David Starks, Bin Yang, and Peter G. Rose
In the field of gynecologic oncology, the various diagnostic modalities available serve as invaluable tools in the diagnosis, management, staging, treatment, and monitoring of gynecologic malignancies. Technological advances in existing modalities such as ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) have furthered their utility as diagnostic and management instruments, while the indications for newer imaging modalities such as 2-(18F)-fluoro-2-deoxy-D glucose (FDG) positron emission tomography (PET)/CT continue to expand. The use of tumor markers in identifying disease and molecular pathology in confirming which specific disease exists is presented. Because of the important role played by these diagnostic tools, the gynecologic oncologist must possess at least a passing familiarity with the basic science underlying these diagnostic instruments, as well as understand their advantages and limitations in imaging the spectrum of gynecologic cancers. Not all diagnostic modalities are useful or appropriate in evaluating the different and varied gynecologic cancers. Furthermore, the impact of diagnostic studies in the field of gynecologic oncology is ever expanding as newer technological developments become available to the clinician who must understand how to translate these advancements into improved patient care. Finally, as cost-effectiveness becomes an ever more important driver of health care decision making, it behooves the gynecologic oncologist to understand the various diagnostic tools in his armamentarium in order to use them to maximal effect.
IMAGING MODALITIES
Diagnostic imaging is an expanding field that has replaced radiology and now encompasses numerous new and varied technologies.
Ultrasound
Ultrasound is the most widely used imaging modality in the field of gynecology and is often the initial radiologic study used in the evaluation of pelvic abnormalities. Ultrasound technology uses a hand-held transducer containing piezoelectric crystals capable of emitting high-frequency sound waves that are projected into the patient’s body. Emitted frequencies range from 7.0 to 8.0 MHz, used in transabdominal scanning, and up to 9.0 MHz is generally used in transvaginal ultrasound (TVUS). Higher-frequency sound waves result in improved image resolution, but reduced tissue penetration. The piezoelectric crystals serve as both the emitter and receiver of the sound waves. As the wave encounters tissue surfaces, it is both reflected and transmitted. The reflected wave returns to the transducer, where it is converted into an electrical signal, which is termed an echo, and the signal is amplified and converted into different shades of gray based on the degree of amplification. Stronger echoes are perceived as whiter shades, whereas weaker echoes are assigned darker shades.
Doppler ultrasonography can be added to basic ultrasound studies in order to evaluate vascular structures and blood flow. Doppler ultrasonography uses the principles of the Doppler Effect, which states that a moving object will emit a wavelength with differing frequencies and lengths based on whether the object is moving toward or away from the source emitting the sound wave. The sound waves emitted by the ultrasound transducer are reflected by vascular structures being studied, and objects moving toward the transducer emit a high-frequency, short-wavelength echo, whereas objects moving away from the transducer emit a low-frequency, long-wavelength echo. Based on these frequencies and wavelengths, the ultrasound transducer is able to determine the velocity of flow in the vascular structure and generate a Doppler waveform. Color-flow Doppler, by convention, assigns flow toward the transducer as red and flow away from the transducer as blue, but this assignment is arbitrary and can be reversed by the ultrasonographer.
The strengths of ultrasound technology includes its relative ubiquity and low cost, as well as a high safety profile due to the absence of ionizing radiation. The weaknesses of ultrasound include its reliance on the skill and experience of the operator, the need for a high degree of training in order to obtain a necessary degree of competence, the inability of the sound waves to penetrate gas or bone, and difficulty in visualizing midline organs due to the obscuring effect of overlying bowel gas or patient obesity and body habitus.
Indications for performing ultrasound studies in the field of gynecologic oncology include the initial evaluation of the endometrial lining in the setting of postmenopausal bleeding; evaluating and describing the nature of pelvic and adnexal masses, identifying and diagnosing gestational trophoblastic disease, and playing an important role in the performance of ultrasound-guided biopsies and percutaneous drainage procedures conducted by interventional radiologists.
Computed Tomography
High-resolution CT scans are an invaluable tool for the gynecologic oncologist, and use of CT scans in the field is extensive. Indeed, CT scans are probably the most frequently used imaging technique ordered by gynecologic oncologists and the second most common imaging study after ultrasound in the field of gynecology. CT scans offer the oncologist a noninvasive means of evaluating the extent and spread of metastatic disease and provide guidance in surgical planning, detecting disease recurrence and progression, and detecting lymph node involvement. CT scans also play a role in interventional radiology, permitting accurate biopsies and drainage procedures to occur.
CT scanners use x-ray beams that are rotated about a patient in a 180-degree arc. Laying directly opposite of the beam emitters are sets of crystal detectors, 2 to 10 mm in size, that capture the emitted photons and measure tissue absorption. This information is then processed by a computer that generates a 2-dimensional cross-sectional representation of the anatomic structure under radiologic evaluation. In order to generate clearer and more diagnostic images, contrast medium is often used to enhance the discrepancy in tissue absorption and densities. Contrast medium may be given to patients either orally, intravenously, or rectally, and care must be used when administering contrast to patients with a known iodine allergy or renal dysfunction.
The advantages of CT scans include a short period of time required for scanning; minimization of the dependency on operator skill, thus leading to a high degree of reproducibility, and a high degree of spatial and anatomic resolution. More recent modifications in CT technology have led to further improvement in image resolution and visualization. For example, helical CT combines continuous patient transport through a scanner with a single row detector array that generates a spiraling projection of x-rays. This is a dynamic modification of the traditional CT, in which the patient remains stationary while being scanned. Helical CT decreases the scanning time required for a patient, decreases the volume of contrast dye that needs to be administered for improved resolution, and provides more images from several different angles, thus generating higher image quality. Helical CT also has improved detection of smaller lesions that are difficult to detect with conventional CT. Movement artifacts such as peristalsis in the intestines or interference from patient breathing are minimized, allowing for improved image resolution. Another modification of conventional CT is the multidetector CT scanner (MDCT), which uses the same basic principles of helical scanning, but uses a multiple-row detector array instead of a single-row detector array. Current scanners use 16-, 32-, or 64-slice systems that create a cross-sectional slice thickness of 1 to 2 mm, yielding images with even higher spatial resolution and improved image quality.
The disadvantages of CT scans include the use of ionizing radiation, which has led to a growing concern of an increased risk of radiation-associated cancers. A recent study concluded that for a 40-year-old female patient being evaluated by a routine CT of the abdomen and pelvis with contrast, the chance of developing cancer is 1 in 870, as compared with 1 in 470 for a woman 20 years of age and 1 in 1400 for a woman 60 years of age.1 Caution must be used in interpreting these findings and those from similar studies. The risk of cancer to the individual patient from having a CT scan performed is small, even with the use of high-dose radiation. This risk is often outweighed by the benefits accrued from CT scans that are truly indicated, but the growing awareness in the medical community of this risk of cancer secondary to radiation has led to an effort to minimize nonindicated or multiple scans. Other disadvantages of CT include the fact that patient body habitus, as well as implanted metallic devices and prostheses, can obscure and degrade image quality, leading to the generation of a nondiagnostic scan.
Magnetic Resonance Imaging
MRI scanners produce a magnetic field that aligns hydrogen nuclei within the patient. An intermittent radiofrequency pulse is emitted from the scanner, which alters the alignment of these hydrogen nuclei. When the radiofrequency pulse is discontinued, the hydrogen nuclei return to their original alignment, releasing a quantity of energy in the process. The amount and rate of energy released is wholly dependent on the property of the tissues containing the hydrogen nuclei. The longitudinal relaxation time is termed T1, and in T1-weighted images, fluid appears dark and fat appears white. In the transverse relaxation time, termed T2, fluid has a white appearance. In contrast to CT, MRI does not use ionizing radiation, but instead relies on magnetic fields and radio-frequencies. The paramagnetic element gadolinium is used as a contrast agent in MRI, and in comparison with the iodine-containing contrast agents used in CT, gadolinium has a much lower rate of adverse reactions and allergies. Recent studies of gadolinium have established an association with nephrogenic systemic fibrosis (NSF) in patients with a history of acute renal failure and end-stage renal disease. The American College of Radiology (ACR) has published guidelines for the use of gadolinium in patients at risk for the development of NSF.2 The ACR recommends that before the administration of gadolinium, a recent glomerular filtration rate (GFR) (in the last 6 weeks) be obtained for any patient with a history of renal disease, age greater than 60 years, history of hypertension or diabetes, or history of severe hepatic disease or liver transplantation. Although patients with stage I or stage II chronic renal disease do not require any special consideration, patients with stage III or V disease (GFR < 30 mL/min/1.73 m2) need to be referred to a nephrologist for evaluation before gadolinium administration and potential dialysis after the performance of an MRI.
The advantages of MRI include the use of nonionizing radiation and non–iodine-containing contrast agents. MRI can penetrate calcified material such as bone without significant attenuation in the signal or loss of image resolution. MRI also has remarkable soft tissue resolution. Recent advances have shortened the imaging time required to obtain scans, as well as created techniques for imaging the heart and blood vessels without the need for contrast agents.
The disadvantages of MRI include the expense of scans, issues with motion artifacts, and the inability of patients with metallic implants and prosthesis, such as pacemakers and artificial joints, to be placed inside an MRI scanner. Many patients experience episodes of anxiety or claustrophobia inside MRI scanners and may require anxiolytics to be fully compliant during scanning.
The role of MRI in gynecologic oncology is still being elaborated, but it can serve as a useful adjunct in determining the extent of local and distant tumor spread in gynecologic malignancies and also has great sensitivity and specificity in the evaluation of vaginal and vulvar cancers.
Positron Emission Tomography
PET scans have developed an important role in the diagnosis and management of a number of different malignancies, including gynecologic cancers. PET scans use a radiochemical tracer, most commonly FDG, which is preferentially taken up by malignant cells due to their increased rate of glycolysis. This means that PET scans are capable of detecting the early biochemical abnormalities associated with malignancy, before the development of the structural and tissue changes caused by malignancies that are necessary for cancer detection by other imaging techniques.3 Combining FDG-PET with CT in a single scanning device has allowed the images obtained by both modalities to be fused, generating images in which areas of increased FDG uptake are superimposed on CT images, providing anatomical localization.
The advantages of PET scans are largely due to their ability to detect the early abnormalities associated with tumor growth and recurrence. Their disadvantages include the current high cost of both the PET scanner, as well as the cost of the scan. These costs are often not covered by a patient’s insurance, except in a few indicated conditions such as staging cervical cancer. PET scans have a high false-negative rate in evaluating lesions that are less than 1.0 cm in size or in detecting malignancies with low metabolic activity. Furthermore, areas of inflammation can result in false-positive results.
The role of FDG-PET CT in the management of gynecologic cancers is still being elaborated. It currently has a role in the staging of cervical cancer, as well as monitoring for recurrence of cervical cancer. PET may also be useful in the detection of recurrent ovarian and endometrial cancer, but more study is needed before greater clinical application is undertaken.
Image-Guided Percutaneous Biopsies
The use of percutaneous biopsy, performed by interventional radiology using either CT or ultrasound guidance, to assist in the diagnosis of a pelvic mass continues to develop and is not without controversy. Biopsies may be obtained with the use of 16- or 18-gauge needles or through the use of 1.0- to 1.8-mm Surecut needles (UK Biopsy Ltd., Halifax, Great Britain) in order to obtain a core biopsy. Clinicians have raised concerns that biopsies of cystic ovarian lesions may have a high false-negative rate and a low diagnostic accuracy, while increasing a patient’s risk for procedure-related tumor seeding and contamination of the peritoneum due to rupture or leakage of the cystic mass. It may be that the concern for peritoneal seeding of malignancy is theoretical at best, but the current clinical consensus in the management of a cystic ovarian lesion is for surgical management of the lesion, either by laparoscopy or laparotomy, rather than through percutaneous biopsies. The use of percutaneous biopsies to aid in the diagnosis of solid pelvic masses is less controversial and may be of benefit, especially in cases of widespread meta-static disease. The concern for intraperitoneal seeding does not appear to have the same risk for tumor leakage that seems inherent in performing a biopsy in a cystic lesion. Ascites is often a common finding in the setting of advanced malignancy, but the sensitivity of performing cytology on ascites appears to be 60%. With solid tumors, it is possible to obtain larger tissue samples by core biopsy, permitting immunohistochemical studies and molecular profiling to be performed. A recent study by Hewitt et al4 examined 149 women with suspected ovarian cancer undergoing a biopsy of an adnexal mass by either CT or ultrasound guidance using an 18-gauge needle. The diagnostic rate was 90% for CT and 91% by ultrasound, with only 1 hemorrhagic complication documented in the series. The authors argued that percutaneous biopsy was safe and could be considered as a replacement for surgical intervention in the initial management and diagnosis of malignancy. Such conclusions are currently preliminary at best, and surgical management and diagnosis of pelvic masses is still considered to be the standard of care.
IMMUNOHISTOPATHOLOGY
Immunohistochemistry is a method for localizing specific antigens in tissue or cells based on antigen-antibody recognition. In the past 3 decades, a laundry list of antibodies has been developed with tissue specificity. Immunohistochemistry has enormous impact on the accuracy of pathologic diagnosis. We briefly summarize the current application of immunohistochemistry in facilitating the accurate diagnosis of gynecologic neoplasms, with focus on malignant epithelial neoplasms.
Cervix: p16 Is a Surrogate Biomarker for High-Grade Cervical Dysplasia
High-risk human papilloma virus (HPV), most commonly 16 or 18, is responsible for the majority of cervical cancers. Development of invasive cervical cancer is preceded by HPV-related cervical dysplasia, known as cervical intraepithelial neoplasia (CIN). CIN can be low grade (CIN1) or high grade (CIN2-3). Detection of high-grade cervical dysplasia by Pap smear and subsequent tissue biopsy is critical in the identification and treatment of precursor lesions for the prevention of cervical cancer. Unfortunately, the reproducibility of diagnosis of CIN2 in pathologic samples is not good. Therefore, the identification of surrogate markers for HPV infection and, more importantly, evidence of molecular changes leading to cervical cancer would be of vital importance in the discrimination between low-grade and high-grade dysplasia.
These HPV types encode 2 proteins, E6 and E7, that are oncogenic by their inhibition of tumor suppressor genes. This results in uninterrupted cellular replication and malignant transformation of some infected cervical cells.5Integration of HPV DNA into the cells is the prerequisite step for the expression of the oncoproteins E6 and E7, which subsequently degrades tumor suppressor proteins such as p53 and pRb. HPV E7 protein expression results in degradation of pRb protein, which normally inhibits the transcription of p16 (CDKN2A).
Diffuse p16 expression has reliably been shown to be a surrogate biomarker for CIN2-3. However, there are approximately less than 30% of CIN1 lesions also expressing p16 with a expressing p16 with a focal and patchy pattern. Recent studies indicated that approximately 20% of p16-positive CIN1 progress to high-grade lesions as compared with none of the p16-negative CIN1 lesions within a 12-month follow-up period. Diffuse and full thickness of the p16 immunostaining pattern is a hallmark of high-grade CIN and is a very useful ancillary tool in those challenging cases in differentiating CIN2 from CIN1 and from immature squamous metaplasia.
Diffuse expression of p16 is also seen in adeno-carcinoma in situ (AIS) of the cervix. Again, this expression correlates with the involvement of high-risk HPV in the development of these lesions. Detection of p16 is helpful in the diagnosis of cervical AIS to distinguish AIS from endometriosis and tuboendometrial metaplasia. The latter has a focal and discontinuous staining pattern that is distinct from the disuse and continuous staining pattern in AIS lesions. Additionally, p16 aids in the differential diagnosis of endometrioid adenocarcinoma of the endometrium with endocervical adenocarcinoma. A diffuse p16staining pattern is typically seen in endocervical adenocarcinoma, but is rarely seen in endometrioid adenocarcinoma. However, it should be emphasized that p16 immunostaining alone has no role in the differential diagnosis between endocervical adeno-carcinoma and serous adenocarcinoma of the endometrium because diffuse p16 staining pattern can be seen in both types of cancer.
Vulva: p16 and p53 Expression in Vulvar Intraepithelial Neoplasia
There are 2 types of vulvar intraepithelial neoplasia (VIN), classic (usual) and simplex (differentiated) types, based on histopathologic features and distinct molecular pathogenetic pathways. The most common precursor for vulvar squamous cell cancers is the classic or usual type of VIN, which is associated with HPV infection. These tumors are seen in younger patients who often have a history of cervical HPV infection. It has shown that tumor suppressor protein p16 has been overexpressed in the majority of high-grade VINs. The staining pattern is diffuse and full thickness of the dysplastic epithelium. The molecular mechanism for overexpression of p16 in the classic type of VIN is analogous to that seen in HPV-associated cervical CIN.6 Furthermore, unlike simplex VIN, classical VIN is rarely associated with p53 overexpression.
Simplex (differentiated) type of VIN is less frequently seen clinically and tends to be seen in older patients with no association of HPV infection. These lesions often arise in a background of lichen sclerosis. Histopathologically, recognition of simplex VIN and differentiating it from benign squamous hyperplasia can be challenging. Furthermore, these lesions are not as commonly detected before the development of invasive disease. This is thought to reflect both the difficulty in clinical and pathologic diagnosis of this lesion and the fact that it is thought to have a short time to progression to invasive disease.
Simplex VIN and keratinizing squamous cell vulvar cancers have consistently been strongly associated with p53 mutations. It has been shown that approximately two-thirds of simplex VIN lesions display overexpression of p53immunohistochemically. p53 immunostaining is a useful ancillary tool in making the distinction between simplex VIN and benign vulvar lesions. However, caution must be taken when dealing with a lesion with morphology and a p53 immunostaining discrepancy. Because p53 deletion is seen in some of the simplex VINs, a negative p53 immunostaining should not prevent the diagnosis if morphologically convincing.
Immunoprofile of Extramammary Paget Disease
Extramammary Paget disease (EMPD) is an unusual diagnosis characterized by the presence of Paget cells proliferating within the intraepidermis. Vulvar Paget disease can be primary or secondary. Primary disease is that which originates from the epidermis or skin appendages, and secondary disease is that which represents extension of a visceral carcinoma to the vulva, most commonly rectal or urologic carcinomas. The clinical and histologic appearance of primary and secondary EMPD is similar, and thus differentiation can be a challenge. The prognostic implications between the 2 diagnoses are significant. Primary EMPD is usually a locally confined lesion, and the clinical outcomes are substantially better. Secondary EMPD, on the other hand, represents the spread of visceral tumor onto the vulvar skin and has a substantially worse prognosis. Therefore, accurate clinical diagnosis is very important in this disorder. Molecular markers have recently been found to be of utility in the distinction between primary and secondary EMPD and may have utility in aiding with diagnosing these lesions.
Cytokeratins (CKs) have utility in the detection of certain cancers. CK20 is associated with colorectal adenocarcinomas. It has also been correlated with the presence of a primary colorectal adenocarcinoma associated with secondary EMPD. Furthermore, CK20 staining was not seen in primary EMPD. CK7, on the other hand, has been consistently seen in primary EMPD and sometimes in secondary EMPD, although not as frequently. Therefore, a vulvar Paget disease possessing the immunoprofile of CK7–/CK20/+should prompt an aggressive search for an underlying malignancy, such as colorectal cancer.
Gross cystic disease fluid protein (GCDFP)-15 is a glycoprotein expressed in apocrine epithelial cells. GCDFP-15 has been found in primary EMPD, and its absence is correlated with the presence of secondary EMPD.7 The immunohistochemical detection of carcinoembryonic antigen (CEA) has also been found to be of value in EMPD. The value of CEA staining seems to be in differentiating EMPD from superficial spreading melanoma and not in the separation of primary and secondary EMPD, although negative CEA staining is seen more commonly in secondary EMPD.
Endometrium
Based on the degree of malignancy and prognosis, endometrial carcinoma is divided into type 1 and type 2 cancers. Type 1 cancer includes endometrioid and mucinous adenocarcinoma, whereas type 2 encompasses serous and clear cell adenocarcinoma.
Loss of PTEN in Type 1 Endometrial Carcinoma
In type 1 cancers, loss of expression of pTEN protein due to point mutations and promoter methylation is the most frequent genetic alterations observed. PTEN is a tumor suppressor gene located on chromosome 10q23. The PTENgene encodes a dual-specificity phosphatase with a role in cell cycle arrest and promotion of apoptosis via phosphatidylinositol-(3,4,5)-triphosphate (PIP3). Immunohistochemically, the majority of endometrioid and mucinous adenocarcinomas have negative immunoreactivity to pTEN antibody compared with adjacent benign endometrium. Because loss of pTEN is an early molecular event during endometrial carcinogenesis, lack of pTEN immunoreactivity has in recent years become the important biomarker in identifying the precursor lesion of endometrial intraepithelial neoplasia and a small subset of higher-grade endometrioid adenocarcinomas acquiring p53mutations at a late stage.8 Therefore, p53 immunostaining alone does not distinguish between serous adenocarcinoma and grade 3 endometrioid adenocarcinoma harboring a p53 alteration.
p53, p16, and IMP3 Expression in Type 2 Endometrial Carcinoma
Type 2 cancers are often characterized by p53 mutations. p53 is a tumor suppressor gene and is the most commonly mutated gene in human cancers. p53 protein product binds to DNA and upregulates transcription of genes, which act to halt the cell cycle and assist with DNA repair or initiate apoptosis if repair is not possible. Anti-p53 antibody reacts with both wild-type and mutant p53 proteins. However, because the half-life of wild-type p53 protein in cells is only less than 20 minutes, it rarely detects p53 immunoreactivity in most normal cells. In contrast, because the majority of mutant p53 proteins have greater than 16 hours of half-life, its accumulation can be easily seen immunohistochemically in malignant cells. Approximately 90% of serous adenocarcinomas demonstrate p53 mutation and accumulation of p53 immunohistochemically. Different from type 1 cancers, p53 mutation is an early genetic event in serous adenocarcinoma. It has been shown that up to 80% of endometrial intraepithelial carcinomas (EIC) contain mutations in p53. Therefore, immunohistochemical detection of p53 overexpression is a very useful biomarker in identifying and confirming early precursors of EIC in endometrial biopsy or curettage specimens. Approximately 30% to 40% of clear cell adenocarcinomas harbor p53 mutations.
Overexpression of p16 protein is also seen in serous adenocarcinoma of the endometrium or the ovary. Overexpression of p16 in serous adenocarcinoma is not linked to HPV infection. The underlying molecular mechanism is still largely unresolved. Because both cervical adenocarcinoma and serous adenocarcinoma can share high nuclear grade histopathologically, and overexpression of p16 is found in both cervical adenocarcinoma and serous adenocarcinoma, another layer of challenge in the differential diagnosis of the 2 is added. When the issue arises, application of both p16 and p53 immunostains helps in resolving the issue. Serous adenocarcinoma will be strongly positive both for p53 and p16, whereas cervical adenocarcinoma will be positive for p16 but negative for p53.
IMP3 is an oncoprotein that is mainly expressed in fetal and malignant tissues and rarely in adult benign tissues. IMP3 has been found to be involved in cell growth, adhesion, and migration. Strong IMP3staining has been demonstrated in 86% to 94% of serous adenocarcinoma, as opposed to only 3% to 28% of endometrioid adenocarcinoma. Furthermore, 50% of clear cell cancers also stain positive for IMP3. Expression of IMP3 is also seen in 89% of EICs, indicating its involvement in early carcinogenesis.
Ovarian Epithelial Cancers
There are 5 types of ovarian epithelial cancer, each with distinct cell types and clinicopathologic features and treatment options: (1) papillary serous carcinoma (PSC), (2) clear cell carcinoma, (3) endometrioid carcinoma, (4) mucinous carcinoma, and (5) transitional cell carcinoma. Each type has its distinct pathogenesis and immunophenotypes.
p53 and p16 Expression in Papillary Serous Carcinoma
p53 mutations are seen in approximately 70% of high-grade PSCs, but are rarely seen in low-grade PSCs.9 p16 is also expressed in high-grade PSC. The similar expression pattern of p53 and p16 in both endometrial and ovarian serous adenocarcinoma suggests an analogous pathway for carcinogenesis of these tumors.
Immunoprofile for Mucinous Adenocarcinoma
Mucinous adenocarcinoma, especially intestinal type, of the ovary can be morphologically indistinguishable from those derived from gastrointestinal tract. One of the most important issues clinically is to know whether a mucinous adenocarcinoma is primary ovarian cancer or secondary from other sites. Earlier studies indicate that the majority of colorectal cancer is CK7 negative and CK20 positive, whereas primary ovarian mucinous carcinoma is positive for both CK7 and CK20. Therefore, a CK7-negative mucinous adenocarcinoma is likely metastasized from the colorectum, and a CK-positive mucinous adenocarcinoma is likely an ovarian primary, if endocervical adenocarcinoma is excluded. The recent discovery of CDX2 expression in most of gastrointestinal cancers further facilitates the differential diagnoses. However, recent evidence of expression of CDX2 in some of the primary mucinous adenocarcinoma of the ovary and the expression of CK7 in some right-sided colon cancers further complicated the case. Therefore, although immunohistochemistry is in many situations helpful in the differential diagnosis, it is by no means a magic bullet. Clinicopathologic correlation is still crucial in rendering the correct diagnosis. Furthermore, CK7-negative and CK20-positive immunoprofile is also seen in mucinous adenocarcinomas of the lung, breast, pancreas, and stomach. Therefore, immunohistochemical findings must be put into a clinical context, and the expression of CK7 does not exclude that the ovarian tumor is secondary.
TUMOR MARKERS
Tumor markers are serologic substances that are produced by a malignancy or are abnormally elevated in response to the presence of a malignancy. They can be enzymes, growth factors, hormones, tumor antigens, receptors, and glycoconjugates. They play a role in screening, diagnosis, monitoring treatment response, and detecting disease recurrence. A large number of tumor markers have been investigated in recent years, but few have entered into clinical practice. This has largely been due to the fact that many of these tumor makers have poor specificity, which is defined as the proportion of patients without a cancer who have a negative test. Many of the current markers can be elevated in a number of conditions, both benign and malignant, contributing to their lack of specificity. With the development of new high-throughput approaches such as proteomics, which uses mass spectrometry (MS) techniques, a new interest has developed in investigating the patterns of tumor marker expression. Identification and describing these patterns offers the promise of the development of more sensitive and specific assays for various cancers and their histologic subtypes.
Cervical Cancer
No serum tumor markers exist or are being researched to screen for cervical cancer due to the success of Pap and HPV DNA-based screening programs. The few markers that were historically investigated were abandoned due to their low sensitivity and specificity. Several serologic markers have been assessed to play a potential role in determining prognosis, detecting recurrent disease, and monitoring treatment response. For example, squamous cell carcinoma antigen (SCCA) has moderate sensitivity when elevated in the setting of cervical cancer, but unfortunately has a low specificity, as SCCA levels may be elevated in a number of other squamous cell cancers, such as carcinoma of the head, neck, and lung. SCCA can also be elevated in benign conditions as well, such as psoriasis and eczema. However, there does appear to be a correlation of SCCA with prognosis and the clinical response of cervical cancer to treatment. Levels of SCCA above 1.1 ng/mL are associated with a poor prognosis. A recent study demonstrated that elevated levels of SCCA immediately after treatment with chemoradiation was predictive of distant recurrence.10 Another study demonstrated that in patients being treated with chemoradiation, previously elevated SCCA levels normalized in 93% of patients at 1 month after treatment and in 96% of patients with a complete remission at 1 month. Although used in Europe and Japan, no US company has pursued licensing of this assay for use in the United States.
CA-125 is elevated in only approximately 21% of women with squamous cell carcinoma of the cervix and may correspond to prognosis, particularly if there is a decrease in preoperative CA-125 after treatment. The addition of other markers to CA-125, such as CEA and CA19-9, can increase the sensitivity of a tumor marker panel for detecting cervical cancer, but currently, obtaining these 3 tumor markers to manage a diagnosed cervical cancer is not considered to be standard of care.
CA-125 is elevated in 20% to 75% of patients with cervical adenocarcinoma and may reflect tumor stage, size, grade, presence of lymphovascular space involvement and lymph node involvement. A recent study of patients with adenocarcinoma of the cervix observed that in multivariate analysis, CA-125 was an independent prognostic factor for disease-free survival. The investigators also demonstrated that tumor necrosis factor receptor type I may potentially be the most useful marker in evaluating the prognosis of adenocarcinoma, particularly in early-stage disease.11 Further research into the use of these particular tumor markers is warranted before their use can become widespread.
Endometrial Cancer
CA-125 is often elevated in patients with uterine serous carcinomas and advanced-stage endometrioid cancers, and obtaining CA-125 preoperatively has demonstrated a correlation with metastatic disease and extrauterine spread. However, CA-125 can be falsely positive, particularly after pelvic or abdominal radiation. Several other serologic markers, including CA15-3, CA19-9, CA72-4, cancer associated serum antigen, CEA, squamous-cell carcinoma antigen (SCCA), gamma-GT, urinary gonadotropin fragment (UGF), placental protein 4 and others, have been investigated as potential tumor markers but have been largely abandoned as viable candidates for either screening or clinically managing endometrial cancer. Other novel makers currently under investigation include the glycoprotein YKL-40, which, obtained preoperatively, may detect endometrial cancer and aid in determining prognosis. Higher levels of serum inhibins, particularly inhibin β-B (INH-β-B) was observed in grade 3 endometrial cancers compared with grade 2 cancers. In one study, inhibin α (INH-α) was an independent prognostic factor for progression-free survival, cause-specific survival, and overall-survival. Elevations in human kallikrein 6 may be overexpressed in patients with papillary serous endometrial cancer. Finally, pyruvate kinase M2, chaperonin 10, and α-1-antitrypsin performed well as a panel of biomarkers demonstrating a high sensitivity, specificity, and positive predictive value (PPV) in detecting endometrial cancer.12 Testing of current biomarker candidates and the development of better tumor markers is ongoing.
Ovarian and Fallopian Tube Cancers
CA-125 is a 200-kilodalton (kDa) glycoprotein that is recognized by the OC-125 murine monoclonal antibody. It has 2 important antigenic domains: A is the domain-binding monoclonal antibody OC125; B is the domain binding monoclonal antibody M11. A number of assays exist for detecting serum CA-125; one of the most widespread is the second-generation heterologous CA-125 II assay, which uses both OC125 and M11 antibodies. The upper limit of normal of serum CA-125 of 35 U/mL was chosen because only 1% of healthy women had a value above this point. Levels can fluctuate, however, based on the phase of the menstrual cycle and is more often elevated if the woman is pre- versus postmenopausal. Eighty-five percent of women with epithelial ovarian cancer have a CA-125 level greater than 35 U/mL, with 25% to 50% of stage I patients having an elevated level and 90% of women with advanced-stage disease demonstrating an elevation in CA-125. The sensitivity of CA-125 is approximately 78%, with a specificity of 95% and a PPV of 82% based on both prospective and retrospective data. CA-125 is nonspecific and can be elevated in a number of other malignancies (cancers of the breast, colon, lung, and pancreas), benign conditions, endometriosis, pelvic inflammatory disease, and pregnancy. Finally, CA-125 is not a marker for nonepithelial ovarian malignancies, and low levels are common in borderline, endometrioid, clear cell, and mucinous epithelial tumors.
CA-125 is useful to measure treatment response, with the levels and pattern of CA-125 being monitored over time. In assessing treatment response, a decrease in CA-125 by 50% correlates with disease responsive to chemotherapy, whereas a doubling of CA-125 from baseline constitutes treatment failure and disease progression. A recent study found that CA-125 half-life and nadir concentration had independent prognostic value for disease-free and overall survival. The same investigative group recently demonstrated that a bi-exponential CA-125 decay was an indicator of poor prognosis after primary chemotherapy.13 However, CA-125 is a poor marker of small-volume disease and can be falsely negative, as demonstrated in studies of second-look laparotomy after chemotherapy, which found active disease despite low or normal CA-125 levels.
CA-125 is a strong predictor of disease recurrence as well, and serial measurements are used to monitor patients for this indication. Serum elevations may be detectable 2 to 6 months before evidence of disease recurrence becomes visible on imaging studies.14 A recent study demonstrated that in patients with a complete clinical remission, a progressive low-level increase in serum CA-125, from a baseline nadir, with an absolute increase of 5 to 10 U/mL, was strongly associated with disease recurrence. However, a recent randomized trial of early detection of recurrent disease with CA-125 failed to improve overall survival.15 The only other marker with current US Food and Drug Administration approval for detection of recurrent ovarian cancer is HE4. A recent study demonstrated that HE4 correlated with a patient’s clinical status in 76.2% of cases, which was not inferior to CA-125’s correlation of 78.8%.16Use of both CA-125 and HE4 in a combined assay appears to increase sensitivity and maintain specificity in the detection of primary ovarian cancer in women with a newly diagnosed pelvic mass.
Although playing an important role in ovarian cancer surveillance strategies, the use of CA-125 as a screening tool remains investigational. Because of its lack of specificity, use of CA-125 alone as a screening tool is controversial. It has been reported that screening specificity can be increased with the addition of TVUS, improving specificity to 99.9% and PPV to 26.8% in postmenopausal women. However, the recent Prostate, Lung, Colorectal and Ovarian (PLCO) screening trial demonstrated the limitations of CA-125 and TVUS in screening for ovarian cancer. In this trial, 34,261 postmenopausal women were randomized to receive annual CA-125 and TVUS screening for 3 years, followed by 2 additional years of CA-125 monitoring. Women were referred to a gynecologic oncologist if either the CA-125 or TVUS were abnormal. The PPV for CA-125 was 3.7% and for TVS was 1%. If both were abnormal, the PPV was 23.5%, but more than 60% of ovarian malignancies would have been missed using this strategy. Finally, the sensitivity for detecting early-stage disease was notably low, with only 21% of detected ovarian cancers being either stage I or II.17
Improved screening PPV and specificity can be obtained by abandoning the use of a fixed cutoff of 35 U/mL and using the statistical Risk of Ovarian Cancer Algorithm (ROC), which uses a woman’s age-specific incidence of ovarian cancer and CA-125 behavior over time to estimate a woman’s risk of ovarian cancer. A prospective randomized trial of the ROC algorithm in 13,582 postmenopausal women demonstrated a high specificity of 99.8% and a PPV of 19% in detecting primary invasive epithelial ovarian cancer.18 The ROC algorithm is currently being used in the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) study, which is ongoing at this time. In addition to CA-125, a large number of serum markers have been studied and evaluated, but very few have any relevance to the clinical realm due to poor sensitivity and specificity. CEA is elevated in endometrioid and Brenner tumors and occasionally in mucinous tumors. CA19-9 is expressed by mucinous ovarian cancers and cancers of the colon, but has low levels of expression in other epithelial ovarian cancers. A few of the tumor markers used in the diagnosis and management of nonepithelial ovarian cancers are detailed in Table 3-1.
Serum inhibin levels are elevated in sex–cord stromal tumors, especially granulosa cell tumors. Inhibin plays a role in the regulation of follicle-stimulating hormone secretion by the pituitary. It is composed of an α subunit and 1 of β subunits (BA or BB). Although inhibin A and inhibin B levels can both be elevated in patients with granulosa cell tumors, an inhibin B level is elevated in a higher proportion of these tumors. A recent study evaluating the use of serum inhibin levels in 30 women with granulosa cell tumors demonstrated that the sensitivities and specificities for inhibin A were 67% and 100% and for inhibin B were 89% and 100%, respectively. The investigators also noted that inhibin A level was elevated before or at the time of first clinical recurrence in 58% of patients, whereas inhibin B level was elevated in 85%. The lead time from elevation of inhibin levels to clinical recurrence was estimated to be 11 months. Inhibin A and B levels were not elevated in any of the 17 patients who were postoperatively disease-free.19
Vulvar and Vaginal Cancers
The rarity of vulvar and vaginal cancers has made the investigation and development of useful tumor markers considerably difficult. There are a couple of novel markers under investigation, including carbonic anhydrase IX (CAIX), which may be elevated in vulvar cancer and may be correlated with recurrence-free survival compared with CAIX-negative tumors. CAIX overexpression may also correspond to tumor progression and inguinal lymph node metastasis. Overexpression of another marker, COX-2, may correspond with disease-specific survival in vulvar cancer. Other markers that have been investigated include tissue polypeptide-specific antigen, SCCA, and urinary gonadotropin fragment. No serologic marker to date has demonstrated sufficient sensitivity or specificity to play a role in screening or in detecting recurrent disease, or in directing clinical management. Due to the rarity of these 2 cancers, the creation of large trials sufficient to validate any tumor marker under current investigation seems doubtful.
IMAGING FOR SPECIFIC GYNECOLOGIC CANCERS
Cervical Cancer
Cancer of the uterine cervix is the third most common gynecologic malignancy in the United States, with 12,200 new cases and 4210 deaths expected in 2010. However, worldwide cervical cancer is the most common gynecologic cancer and among women is second only to breast cancer as the most common cancer. Staging of cervical cancer remains clinical, rather than surgical, due the high prevalence of disease in developing countries, where access to imaging technology and resources are limited. In the United States, radiologic studies are used for assistance in evaluating the extent of disease, monitoring treatment response, and detecting recurrent disease. Because most cases of cervical cancer are detected on physical examination or Pap smear, imaging studies play a limited role in the screening of cervical cancer.
Ultrasound
The role of ultrasound in the detection of primary cervical cancer is limited. Transrectal (TRUS) ultrasound permits better visualization of the cervix and detection of cervical malignancies than transabdominal or transvaginal ultrasound. Abnormal findings due to cervical canal obstruction by a mass, such as hematometra, can be detected on ultrasound, but overall its use in evaluating newly diagnosed cervical cancer is marginal.
TVUS or TRUS does not play a significant role in the evaluation of initial disease extent in cervical cancer. Ultrasound may be able to detect parametrial, pelvic sidewall, and bladder involvement by a cervical malignancy. However, the limited ability of ultrasound to appropriately visualize soft tissue planes, and its limited field of view compared with cross-sectional imaging modalities such as CT or MRI, has inhibited any meaningful use of ultrasound in the staging of cervical cancer.
The role of ultrasound in the detection of recurrent disease is obviously limited as well. Ultrasound’s greatest utility in the setting of suspected recurrence is as a guidance modality for tissue biopsies in confirming disease recurrence. Transabdominal ultrasound can also detect hydronephrosis, indicating the need for ureteral stent placement in order to protect renal function. Ultrasound may also assist in the detection and management of complications related to disease status or treatment such as lymphocysts, abscesses, and fistulous tracts.
Computed Tomography
CT has no role in the primary detection of cervical cancer. CT has difficulty in direct tumor visualization and in distinguishing the interfaces between normal tissue and tumor. However, contrast-enhanced CT has had a long and well-established role in evaluating the extent of disease spread in cervical cancer. Although the staging of cervical cancer is still clinical, a staging system by CT has been proposed and correlated with the International Federation of Gynecology and Obstetrics (FIGO) staging system.20 Of note, lymph node involvement is not considered to be part of the FIGO staging system, but detection of pelvic or paraaortic lymph node involvement, with lymph nodes larger than 1 cm in diameter in the short axis, correlates with stage IVB disease when using this system of evaluation with CT. This is important because the detection of lymphadenopathy has been shown both to affect prognosis and to alter potential clinical interventions. The reported accuracy of CT in detecting lymph node invasion is 83% to 85%, with a reported sensitivity ranging from 24% to 70%. CT can also be used to direct biopsies in the case of enlarged or suspicious nodes. Because cervical cancer is clinically staged, this text refers to evaluation of the extent of disease spread, rather than “staging,” because according to FIGO criteria, no gynecologic cancer is staged through imaging modalities alone.
It is important to note that the sensitivity and specificity of CT in evaluating the spread of disease is constrained due to its inability to detect small tumors, invasion of the parametria (76%-80% accuracy), and early invasion of the rectum or bladder. Fifty percent of cervical tumors are isodense to the cervical stroma on CT, rendering the tumor practically invisible. When a cervical tumor is visible, it has a hypodense appearance due to necrosis and diminished vascularity (Figures 3-1 and 3-2). In a study of 172 patients, CT was compared with MRI and FIGO clinical staging in the pretreatment evaluation of invasive cervical cancer. In cases of advanced cervical cancer (stage ≥ IIB), the sensitivity of CT was 42%, specificity was 82%, and negative predictive value (NPV) was 84%.21 Although the accuracy of CT in evaluating disease spread is better than that of clinical examination, it appears to be greatly inferior to MRI. Multidetector CT (MDCT) may improve the diagnostic accuracy of conventional CT due to enhanced spatial and contrast resolution, but further investigation is warranted before MDCT could play a larger role in the staging process.
FIGURE 3-1. Cervical cancer on CT. Stage IB1 adenocarcinoma of the cervix. Enlarged cervical mass noted without invasion into the parametria or sidewall.
FIGURE 3-2. Cervical cancer on CT. Stage IB2 squamous cell carcinoma of the cervix. Enlarged aortocaval lymph node is noted, consistent with metastatic lymph node involvement.
The most important and established role for CT is in the setting of post-treatment surveillance and the detection of recurrent disease. Currently CT is the imaging modality of choice for routine surveillance. Early detection of cervical cancer is essential in order to maximize the potential for salvage therapy such as radiotherapy or pelvic exenteration. The majority of recurrences are in the pelvis, including the vaginal cuff, parametrium, and the pelvic sidewall. Serial CT scans used in disease surveillance have a higher sensitivity and specificity in the detection of recurrent malignancy as compared with primary disease detection. CT also has improved detection and visualization of extrapelvic involvement by cervical cancer, including metastasis to the para-aortic nodes, liver, lungs, and abdomen. CT continues to have limitations in discriminating between postradiation changes, post-surgical fibrosis, and disease recurrence. In cases in which there is a high suspicion for disease recurrence, equivocal findings on CT may require either an MRI or obtaining a tissue specimen by fine-needle aspiration (FNA) or a core biopsy for a pathologic diagnosis of cervical cancer recurrence.
Magnetic Resonance Imaging
MRI is currently considered to be the most accurate imaging modality in the management of cervical cancer, playing key roles in pretreatment assessment, particularly in determining the tumor size and extent of local invasion, detecting lymph node involvement, evaluating response to therapy, and detecting disease recurrence. Although MRI is not routinely used in the primary screening of cervical cancer, it is considered to be the imaging modality of choice performed after the detection of cervical cancer by Pap smear and physical examination due to its superior soft tissue resolution capabilities (Figures 3-3, 3-4, and 3-5). T2-weighted images are particularly useful in the detection of cervical cancer, and it has been noted that the tumor has the appearance of a mass with an intermediate-signal intensity. Contrast-enhanced T1-weighted images can be helpful in detecting bladder or rectal wall invasion. However, non–contrast-enhanced T1-weighted images can pose difficulty in detecting the primary cervical lesion, which can be isointense to the surrounding normal cervix. The addition of gadolinium increases early enhancement of the primary tumor on T1-weighted images, rendering the malignancy more visible.
FIGURE 3-3. Cervical cancer on MRI. Stage IB1 squamous cell carcinoma of the cervix on T2-weighted MRI. Small areas of decreased enhancement are noted in the lower cervical stroma. This mass is confined to the cervical stroma, with no invasion of the parametria.
FIGURE 3-4. Cervical cancer on MRI. Stage IB1 squamous cell carcinoma of the cervical stump after a supracervical hysterectomy. T2-weighted MRI reveals a 3.7 cm heterogeneous enhancing mass. The mass is noted to disrupt the cervical stroma and invade the parametria on imaging.
FIGURE 3-5. Cervical cancer on MRI. Stage IIB squamous cell carcinoma of the cervix on T2-weighted imaging. A 2.5 × 4.5 × 4.3 cm enhancing circumferential lobulated mass is noted to replace the cervix, with parametrial invasion noted on this T2-weighted MRI.
MRI appears to be superior to ultrasound, CT, and clinical examination in the evaluation of disease extent of cervical cancer. Like CT, MRI staging conventions correlate with the FIGO staging system. The overall staging accuracy of MRI is reported to be 77% to 93%, as compared with 69% in CT. However, a recent prospective study conducted by the ACR Imaging Network and the Gynecologic Oncology Group demonstrated equivalence between MRI and CT in the preoperative staging of patients with early invasive cervical cancer. MRI was superior in the detection of parametrial invasion and visualization of the primary tumor. Furthermore, CT demonstrated greater interobserver variability compared with MRI.21
MRI is superior to clinical evaluation in determining the size of the primary tumor and has measurement capabilities comparable to those of surgical measurements. A retrospective study demonstrated size determination by MRI to be within 8 mm of histologic size in 95% of tumors larger than 10 mm. The accuracy of MRI in the detection of lymphadenopathy is reported to range from 76% to 100%, with a sensitivity ranging from 36% to 89.5%. This lower sensitivity is due to the inability of MRI to detect micrometastasis within normal-appearing lymph nodes. Nodes greater than 1.0 mm in the short-axis diameter are considered to be positive, as are findings of central necrosis, extracapsular extension, round shape, and soft tissue in the node with the same signal intensity as the tumor. A study using lymph node–specific contrast agent (ferumoxtran-10, an ultrasmall particles of iron oxide [USPIO]) increased the sensitivity of MRI in detecting lymph node involvement from 29% to 82% to 93% on a node-by-node basis and from 27% to 91% to 100% on a patient-by-patient basis without a loss in specificity. The accuracy of MRI in detecting parametrial involvement has been reported to range from 88% to 97%, with a sensitivity of 44% to 100% and a specificity of 80% to 97%. Finally, use of MRI as a staging modality has been found to be cost-effective because its use precludes the need for further imaging studies, diagnostic tests, or surgical procedures.22
Like CT, MRI plays an important role in the detection of disease recurrence. Recurrent tumor has an intermediate to high signal on T2-weighted images. MRI has several advantages over CT in the setting of post-treatment recurrence because of its ability to distinguish recurrence from post-treatment fibrosis due to surgery and radiation. One year after treatment, post-treatment fibrosis has a low-signal intensity on T1- and T2-weighted images, compared with the intermediate- to high-signal intensity of tumor on T2-weighed images. MRI has a reported sensitivity of 86% and specificity of 94% in detecting recurrent cervical cancer a year after treatment. Dynamic contrast-enhanced MRI is reported to have an accuracy of 85%, as compared with 64% to 68% in non–contrast-enhanced T2-weighted MRI. Functional imaging such as dynamic multiphase contrast-enhanced MRI and diffusion-weighted imaging have demonstrated an increased ability to distinguish recurrence from post-treatment changes even as early as in the first 6 months after treatment and may play an important role in detecting cervical cancer recurrence in the future.23
Positron Emission Tomography
There is no established role for FDG-PET in screening for primary cervical cancer. However, the role of FDG-PET in the staging and detection of recurrent disease has been well studied, and its use has become increasingly widespread, particularly after coverage for the initial staging of cervical cancer was approved by the Centers for Medicare and Medicaid Services. Lymph node involvement on imaging correlates with stage IVB disease, and one of the strengths of FDGPET is its ability to detect lymphadenopathy. Prospective studies have demonstrated FDG-PET/CT to have a sensitivity of 75% to 100% and a specificity of 87% to 100% (Figures 3-6, 3-7, and 3-8). In early cervical cancer, FDG-PET/CT is reported to have a sensitivity of 72%, specificity of 99.7%, and a diagnostic accuracy of 99.3% (Figures 3-9and 3-10). All undetected nodes were smaller than 0.5 cm in diameter; for nodes greater than 0.5 cm, the sensitivity of FDG-PET/CT was 100% and specificity was 99.6%. FDG-PET also improves initial staging because it is also able to detect distant extrapelvic disease, such as metastasis to the supraclavicular nodes. A meta-analysis of 15 studies on the use of FDG-PET in patients with cervical cancer reported a pooled sensitivity of 84% and specificity of 95% in the detection of aortic lymph node involvement and 79% to 99% in the detection of pelvic lymph node involvement. This improvement in staging has been shown to significantly alter clinical and treatment decisions, including changing the therapeutic approach or initiating a treatment that was not previously planned.24
FIGURE 3-6. Cervical cancer on PET/CT. Stage IIIB squamous cell carcinoma of the cervix with marked hypermetabolic uptake of soft tissue lesion in the cervix.
FIGURE 3-7. Cervical cancer on PET/CT. Exophytic, hypermetabolic mass arising from the anterior lip of the cervix and projecting into the upper vagina. The parametrium is normal without any obvious tumor involvement, and there are no pelvic side wall lymph nodes. The uterus is enlarged and contains a gestational sac with a single fetus. The patient was approximately 12 weeks pregnant at the time of this imaging.
FIGURE 3-8. Cervical cancer on PET/CT. Right upper pelvic retroperitoneal mass noted with hypermetabolic uptake. Also noted is a hypermetabolic periaortic lymph node slightly caudal to the level of the renal veins, consistent with nodal metastasis.
FDG-PET can also be used to monitor disease response to treatment and detect recurrent disease. A demonstration of a complete metabolic response post-treatment has been positively associated with survival in both prospective and retrospective studies. In a recent prospective study of patients treated with chemoradiation, the 3-year survival rate was 78% for patients who had a complete metabolic response on FDG-PET. The 3-year survival rate was 33% for a partial metabolic response and 0% for patient who had documented disease progression.25 FDG-PET has a documented sensitivity of 80% for the detection of recurrent disease in asymptomatic women, as compared with a sensitivity of 100% in symptomatic women. A recent study evaluating the utility of FDG-PET in the detection of recurrence reported a sensitivity of 96%, specificity of 84%, and a diagnostic accuracy of 92%. Revised Response Evaluation Criteria In Solid Tumor (RECIST) that have been created, which include an interpretation of FDG-PET scans, demonstrating the widespread acceptance of FDG-PET in monitoring tumor response.26 Further evaluation and larger prospective studies are still needed to fully elaborate the utility of FDG-PET in surveillance and recurrence detection.
Endometrial Cancer
Endometrial cancer is the most common gynecologic malignancy affecting women in the developed world. There are an estimated 43,470 new cases of cancer of the uterus and 7950 deaths expected in 2010. Most cancers of the uterus are detected at an early stage due to abnormal or postmenopausal bleeding, and when detected early, the overall survival rate is quite high. The symptoms of abnormal bleeding usually lead to a diagnosis by either endometrial biopsy or dilation and curettage (D&C). Often ultrasound is used in the initial assessment of patients with abnormal uterine bleeding as a means of evaluating the thickness of the endometrial lining, providing adjunctive information about the uterine anatomy. CT, MRI, and FDG-PET/CT have a very limited role in the screening of primary endometrial cancer and are generally used to assist in evaluating disease extent, monitoring treatment response, or in the detection of recurrent disease.
Ultrasound
Transvaginal ultrasound (TVUS) is the most common imaging modality ordered in the setting of peri- and postmenopausal bleeding. TVUS evaluation is invaluable in ruling out other possible causes of abnormal bleeding such as fibroids and polyps. TVUS also allows evaluation of the endometrial lining, which in a post-menopausal woman should be less than 5 mm in thickness. When a normal endometrial stripe is documented on TVUS, the risk of there being an underlying endometrial cancer ranges from 1% to 5.5%. A finding of an endometrial stripe with a thickness greater than 5 mm often leads to the performance of an endometrial biopsy (EMB) or D&C. However, in an asymptomatic patient, clinical judgment must be exercised, and thus clinical judgment must be exercised in evaluating such findings. Other studies have concluded that TVUS is much more sensitive in the detection of endometrial cancers and in triaging women into a subgroup of low-risk patients who do not require invasive follow-up studies such as EMB or D&C. However, the aggressive histologies of endometrial cancer (serous and clear cell) arise from atrophic endometrium and may have normal endome-trial thickness necessitating histologic evaluation.
Saline infusion sonohysterography (SHG) appears to improve the diagnostic accuracy of TVUS in detecting endometrial cancer. The sensitivity of SHG is 89%, the specificity 46%, the PPV is 16%, and the NPV is 97%. However, there is a potential concern for seeding the peritoneal cavity with malignant cells, which in some cases has been reported in up to 7% of cases. The role of SHG in the evaluation of patients with abnormal uterine bleeding continues to evolve, but it is currently recommended as an adjunctive study when TVUS demonstrates a thickened endometrial lining but an EMB has a negative result for malignancy. Doppler ultrasound of uterine malignancies can demonstrate increased vascularity with vessels originating from multiple sources, in contrast to benign conditions such as polyps, which obtain their blood supply from a single artery. The vascular findings, which can also include an increase in the resistance index, are not considered to be particularly accurate in distinguishing benign from malignant endometrial conditions. Currently, Doppler ultrasound findings do not play any significant role in the primary screening for cancer of the uterus.
After the initial diagnosis of endometrial cancer has been made through a pathologic tissue diagnosis, the role of ultrasound in the further evaluation of endometrial cancer is limited. TVUS is significantly less accurate than MRI in detecting myometrial invasion, with an overall diagnostic accuracy of 60% to 76%. SHG may be slightly more accurate than TVUS in detecting myometrial invasion and cervical involvement by endometrial cancer. However, neither modality has found widespread acceptance or use in the staging or detection of recurrent endometrial cancer.
Computed Tomography
CT has no established role in the primary screening of endometrial cancer, although disease may be found incidentally on CT scans obtained for other indications. Detection of uterine cancer on noncontrast CT is difficult due to endometrial cancer’s similar attenuation to the myometrium on imaging, rendering evaluation of the malignancy rather difficult. The administration of intravenous (IV) contrast improves the ability of CT to detect endometrial cancer because the tumor will demonstrate a low attenuation signal compared with the surrounding myometrium. Uterine cancer can appear as a hypodense mass with a smooth interface identified between the mass and the myometrium (Figure 3-9). Several studies have found that helical CT has improved sensitivity and specificity in the detection of myometrial and cervical invasion compared with conventional CT, although use of helical CT for this indication is not currently widespread.
FIGURE 3-9. Uterine cancer on CT. Carcinosarcoma of the uterus demonstrating a heterogenous soft tissue mass filling the uterine cavity.
Although the staging of endometrial cancer is surgical, CT can play an important role in the evaluation of the extent of disease. Criteria for the staging of endometrial cancer by CT have been proposed and are similar to FIGO classifications. CT is often used to detect myometrial invasion, pelvic sidewall or parametrial involvement, and lymphadenopathy. CT has a reported accuracy of 86% in the detection of extrauterine spread. Lymph nodes greater than 1 cm in the short axis or rounded nodes with a short-axis to long-axis ratio of > 0.8 are considered to have tumor involvement. CT is also extremely accurate in the detection of malignant ascites, peritoneal implants, liver parenchyma involvement, and lung metastasis.
Endometrial cancer tends to recur most commonly at the vaginal cuff in women treated with surgery alone. Serial imaging is generally unnecessary in such patients, because they are subsequently treated with radiation therapy after the detection of disease recurrence. For patients previously treated with radiation, however, recurrence happens more commonly along the pelvic side wall or distantly. CT can play an important role in the early detection of distant recurrence, and because CT is commonly found in most medical centers, and is relatively inexpensive compared with MRI or FDG-PET/CT, it is the primary imaging modality used in the evaluation of recurrent endometrial cancer.
MRI
MRI does not have an accepted role in the initial evaluation of patients with abnormal uterine bleeding. Because of its lower cost, ultrasound is the initial imaging modality of choice. However, MRI is considered to be the most accurate imaging modality in the staging of endometrial cancer due to its excellent resolution of the soft tissues of the pelvis (Figure 3-10). Although surgical staging is the gold standard in endometrial cancer, MRI can be used to evaluate the extent of disease in patients who are poor surgical candidates, in cases of type 2 endome-trial cancer in which there is a high probability of nodal involvement, and in cases of suspected cervical involvement. Compared with TVUS or CT, it appears that MRI is currently the best imaging modality for the evaluation of cervical or parametrial extension of disease that warrants the performance of a radical hysterectomy. Several studies have demonstrated that contrast MRI has an overall staging accuracy higher than that of CT and TVUS. MRI has an 87% sensitivity and 91% specificity in evaluating myometrial invasion, an 80% sensitivity and 96% specificity in determining cervical involvement, and a 50% sensitivity and 95% specificity in determining lymph node involvement by endometrial cancer.
FIGURE 3-10. Uterine cancer on MRI. Stage IC endometrial adenocarcinoma of the uterus. Heterogenous thickened endometrium with partial enhancement on T2-weighted imaging.
Like CT, MRI can also be used in the evaluation of recurrent endometrial cancer. After surgery alone, without adjuvant radiation therapy, vaginal recurrence is the most frequent site of disease recurrence. Vaginal recurrence appears as a high-signal-intensity mass on T2-weighted images. MRI can accurately determine vaginal tumor size and extravaginal disease extent, which correlates with a poor prognosis.27
FDG-PET/CT
The role of FDG-PET/CT in endometrial cancer screening, evaluation of disease extent or monitoring for recurrence, has not been established. Case reports of PET scans obtained in the evaluation of other cancers have incidentally detected uptake in uterus, which has led to the diagnosis of endometrial cancer. However, endometrial FDG uptake can be physiologic and is increased in the proliferative phase of the menstrual cycle. Despite the possibility of physiologic uptake, incidental abnormal uterine uptake warrants histologic evaluation. In evaluating disease extent, FDG-PET was compared with MRI in evaluating the depth of myometrial invasion in 22 patients with stage I disease, but neither was clearly superior. A recent study demonstrated that FDG-PET/CT had a similar sensitivity, specificity, and accuracy to MRI in the detection of primary disease, nodal involvement, and distant metastasis (Figure 3-11). Preoperative evaluation with FDG-PET/CT compared with MRI in patients who were subsequently surgically staged demonstrated a sensitivity of 90% for PET/CT compared with 92% for MRI; a specificity of 51% compared with 33%, respectively; and an accuracy of 85% for both imaging modalities. PET/CT had a nonsignificant trend in the increased detection of nodal metastasis over MRI and had 100% sensitivity and 93% specificity in detecting metastatic disease. The authors concluded that the primary benefit of PET/CT is in the detection of meta-static disease, although this imaging modality may play a future role in evaluating lymph node involvement in patients who are poor surgical candidates.28
FIGURE 3-11. Uterine cancer on PET/CT. Stage IIIC endometrial adenocarcinoma with a soft tissue density noted in the distal portion of the right rectus muscle, consistent with metastatic disease.
There currently is no accepted role for FDG-PET in the preoperative evaluation of endometrial cancer. Its utility in the detection of recurrent disease is also an area of current investigation. A recent study by Chung29retrospectively evaluated 31 patients who had FDG-PET/CT performed for evaluation of recurrent endometrial cancer. They described the sensitivity of imaging to be 100%, with a specificity of 95% and an accuracy of 93%. These findings appeared to be similar in the detection of recurrent uterine sarcomas as well. A study demonstrated a sensitivity of 93%, specificity of 100%, and an accuracy of 95% for surveillance PET in evaluating symptomatic patients with a diagnosis of sarcoma, and a sensitivity of 88%, a specificity of 96%, and an accuracy of 93% in asymptomatic patients with a history of sarcoma. Evaluation with PET led to changes in clinical decision making in 33% of the patients in the study.30
Gestational Trophoblastic Disease
Ultrasound
Gestational trophoblastic disease (GTD) comprises a spectrum of rare placental abnormalities that includes hydatidiform mole, invasive mole, choriocarcinoma, placental site trophoblastic tumor, and epithelioid trophoblastic tumor. The primary lesion, either a complete or incomplete mole, is usually detected by ultrasound, and ultrasound remains the imaging modality of choice for detection of primary lesions. Ultrasound also plays an important role in the subsequent monitoring of patients for persistent or recurrent disease. Molar pregnancies can be subdivided into complete and incomplete moles based on their genetic composition. Complete moles have a classical vesicular appearance on ultrasound described as a “snowstorm” appearance due to swelling of the chorionic villi. Fetal parts are largely absent. Detection of complete moles by ultrasound is best accomplished in the late first trimester or early second trimester, when the detection rate can be as high as 80%. However, improvements in TVUS have led to enhanced detection of molar pregnancies in the early first trimester as well.
Detection of partial moles by ultrasound is more difficult than detection of complete moles due to their less classic appearance on imaging. A recent retrospective study suggested that although the sensitivity and PPV of ultrasound in detecting complete moles was 95% and 40%, respectively, it was only 20% and 22%, respectively, for partial moles.31 Occasionally a fetus with multiple congenital anomalies may be present, along with a placenta that may be enlarged and contain hydropic foci, but often the scan by itself is nondiagnostic.
The presence of bilateral theca-lutein cysts is another classic ultrasound finding in GTD that is detected on ultrasound. However, because of improved ultrasound technology and the earlier diagnosis of GTN, enlarged theca-lutein cysts are becoming rare. Ultrasound is used to follow patients who are suspected of having persistent GTD, as well as in the evaluation of choriocarcinoma or placental site trophoblastic tumor, and to look for retained trophoblastic tissue or local spread to the pelvis.
Computed Tomography
CT does not have a role in the primary evaluation of patients with suspected GTN (Figure 3-12). Its major use is in the detection of metastatic disease. Metastasis to the lungs is the most common site of spread in GTN due to embolization of trophoblastic tissue, and chest imaging is a standard component of a meta-static work-up. Chest imaging, either by chest x-ray or by CT, is required to rule out metastasis to the lungs. Although chest CT is more sensitive than chest x-ray in the detection of pulmonary metastasis, it is not predictive of outcomes. FIGO currently still recommends obtaining a chest x-ray, rather than a chest CT, in the evaluation of metastatic disease.32 GTN can also metastasize to the liver, and metastases to the liver are often multiple, hypointense, and hemorrhagic. Metastasis to the brain may present as single or multiple lesions, which are often hemorrhagic, and appear hyperintense on noncontrast CT scans. It is our practice to obtain CT scans of the head in patients with pulmonary metastasis.
FIGURE 3-12. Gestational trophoblastic disease on CT. Enhancing heterogenous material noted in an enlarged uterus found to be a molar pregnancy on D&C. The imaging also demonstrates an exophytic lesion, consistent with a fibroid uterus.
Magnetic Resonance Imaging
MRI does not have a role in the evaluation and management of suspected GTN, and often MRI findings in primary disease can be nonspecific, making it difficult to distinguish GTN from an incomplete miscarriage. On MRI, GTN obliterates the normal uterine zonal anatomy, and tumors have a highly vascular appearance on T1- and T2-weighted images. Engorgement of the internal iliac vessels and increased tumor enhancement is noted with the use of gadolinium contrast. MRI can also detect extrapelvic extension, including spread of disease to the adnexa, parametria, and the vagina. It is the preferred imaging modality for the evaluation of metastatic disease to the vagina. Finally, patients treated with chemotherapy will have a noted decrease in uterine volume and tumor vascularity that is detectable with MRI. With treatment, normal uterine zonal anatomy will reappear on T2-weighted images, and intra-lesional hemorrhage may develop. By 6 to 9 months after treatment, the uterus should appear normal on both T1- and T2-weighted images, although persistent adnexal cysts may be evident.
Positron Emission Tomography
The data on the role of PET in the setting of GTN is limited to case reports and a few case series. FDG-PET may be useful in the diagnosis of metastatic disease, particularly when conventional imaging modalities fail to identify an occult malignancy. FDG-PET/CT was reported in one study to be more beneficial in 7 of 16 patients (43.8%) compared with CT alone in terms of detecting chemotherapy-resistant lesions, excluding false-positive lesions found on CT, and confirming complete responses to chemotherapy.32 However, 2 scans were falsely negative, 1 scan was indeterminate, and 6 scans were of no benefit. Larger studies are required before FDG-PET can play a meaningful clinical role in the management of GTN.
Ovarian Cancer
Although ovarian cancer is the second most common gynecologic malignancy in the United States, it has the highest mortality rate of all gynecologic cancers. This is due to the fact that most ovarian cancers present at an advanced stage, usually stage III or stage IV. Early detection of ovarian cancer has continued to elude clinicians due to the current absence of any reliable screening modalities. The role of imaging in ovarian cancer is largely one of detection, staging, preoperative planning, and post-treatment surveillance.
Ultrasound
TVUS is the imaging modality of choice in the initial evaluation of a pelvic mass. Transabdominal ultrasound can play a limited role, especially in the evaluation of large masses that are displaced outside the pelvis, as well as in the documentation of ascites and hydronephrosis. Classic morphologic findings in pelvic masses that are concerning for malignancy include thick septations (> 3 mm), internal papillations, loculations, solid masses, cystic lesions with solid components, and smaller cysts incorporated into the structure of larger cysts. These morphologic descriptors have a high sensitivity for malignancy, but a low specificity because some of these findings are also found in benign lesions and borderline tumors as well as invasive malignancy. Multiple studies have shown that when strict morphologic criteria are used in the evaluation of an adnexal mass, the PPV of ultrasound is 95% and the NPV is 99% in excluding malignancy.
The addition of color Doppler to the traditional ultrasound morphologic assessment may increase the diagnostic capability of ultrasound in evaluating an adnexal mass, but the benefit of the addition of duplex scanning remains controversial. The neo-vascularization associated with ovarian malignancy has led to the evaluation of the blood flow and vascular structures that support an adnexal mass. Several studies have demonstrated that malignancies often have an increase in vessel density and tortuosity, as well as low-resistance waveforms due to the absence of smooth muscle in the newly formed vascular support structures. It was initially thought that documentation of arteriovenous shunting and finding a resistance index cutoff of 0.4 and a pulsatility index of 1 were highly sensitive and specific for malignancy. However, such waveform findings have been frequently noted in benign conditions as well, especially in premenopausal women. However, a finding of low-resistance blood flow in a postmenopausal woman must always be viewed with suspicion. A combination of morphologic assessment with Doppler ultrasound may lead to an increase in the ability to discriminate between benign and malignant lesions. The addition of color Doppler to conventional ultrasound appeared to improve the specificity from 82% to 92% and the PPV from 63% to 97%. Doppler evaluation can provide adjunctive information to help classify a lesion as malignant rather than benign.
Because of the low prevalence of ovarian cancer in the general population, large-scale screening programs have proven to be unsuccessful and not cost-effective. The focus of most screening programs have been on women at high risk for developing ovarian cancer, especially those with a family history or genetic predisposition (BRCA1 and BRCA2, HNPCC). Besides the low prevalence, another difficulty in creating an effective screening program is that ovarian cancer may have rapid growth and early spread outside of the ovary, making detection of early disease extremely difficult. A recent prospective study enrolled 25,327 women to receive an annual screening TVUS. The women enrolled were either asymptomatic women older than 50 years or women older than 25 years with a family history of ovarian cancer. The PPV for ultrasound was 27% and the sensitivity was 85%, and for those women detected by screening, the 5-year-overall survival was 77% compared with 49% for historical controls from the same institution.33 Several issues exist with this study, including its lack of randomization, lack of a control group, and inclusion of many high-risk women, which argues that the PPV would presumably be lower in a group of average-risk women. Additionally, the percentage of expected BRCA-mutated patients is likely increased among high-risk women, and the improved survival of BRCA-mutated ovarian cancer patients is well known.
Screening protocols have been developed that attempt to combine a number of modalities, including ultrasound, serologic markers (CA-125), and physical examination. Two recent large trials have been conducted evaluating the impact of ovarian cancer screening on survival. The PLCO Cancer Screening Trial enrolled 34,261 healthy women with an age range of 55 to 74 years and randomly assigned them to have an annual CA-125 plus TVUS or “usual care.” During the 4 years of screening, the PPV of a positive screening test was 1% to 1.3% for an abnormal TVUS, 3.7% for an abnormal CA-125, and 23.5% if both CA-125 and TVUS were abnormal.34 The United Kingdom Collaborative Trial of Ovarian Cancer Screening enrolled 202,638 postmenopausal women with an age range of 50 to 74 years at an average risk for ovarian cancer.35 These women were randomly assigned to a control group that received a pelvic examination only, an annual TVUS, or an annual CA-125 plus TVUS if an abnormal CA-125 was detected. The multimodality group had a significantly greater specificity (99.8%) and PPV (35.1%) compared with ultrasound alone (specificity of 98.2% and PPV of 2.8%). The sensitivity did not differ between the 2 groups. The effect of screening on mortality has not yet been reported.
In the United States, ultrasound is not typically used in the evaluation of disease extent of ovarian cancer. Transabdominal ultrasound is useful in the detection of ascites and in performing ultrasound-guided paracentesis. Ultrasound also has a limited role in the detection of recurrent ascites. Ultrasound may be used in suspected recurrent disease, assisting interventional radiologists in performing imaging-guided biopsies to obtain a tissue for cytologic diagnosis.
Computed Tomography
CT does not have a role in the primary screening of ovarian cancer, although incidental detection of adnexal malignancies is not an infrequent event. Ultrasound is considered to be the initial study of choice in the evaluation of adnexal masses, but CT can provide additional information when the initial ultrasound evaluation is indeterminate. One study demonstrated that in the case of an indeterminate ultrasound, CT had a sensitivity of 81% and a specificity of 87% in predicting a diagnosis of ovarian cancer. Advanced stage ovarian cancer on contrast-enhanced CT demonstrates cystic lesions that may have irregular borders, thickened walls, papillary projections, calcifications, and septations (Figure 3-13). Peritoneal implants, carcinomatosis, ascites, and pelvic organ and sidewall involvement can also be described on CT. The clinician must always have a high suspicion for metastatic tumor involvement from the gastrointestinal tract or the breast when evaluating CT scans demonstrating carcinomatosis, as these tumors may have a similar radiographic appearance.
FIGURE 3-13. Ovarian cancer on CT. 13.9 × 9.3 × 10.9 cm fluid-filled mass found to be consistent with a high-grade ovarian cancer.
Ovarian cancer is surgically staged, but CT can also assist in the process. CT can provide evidence of metastatic spread, involvement of abdominal and pelvic organs, and lymphadenopathy and permit some determination of the likelihood of performing optimal surgical cytoreduction. Numerous studies have suggested that CT findings can predict suboptimal cytoreduction (Table 3-1). However the predictive value of CT varies by institution due to variations in surgical practice and aggressiveness.36
Table 3-1 Tumor Markers Useful in Germ Cell Tumors of the Ovary
Lesions located in the mesentery, porta hepatis, liver, pelvic sidewall, and lymphadenopathy superior to the celiac axis may not be optimally resectable. Patients who have disease that may not be amenable to optimal debulking are often treated with neoadjuvant chemotherapy.
Although CT scans are routinely used in clinical trials in post-treatment surveillance, the optimal tumor marker and imaging algorithm is not established. According to the 2009 National Comprehensive Cancer Network guidelines, CT scans can be used as a surveillance tool whenever clinical indications such as patient symptoms or a rising CA-125 level warrant. CT has an estimated sensitivity of 59% to 83% and a specificity of 83% to 88% in detecting residual tumor and recurrence. Recurrence often presents as a pelvic mass, malignant ascites, peritoneal implants, carcinomatosis, and lymphadenopathy. It is important to note that CT has difficulty with visualization of lesions smaller than 1cm in size, and CT sensitivity decreases to 25% to 50% in such cases. Thus the clinician should consider the possibility of a false-negative CT scan in cases of symptomatic patients or in those patients with rising CA-125 levels. FDG-PET/CT may aid in evaluating such situations.37
Magnetic Resonance Imaging
MRI does not have a role in the primary screening of ovarian cancer. However, contrast-enhanced MRI does appear to have potential utility in the evaluation of adnexal masses that produced an indeterminate ultrasound study (Figure 3-14). In a meta-analysis examining patients with an indeterminate TVUS study, MRI with gadolinium contrast had the greatest accuracy of diagnosing ovarian cancer as compared with CT, Doppler ultrasound, or MRI without gadolinium contrast. The sensitivity of contrast-enhanced MRI was 81%, and the specificity was 98%. A prospective study of contrast-enhanced MRI used as a second imaging modality after an indeterminate ultrasound study of a pelvic mass demonstrated a sensitivity of 100% and specificity of 94%.
FIGURE 3-14. Ovarian cancer on MRI. Enlarged, complex, septated, adnexal mass on T2-weighted imaging.
Surgical staging is considered the gold standard for management of early-stage ovarian cancer. Like CT, MRI can be used for pretreatment planning and assistance with determining whether optimal debulking is feasible. MRI can also be used in suboptimally debulked patients to evaluate multisite disease or areas of unresectable tumor. The accuracy of MRI in staging ovarian cancer, which is approximately 70% to 90%, is comparable to that of CT.
The use of MRI for post-treatment surveillance is not considered routine in most medical centers. MRI offers certain advantages over CT; it is safe to use in patients with iodine allergies, and MRI is better able to visualize disease recurrence in the vaginal vault, cul-de-sac, and the bladder base.
Positron Emission Tomography
FDG-PET currently does not have an accepted role in the primary evaluation of pelvic masses due to its expense. However, the ability of FDG-PET to detect early metabolic changes in tumors before the development of structural changes offers a promise for earlier detection of malignancy. A prospective study evaluated 101 patients with adnexal masses and a Risk of Malignancy Index > 150 based on CA-125, ultrasound, and menopausal status who were referred for evaluation by FDG-PET/CT. The authors found that the sensitivity of FDG-PET/CT in the diagnosis of ovarian cancer was 100% and the specificity was 93%. They concluded that PET/CT should be considered an imaging modality of choice after a finding of a pelvic mass by ultrasound.38
Again, due to its cost, there is currently no established role for FDG-PET/CT in the initial evaluation of disease extent in ovarian cancer. However, a recent study demonstrated that the addition of FDG-PET to CT appears to increase the staging accuracy of CT from 89.7% to 94%, the sensitivity from 37.6% to 69.4%, and the specificity from 97.1% to 97.5% (Figures 3-15 and 3-16). Despite the more common use of CT scans in detecting recurrent disease, it appears that FDG-PET/CT may be the most accurate imaging technique. A study by Nam et al39 found that preoperative PET/CT was superior to pelvic ultra-sound, CT, and pelvic MRI in the diagnosis of ovarian cancer and in the detection of metastatic disease. A meta-analysis comparing FDG-PET/CT with CT and MRI in the detection of recurrent disease found that FDG-PET/CT was more accurate than CT or MRI. The sensitivity of FDG-PET/CT was 91% and the specificity was 88%, whereas the sensitivity and specificity of CT was 79% and 84% and for MRI was 75% and 78%, respectively.40 Although there appears to be evidence that FDG-PET/CT may be comparable, if not superior, to CT and MRI in the detection of recurrent ovarian cancer, its routine use is not currently standard of care.
FIGURE 3-15. Ovarian cancer on PET/CT. Stage IIIC papillary serous ovarian cancer after multiple courses of chemotherapy. Patient noted to have hypermetabolic uptake in 2 pelvic lesions, consistent with recurrent disease.
FIGURE 3-16. Ovarian cancer on PET/CT. Stage IIIC papillary serous ovarian cancer. Hypermetabolic nodule measuring 1.1 × 0.9 cm noted in the mesentery, consistent with metastatic disease.
Vulvar Cancer
Due to their rarity in Western populations, large prospective studies investigating the utility of various imaging modalities in vulvar and vaginal cancer are nonexistent. Vulvar cancer, with an incidence of 5% of gynecologic cancers, was by convention clinically staged until FIGO adopted a surgical staging system in 1988. Because of the emphasis placed by the FIGO system on the surgical-pathologic approach to staging vulvar cancer, imaging modalities have traditionally had a minimal role in the detection and evaluation of primary malignancies. Barium enemas, intravenous pyelograms, and, later, CT have been the traditional tools used to assist in management of especially large vulvar tumors that may have extended into adjacent organs, in cases of suspected metastatic disease, and in select cases of recurrent disease.
Ultrasound
The role of ultrasound in the detection and management of vulvar cancer has been extremely limited to date. A recent study demonstrated that ultrasound combined with fine-needle aspiration (FNA) was superior to CT in the detection of groin node metastasis. Ultrasound combined with FNA had a sensitivity of 80%, specificity of 100%, PPV of 93%, and NPV of 100%.
Magnetic Resonance Imaging
MRI has been shown to have some value in differentiating vulvar cancer recurrence from postradiation changes. In the setting of primary and recurrent vulvar cancer, MRI accurately determined the size of the vulvar lesion in 83% of patients.41 Accuracy in staging of primary vulvar cancers was only 69.4%, and detection of groin lymph node metastasis was between 85% and 87%. The authors concluded that MRI might have a useful role as an adjunct in determining the size of a vulvar lesion and in detecting the presence of lymph node metastasis.
Positron Emission Tomography
The role of FDG-PET in vulvar cancer has yet to be fully elucidated, but several small, early studies have indicated some promise. On a per-patient basis FDGPET had a sensitivity of 80%, specificity of 90%, PPV of 80%, and a NPV of 90% for detecting groin node metastases. On a groin-to-groin basis, the sensitivity was 67%, specificity 95%, PPV 86%, and NPV 86%. FDG-PET was more accurate in the detection of extranodal disease than disease found only in the nodes. The authors concluded that PET was relatively insensitive in the detection of vulvar cancer metastatic to the inguinal lymph nodes and was not an adequate substitute for traditional lymph node dissection. However, FDG-PET may have an as yet undefined role in aiding radiation planning or as an adjunct to sentinel lymph node dissection. More research into the application and utility of FDT-PET in vulvar cancer is indicated based on the initial promise of this study.
Vaginal Cancer
Magnetic Resonance Imaging
Vaginal cancer has an incidence of 3% of gynecologic malignancies and is clinically staged. Both CT and MRI can be used for the detection of pelvic adenopathy and metastatic disease. Additionally, MRI provides information regarding tumor size and invasion of adjacent organs. This is important for radiation treatment planning and evaluating response to therapy. There appears to be no role for ultrasound in the setting of vaginal cancer.
Positron Emission Tomography
There are few studies examining the use of FDG-PET in the diagnosis and management of vaginal cancer. FDG-PET detected 100% of the primary vaginal lesions as compared with 43% detected by CT. FDG-PET detected metastasis to the groin and pelvic lymph nodes in 35% of patients as compared with 17% detected by CT. Although FDG-PET was superior to CT in the detection of vaginal lesions and metastasis, this was not correlated with a survival advantage. Further research into the utility of FDG-PET in the setting of vaginal cancer is warranted.
FUTURE DIRECTIONS
Gynecologic malignancies are clinically or surgically staged by established (periodically revised) systems. Although diagnostic imaging can play a role in the pretreatment determination of the extent of disease, “staging” by imaging is not considered to be standard of care. Ultrasound, as the least expensive imaging modality in gynecology, continues to play an important role in the detection and diagnosis of primary diseases, in particular endometrial and ovarian cancers, as well as in the management and surveillance of gestational trophoblastic diseases. CT continues to play the predominant role in the evaluation of most gynecologic malignancies, including determining the extent of metastatic disease and detection of recurrent disease. Newer imaging techniques, such as helical CT, have improved both the speed and imaging resolution of the traditional CT scanner. Concern has been raised in recent years about the level of radiation exposure to patients, particularly cancer patients who are subjected to repetitive CT scans as part of their post-treatment surveillance.
Due to the lack of ionizing radiation, its superior soft tissue resolution, and its multiplanar imaging capabilities, the use of MRI in the evaluation and management of pelvic malignancies continues to increase. Cost, slower scanning times, and patient anxiety during scans were some of the initial barriers to the more widespread use of MRI and are starting to be overcome. In particular, MRI appears to be superior in staging endometrial and cervical cancers as compared with CT and TVUS. Although CT and MRI both provide important information about nodal and distant metastasis, MRI provides more information about local disease extent, whereas CT provides more information about distant spread of disease. The addition of PET to CT provides information about biologic activity with the ability to provide anatomic localization, providing an additional component of information beyond simple anatomic dimensions. Further investigation with larger prospective trials in all pelvic malignancies is required to fully elaborate the role of FDG-PET, but the initial findings in a number of clinical settings—primary detection, staging, surveillance, and recurrence—have been extremely promising. Despite improvements in diagnostic imaging, serum tumor markers remain a safe and cost-effective way to monitor disease status and aid in determining when further diagnostic modalities may be indicated.
As we move into a more structured practice environment, issues of cost and cost-effectiveness will become ever more important considerations for the clinician. One of the ways in which cost-effectiveness will factor into the practice of gynecologic oncology will be in selecting the most accurate diagnostic modalities for evaluating and following patients. Although CT is the most commonly used imaging modality for diagnosing and following patients with gynecologic cancers, it has a number of limitations, including its reliance on ionizing radiation and reduced sensitivity and specificity in the detection of lymph node metastasis and early recurrent disease. MRI and PET-CT can compensate for some of the deficiencies inherent in CT technology, but their current expense and lack of coverage by most insurance companies limit their routine use except in select circumstances. Further prospective studies will be needed to determine how best to leverage these imaging modalities in the interest of improved patient care. In particular, MRI and PET-CT herald the advent of functional imaging and the promise of earlier detection of primary disease as well as recurrence. Such early detection may lead to a significant impact on disease incidence and mortality if earlier detection can be coupled with more effective treatment. Additional comparative studies will be necessary to determine the most accurate and cost-effective diagnostic algorithm if such potential benefits are to be realized.
REFERENCES
1. Smith-Bindman R, Lipson J, Marcus R, et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch intern Med.2009;169(22):2078-2086.
2. Kanal E, Barkovich AJ, Bell C, et al. ACR guidance document for safe MR practices: 2007. AJR Am J Roentgenol. 2007; 188(6):1447-1474.
3. Juweid ME, Cheson BD. Positron-emission tomography and assessment of cancer therapy. N Engl J Med. 2006;354(5):496-507.
4. Hewitt MJ, Anderson K, Hall GD, et al. Women with peritoneal carcinomatosis of unknown origin: efficacy of image-guided biopsy to determine site-specific diagnosis. BJOG. 2007;114:46-50.
5. Ledwaba T, Dlamini Z, Naicker S, et al. Molecular genetics of human cervical cancer: role of papillomavirus and the apoptotic cascade. Biol Chem. 2004;385(8):671-682.
6. van der Avoort IA, Shirango H, Hoevanaars BM, et al. Vulvar squamous cell carcinoma is a multifactorial disease following two separate and independent pathways. Int J Gynecol Pathol. 2006;25(1):22-29.
7. Liegl B, Leibel S, Gogg-Kamerer M, et al. Mammary and extra-mammary Paget’s disease: an immunohistochemical study of 83 cases. Histopathology. 2007;50(4):439-447.
8. Mutter GL, Lin M-C, Fitzgerald JT, et al. Altered PTEN expression as a diagnostic marker for the earliest endometrial precancers. J Natl Cancer Inst. 2000;92(11):924-930.
9. Singer G, Stohr R, Cope L, et al. Patterns of p53 mutations separate ovarian serous borderline tumors and low- and high-grade carcinomas and provide support for a new model of ovarian carcinogenesis: a mutational analysis with immunohistochemical correlation. Am J Surg Pathol. 2005;29(2):218-224.
10. Hirakawa M, Nagai Y, Inamine M, et al. Predictive factor of distant recurrence in locally advanced squamous cell carcinoma of the cervix treated with concurrent chemoradiotherapy. Gynecol Oncol.2008;108(1):126-129.
11. Kotowicz B, Kaminska J, Fukseiwicz M, et al. Clinical significance of serum CA-125 and soluble tumor necrosis factor receptor type I in cervical adenocarcinoma patients. Int J Gynecol Cancer.2010;20(4):588-592.
12. Dube V, Grigull J, DeSouza LV, et al. Verification of endometrial tissue biomarkers previously discovered using mass spectrometry-based proteomics by means of immunohistochemistry in a tissue microarray format. J Proteome Res. 2007;6(7):2648-2655.
13. Riedinger JM, Eche N, Basuyau JP, et al. Prognostic value of serum CA 125 bi-exponential decrease during first line paclitaxel/platinum chemotherapy: a French multicentre study. Gynecol Oncol.2008;109(2):194-198.
14. Rustin GJ, van der Burg ME. A randomized trial in ovarian cancer (OC) of early treatment of relapse based on CA-125 levels alone versus delayed treatment based on conventional clinical indications: MRC OV05/EORTC 55955 trials. Lancet 2010;276(9747):1155-1163.
15. Rustin GJ, van der Burg ME, Griffin CL. Early versus delayed treatment of relapsed ovarian cancer (MRC OV05/EORTC 55955): a randomised trial. Lancet. 2010;376:1155-1163.
16. Allard WJ, Somers E, Theil R, Moore RG. Use of a novel biomarker HE4 for monitoring patients with epithelial ovarian cancer. J Clin Oncol. 2009;26(suppl). Abstract 5533.
17. Partridge E, Kreimer AR, Greenlee RT, et al. Results from four rounds of ovarian cancer screening in a randomized trial. Obstet Gynecol. 2009;113(4):775-782.
18. Menon U, Skates SJ, Lewis S, et al. Prospective study using the risk of ovarian cancer algorithm to screen for ovarian cancer. J Clin Oncol. 2005;23:7919-7926.
19. Mom CH, Engelen MJ, Willemse PH, et al. Granulosa cell tumors of the ovary: the clinical value of serum inhibin A and B levels in a large single center cohort. Gynecol Oncol. 2007;105(2):365-372.
20. Hancke K, Heilmann V, Straka P, et al. Pretreatment staging of cervical cancer: is imaging better than palpation? Role of CT and MRI in preoperative staging of cervical cancer: single institution results for 255 patients. Ann Surg Oncol. 2008;15(10):2856-2861.
21. Hricak H, Gatsonis C, Coakley FV, et al. Early invasive cervical cancer: CT and MR imaging in preoperative evaluation-ACRIN/GOG comparative study of diagnostic performance and interobserver variability. Radiology.2007;245(2):491-498.
22. Rockall AG, Ghosh S, Alexander-Sefre F, et al. Can MRI rule out bladder and rectal invasion in cervical cancer to help select patients for limited EUA? Gynecol Oncol. 2006;101(2):244-249.
23. Padhani AR, Liu G, Koh DM, et al. Diffusion weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia. 2009;11(2):102-125.
24. Chao A, Ho KC, Wang CC, et al. Positron emission tomography in evaluating the feasibility of curative intent in cervical cancer patients with limited distant lymph node metastases. Gynecol Oncol.2008;110(2):172-178.
25. Schwarz JK, Siegel BA, Dehdashti F, Grigsby PW. Association of posttherapy positron emission tomography with tumor response and survival in cervical carcinoma. JAMA. 2007;298:2289-2295.
26. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45:228-247.
27. Sohaib SA, Houghton SL, Meroni R, et al. Recurrent endometrial cancer: patterns of recurrent disease and assessment of prognosis. Clin Radiol. 2007;62:28-34.
28. Park JY, Kim EN, Kim DY, et al. Comparison of the validity of magnetic resonance imaging and positron emission tomography/computed tomography in the preoperative evaluation of patients with uterine corpus cancer. Gynecol Oncol. 2008;108:486-492.
29. Chung HH, Kang WJ, Kim JW, et al. The clinical impact of ((18)F)FDG PET/CT for the management of recurrent endometrial cancer: correlation with clinical and histological findings. Eur J Nucl Med Mol Imaging.2008;35:1018-1088.
30. Par JY, Kim NE, Kim DY, et al. Role of PET or PET/CT in the post-therapy surveillance of uterine sarcoma. Gynecol Oncol. 2008;109:255-262.
31. Kirk E, Papageorghiou AT, Condous G, et al. The accuracy of first trimester ultrasound in the diagnosis of hydatidiform mole. Ultrasound Obstet Gynecol. 2007;29:70-75.
32. Allen SD, Lim AK, Seckl MJ, Blunt DM, Mitchell AW. Radiology of gestational trophoblastic neoplasia. Clin Radiol. 2006; 61:301-313.
33. Van Nagell JR Jr, Depriest PD, Reedy MB, et al. Ovarian cancer screening with annual transvaginal sonography: findings of 25,000 women screened. Cancer. 2007;109:1887-1896.
34. Partridge E, Kreimer AR, Greenlee RT, et al. Results from four rounds of ovarian cancer screening in a randomized trial. Obstet Gynecol. 2009;113:775-782.
35. Menon U, Gentry-Majaraj A, Hallett R, et al. Sensitivity and specificity of multimodal and ultrasound screening for ovarian cancer, and stage distribution of detected cancers: results of the prevalence screen of the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS). Lancet Oncol. 2009;10(4):327-340.
36. Axtell AE, Lee MH, Bristow RE, et al. Multi-institutional reciprocal validation study of computed tomography predictors of suboptimal primary cytoreduction in patients with advanced ovarian cancer. J Clin Oncol. 2007;25:384-389.
37. Fulham MJ, Carter J, Baldey A, et al. The impact of PET-CT in suspected recurrent ovarian cancer: a prospective multi-centre study as part of the Australian PET Data Collection Project. Gynecol Oncol.2009;112(3):462-468.
38. Risum S, Hogdall C, Loft A, et al. The diagnostic value of PET/CT for primary ovarian cancer: a prospective study. Gynecol Oncol. 2007;105:145-149.
39. Nam EJ, Yun MJ, Oh YT, et al. Diagnosis and staging of primary ovarian cancer: correlation between PET/CT, Doppler US, and CT or MRI. Gynecol Oncol. 2010;116:389-394.
40. Gu P, Pan LL, Wu SQ, Sun L, Huang G. CA 125, PET alone, PET-CT, CT and MRI in diagnosing recurrent ovarian carcinoma a systematic review and meta-analysis. Eur J Radiol. 2009;71:164-174.
41. Kataoka MY, Sala E, Baldwin P, et al. The accuracy of magnetic resonance imaging in staging of vulvar cancer: a retrospective multi-centre study. Gynecol Oncol. 2010;117(1):82-87.