotherapeutic Agents
Other active oral agents associated with response rates of 20% to 25% include hexamethamelamine (418,419,420,421,422), capecitabine (424), ifosphamide with mesna (425), andtrabectedin (440).
Tamoxifen has been associated with response rates of 15% to 20% in well-differentiated carcinomas of the ovary (426,427,428,429,430). The gonadotropin-agonist leuprolide acetate (Lupron) has been shown to produce a response rate of 10% in one series (431). Trials combining tamoxifen and leuprolide acetate, and tamoxifen and combination chemotherapy are being conducted (432). Aromatase inhibitors (e.g., letrozole, anastrozole, and exemestane, which have been shown to have activity in metastatic breast cancer) are being studied in relapsed ovarian cancer (433). One principal advantage of this class of agents is the very low toxicity (434).
A new era of cancer treatment is being entered in which knowledge of molecular pathways within normal and malignant cells has led to the development of agents with specific molecular targets. The greatest success to date in ovarian cancer has been in targeting angiogenesis, in particular VEGF, which has been found to play a major role in the biology of epithelial ovarian cancer (441). There are three main approaches to targeting angiogenesis: The first is to target VEGF itself, the second to target the VEGF receptor, and the third to inhibit tyrosine kinase activation and downstream signaling with small molecules that work at the intracellular level.
Bevacizumab is the first targeted agent to show significant single-agent activity in ovarian cancer. It is a humanized monoclonal antibody that targets angiogenesis by binding to VEGFA, thereby blocking the interaction of VEGF with its receptor. There have been a number of phase II studies reported using bevacizumab in patients with platinum-sensitive and platinum-resistant ovarian cancer, with response rates ranging from 16% to 22% in both platinum-sensitive and platinum-refractory patients (442,443). Furthermore, as many as 40% of patients have had stabilization of disease for at least 6 months.
A study of 70 patients with recurrent ovarian cancer using low-dose metronomic chemotherapy with 50 mg of cyclophosphamide daily and bevacizumab 10 mg/kg intravenously every 2 weeks showed significant activity (444). The primary end point was progression-free survival at 6 months. The probability of being alive and progression free at 6 months was 56%. A partial response was achieved in 17 patients (24%). Median times to progression and survival were 7.2 and 16.9 months, respectively. The side effects of bevacizumab are now well recognized and include hypertension, fatigue, proteinuria, and gastrointestinal perforation or fistula. Uncommonly, vascular thrombosis and CNS ischemia, pulmonary hypertension, bleeding, and wound healing complications may occur. The most common side effect is hypertension. This is grade 3 in 7% of patients and is usually readily treatable, whereas the most concerning side effect is bowel perforation. The study by Cannistra was stopped after recruiting 44 patients because of an 11% incidence of perforation of the bowel(441,442,443,444).
It has been suggested that the complication of bowel perforation can be avoided by careful screening of patients. Simpkins et al. (445) limited bevacizumab treatment to patients without clinical symptoms of bowel obstruction, evidence of rectosigmoid involvement on pelvic examination, or bowel involvement on CT scan. Their study included 25 patients with platinum-resistant ovarian cancer, and all had been heavily pretreated. They observed a response rate of 28% and had no bowel perforations or any other grade 3 or 4 toxicity (445). This highlights the importance of patient selection and also suggests that as clinicians become more experienced with these agents, there should be less toxicity.
VEGF-Trap functions as a soluble decoy receptor, soaking up a ligand before it can interact with its receptor. It is also currently being evaluated in phase II trials in patients with recurrent ovarian cancer. There are also a number of other oral agents that target angiogenesis through tyrosine kinase inhibition that are currently in clinical trial (446).
Dose-Intense Second-Line Chemotherapy
In patients with minimal residual (≤5 mm) or microscopic disease confined to the peritoneal cavity, intraperitoneal chemotherapy or immunotherapy has been used (447,448,449,450,451,452,453,454,455,456,457,458,459,460,461,462,463,464). Cytotoxic chemotherapeutic agents such as cisplatin, paclitaxel, 5-fluorouracil (5-FU), etoposide (VP-16), and mitoxantrone, have been used as single agents in patients with persistent epithelial ovarian cancer (447,448,449,450,451,452), and complete responses have been seen in patients who start their treatment with minimal residual disease. The surgically documented response rates reported with this approach are approximately 20% to 40% for carefully selected patients, and the complete response rate is 10% to 20%. Although it has been suggested that this approach produces a significant subsequent improvement in survival (455), there are no prospective phase III data that demonstrate this, and the patients so treated tend to be those with a more favorable prognosis regardless of subsequent therapy.
Of historical interest is intraperitoneal immunotherapy with alpha interferon (α-interferon), γ-interferon, tumor necrosis factor, and interleukin 2, which were shown to have some activity in patients with minimal residual disease (456,457,458,459,460,461,462,463) (see Chapter 2). The response rate for the intraperitoneal cytokines, α-interferon and γ-interferon, is the same as that for the cytotoxic agents: approximately 28% to 50% (456,457,458,461,463). The combination of cisplatin and α-interferon produced a surgically documented 50% complete response rate, which was greater than that produced by either single agent alone (458). However, because interferons are not FDA approved for use in ovarian cancer, they are no longer used.
Second-line intraperitoneal treatment is not suitable for most patients because they often have extensive intraperitoneal adhesions or extraperitoneal disease. Therefore, second-line intraperitoneal chemotherapy and immunotherapy should be considered experimental.
High-Dose Chemotherapy and Autologous Bone Marrow Transplantation The use of high-dose chemotherapy and either autologous bone marrow transplantation or peripheral stem cell protection has been tested in patients with advanced ovarian cancer (465,466,467). In one trial of high-dose carboplatin with autologous bone marrow transplantation, 7 of the 11 patients with extensive refractory disease had an objective response. The maximum tolerated dose of high-dose carboplatin was 2 g/m2 (465).
A phase III randomized trial of 57 patients treated with high-dose chemotherapy (cyclophosphamide 6,000 mg/m2 and carboplatin 1,600 mg/m2) with peripheral blood stem cell support as consolidation versus 53 patients treated with conventional dose maintenance (cyclophosphamide 600 mg/m2 and carboplatin 300 mg/m2) has been reported (467). Only 43 of the 57 women (75%) completed the high-dose therapy, while 48 of 53 (92%) completed the standard dose regimen. There was no statistically significant difference in progression-free and overall survival between the two groups of patients.
A prospective randomized clinical trial of a combination, very high-dose chemotherapy supported with autologous bone marrow transplantation versus standard-dose chemotherapy with paclitaxel and carboplatin was initiated by the Gynecologic Oncology Group, but the trial was discontinued because of poor accrual.
A European study of high-dose chemotherapy was recently reported (468). One hundred forty-nine patients with untreated ovarian cancer were randomly assigned after debulking surgery to receive standard combination chemotherapy or sequential high-dose treatment with two cycles of cyclophosphamide and paclitaxel followed by three cycles of high-dosecarboplatin and paclitaxel with stem cell support. High-dose melphalan was added to the final cycle. After a median follow-up of 38 months, the progression-free survival was 20.5 months in the standard arm and 29.6 months in the high-dose arm. The median overall survival was 62.8 months in the standard arm and 54.4 months in the high-dose arm. This is the first randomized trial comparing sequential high-dose with standard-dose chemotherapy in first-line treatment of patients with advanced ovarian cancer, and no statistically significant difference in progression-free survival or overall survival was observed. The investigators concluded that high-dose chemotherapy does not appear to be superior to conventional-dose chemotherapy.
Whole-abdominal radiation therapy given as a second-line treatment has been shown to be potentially effective in a small subset of selected patients with microscopic disease, but it is associated with a relatively high morbidity. The principal problem associated with this approach is the development of acute and chronic intestinal morbidity. As many as 30% of patients treated with this approach develop intestinal obstruction, and this may necessitate potentially morbid exploratory surgery (469). Because there are many new chemotherapeutic agents available for the treatment of relapsed ovarian cancer, most centers have stopped using second-line whole-abdominal radiation therapy.
Patients with epithelial ovarian cancer often develop intestinal obstruction, either at the time of initial diagnosis or, more frequently, in association with recurrent disease (470,471,472,473,474,475,476,477,478). Obstruction may be related to a mechanical blockage or to carcinomatous ileus. Correction of the intestinal blockage can be accomplished in most patients whose obstruction appears at the time of initial diagnosis. However, the decision to perform an exploratory procedure to palliate intestinal obstruction in patients with recurrent disease is more difficult. In patients whose life expectancy is very short (e.g., less than 2 months), surgical relief of the obstruction is not indicated (470). In those whose projected life span is longer, features that predict a reasonable likelihood of correcting the obstruction include young age, good nutritional status, and the absence of rapidly accumulating ascites.
For most patients with recurrent ovarian cancer who present with intestinal obstruction, initial management should include radiographic documentation of the obstruction, hydration, correction of any electrolyte disturbances, and parenteral alimentation (471,472,473,474,475). In many patients, the obstruction may be alleviated by this conservative approach. A preoperative upper gastrointestinal series and a barium enema will define possible sites of obstruction.
If exploratory surgery is deemed appropriate, the type of operation to be performed will depend on (i) the site and (ii) the number of obstructions. Multiple sites of obstruction are not uncommon in patients with recurrent epithelial ovarian cancer. More than one-half of the patients have small-bowel obstruction, one-third have colonic obstruction, and one-sixth have both (474,475,476). If the obstruction is principally contained in one area of the bowel (e.g., the terminal ileum), then this area can either be resected or bypassed, depending on what can be accomplished safely. Intestinal bypass is generally less morbid than resection; in patients with progressive cancer, the survival time after these two operations is the same (474,475,476).
If multiple obstructions are present, then resection of several segments of intestine is usually not indicated, and intestinal bypass or colostomy should be performed. A gastrostomy may occasionally be useful in this circumstance (476), and this can usually be placed percutaneously (477).
Surgery for bowel obstruction in patients with ovarian cancer carries an operative mortality of approximately 10% and a major complication rate of some 30%(470,471,472,473,474,475,476). The need for multiple reanastomoses and preceding radiation therapy increase the morbidity, which consists primarily of sepsis and enterocutaneous fistulae. The median survival ranges from 3 to 12 months, although approximately 20% of such patients survive longer than 12 months (471).
There is a trend toward improved survival for ovarian cancer (1,447). FIGO has reported a statistically significant improvement in survival for all stages from 29.8% for the interval 1976 to 1978 to 49.7% for the interval 1999 to 2001 (479). In the United States, the 5-year survival for all stages combined increased from 37% in 1975-1977 to 45% in 1996-2003 (1). The death rate decreased from 9.51 per 100,000 to 8.75 per 100,000 from 1991 to 2003, an 8% decline (1).
The 5-year survival rate for carefully staged patients with stage IA disease is 89.6%, whereas it is 86.1% for stage IB and 83.4% for stage IC (479). The 5-year survival for stage II disease is 70.7%, 65.5%, and 71.4% for stages IIA, IIB, and IIC, respectively. For stage IIIA, 5-year survival is 46.7%, whereas it is 41.5% for stage IIIB, 32.5% for stage IIIC, and 18.6% for stage IV (479). The proportion in each stage at the time of diagnosis is shown in Fig. 11.18, and the survival by substage is presented in Fig. 11.19.
The 5-year survival of patients with stage III disease with microscopic residual disease only at the start of treatment is 63.5% compared with 32.9% for those with optimal residual disease (≤2 cm), and 24.8% for those with suboptimal residual disease (>2 cm) (279) (Fig. 11.19).
Including patients at all stages, patients younger than 50 years of age have a 5-year survival rate of approximately 40%, compared with some 15% for patients older than 50 (279). Patients whose Karnofsky's index (KI) is low (<70) have a significantly shorter survival than those with a KI >70 (169,174).
Regarding patients with invasive cancer, for stages I and II disease, the 5-year survival rate for grade 1 epithelial ovarian cancer is approximately 90.3%, compared with approximately 79.7% for grade 2 and 75% for grade 3 (479) (Fig. 11.20). Examining stages III and IV patients, the 5-year survivals for grades 1, 2, and 3 are 57.2%, 31%, and 28.5%, respectively (479) (Fig. 11.21).
Survival of patients with borderline tumors is excellent, with stage I lesions having a 98% 10-year survival (8,10,185,192,479). When all stages of borderline tumors are included, the 5-year survival rate is 87% (479) (Fig. 11.22).
Figure 11.18 Survival of patients with epithelial ovarian cancer by stage. The percentage of patients diagnosed at a particular stage is shown next to the 5-year survival by stage. (Data from Heintz APM, Odicino F, Maisonneuve P, Quinn MA, Benedet JL, Creasman WT, et al. Carcinoma of the ovary. In: Pecorelli S, ed. Twenty-Sixth Annual Report on the Results of Treatment in Gynaecological Cancer. Int J Gynecol Oncol 2006;95(suppl 1):S161-S192, with permission.)
Figure 11.19 Survival of patients with epithelial ovarian cancer by substage. (From Heintz APM, Odicino F, Maisonneuve P, Quinn MA, Benedet JL, Creasman WT, et al. Carcinoma of the ovary. In: Pecorelli S, ed. Twenty-Sixth Annual Report on the Results of Treatment in Gynaecological Cancer. Int J Gynecol Oncol 2006;95(suppl 1):S161-S192, with permission.)
Figure 11.20 Survival of patients with FIGO stages I and II epithelial ovarian cancer by grade of the tumor. (From Heintz APM, Odicino F, Maisonneuve P, Quinn MA, Benedet JL, Creasman WT, et al. Carcinoma of the ovary. In: Pecorelli S, ed. Twenty-Sixth Annual Report on the Results of Treatment in Gynaecological Cancer. Int J Gynecol Oncol 2006;95(suppl 1): S161-S192, with permission.)
Figure 11.21 Survival of patients with FIGO stages III and IV epithelial ovarian cancer by grade of the tumor. (From Heintz APM, Odicino F, Maisonneuve P, Quinn MA, Benedet JL, Creasman WT, et al. Carcinoma of the ovary. In: Pecorelli S, ed. Twenty-Sixth Annual Report on the Results of Treatment in Gynaecological Cancer. Int J Gynecol Oncol 2006;95(suppl 1):S161-S192, with permission.)
Figure 11.22 Survival of patients with borderline versus invasive epithelial ovarian cancer. (From Heintz APM, Odicino F, Maisonneuve P, Quinn MA, Benedet JL, Creasman WT, et al. Carcinoma of the ovary. In: Pecorelli S, ed. Twenty-Sixth Annual Report on the Results of Treatment in Gynaecological Cancer. Int J Gynecol Oncol2006;95(suppl 1):S161-S192, with permission.)
Fallopian Tube Cancer
Carcinoma of the fallopian tube currently accounts for 0.3% of all cancers of the female genital tract (99,480,481,482,483,484,485). Molecular and genetic data suggest that many serous ovarian cancers arise in the distal fallopian tube (2). The definitions of ovarian and fallopian tube malignancies may need to be revised as the distinction between these two entities is reassessed.
In histologic features and clinical behavior, fallopian tube carcinoma is similar to ovarian cancer; thus, the management is essentially the same. Almost all fallopian tube cancers are of “epithelial” origin, most frequently of serous histology. Rarely, sarcomas have also been reported. The fallopian tubes are frequently involved secondarily from other primary sites, most often the ovaries, endometrium, gastrointestinal tract, or breast (99). They may be secondarily involved in primary peritoneal cancer.
Tubal cancers are seen most frequently in the fifth and sixth decades, with a mean age of 55 to 60 years (99,480,481,482,483,484). Women who have germ-line mutations in BRCA1and BRCA2 are at substantially higher risk of developing fallopian tube carcinoma; therefore, prophylactic surgery in these women should include a complete removal of both fallopian tubes along with the ovaries (90,486).
Symptoms and Signs
The classic triad of symptoms and signs associated with fallopian tube cancer is (i) a prominent watery vaginal discharge—that is, hydrops tubae profluens; (ii) pelvic pain; and (iii) a pelvic mass. However, this triad is noted in fewer than 15% of patients (90).
Either vaginal discharge or bleeding is the most common symptom reported and is documented in more than 50% of patients (302,482). Lower abdominal or pelvic pressure and pain also are noted in many patients. However, the presentation may be rather vague and nonspecific. In perimenopausal and postmenopausal women with an unexplained or persistent vaginal discharge, even in the absence of bleeding, the clinician should be concerned about the possibility of an occult tubal cancer. Fallopian tube cancer is often found incidentally in asymptomatic women at the time of abdominal hysterectomy and bilateral salpingo-oophorectomy.
On examination, a pelvic mass is present in approximately 60% of patients, and ascites may be present if advanced disease exists. Patients with tubal carcinoma will have a negative dilation and curettage (483,487), although abnormal or adenocarcinomatous cells may be seen in cytologic specimens obtained from the cervix in 10% of patients.
Tubal cancers spread in much the same manner as epithelial ovarian malignancies, principally by the transcoelomic exfoliation of cells that implant throughout the peritoneal cavity. In approximately 80% of the patients with advanced disease, metastases are confined to the peritoneal cavity at the time of diagnosis (482).
The fallopian tube is richly permeated with lymphatic channels, and spread to the paraaortic and pelvic lymph nodes is common. Metastases to the paraaortic lymph nodes have been documented in at least 33% of the patients with all stages of disease (482).
Fallopian tube cancer is staged according to the International Federation of Gynecology and Obstetrics (480). The staging is based on the surgical findings at laparotomy (Table 11.8). According to this system, approximately 29% of patients have stage I disease, 23% have stage II, 39% have stage III, and 7% have stage IV at the time of diagnosis (480). A somewhat lower incidence of advanced disease is seen in these patients than in patients with epithelial ovarian carcinomas, which may result from many advanced-staged fallopian tube cancers being designated ovarian or peritoneal cancers. The 2008 FIGO staging for carcinoma of the fallopian tube is unchanged.
Table 11.8 FIGO Staging for Carcinoma of the Fallopian Tube
The treatment of this disease is identical to that of epithelial ovarian cancer.
Patients with tubal carcinoma should undergo total abdominal hysterectomy and bilateral salpingo-oophorectomy (99,484,488,489,490,491,492,493). If there is no evidence of gross tumor spread, a staging operation is performed. The retroperitoneal lymph nodes should be adequately evaluated, and peritoneal cytologic studies and biopsies should be performed, along with an infracolic omentectomy.
In patients with metastatic disease, an effort should be made to remove as much tumor bulk as possible. Extrapolation from the experience with epithelial ovarian cancer suggests that the best outcome should be achieved in patients in whom all macroscopic disease can be resected (459).
As with epithelial ovarian cancer, the most active agents are platinum and the taxanes. Cisplatin or carboplatin plus paclitaxel has response and survival outcomes that are similar to that for epithelial ovarian cancer (99,489,490,491,492,493). Therefore, the same treatments used for epithelial ovarian cancer should be used in patients with epithelial tubal malignancies (490,493).
A variety of other chemotherapeutic agents that are effective against ovarian cancer appear to be active in fallopian tube carcinomas as well. These agents include docetaxel, etoposide, topotecan, gemcitabine, and liposomally encapsulated doxorubicin (494,495,496,497,498).
Data on well-staged lesions are scarce, so it is unclear whether patients with disease confined to the fallopian tube (i.e., a stage IA, grade 1 or 2 carcinoma) benefit from adjuvant therapy.
The role of radiation in the management of the disease remains unclear because patients have not been treated in any consistent manner and the small numbers treated preclude any meaningful conclusions (485,499,500,501). In advanced stage disease, chemotherapy appears to be more effective (501). Whole-abdominal radiation has been used in patients with no evidence of gross disease in the abdomen (i.e., completely resected disease or microscopic metastases only) but was ineffective (499). As with epithelial ovarian cancer, there may be a role in properly selected patients.
The overall 5-year survival for patients with epithelial tubal carcinomas is 56% (480). This number is higher than for patients with ovarian cancer and reflects the somewhat higher proportion of patients with early stage disease. The reported 5-year survival rate for patients with stage I disease is 81%, 67% for stage II disease, 41% for stage III disease, and 33% for stage IV (480).
Tubal sarcomas, mostly malignant mixed mesodermal tumors, have been described but are rare (502). Leiomyosarcomas have been reported (503). They occur mainly in the sixth decade and are typically advanced at the time of diagnosis (2). An attempt should be made to resect all gross disease, and platinum-based combination chemotherapy should be given (502). These lesions need to be distinguished from extragastrointestinal stromal tumors, which are treated with imatinib (467). Survival of patients with primary tubal sarcomas is generally poor, and most die of their disease within 2 years (2).