Despite steady improvements in diagnostic and imaging techniques during the past decades, a reasonable evaluation will not identify the source of cancer in a small number (less than 5%) of all patients presenting with metastatic cancer. This heterogeneous group of patients is classified as having cancer of unknown primary (CUP). As diagnostic imaging continues to improve, the frequency of CUP diagnoses is decreasing.
On identification of metastatic cancer, a full medical history and physical examination should be obtained, as well as contrast-enhanced CT of the chest, abdomen, and pelvis. Histologic evaluation of the most accessible tumor mass should include a limited number of immunohistochemical stains to assess the nature of the tumor and to identify or exclude treatable histologies (such as lymphoma or germ cell tumor). Patterns of clinical presentation, or subgroups (discussed later), that also suggest a potential for better outcome should be sought. In addition, specific symptom or presentation-related evaluations may be pursued, such as upper endoscopy and colonoscopy in patients with symptoms or evidence of gastrointestinal bleeding. Patients with regional lymphadenopathy require focused evaluation, such as pan-endoscopic evaluation of the head and neck for patients with isolated or dominant cervical lymphadenopathy or anoscopy for those with isolated inguinal lymphadenopathy. In female patients, breast examination and mammography should be done to search for a breast primary cancer, and a gynecologic evaluation should be performed to look for an ovarian primary. Male patients require a testicular examination and, in those with bone metastases, a prostate examination and serum prostate-specific antigen to evaluate for prostate cancer. Nonspecific tumor markers, such as serum carcinoembryonic antigen, CA-19-9, CA-15-3, or CA-125, are not definitive for identifying a specific site of origin and are not routinely recommended. PET may in some patients suggest the possible primary location, but false-positive results are significant and PET scan findings are not apt to change the treatment plan. The use of gene expression arrays has been commercially promoted, but the clinical utility of these tests to identify more effective therapy is unknown. As these add expense without clear benefit, their routine use in the evaluation of CUP is not recommended.
Ultimately, CUP is a diagnosis of exclusion after evaluation has failed to identify the primary tumor, which either may be too small to be detected or may have been destroyed immunologically and is no longer present. Both patients and clinicians tend to overemphasize the importance of identifying the primary site. Such efforts often deflect focus from the major issue, which is the presence of metastatic cancer. Once the more treatable possibilities have been excluded, specific identification of the site of origin is very unlikely to improve treatment options or clinical outcome. At that point, it should be determined if the metastatic cancer has a favorable or unfavorable prognosis and if there is a specific, efficacious therapy for that patient.
For patients with CUP, the identification of a favorable prognostic subgroup allows selection of specific surgical, radiation, or chemotherapy to which patients are more likely to respond and on occasion achieves long-term remission and cure.
Poorly differentiated cancers are usually more aggressive than well-differentiated cancers, and metastatic poorly differentiated adenocarcinoma has a poor prognosis; however, some patients with poorly differentiated CUP that is not definitively an adenocarcinoma may have specific treatment options as a result of their histology. In particular, young men with poorly differentiated carcinoma that is predominantly in the midline, such as those with large retroperitoneal or mediastinal lymphadenopathy, or both, should be carefully evaluated for the possibility of a germ cell tumor. Serum α-fetoprotein and β-human chorionic gonadotropin levels should be measured, and a testicular examination and ultrasonography should be performed. Even if these evaluations have negative findings, an unrecognized germ cell tumor may still exist, and these patients should be treated for this possibility with a platinum-based chemotherapy regimen.
Patients with poorly differentiated neuroendocrine tumors also warrant careful consideration. These tumors frequently both metastasize like small cell lung cancers and respond similarly to platinum-based chemotherapy.
A group of patients with CUP who may have a more favorable prognosis is women found to have adenocarcinoma in isolated axillary lymphadenopathy. These patients should be presumptively considered to have locoregional breast cancer. If mammography is unrevealing, a breast MRI should be performed. If the MRI scan is negative, the patient is still assumed to have a presumptive stage II breast cancer. Given the inability to identify the primary, a mastectomy or whole breast radiation therapy is recommended. These patients should all receive adjuvant treatment consistent with a stage II breast cancer diagnosis. Patients with isolated or dominant cervical lymphadenopathy should undergo full endoscopic examination of the upper aero-digestive tract to evaluate for a head and neck primary. Even if a primary is not identified, treatment along a head and neck paradigm with chemotherapy and radiation therapy is often appropriate. In particular, patients with high cervical lymphadenopathy with squamous cell cancer occasionally achieve cure. Supraclavicular lymphadenopathy or adenocarcinoma makes a head and neck primary far less likely, and therapy is less efficacious.
Isolated inguinal lymphadenopathy should prompt a careful examination of the anal, perineal, and genital regions that includes anoscopy. Even in the absence of a defined primary tumor, definitive resection or radiation to inguinal or other isolated solitary or regional lymph nodes may provide long-term tumor control and cures in rare circumstances.
Women who have adenocarcinoma with abdominal carcinomatosis and ascites should be presumptively treated for ovarian cancer. Ovarian cancer paradigms, including initial cytoreductive surgery and ovarian cancer chemotherapy regimens, should be used.
Therapy for CUP that does not fall into one of the favorable subgroups is empirically directed with chemotherapy and radiation therapy based on the pattern of presentation. CUP presenting above the diaphragm should be evaluated and managed as metastatic lung cancer. CUP that is predominantly below the diaphragm should be managed as gastrointestinal cancer.
Chronic medical comorbidities and performance status of the patient greatly influences the range of treatment options. As with other solid tumors, patients with several comorbidities and poor performance status are far less likely to benefit from aggressive chemotherapy and are far more likely to experience serious or life-threatening toxicity. Palliative and hospice care should be considered in such patients. Clinical trials that may demonstrate tumor response are typically restricted to patients with normal organ function and good performance status. The results of those trials are unlikely to be informative regarding the response to therapy or prognosis for patients who are not well enough to have qualified for entry into those trials.