The Washington Manual of Oncology, 3 Ed.

Cancer of Unknown Primary Site

Siddartha Devarakonda • Danielle Carpenter • Daniel Morgensztern

  I. DEFINITION. Carcinoma of unknown primary (CUP) is defined as a biopsy-proven metastatic malignant tumor whose primary site cannot be identified during pretreatment evaluation that includes a thorough history and physical examination, standard laboratory and radiographic studies, and a detailed histologic investigation.

1. PRESENTATION
2.  Subjective. Although CUPs comprise a heterogeneous group of tumors with different natural histories, there are still some typical characteristics. Some of the clinical features include a short history of local symptoms related to the metastatic sites (pain, swelling, and cough) and constitutional symptoms (weight loss, fatigue, and fever).
3.  Objective. The physical examination is frequently abnormal with findings such as effusions, lymphadenopathy, and hepatomegaly, indicating the site of metastatic involvement. Patients should undergo a thorough examination of the skin to rule out the presence of melanoma or nonmelanoma tumors, breast, rectum, pelvis, and genitals. The most common sites involved are lymph nodes, liver, bone, lungs, and pleura. Most patients present with multiple metastatic sites because of early dissemination and, unlike known primary tumors, the pattern is usually unpredictable.

 III. DIAGNOSIS. The diagnosis is made by biopsy. Since several studies may need to be performed, it is important to consult with the pathologist to determine whether the specimen is sufficient, as the commonly used fine-needle aspiration contains limited tissue and does not provide information on tissue architecture.

 IV. WORKUP

1.  Initial assessment. With the histologically proven diagnosis of malignancy, patients should undergo a limited clinical investigation to identify the primary site and favorable subsets. This evaluation should include a complete history of physical examination including pelvic and rectal examination, complete blood count (CBC), chemistry profile, urinalysis, occult blood stool testing, chest radiography, computer tomography (CT) of the abdomen and pelvis, and symptom-oriented endoscopy. Subsequent diagnostic tests are based on the clinical presentation, gender, and histopathologic findings. Comprehensive and exhaustive radiographic and endoscopic tests should not be performed because even with extensive workup, the primary site becomes evident in less than 25% of the patients. Up to 80% of primaries can be found in autopsy series, most commonly in the lungs and pancreas.
2.  Imaging. The initial radiological evaluation may be limited to chest radiograph and CT scans of the chest, abdomen, and pelvis. Chest radiograph is usually performed during the initial evaluation, even in the absence of respiratory symptoms, since a large number of patients will eventually have the diagnosis of lung cancer. Contrast radiographic studies have a low yield and should be reserved for patients with findings related to the organ to be examined. CT scan of the abdomen and pelvis may detect the primary site in approximately one-third of patients. It can also be particularly useful in the detection of occult pancreatic carcinomas. Mammogram is indicated in the diagnostic investigation of all women with CUP, particularly in the cases of adenocarcinoma metastatic to axillary lymph nodes. Breast magnetic resonance imaging (MRI) may be indicated in cases where the suspicion for primary breast remains high despite a negative mammogram. The experience with fluorodeoxyglucose positron emission tomography (FDG-PET) scan in CUP has been limited so far, and larger prospective series are needed before its routine use. Some of the problems associated with the use of PET include the high cost, elevated false-positive rate, and lack of improved survival after the identification of the primary tumor. Nevertheless, PET may be particularly useful in patients with squamous cell carcinoma in the cervical lymph nodes, where it may allow the detection of a primary site in the head and neck area in approximately one-third of the cases, and in patients with single metastatic site, where additional metastases may influence the treatment.
3.  Endoscopy. Endoscopy cannot be recommended during the routine workup for patients with CUP in asymptomatic patients. Instead, it should be used according to the clinical presentation. Therefore, ENT endoscopy should be performed in patients with isolated cervical lymph node involvement by squamous cell carcinoma, bronchoscopy in patients with pulmonary symptoms, gastrointestinal endoscopies in patients with abdominal symptoms or occult fecal blood, and proctoscopy or colposcopy in patients with inguinal lymph node involvement.
4.  Pathology
5.  Light microscopy. Routine light microscopy of the tissue specimen after staining with hematoxylin and eosin may identify the five major histologic subtypes of CUP: adenocarcinoma (50% to 60%), poorly differentiated carcinomas or adenocarcinomas (30%), squamous cell carcinomas (5% to 15%), undifferentiated malignant neoplasms (5%), and neuroendocrine carcinomas (1%).
6.  Immunohistochemistry. Immunohistochemistry (IHC) represents the most widely available specialized technique for the classification of neoplasms and may help identify the tumor lineage by the use of peroxidase-labeled antibodies against specific tissue antigens. Immunoperoxidase (IP) can be used on formalin-fixed specimens, which usually makes repeated biopsy unnecessary and may identify several cell components, resulting in the narrowing of diagnostic possibilities (Table 33-1). IHC can aid in determining whether a poorly differentiated neoplasm is a carcinoma, sarcoma, lymphoma, or melanoma, and if a carcinoma is an adenocarcinoma, squamous cell carcinoma, germ cell tumor, or neuroendocrine carcinoma. Cytokeratins (CKs) are a family of intermediate filaments characteristic of carcinomas. The CK profile may be useful in the identification of the primary tumor site, and the ones most commonly used in patients with CUP are CK7 and CK20. CK7 is more commonly present in tumors of the lung, ovary, endometrium, and breast, and absent in lower gastrointestinal tumors. CK20 is expressed in the gastrointestinal and urothelial cells. Therefore, the CK7/CK20 phenotype can be very useful in narrowing the differential for identification of the primary site, particularly for adenocarcinomas (Table 33-2). Patients with CK7+ and TTF-1 positive are likely to have lung cancer, whereas those with CK7−/CK20+/CDX2+ are likely to have lower gastrointestinal cancer.
7.  Electron microscopy. Electron microscopy (EM) allows the visualization of the ultrastructural features of the tumors such as cellular organelles, granules, and cell junctions. It may be useful in the identification of neuroendocrine tumors (neurosecretory granules), melanoma (premelanosomes), and poorly differentiated sarcomas. It may also help in differentiating between lymphoma and carcinoma or adenocarcinoma and squamous cell carcinoma, although it does not localize the primary site of the malignancy. Since EM is expensive, time consuming, and not widely available, its use should be reserved for the cases with unclear lineage after light microscopy and IHC.
8.  Tumor markers. Commonly used serum tumor markers such as CEA, CA 19-9, and CA 125 are of limited value in the diagnosis of patients with CUP. Thyroglobulin may be increased in patients with bone metastases, suggesting an occult thyroid primary. CA 125 may be helpful in women with peritoneal papillary adenocarcinomatosis. A serum beta-human chorionic gonadotropin (β-HCG) and alpha-fetoprotein (AFP) levels in younger males, and prostate-specific antigen (PSA) in older men should be tested to exclude testicular and prostate cancer, respectively.

AFP, α-fetoprotein; CK, cytokeratin; CLA, common leukocyte antigen; EMA, epithelial membrane antigen; ER, estrogen receptor; β-HCG, beta-human chorionic gonadotropin; HMB, human melanoma black; MTC, medullary thyroid carcinoma; NSE, neuron-specific enolase; PLAP, placental alkaline phosphatase; PR, progesterone receptor; PSA, prostate-specific antigen; TTF-1, thyroid transcription factor 1.

5.  Genetics. Genetic analyses of the biopsy specimen may provide further characterization regarding the origin of the malignancy since a large number of tumors display characteristic cytogenetic abnormalities (Table 33-3).
6.  Expression and micro-RNA profiling. Molecular profiling including gene expression and micro-RNA can aid in the identification of the primary site of a CUP, as molecular profiles vary across different cancers and are usually comparable to the profiles of their underlying normal tissue of origin. Molecular profiling has the potential to improve outcomes in patients with CUP, especially when the results allow site-specific therapy for favorable tumor types. Studies have demonstrated good agreement between the sites of origin predicted by molecular profiling and suspected primary sites based on clinicopathological findings. These results were, however, shown to be less accurate when IHC suggested two or more sites. While there is little evidence describing the impact of molecular profiling on outcomes in patients with CUP, molecular profiling can be helpful when clinical and pathological workup fails to reveal the primary site.
7. THERAPY AND PROGNOSIS
8.  Favorable subsets. Following the exclusion of lymphoma and sarcoma by a careful pathologic evaluation, the vast majority of the patients will have the diagnosis of carcinoma. The next step in the investigation is to determine whether they belong to one of the several subsets of CUP patients that require specific treatment approaches that may lead to improved outcomes and possibly cure (Table 33-4).
9.  Women with isolated axillary adenopathy. Patients with CUP and isolated axillary adenopathy are usually females, and the diagnosis is most likely breast cancer. The lymph node specimen should be tested for ER, PR, and HER2/neu. In the case of a negative mammogram, the occult breast primary may be seen on MRI. The primary tumor can be identified after mastectomy in 40% to 80% of cases. Patients with mobile axillary lymph nodes (N1) should be treated as stage IIA breast cancer, whereas patients with fixed nodes (N2) should be treated as stage IIIA disease.

PSA, Prostate-specific antigen.

2.  Women with papillary serous adenocarcinoma of the peritoneal cavity. The presence of ascites and peritoneal adenocarcinoma in women is typical of ovarian carcinoma, although this pattern of spread may also occur in tumors of the lung, breast, and gastrointestinal tract. However, a primary tumor is not found in a large number of these patients. Although the origin of these cells is unknown, histologic features such as papillary configuration or psammoma bodies are typical of ovarian carcinoma. Many of these patients, in whom no ovarian or abdominal primary is obvious on laparotomy, are believed to have primary peritoneal carcinoma. The incidence of primary peritoneal carcinoma is increased in women with a history of ovarian carcinoma and BRCA1 mutations. Patients in this subgroup should be considered to have stage III ovarian carcinoma and treated with cytoreductive surgery followed by platinum-based chemotherapy.
3.  Men with blastic bone metastases and elevated PSA. Elderly men with metastatic adenocarcinoma of unknown primary predominantly involving the bones, and those with increased serum PSA or positive PSA staining in the biopsy specimen should be treated for metastatic prostate cancer.
4.  Men with poorly differentiated carcinoma of midline distribution. Young men with a poorly differentiated neoplasm and predominant midline tumor distribution (mediastinum and retroperitoneum) should be treated as extragonadal germ cell tumors even in the absence of elevated serum levels of AFP or β-HCG. The presence of isochromosome 12p in some tumors allows their classification as germ cell tumors.
5.  Squamous cell carcinoma of the cervical lymph nodes. Patients with mid or upper cervical lymph nodes are usually middle-aged or elderly, with frequent history of tobacco and alcohol abuse. The main suspicion in these patients is of a primary head and neck tumor, and the workup should involve the complete evaluation of the upper airways. In the absence of an identifiable primary site, patients should be considered to have locally advanced head and neck cancer. Patients with lower cervical lymph or supraclavicular lymph nodes may have lung cancer and should undergo fiberoptic bronchoscopy during the workup, particularly in the case of unrevealing head and neck examination and nondiagnostic chest imaging. If no primary site is found, the prognosis for this subset of patients is usually poor.
6.  Isolated inguinal lymphadenopathy from squamous cell carcinoma. Most patients with inguinal lymph nodes have a detectable primary tumor either in the genital or in the anorectal area. Therefore, both the genitalia and the rectum should be evaluated during the initial workup. If the primary cancer cannot be identified, long-term survival may be achieved with inguinal lymphadenectomy with or without adjuvant radiation therapy. Some patients may also benefit from the addition of chemotherapy, either in the neoadjuvant or in the adjuvant settings.
7.  Single metastasis. In a small number of patients, only a single metastatic lesion is identified despite a complete clinical and radiologic evaluation. Although other metastatic sites may become evident within a short period of time, some patients may achieve prolonged disease-free interval with local therapies such as surgery or radiation therapy. Despite the uncertain role, adjuvant chemotherapy may be considered for patients with good performance status.
8.  Low-grade neuroendocrine carcinoma. Metastatic carcinoid tumor and islet cell tumors are considered in this subgroup. Treatment includes the use of long-acting octreotide and local therapy when clinically indicated. Cytotoxic agents including streptozocin may be used in selected cases.
9.  Unselected patients. With the exception of patients in the favorable subsets, most patients with CUP remain relatively resistant to chemotherapy, indicating a very poor prognosis. Although the median survival for patients enrolled into clinical trials ranges between 6 and 10 months, population data from tumor registries report median survivals of 2 to 3 months in unselected patients. A prognostic model proposed by the French study group was based on ECOG (Eastern Cooperative Oncology Group) performance status higher than 1 and abnormal lactate dehydrogenase (LDH). Patients with none, one, or two risk factors had median survivals of 10.8, 6.0, and 2.4 months, respectively (J Clin Oncol 2002;20:4679). Patients with good performance status may benefit from chemotherapy. No single chemotherapeutic regimen has emerged as the treatment of choice, and the most commonly used include a combination of platinum and a taxane (Table 33-5). The role for a third agent such as gemcitabine or etoposide remains unclear.

 VI. BACKGROUND. Metastatic CUP is a common entity, accounting for 2.3% of all cancers reported to the Surveillance, Epidemiology, and End Results (SEER) database between 1973 and 1987. It represents the seventh to eighth most common type of cancer and the fourth commonest cause of death in both men and women. The median age at presentation is approximately 60 years, and it is slightly more prevalent in men.

   The characteristic of CUP is the development of metastases before the primary tumor becomes detectable. These tumors are characterized by early dissemination, unpredictable metastatic spread, and very aggressive behavior.

SUGGESTED READINGS

Fizazi K, Greco FA, Pavlidis N, et al. Cancers of unknown primary site: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2011;22(Suppl 6):i64.

Culine S, Kramar A, Saghatchian M, et al. Development and validation of a prognostic model to predict the length of survival in patients with carcinoma of an unknown primary site. J Clin Oncol 2002;20:4679–4683.

Kamposioras K, Pentheroudakis G, Pavlidis N. Exploring the biology of cancer of unknown primary: breakthroughs and drawbacks. Eur J Clin Invest 2013;43:491.

Massard C, Loriot Y, Fizazi K. Carcinomas of an unknown primary origin—diagnosis and treatment. Nat Rev Clin Oncol 2011;8:701.

Pavlidis N, Briasoulis E, Hainsworth J, et al. Diagnostic and therapeutic management of cancer of unknown primary. Eur J Cancer 2003;39:1990–2005.

Pavlidis N, Fizazi K. Cancer of unknown primary. Crit Rev Onc Hematol 2005;54:243–250.

Pavlidis N, Pentheroudakis G. Cancer of unknown primary site. Lancet 2012;379:1428–1435.

Rubin BP, Skarin AT, Pisick E, et al. Use of cytokeratins 7 and 20 in determining the origin of metastatic carcinoma of unknown primary, with special emphasis on lung cancer. Eur J Cancer Prev 2001;10:77–82.

Van De Wouw AJ, Jansen RLH, Speel EJM, et al. The unknown biology of the unknown primary tumor: a literature review. Ann Oncol 2003;14:191–196.

Varadhachary R. Carcinoma of unknown primary: focused evaluation. J Natl Compr Canc Netw 2011;9:1406.