Cancer from an Unknown Primary Site

Southwestern Internal Medicine Conference Editor Norman M. Kaplan, MD Biff F. Palmer, MD

Cancer from an Unknown Primary Site Author Jonathan E. Dowell, MD

ABSTRACT: Cancer from an unknown primary site (CUP) is frequently encountered in clinical practice. This review is designed to help physicians identify those patients with CUP that benefit from specific therapeutic approaches. The utility of pathologic and diagnostic tests in patients with CUP will also be discussed, as will the prognosis and appropriate treatment of these patients. KEY INDEXING TERMS: Neoplasms; Unknown primary; Chemotherapy; Neoplasm metastasis; Unknown primary. [Am J Med Sci 2003;326(1):35–46.]

I

n most patients with cancer, the organ in which the malignancy developed is readily apparent. However, some patients present with cancer that has metastasized, and the primary tumor site cannot be identified. Cancer from an unknown primary site (CUP) represents 2 to 3% of all cancer diagnoses.1 Certain subgroups of patients with this diagnosis benefit from specific therapeutic approaches. These include patients with tumors with neuroendocrine features, patients with squamous cell carcinoma involving cervical lymph nodes, women with adenocarcinoma that has metastasized to axillary lymph nodes, women with peritoneal carcinomatosis, men with blastic bone metastases, and men with presentations characteristic of extragonadal germ cell tumors.2,3 Unfortunately, these groups represent a distinct minority (5–10%) of patients diagnosed with CUP. Historically, outcomes for the remainder of patients have been disappointing. The goals of this review are to make physicians aware of these treatable, and, in some cases, potentially curable, subsets of patients with CUP and to critically examine the utility of diagnostic and pathologic studies in these patients. In addition, the prognosis and limited therapeutic options that exist

From the Department of Medicine, Division of Hematology/ Oncology, Dallas Veterans Affairs Medical Center, Dallas Texas. Correspondence: Jonathan E. Dowell, M.D., University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas TX 75390 (E-mail: [email protected]). THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES

for the majority of patients with CUP will be discussed. Biology of CUP Although CUP represents a heterogeneous group of malignancies, the presence of early metastasis that defines this entity suggests that these tumors may share unique biological features. Specifically, CUP may represent a useful model for understanding the biology of early tumor invasion and distant spread. In fact, surprisingly little is known about the biology of these tumors. Investigators at the M.D. Anderson Cancer Center attempted to identify the cytogenetic changes present in a series of tumors from patients with CUP. Of the 13 (of 30) tumors that were successfully karyotyped, 12 were found to have abnormalities involving the short arm of chromosome 1.4,5 Although the gene involved has not been identified, similar changes have been reported in advanced malignancies with known primaries, suggesting that this region of chromosome 1 may be involved in the metastatic process.6,7 Other investigators have used immunohistochemistry (IHC) to identify the presence of oncogene and tumor suppressor gene products in these tumors. Mutations of the tumor suppressor gene P53 result in overexpression of the aberrant protein and have been linked to a more aggressive phenotype in some malignancies.8 In 2 small series of patients with CUP, P53 overexpression was identified in 28 and 53% of tumors.9,10 In addition, expression of the antiapoptotic factor bcl-2 was found in 26 of 40 (65%) CUP tumors.9 The HER-2/neu oncogene has also been examined. This gene encodes a transmembrane glycoprotein that shares homology with the epidermal growth factor receptor. The her-2 protein is expressed in a number of adenocarcinomas, and its presence has been associated with poor clinical outcome in breast cancer.11 Two separate groups have identified her-2 expression in 11 and 58% of CUP tumors.12,13 Importantly, none of the aforementioned trials demonstrated any association between p53, bcl-2, or her-2 expression in CUP and survival or response to therapy. Given the small numbers of patients analyzed to date, the precise role that these genes play in the pathogenesis of CUP has not been determined. Hillen and colleagues14 also evaluated microvessel density (MVD) in liver metastases from a small number of patients with either CUP or a known primary site. MVD is a measure of tumor angiogenesis and is typically increased in tumors that are metastatic. In this study, no difference in the MVD of liver metastases was noted between the CUP and known primary patients, suggesting that MVD is not a specific marker for the early metastatic phenotype that characterizes CUP. However, in the patients with CUP, increased MVD was significantly associated with a worse overall survival. The limited available data does not allow for any 35

Southwestern Internal Medicine Conference

conclusions regarding the biology of CUP. Clearly, additional studies are needed to better delineate the molecular alterations that result in the CUP phenotype and to identify promising therapeutic molecular targets in these patients. Pathology CUP represents a heterogeneous collection of tumor types and clinical presentations. Although it is commonly believed to include only carcinomas, poorly differentiated tumors of unclear origin also fall into this group. With sophisticated pathological evaluation, these tumors may be shown to not be epithelial. Correct identification of certain histologies, such as lymphomas and some chemosensitive sarcomas, may permit prolonged patient survival because of appropriate therapy. Therefore, a careful assessment to insure a precise histological diagnosis is a critical first step in the evaluation of patients with CUP. Retrospective series suggest that conventional light microscopy alone can result in a correct determination of the tumor type in approximately two thirds of patients with CUP.2 Tumors without clear histologic differentiation may require more sophisticated pathological techniques, such as IHC staining, electron microscopy, and molecular genetic studies. Immunohistochemistry. IHC is now the most common adjunctive study used in diagnostic pathology. To perform these studies, the tissue is incubated with monoclonal or polyclonal antibody directed against specific cellular elements. After incubation, the tissue section is washed and exposed to a second labeled antibody that recognizes the first. An oxidation-reduction reaction is then generated that precipitates a dye at the site of antibody binding. There are now hundreds of antibodies commercially available, and this technique can therefore identify a myriad of cellular components, including enzymes, structural elements, hormones, hormone receptors, surface proteins, oncofetal antigens, and other miscellaneous markers.15,16 Caution must be used when interpreting the results of these studies, because no single antibody is perfectly sensitive or specific. However, small panels of antibodies can be quite useful in cases with unclear histology. The utility of IHC in CUP has been evaluated to a limited extent. Gamble and colleagues17 retrospectively reviewed 100 patients with metastatic adenocarcinoma for whom they had diagnostic pathology of the primary tumor and at least 1 metastatic site. Using light microscopy and a limited panel of 9 IHC antibodies, a pathologist who was blinded to the primary tumor site attempted to identify the organ of origin based on the histology and IHC staining of the metastasis. The correct site of origin was predicted in 70% of tumors in men and 54% in women. 36

This represents a considerable improvement over a previous series in which the primary site was correctly identified only 26% of the time using morphology alone.18 Hainsworth and colleagues19 retrospectively performed IHC on 87 patients that had been diagnosed with “poorly differentiated carcinoma” or “poorly differentiated adenocarcinoma” from an unknown primary site before the advent of IHC staining. They then correlated the results of staining with clinical outcome in this group of similarly treated patients. Staining confirmed the diagnosis of carcinoma in 49 patients (56%). A new diagnosis was made by IHC in 14 patients (26%), and the staining was equivocal in the remainder. The “new” diagnoses included melanoma (8 patients), lymphoma (4 patients), prostatic carcinoma (1 patient), and yolksac tumor (1 patient). All 4 patients retrospectively diagnosed with lymphoma had a complete response to chemotherapy and were long-term survivors. A virtually identical study performed by investigators in Europe found 1 patient with lymphoma among 41 patients that had been diagnosed with poorly differentiated carcinoma of unknown primary before the availability of IHC staining. This patient also enjoyed long-term disease-free survival after a complete response to chemotherapy.20 More recently, some investigators have attempted to develop structured approaches to the IHC assessment of CUP. One group has developed a complex algorithm using 18 antibodies that is reportedly 67% accurate with regard to identification of the site of origin of CUP.21 In addition, a more limited panel of 2 cytokeratin antibodies (CK7 and CK20) has been shown to be able to discriminate between adenocarcinoma of the colon and lung in a majority of patients.22–26 Although these results are intriguing, the heterogeneity of CUP will probably limit the usefulness of a single algorithmic approach. The present data suggest that IHC represents a powerful tool in patients with CUP. Given the universal availability of IHC studies, treatable histologies such as lymphoma are now routinely sought in most patients with poorly differentiated neoplasms of unknown origin. Although IHC can also accurately determine the primary site in many patients with epithelial malignancies, the inadequacy of therapy for most metastatic carcinomas limits the utility of these designations. As better and more specific therapies are developed, accurate determination of the primary site in these patients may also be of critical importance. Electron Microscopy. Electron microscopy is another ancillary pathologic study that is occasionally used in patients with CUP. It has been performed successfully on both tissue biopsies and fine needle aspirates.16,27–29 With this technique, ultrastructural features of a tumor, such as cellular organelles, granules, and cell junctions, can be easily visualized. These features can indicate a definitive July 2003 Volume 326 Number 1

Dowell

Table 1. Cytogenic Abnormalities in Selected Solid Tumors and Lymphoma Tumor Alveolar rhabdomyosarcoma Alveolar soft part sarcoma Anaplastic large cell lymphoma Burkitt lymphoma Desmoplastic small round cell tumor Ewing sarcoma/PNET Extraskeletal myxoid chondrosarcoma Germ cell tumor Follicular/diffuse large cell lymphoma Mantle cell lymphoma Melanoma of soft parts Papillary renal cell carcinoma Uterine leiomyoma Renal cell carcinoma Retinoblastoma Synovial sarcoma Wilm tumor

Abnormality t(2;13)(q5;q14) der(X)t(X;17)(p11;q25) t(2;5)(p23;q35) t(8;14)(q24;q32) t(11;22)(p13;q11) t(11;22)(q24;q12) t(11;22)(p13;q11) i(12p) t(14,18)(q32;q21) t(11;14)(q13;q32) t(12;22)(q13;q12) t(X;1)(p11.2;q21.2) t(12;14)(q14–15;q23–24) t/del(3)(p11–21) del(13)(q14) t(X;18)(p11;q11) del(11)(p13)

PNET, peripheral neuroendocrine tumor.

histologic diagnosis in some poorly differentiated malignancies. For instance, visualization of premelanosomes suggests a diagnosis of melanoma, and the presence of dense core (neurosecretory) granules is seen in tumors of neuroendocrine origin. Whether tumors diagnosed solely by electron microscopy display a natural history similar to their counterparts diagnosed through more conventional techniques remains a question. Given the expense and technical expertise that is required for these studies, electron microscopy is generally reserved for those tumors whose lineage remains uncertain after evaluation with light microscopy and IHC. Molecular Genetics. Historically, cytogenetic analysis was reserved for patients with hematological malignancies. Over the last several years, it has become increasingly apparent that a number of solid tumors display characteristic cytogenetic abnormalities as well (Table 1). In poorly differentiated neoplasms, the identification of a specific cytogenetic abnormality may allow for a conclusive diagnosis. More detailed molecular studies have also been performed in some patients. DNA analysis with Southern blot allows for the detection of numerical and structural alterations of genes that may be specific to certain tumors. In addition, with fluorescence in situ hybridization (FISH), specific DNA probes are applied to intact cells. This technique overcomes the limitations of traditional cytogenetics, which requires fresh tissue to establish a cell culture and adequate metaphase preparations.30 Motzer and colleagues31 applied these techniques to a selected group of 40 patients with poorly differentiated tumors from an unknown primary site. Patients were selected because features of their clinical presentations were suggestive THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES

of a possible germ cell malignancy, although the histology of their tumors was not diagnostic. It has recently been recognized that i(12p) is a specific chromosomal marker for germ cell malignancies. In this study, all patients underwent conventional cytogenetic analysis. In addition, 4 of the patients underwent Southern blot analysis for 12p copy number, and 29 patients had FISH performed for the identification of i(12p) and chromosome 12 aneuploidy. In 17 (42%), a specific diagnosis was suggested by 1 of the genetic studies. In 12 of these patients, the diagnosis was germ cell tumor. In 5 patients, another diagnosis was suggested, including lymphoma, melanoma, and 3 rare sarcomas. The majority of patients were treated empirically with regimens directed at germ cell tumors. Importantly, both response to therapy and overall survival were dramatically better in those patients who had a presumptive diagnosis of germ cell tumor made by genetic analysis and were similar to that seen in patients with classical germ cell malignancies. This suggests that the molecular genetic techniques were accurately identifying those patients with poorly differentiated tumors that actually had germ cell malignancies. The polymerase chain reaction has also been used to a limited extent in patients with CUP. This technique results in a marked amplification of DNA and allows for molecular detection and diagnosis from minute amounts of tissue. Using this technique, investigators were able to identify Epstein-Barr viral genome sequences in carcinoma cells of 2 patients with cervical lymph node metastases from an unknown primary.32 In both cases, this predicted the development of a nasopharyngeal carcinoma, which is commonly associated with latent Epstein Barr infection. Additional series have now confirmed that identification of Epstein Barr viral DNA in cervical lymph node metastases is a specific predictor for the presence of a nasopharyngeal primary tumor.33,34 Califano and colleagues,35 using microsatellite analysis, examined 18 patients with squamous cell carcinoma of unknown primary with cervical lymph node metastases for loss of heterozygosity at certain genetic loci that are commonly involved in head and neck tumors. All patients also had biopsies of benign-appearing mucosal sites that are frequently implicated in unknown primary of the head and neck (tonsil, base of tongue, piriform sinus, nasopharynx). In more than half of these patients, at least 1 of these histopathologically benign biopsies displayed a genetic alteration identical to that seen in that patient’s cervical lymph node metastases. Three of these patients went on to develop primary tumors with the same genetic changes at identical or adjacent head and neck sites. Although not conclusive, these data suggest that benign-appearing mucosa of the head and neck may harbor foci of clonal 37

Southwestern Internal Medicine Conference

cells that are presumptively the site of origin of unknown primary tumors of the head and neck. The microsatellite analysis used in this study may prove to be a useful tool for identifying these sites. In the near future, surgical pathologists may also use gene expression data generated from cDNA microarray technology to help identify the primary site in patients with CUP. Using this technology, thousands of genes can be analyzed simultaneously and can potentially provide an organ-specific genetic expression profile. Two independent investigators have already shown that this technology can be used to accurately predict the site of origin of primary and metastatic tumor samples.36,37 This preliminary data suggests that gene expression profiling may soon be a routine part of the pathological evaluation of CUP patients. The burgeoning array of available molecular genetic techniques offers the hope of eventually improving the ability to correctly determine the histology and site of origin of unknown primary tumors. Although some of the data presented above is preliminary and requires confirmation, in selected instances, genetic analysis may be quite useful in patients with CUP. Specifically, in those patients with a clinical presentation consistent with an extragonadal germ cell tumor, serious consideration should be given to performing cytogenetics or FISH (even if it requires repeat biopsy to obtain additional material), because these studies have the potential to identify a curable subset of patients with CUP. Treatable Subsets of CUP After a careful review by pathology to exclude treatable histologies such as lymphoma, the vast majority of patients with CUP will be given a diagnosis of carcinoma. The next step in these patients’ evaluation should be to determine whether they fall into 1 of the following treatable subsets of patients with carcinoma from an unknown primary site. Women with Peritoneal Carcinomatosis. A woman with CUP will occasionally present with malignant ascites and extensive peritoneal involvement with adenocarcinoma. Although this presentation can occur with lung, gastrointestinal, and breast cancers, it is most characteristic of ovarian cancer. However, some of these women do not have evidence of an ovarian tumor or other obvious primary at surgical exploration. This syndrome has gone by several names, including papillary carcinoma of the peritoneum, extraovarian papillary serous carcinoma, serous surface papillary carcinoma, psammomacarcinoma, and now, most commonly, primary peritoneal carcinoma.38 The cell of origin is unclear. The germinal epithelium of the ovary and the mesothelium of the peritoneum have the same embryologic origin. It is thought that the mesothelium may retain the multipotentiality of the Mülle38

rian system and that this allows for the development of a peritoneal carcinoma.39 The histology is that of a serous carcinoma and is often indistinguishable from ovarian carcinoma.40,41 Although small deposits of tumor may be seen on the surface of the ovaries, the key to the diagnosis is the fact that the body of the ovaries are uninvolved.42 The syndrome can be seen in women that have previously had their ovaries removed.43– 45 In addition, 2 small series have shown that some women with primary peritoneal carcinoma harbor BRCA1 mutations, and this tumor is now included on the list of those seen in the familial breast and ovarian cancer syndromes.46,47 Women with this clinical presentation should be referred to a gynecologic oncologist for evaluation and eventual surgical exploration. Numerous small series have now demonstrated that these patients benefit from aggressive surgical tumor debulking followed by systemic chemotherapy, as is generally done for ovarian carcinoma. With this approach, most series report outcomes similar to that seen in ovarian cancer of an equivalent stage. Median survivals range from 11 to 24 months and 5-year survival rates of 15 to 20% are reported.42– 44,48 –55 Thus, it seems that some of these patients may be cured with appropriate evaluation and treatment. Women with Metastatic Carcinoma in Axillary Lymph Nodes. Occasionally a woman will present with unilateral axillary lymph nodes as the only site of metastatic carcinoma. Although this pattern of spread can be seen with lung, thyroid, head and neck, and gastrointestinal tumors, it is most suggestive of a breast primary. In the absence of another obvious primary tumor site, these patients are typically treated as breast cancer. Initial evaluation of these patients should include careful breast exam and mammography. Retrospective series have illustrated that 25 to 50% of these patients will not have a lesion identified, even with modern mammographic techniques.56 Several small studies have looked at the utility of magnetic resonance imaging (MRI) of the breast in identifying the primary lesion in these patients. The results of these small series suggest that MRI may be a useful test in patients with no identifiable lesion on mammography.57,58 However, the exact role that it should play in the evaluation of these patients remains to be determined. The optimal local treatment for these patients remains controversial. Historically, the treatment these women typically received was mastectomy. Despite this aggressive surgical approach, careful pathologic review of the mastectomy specimen failed to reveal a breast tumor in 33 to 47% of these patients.59 – 63 The prognosis of these patients seems to be quite similar to patients with node positive breast cancer and is not affected by whether a primary tumor is identified at surgery. As with breast July 2003 Volume 326 Number 1

Dowell

cancer, the most important prognostic feature is the number of lymph nodes involved with tumor.59,60 As less aggressive surgery and breast conservation have become the treatment of choice in breast cancer, this approach has also been evaluated to a limited extent in patients with axillary metastasis from an unknown primary. Most retrospective series evaluating axillary dissection either alone or with breast irradiation have shown that the overall survival and local recurrence rate are similar to that seen with mastectomy.64 – 68 A few have suggested that local recurrence may occur more frequently in those patients that do not receive radiation.64,67,68 To date, no prospective trial has compared the 2 approaches and therefore both represent reasonable options. Most of these patients also receive systemic chemotherapy (as would patients with node-positive breast cancer), although this approach has not been validated in any prospective study. Patients with Squamous Cell Carcinoma in Cervical Lymph Nodes. Patients that present with squamous cell carcinoma metastatic to cervical lymph nodes represent another treatable subset of patients with CUP. This presentation accounts for 1 to 2% of all head and neck malignancies.69 Lung and esophageal cancer can also present in this fashion, especially in those patients that have lymphadenopathy in the low neck. These patients are typically referred to an otolaryngologist, where the evaluation consists of computed tomography of the head and neck and panendoscopy (laryngoscopy, bronchoscopy, and esophagoscopy) in an effort to identify a primary lesion. This typically also includes blind biopsies of common potential primary sites, including the nasopharynx, tonsils, base of the tongue, piriform sinuses, and any suspicious mucosal areas. Ipsilateral tonsillectomy is often performed because it will be found to harbor an occult primary in 10 to 25% of patients.70 As discussed in more detail below, positron emission tomography (PET) has also been evaluated to a limited extent in this setting, and it can successfully locate the primary tumor in some patients.71 Despite this aggressive diagnostic approach, a primary site will not be identified in approximately two thirds of patients.70 The appropriate treatment of these patients is controversial. The typical approach includes neck dissection followed by radiation therapy, although some series suggest that patients with limited neck disease may be adequately treated with neck dissection or radiation alone.69,72 There is also disagreement regarding the extent of radiation. Some centers advocate bilateral neck and total mucosal irradiation.73 The radiation field for this approach extends from the nasopharynx to the clavicles bilaterally and results in a high incidence of significant chronic xerostomia. Other centers have felt that radiation to the ipsilateral neck alone results in adequate tumor control with less morbidiTHE AMERICAN JOURNAL OF THE MEDICAL SCIENCES

ty.69 In a retrospective nonrandomized comparison, local control and survival were improved with the more aggressive approach.73 However, the confounding effects of patient selection and imbalances in other prognostic factors limit the ability to interpret this data. Chemotherapy has been evaluated in this setting to a very limited degree, and its role in the treatment of these patients remains to be defined.74 As in patients with head and neck cancer, the prognosis of patients with cervical lymph node metastases from an unknown primary site is mostly determined by the extent of the lymph node involvement. With the treatment approaches outlined above, long-term local control is achieved in 50 to 75% of patients, and 5-year survival rates of 40 to 60% are reported.69,70,72,73,75,76 Patients with Squamous Cell Carcinoma in Inguinal Lymph Nodes. A rare patient will present with metastatic poorly differentiated or squamous cell carcinoma involving inguinal lymph nodes and no obvious primary tumor site. Evaluation of these patients should include careful genital/ pelvic exam and anoscopy to exclude an anogenital primary. Computed tomography of the pelvis may also be useful to evaluate the extent of lymph node involvement. In patients without an identifiable primary site or tumor that has spread beyond the inguinal region, small uncontrolled series suggest that surgery either alone or followed by radiation can result in long term survival in approximately 25% of patients.77,78 Men with Possible Prostate Cancer. Older men with adenocarcinoma of unknown primary, especially those with predominant bone metastases, should be evaluated for the presence of prostate cancer. The evaluation should include serum prostate specific antigen (PSA) and IHC staining of the biopsy material for PSA and other prostate-specific markers. Two small case series suggest that the presence of an elevated serum PSA or a positive IHC stain for PSA, even in the setting of an atypical clinical presentation for prostate cancer, predicts for response to hormonal therapy.79,80 Some investigators advocate an empiric trial of hormonal therapy in all male patients with osteoblastic bone metastases regardless of the PSA results.2 Neuroendocrine Carcinoma from an Unknown Primary Site Neuroendocrine carcinoma from an unknown primary represents a heterogeneous group of clinical presentations. Greco and Hainsworth3 have described 3 subsets of neuroendocrine carcinoma. Each of these can present with metastatic disease in the absence of an obvious primary site. The first group presents with histology similar to that of typical carcinoid or pancreatic islet cell tumors. The patients often present with metastatic disease to the liver and may or may not have clinical evidence of hormone production. These tumors are typically indolent and progress slowly over 39

Southwestern Internal Medicine Conference

years. Chemotherapy seems to have limited efficacy in this setting, and treatment options are similar to those used in carcinoid and islet cell tumors with known primaries. Octreotide is potentially useful for those patients with symptomatic hormone production. A second group of patients with neuroendocrine carcinoma from an unknown primary includes those with a histologic diagnosis of small-cell carcinoma. Even in the absence of an obvious lung primary, the natural history and response to chemotherapy are similar to that seen with small cell lung cancer.81– 84 Thus, these patients are generally treated with platinum-based chemotherapy similar to that used in small-cell lung cancer. With this approach, rare long-term survival has been reported. The final subset of neuroendocrine carcinoma from an unknown primary site includes those patients with poorly differentiated carcinoma that has neuroendocrine features identified only by electron microscopy or IHC. Hainsworth and colleagues85 have reported the only series of such patients in the literature. In their hands, this subset represented a particularly chemosensitive group of patients with CUP. They reported a response rate to platinumbased chemotherapy of more than 60% and longterm disease-free survival in approximately 10%. The site of origin of these tumors remains uncertain in most patients. In the Hainsworth series, a specific diagnosis was made during follow up or at autopsy in 4 of the 29 patients. Two patients were classified as having “undifferentiated” carcinoid. One was found to have small-cell lung cancer and another was diagnosed with “extragonadal germ cell tumor with predominant neuroendocrine differentiation.” Whether this subgroup of neuroendocrine carcinoma truly represents a distinct and treatable subset of CUP ideally requires confirmation in larger prospective series. The Extragonadal Germ Cell Tumor Syndrome. The “extragonadal germ cell tumor syndrome” was first described it 1981.86 This syndrome includes patients that have clinical presentations consistent with metastatic extragonadal germ cell tumor but lack the definitive histology. These patients are typically given a histological diagnosis of poorly differentiated carcinoma. The clinical features of the syndrome are listed in Table 2. Most patients do not display all of the features of the syndrome. As discussed above, some of these patients can now be definitively diagnosed with sophisticated molecular genetic techniques that can identify chromosomal derangements specific to germ cell tumors. In the absence of this, patients that display features of the syndrome should be treated with platinumbased chemotherapy regimens that are used for germ cell tumors. With this approach, the majority 40

Table 2. Features of the Extragonadal Germ Cell Tumor Syndrome Clinical Features Occurs in men ⬍50 Patients present with tumors located predominantly in the midline (retroperitoneum, mediastinum) and/or with multiple pulmonary nodules The duration of symptoms before seeking medical attention is short (⬍3 months) and/or there is a history of rapid tumor growth Serum levels of AFP, ␤-HCG, or both are elevated A good response to previously administered chemotherapy or radiotherapy has been demonstrated Adapted from Ref. 3.

of patients will respond to treatment and long-term disease free survival can be achieved.86,87 Patients with Poorly Differentiated Carcinoma or Poorly Differentiated Adenocarcinoma. Whether patients with poorly differentiated carcinoma (PDC) or poorly differentiated adenocarcinoma (PDA) from an unknown primary site represent a treatable subset of patients with CUP is controversial. A series reported by Hainsworth and colleagues suggested that patients with metastatic PDA and PDC were particularly responsive to platinum-based chemotherapy; in that series, approximately 15% of these patients were long-term survivors with apparent cure.88 The series was initially designed to include only patients with features of the extragonadal germ cell tumor syndrome (see above) but was later modified to include all patients with PDA or PDC and no obvious primary tumor site. Approximately 1 quarter of the patients in this series had clinical features consistent with an extragonadal germ cell tumor. Another confounding factor is that many of the patients on this study were enrolled before the routine use of IHC. A retrospective analysis of these patients’ diagnostic material using more contemporary IHC techniques discovered that several of the long-term survivors actually had other diagnoses that are particularly sensitive to chemotherapy, such as lymphoma and germ cell tumor. A more recent and larger review of patients with poorly differentiated carcinoma from an unknown primary site was performed by investigators at the M.D. Anderson Cancer Center.89 In this review of all 1400 patients with CUP seen between 1987 and 1994, all patients had extensive pathological analysis of their biopsy specimens, including IHC, and appropriate clinical and radiological evaluation. Using these techniques, 5% of patients were found to have specific treatable diagnoses, including lymphoma, germ cell tumor, and breast and ovarian carcinomas. An analysis of the remaining patients found no survival advantage or particular sensitivity to chemotherapy for the subset with PDC or July 2003 Volume 326 Number 1

Dowell

PDA. It now seems that with the advent of IHC to identify treatable histologies in patients previously called PDA or PDC, and with the ability to recognize patients with features of the extragonadal germ cell tumor, the remainder of patients with PDC or PDA probably do not represent a specific treatable subset of patients with CUP. The Remainder of Patients with CUP Clinical Evaluation. As discussed above, patients that fall into a treatable subset of CUP may benefit from specific diagnostic and clinical examinations. For the remainder of patients, what constitutes an appropriate clinical evaluation has been debated for many years. The natural inclination for many patients and physicians is that an extensive search for a primary site should be undertaken. The rationale for this approach is 3-fold.90 The absence of a primary generates anxiety in both patients and their physicians and results in feelings that the evaluation has been inadequate. In addition, physicians are used to making predictions about tumor behavior and prognosis and decisions about therapy based on knowledge of the primary site. Lastly, there is often a feeling that the prognosis could be improved if the primary were found.90 It has been estimated that the typical evaluation for a patient with CUP costs between $4,500 and $18,000 per patient.90,91 Total annual costs in the United States may be as high as $1.5 billion. Unfortunately, the data do not confirm the value of this exhaustive approach. Several large series in the 1970s and 1980s evaluated the utility of conventional contrast roentgenographic studies in this setting. Two of the series evaluated the routine use of chest radiograph, barium enema, upper gastrointestinal series, and intravenous pyelography in patients with CUP. Primary sites were successfully demonstrated in only 8 to 9% of patients.92,93 Notably, the false positive rate equaled or exceeded the true positive rate for each of the procedures evaluated.94 The value of computed tomography (CT) in CUP has also been examined. In retrospective series from the Mayo Clinic and the University of Kansas, CT of the abdomen and/or chest correctly identified the primary tumor site in 35 and 32% of patients, respectively.95,96 The most common primary sites identified were pancreas, kidney, and lung. In the Mayo series, 10% of positive scans were ultimately found to be false positives, and approximately one third of negative scans were shown to be falsely negative. In the 2 series, 23 and 65% of patients were found to have additional sites of metastatic disease by CT. A complete picture of the accuracy of CT in this setting could not be determined given that 36% of patients in the Mayo series and 54% of patients in the UniTHE AMERICAN JOURNAL OF THE MEDICAL SCIENCES

versity of Kansas series never had a primary site identified. The M.D. Anderson Cancer Center reported their experience with a more limited clinical evaluation of 879 consecutive patients with CUP.90 Their assessment included history and physical examination as well as routine chemistries and a complete blood count. All patients also had a chest radiograph and CT of the abdomen and pelvis. Men also had studies to measure the serum prostate specific antigen, ␣-fetoprotein (AFP) and ␤-human chorionic gonadotropin (␤-HCG). All women had mammography. Additional testing was performed only if clinically indicated. Using this approach, a primary site was identified in 20% of patients. One third of these diagnoses were suggested by unique histologic features that were identified on careful pathologic review of the biopsy material. The survival of this group with unique histologies was significantly prolonged compared with the rest of the group, and this survival benefit was believed to be mostly attributable to the patients identified with lymphoma. Importantly, if one examined only the patients with carcinoma after pathologic review, there was no survival difference seen between the group with an identified primary site and the group with occult primary. This confirms the findings of a prior smaller series from Australia that also concluded that identification of the primary tumor had no impact on the survival of patients with CUP.93 Another significant finding from the M.D. Anderson review was that 14 women were diagnosed with either breast or ovarian cancer by mammography or abdominal CT. Even in the setting of metastatic disease, these 2 malignancies are relatively treatable; not unexpectedly, these patients’ survival was found to be superior to the rest of the group. The authors’ conclusions were that the most important features of the evaluation of CUP patients are a careful pathologic review and specific studies in women to exclude breast and ovarian malignancies. Despite the above recommendations, the merits of routine mammography in this setting are still debated. Two large retrospective series found that mammography was “positive” in 5 to 10% of women with CUP; in all cases, however, these were demonstrated to be false positive results.97,98 False negative rates of 5 to 20% were also quoted in these series. In the M.D. Anderson review, only 4 women of 304 were found to have positive or suspicious mammography that ultimately resulted in a diagnosis of breast cancer.90 The investigators did not report the false positive rates observed in their series and the attendant costs and morbidity (from additional diagnostic tests and biopsies) that result from such findings. Another area that deserves mention is the routine use of tumor markers in patients with CUP. Except for the special circumstances discussed above, when 41

Southwestern Internal Medicine Conference

Table 3. Suggested Initial Clinical Evaluation for the Patient with CUP Medical history (including) Family history of malignancy Social history (tobacco use, carcinogen exposure, HIV risk factors) Physical examination (including) Integument exam Lymph node survey Breast exam Rectal and genital/pelvic examination Laboratory evaluation Complete blood count Complete metabolic profile Radiological evaluation Chest radiograph or computed tomography (CT) CT of the abdomen and pelvis

male patients present with features consistent with possible germ cell or prostatic malignancy, the routine use of a battery of tumor markers should be discouraged. Although carcinoembryonic antigen (CEA) is frequently elevated in colon cancer and CA 19-9 in pancreatic cancer, they may also be increased in a number of other adenocarcinomas, including lung and breast cancers.15 Although profound elevations of a certain marker may be highly suggestive of a diagnosis in the appropriate clinical setting, their specificity in CUP has not been validated. Pavlidis et al99 reported a retrospective review of 6 tumor markers (CEA, CA 19-9, CA 15-3, CA-125, AFP, and ␤-HCG) in 85 patients with CUP. With the exception of AFP and ␤-HCG, which were elevated in 17 and 44% of patients, respectively, each of the remaining markers was increased in more than 60% of subjects. This clearly illustrates the low specificity of tumor markers in this setting. In addition, no association between a particular marker elevation and response to therapy or survival was noted. Although elevated markers may be followed as surrogates for response to treatment, their diagnostic utility in this setting is limited. Newer diagnostic technologies, such as PET, have also been evaluated to a limited extent in CUP. The available data suggests that PET can correctly identify a primary site of disease in some of these patients and that it may be particularly helpful in localizing head and neck tumors in patients with cervical adenopathy from an unknown primary.71,100 –103 However, the published series have included only a handful of patients, and some have suggested that the false positive rate of PET in this setting may be distressingly high.100 Larger prospective series that include cost analyses will be needed to better define the role of PET in patients with CUP. The appropriate evaluation for the patient with CUP begins with a thorough history (Table 3). The history of the present illness should focus on the 42

presence of symptoms that suggest a specific primary site. A review of the past medical history should include questions regarding prior surgeries for “benign” skin lesions or moles, because further investigation may reveal these to be the source of the current metastatic disease. In addition, a complete family history can be helpful. Clustering of certain tumor types within a family, such as breast, ovarian, or colorectal cancers, may suggest the presence of an underlying heritable cancer syndrome and an increased risk for specific malignancies. Lastly, a social history focused on prior exposures (such as tobacco or asbestos) and risk factors for HIV infection may suggest an underlying risk for certain tumors. In the patient with CUP, a careful physical examination can also identify the primary site. Certain aspects of the physical exam that are frequently omitted are important in this setting. All patients should have a thorough skin exam in an effort to find a possible primary melanoma or nonmelanoma skin cancer. Examination of the oral and nasal cavities and a complete lymph node survey may also be helpful. In some instances, the pattern of lymph node involvement may indicate a possible primary site, such as axillary lymphadenopathy, suggesting a breast primary, or inguinal lymphadenopathy, pointing to a genital, anal, or lower extremity tumor. In addition, all patients should have breast, rectal, and pelvic/genital examinations. In addition to a complete history and physical examination, it is recommended that patients with CUP have a basic laboratory evaluation. This typically includes serum electrolytes, creatinine, liver function tests, and a complete blood count. Although these tests are unlikely to identify the primary tumor, they may provide information regarding potential unrecognized sites of metastatic disease. In addition, they offer an assessment of organ function that can be used to predict the patient’s ability to tolerate chemotherapy. It is also recommended that patients with CUP have a chest radiograph or CT as well as CT of the abdomen and pelvis as part of their initial evaluation. As discussed above, the ability of these studies to clearly delineate a primary tumor is modest. However, they can identify additional sites of metastasis and provide a measure of tumor bulk that can be used to monitor response to therapy. Mammography should be considered in women with a presentation consistent with a possible breast primary. Additional radiological or endoscopic procedures should be performed only if the standard indications for those procedures are present. Prognosis and Treatment. Unfortunately, the vast majority of patients with CUP do not fall into 1 of the treatable subsets discussed above. The prognosis for this group is poor. Hess and colleagues104 reported on their single institution experience with approximately 1000 patients with CUP. The median July 2003 Volume 326 Number 1

Dowell

survival for this population, which included some patients with treatable subsets of CUP, was 11 months, and only 11% survived 5 years. A multivariate analysis demonstrated that male gender, the presence of liver metastases, and an increasing number of organs involved with tumor were all independent predictors for an especially poor outcome.105 For example, CUP patients with liver metastases had a median survival of 7 months, and the median survival of patients ⬎61.5 years old with liver metastases was only 5 months.106 Other investigators have shown that, as with most cancers, the patient’s performance status is also a critical prognostic variable.107 For those patients who do not fall into a treatment-responsive subset of CUP, the utility of chemotherapy is unproven. No large-scale randomized trial has been undertaken to establish a beneficial survival or palliative effect from treatment.108 Numerous phase II studies in the 1980s evaluated treatment regimens that included 5-fluorouracil, cisplatin, doxorubicin, or mitomycin.109 –114 Typical objective tumor response rates ranged from 0 to 40% with resultant median survivals of 3 to 8 months. These modest results often came at the expense of considerable toxicity. More recently, some of the newer chemotherapeutic agents have been incorporated into treatment regimens for these patients.115– 123 One particularly encouraging regimen includes a combination of carboplatin, paclitaxel, and oral etoposide. Hainsworth et al122 reported a response rate of 47% and a median survival of over 13 months in 53 CUP patients that received this treatment. Unfortunately, other trials using similar (but not identical) drug combinations have produced much less impressive results.115–117,121 Consequently, although many oncologists have adopted the “Hainsworth regimen,” there is no standard treatment for these patients. The hope is that selected patients with reasonable performance status may derive some benefit from chemotherapy. However, supportive care without chemotherapy is always a reasonable option in this group. Conclusions Patients with CUP are frequently encountered in everyday practice, with between 20,000 and 50,000 new cases diagnosed in the United States each year. Improvements in molecular diagnostics may soon significantly reduce this number; until then, physicians should familiarize themselves with this entity. Approximately 5 to 10% of patients with CUP may benefit from a specific treatment approach. Close communication between the pathologist and the primary physician is critical to insure that a treatable histology is not overlooked. In addition, physicians should be familiar with the clinical presentations that suggest a potentially treatable subset of CUP. THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES

For the majority of patients with CUP, an extensive search for a primary has little value. Treatment for these patients is palliative, and the optimal approach has not been defined. Acknowledgements Special thanks to Kristi Willis for her invaluable assistance with the preparation of the manuscript. References 1. Jemal A, Thomas A, Murray T, et al. Cancer statistics, 2002. CA Cancer J Clin 2002;52:23– 47. 2. Hainsworth JD, Greco FA. Treatment of patients with cancer of an unknown primary site. N Engl J Med 1993;329: 257– 63. 3. Greco FA, Hainsworth JD. Cancer of unknown primary site. In: Devita VT, Hellman S, Rosenberg S, editors. Cancer: principles and practice of oncology. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 2537– 60. 4. Abbruzzese JL, Lenzi R, Raber MN, et al. The biology of unknown primary tumors. Semin Oncol 1993;20:238 – 43. 5. Bell CW, Pathak S, Frost P. Unknown primary tumors: establishment of cell lines, identification of chromosomal abnormalities, and implications for a second type of tumor progression. Cancer Res 1989;49:4311–5. 6. Atkin NB. Chromosome 1 aberrations in cancer. Cancer Genet Cytogenet 1986;21:279 – 85. 7. Rowley JD. Abnormalities of chromosome No. 1: significance in malignant transformation. Virchows Arch B Cell Pathol 1978;29:139 – 44. 8. Belluco C, Guillem JG, Kemeny N, et al. p53 Nuclear protein overexpression in colorectal cancer: a dominant predictor of survival in patients with advanced hepatic metastases. J Clin Oncol 1996;14:2696 –701. 9. Briasoulis E, Tsokos M, Fountzilas G, et al. Bcl2 and p53 protein expression in metastatic carcinoma of unknown primary origin: biological and clinical implications. A Hellenic Co-operative Oncology Group study. Anticancer Res 1998;18:1907–14. 10. Bar-Eli M, Abbruzzese JL, Lee-Jackson D, et al. p53 gene mutation spectrum in human unknown primary tumors. Anticancer Res 1993;13:1619 –23. 11. Paik S, Hazan R, Fisher ER, et al. Pathologic findings from the National Surgical Adjuvant Breast and Bowel Project: prognostic significance of erbB-2 protein overexpression in primary breast cancer. J Clin Oncol 1990;8:103– 12. 12. Pavlidis N, Briassoulis E, Bai M, et al. Overexpression of C-myc, Ras and C-erbB-2 oncoproteins in carcinoma of unknown primary origin. Anticancer Res 1995;15:2563–7. 13. Hainsworth JD, Lennington WJ, Greco FA. Overexpression of Her-2 in patients with poorly differentiated carcinoma or poorly differentiated adenocarcinoma of unknown primary site. J Clin Oncol 2000;18:632–5. 14. Hillen HF, Hak LE, Joosten-Achjanie SR, et al. Microvessel density in unknown primary tumors. Int J Cancer 1997;74:81–5. 15. Ruddon RW, Norton SE. Use of biological markers in the diagnosis of cancers of unknown primary tumor. Semin Oncol 1993;20:251– 60. 16. Mackay B, Ordonez NG. Pathological evaluation of neoplasms with unknown primary tumor site. Semin Oncol 1993;20:206 –28.

43

Southwestern Internal Medicine Conference

17. Gamble AR, Bell JA, Ronan JE, et al. Use of tumour marker immunoreactivity to identify primary site of metastatic cancer. BMJ 1993;306:295– 8. 18. Sheahan K, Abramowitz A, Carlson J. Accuracy of pathologic evaluation of metastatic carcinoma of unknown primary site. An analysis using a computerized image storage system [abstract]. Lab Invest 1990;62:112A. 19. Hainsworth JD, Wright EP, Johnson DH, et al. Poorly differentiated carcinoma of unknown primary site: clinical usefulness of immunoperoxidase staining. J Clin Oncol 1991;9:1931– 8. 20. van der Gaast A, Verwij J, Planting AS, et al. The value of immunohistochemistry in patients with poorly differentiated adenocarcinomas and undifferentiated carcinomas of unknown primary. J Cancer Res Clin Oncol 1996;122:181–5. 21. DeYoung BR, Wick MR. Immunohistologic evaluation of metastatic carcinomas of unknown origin: an algorithmic approach. Semin Diagn Pathol 2000;17:184 –93. 22. Tot T. Adenocarcinomas metastatic to the liver: the value of cytokeratins 20 and 7 in the search for unknown primary tumors. Cancer 1999;85:171–7. 23. Moll R, Lowe A, Laufer J, et al. Cytokeratin 20 in human carcinomas. A new histodiagnostic marker detected by monoclonal antibodies. Am J Pathol 1992;140:427– 47. 24. Ramaekers F, van Niekerk C, Poels L, et al. Use of monoclonal antibodies to keratin 7 in the differential diagnosis of adenocarcinomas. Am J Pathol 1990;136:641–55. 25. Wauters CC, Smedts F, Gerrits LG, et al. Keratins 7 and 20 as diagnostic markers of carcinomas metastatic to the ovary. Hum Pathol 1995;26:852–5. 26. Loy TS, Calaluce RD. Utility of cytokeratin immunostaining in separating pulmonary adenocarcinomas from colonic adenocarcinomas. Am J Clin Pathol 1994;102:764 –7. 27. Mackay B, Fanning T, Bruner JM, et al. Diagnostic electron microscopy using fine needle aspiration biopsies. Ultrastruct Pathol 1987;11:659 –72. 28. Hammar S, Bockus D, Remington F. Metastatic tumors of unknown origin: an ultrastructural analysis of 265 cases. Ultrastruct Pathol 1987;11:209 –50. 29. Sehested M, Juul N, Hainau B, et al. Electron microscopy of ultrasound-guided fine-needle biopsy specimens. Br J Radiol 1987;60:351–3. 30. Ilson DH, Motzer RJ, Rodriguez E, et al. Genetic analysis in the diagnosis of neoplasms of unknown primary tumor site. Semin Oncol 1993;20:229 –37. 31. Motzer RJ, Rodriguez E, Reuter VE, et al. Molecular and cytogenetic studies in the diagnosis of patients with poorly differentiated carcinomas of unknown primary site. J Clin Oncol 1995;13:274 – 82. 32. Feinmesser R, Miyazaki I, Cheung R, et al. Diagnosis of nasopharyngeal carcinoma by DNA amplification of tissue obtained by fine-needle aspiration. N Engl J Med 1992;326: 17–21. 33. Macdonald MR, Freeman JL, Hui MF, et al. Role of Epstein-Barr virus in fine-needle aspirates of metastatic neck nodes in the diagnosis of nasopharyngeal carcinoma. Head Neck 1995;17:487–93. 34. Lee WY, Hsiao JR, Jin YT, et al. Epstein-Barr virus detection in neck metastases by in-situ hybridization in fine-needle aspiration cytologic studies: an aid for differentiating the primary site. Head Neck 2000;22:336 – 40. 35. Califano J, Westra WH, Koch W, et al. Unknown primary head and neck squamous cell carcinoma: molecular identification of the site of origin. J Natl Cancer Inst 1999;91:599 – 604.

44

36. Su AI, Welsh JB, Sapinoso LM, et al. Molecular classification of human carcinomas by use of gene expression signatures. Cancer Res 2001;61:7388 –93. 37. Giordano TJ, Shedden KA, Schwartz DR, et al. Organspecific molecular classification of primary lung, colon, and ovarian adenocarcinomas using gene expression profiles. Am J Pathol 2001;159:1231– 8. 38. Bartlett D. Peritoneal carcinomatosis. In: Devita VT, Hellman S, Rosenberg S, editors. Cancer: principles and practice of oncology. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 2561–74. 39. Chu CS, Menzin AW, Leonard DG, et al. Primary peritoneal carcinoma: a review of the literature. Obstet Gynecol Surv 1999;54:323–35. 40. Muggia FM, Baranda J. Management of peritoneal carcinomatosis of unknown primary tumor site. Semin Oncol 1993;20:268 –72. 41. August CZ, Murad TM, Newton M. Multiple focal extraovarian serous carcinoma. Int J Gynecol Pathol 1985;4:11– 23. 42. Kennedy AW, Markman M, Webster KD, et al. Experience with platinum-paclitaxel chemotherapy in the initial management of papillary serous carcinoma of the peritoneum. Gynecol Oncol 1998;71:288 –90. 43. Strnad CM, Grosh WW, Baxter J, et al. Peritoneal carcinomatosis of unknown primary site in women. A distinctive subset of adenocarcinoma. Ann Intern Med 1989;111: 213–7. 44. Dalrymple JC, Bannatyne P, Russell P, et al. Extraovarian peritoneal serous papillary carcinoma. A clinicopathologic study of 31 cases. Cancer 1989;64:110 –5. 45. Tobacman JK, Greene MH, Tucker MA, et al. Intraabdominal carcinomatosis after prophylactic oophorectomy in ovarian-cancer-prone families. Lancet 1982;2:795–7. 46. Schorge JO, Muto MG, Welch WR, et al. Molecular evidence for multifocal papillary serous carcinoma of the peritoneum in patients with germline BRCA1 mutations. J Natl Cancer Inst 1998;90:841–5. 47. Bandera CA, Muto MG, Schorge JO, et al. BRCA1 gene mutations in women with papillary serous carcinoma of the peritoneum. Obstet Gynecol 1998;92:596 – 600. 48. Bloss JD, Liao SY, Buller RE, et al. Extraovarian peritoneal serous papillary carcinoma: a case-control retrospective comparison to papillary adenocarcinoma of the ovary. Gynecol Oncol 1993;50:347–51. 49. Piver MS, Eltabbakh GH, Hempling RE, et al. Two sequential studies for primary peritoneal carcinoma: induction with weekly cisplatin followed by either cisplatin-doxorubicin-cyclophosphamide or paclitaxel-cisplatin. Gynecol Oncol 1997;67:141– 6. 50. Whitcomb BP, Kost ER, Hines JF, et al. Primary peritoneal psammocarcinoma: a case presenting with an upper abdominal mass and elevated CA-125. Gynecol Oncol 1999; 73:331– 4. 51. Ransom DT, Patel SR, Keeney GL, et al. Papillary serous carcinoma of the peritoneum. A review of 33 cases treated with platin-based chemotherapy. Cancer 1990;66: 1091– 4. 52. Lele SB, Piver MS, Matharu J, et al. Peritoneal papillary carcinoma. Gynecol Oncol 1988;31:315–20. 53. Hochster H, Wernz JC, Muggia FM. Intra-abdominal carcinomatosis with histologically normal ovaries. Cancer Treat Rep 1984;68:931–2. 54. Fromm GL, Gershenson DM, Silva EG. Papillary serous carcinoma of the peritoneum. Obstet Gynecol 1990;75:89 – 95.

July 2003 Volume 326 Number 1

Dowell

55. Chen KT, Flam MS. Peritoneal papillary serous carcinoma with long-term survival. Cancer 1986;58:1371–3. 56. Leibman AJ, Kossoff MB. Mammography in women with axillary lymphadenopathy and normal breasts on physical examination: value in detecting occult breast carcinoma. AJR Am J Roentgenol 1992;159:493–5. 57. Orel SG, Weinstein SP, Schnall MD, et al. Breast MR imaging in patients with axillary node metastases and unknown primary malignancy. Radiology 1999;212:543–9. 58. Tilanus-Linthorst MM, Obdeijn AI, Bontenbal M, et al. MRI in patients with axillary metastases of occult breast carcinoma. Breast Cancer Res Treat 1997;44:179 – 82. 59. Patel J, Nemoto T, Rosner D, et al. Axillary lymph node metastasis from an occult breast cancer. Cancer 1981;47: 2923–7. 60. Ashikari R, Rosen PP, Urban JA, Senoo T. Breast cancer presenting as an axillary mass. Ann Surg 1976;183: 415–7. 61. Fieuerman L, Attie JN, Rosenberg B. Carcinoma in axillary lymph nodes as an indicator of breast cancer. Surg Gynecol Obstet 1962;114:5. 62. Owen H, Dockerty MB, Gray HK. Occult carcinoma of the breast. Surg Gynecol Obstet 1954;98:302. 63. Westbrook KC, Gallager HS. Breast carcinoma presenting as an axillary mass. Am J Surg 1971;122:607–11. 64. Vlastos G, Jean ME, Mirza AN, et al. Feasibility of breast preservation in the treatment of occult primary carcinoma presenting with axillary metastases. Ann Surg Oncol 2001; 8:425–31. 65. Morrow M, Schmidt RA, Bucci C. Breast conservation for mammographically occult carcinoma. Ann Surg 1998;227: 502– 6. 66. Baron PL, Moore MP, Kinne DW, et al. Occult breast cancer presenting with axillary metastases. Updated management. Arch Surg 1990;125:210 – 4. 67. Ellerbroek N, Holmes F, Singletary E, et al. Treatment of patients with isolated axillary nodal metastases from an occult primary carcinoma consistent with breast origin. Cancer 1990;66:1461–7. 68. Merson M, Andreola S, Galimberti V, et al. Breast carcinoma presenting as axillary metastases without evidence of a primary tumor. Cancer 1992;70:504 – 8. 69. Nieder C, Gregoire V, Ang KK. Cervical lymph node metastases from occult squamous cell carcinoma: cut down a tree to get an apple? Int J Radiat Oncol Biol Phys 2001;50: 727–33. 70. Mendenhall WM, Mancuso AA, Amdur RJ, et al. Squamous cell carcinoma metastatic to the neck from an unknown head and neck primary site. Am J Otolaryngol 2001; 22:261–7. 71. OS AA, Fischbein NJ, Caputo GR, et al. Metastatic head and neck cancer: role and usefulness of FDG PET in locating occult primary tumors. Radiology 1999;210:177– 81. 72. Coster JR, Foote RL, Olsen KD, et al. Cervical nodal metastasis of squamous cell carcinoma of unknown origin: indications for withholding radiation therapy. Int J Radiat Oncol Biol Phys 1992;23:743–9. 73. Reddy SP, Marks JE. Metastatic carcinoma in the cervical lymph nodes from an unknown primary site: results of bilateral neck plus mucosal irradiation vs. ipsilateral neck irradiation. Int J Radiat Oncol Biol Phys 1997;37:797– 802. 74. de Braud F, Heilbrun LK, Ahmed K, et al. Metastatic squamous cell carcinoma of an unknown primary localized to the neck. Advantages of an aggressive treatment. Cancer 1989;64:510 –5. 75. Grau C, Johansen LV, Jakobsen J, et al. Cervical lymph node metastases from unknown primary tumours. Results

THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES

76.

77.

78. 79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

from a national survey by the Danish Society for Head and Neck Oncology. Radiother Oncol 2000;55:121–9. Colletier PJ, Garden AS, Morrison WH, et al. Postoperative radiation for squamous cell carcinoma metastatic to cervical lymph nodes from an unknown primary site: outcomes and patterns of failure. Head Neck 1998;20:674 – 81. Guarischi A, Keane TJ, Elhakim T. Metastatic inguinal nodes from an unknown primary neoplasm. A review of 56 cases. Cancer 1987;59:572–7. Zaren HA, Copeland EM 3rd. Inguinal node metastases. Cancer 1978;41:919 –23. Tell DT, Khoury JM, Taylor HG, et al. Atypical metastasis from prostate cancer. Clinical utility of the immunoperoxidase technique for prostate-specific antigen. JAMA 1985;253:3574 –5. Gentile PS, Carloss HW, Huang TY, et al. Disseminated prostatic carcinoma simulating primary lung cancer. Indications for immunodiagnostic studies. Cancer 1988;62:711–5. Garrow GC, Greco FA, Hainsworth JD. Poorly differentiated neuroendocrine carcinoma of unknown primary tumor site. Semin Oncol 1993;20:287–91. Van Der Gaast A, Verwey J, Prins E, et al. Chemotherapy as treatment of choice in extrapulmonary undifferentiated small cell carcinomas. Cancer 1990;65:422– 4. Moertel CG, Kvols LK, O’Connell MJ, et al. Treatment of neuroendocrine carcinomas with combined etoposide and cisplatin. Evidence of major therapeutic activity in the anaplastic variants of these neoplasms. Cancer 1991;68:227–32. Kasimis BS, Wuerker RB, Malefatto JP, et al. Prolonged survival of patients with extrapulmonary small cell carcinoma arising in the neck. Med Pediatr Oncol 1983;11:27–32. Hainsworth JD, Johnson DH, Greco FA. Poorly differentiated neuroendocrine carcinoma of unknown primary site. A newly recognized clinicopathologic entity. Ann Intern Med 1988;109:364 –71. Richardson RL, Schoumacher RA, Fer MF, et al. The unrecognized extragonadal germ cell cancer syndrome. Ann Intern Med 1981;94:181– 6. Sumi H, Itoh K, Onozawa Y, et al. Treatable subsets in cancer of unknown primary origin. Jpn J Cancer Res 2001; 92:704 –9. Hainsworth JD, Johnson DH, Greco FA. Cisplatinbased combination chemotherapy in the treatment of poorly differentiated carcinoma and poorly differentiated adenocarcinoma of unknown primary site: results of a 12-year experience. J Clin Oncol 1992;10:912–22. Lenzi R, Hess KR, Abbruzzese MC, et al. Poorly differentiated carcinoma and poorly differentiated adenocarcinoma of unknown origin: favorable subsets of patients with unknown-primary carcinoma? J Clin Oncol 1997;15:2056 – 66. Abbruzzese JL, Abbruzzese MC, Lenzi R, et al. Analysis of a diagnostic strategy for patients with suspected tumors of unknown origin. J Clin Oncol 1995;13:2094 –103. Schapira DV, Jarrett AR. The need to consider survival, outcome, and expense when evaluating and treating patients with unknown primary carcinoma. Arch Intern Med 1995;155:2050 – 4. Nystrom JS, Weiner JM, Wolf RM, et al. Identifying the primary site in metastatic cancer of unknown origin. Inadequacy of roentgenographic procedures. JAMA 1979;241: 381–3. Stewart JF, Tattersall MH, Woods RL, et al. Unknown primary adenocarcinoma: incidence of overinvestigation and natural history. Br Med J 1979;1:1530 –3.

45

Southwestern Internal Medicine Conference

94. Leonard RJ, Nystrom JS. Diagnostic evaluation of patients with carcinoma of unknown primary tumor site. Semin Oncol 1993;20:244 –50. 95. McMillan JH, Levine E, Stephens RH. Computed tomography in the evaluation of metastatic adenocarcinoma from an unknown primary site. A retrospective study. Radiology 1982;143:143– 6. 96. Karsell PR, Sheedy PF 2nd, O’Connell MJ. Computed tomography in search of cancer of unknown origin. JAMA 1982;248:340 –3. 97. Nystrom JS, Weiner JM, Heffelfinger-Juttner J, et al. Metastatic and histologic presentations in unknown primary cancer. Semin Oncol 1977;4:53– 8. 98. Kirsten F, Chi CH, Leary JA, et al. Metastatic adeno or undifferentiated carcinoma from an unknown primary site– natural history and guidelines for identification of treatable subsets. Q J Med 1987;62:143– 61. 99. Pavlidis N, Kalef-Ezra J, Briassoulis E, et al. Evaluation of six tumor markers in patients with carcinoma of unknown primary. Med Pediatr Oncol 1994;22:162–7. 100. Greven KM, Keyes JW Jr, Williams DW 3rd, et al. Occult primary tumors of the head and neck: lack of benefit from positron emission tomography imaging with 2-[F18]fluoro-2-deoxy-D-glucose. Cancer 1999;86:114 – 8. 101. Lassen U, Daugaard G, Eigtved A, et al. 18F-FDG whole body positron emission tomography (PET) in patients with unknown primary tumours (UPT). Eur J Cancer 1999;35: 1076 – 82. 102. Bohuslavizki KH, Klutmann S, Kroger S, et al. FDG PET detection of unknown primary tumors. J Nucl Med 2000;41:816 –22. 103. Kole AC, Nieweg OE, Pruim J, et al. Detection of unknown occult primary tumors using positron emission tomography. Cancer 1998;82:1160 – 6. 104. Hess KR, Abbruzzese MC, Lenzi R, et al. Classification and regression tree analysis of 1000 consecutive patients with unknown primary carcinoma. Clin Cancer Res 1999;5: 3403–10. 105. Abbruzzese JL, Abbruzzese MC, Hess KR, et al. Unknown primary carcinoma: natural history and prognostic factors in 657 consecutive patients. J Clin Oncol 1994;12: 1272– 80. 106. Ayoub JP, Hess KR, Abbruzzese MC, et al. Unknown primary tumors metastatic to liver. J Clin Oncol 1998;16: 2105–12. 107. van der Gaast A, Verweij J, Planting AS, et al. Simple prognostic model to predict survival in patients with undifferentiated carcinoma of unknown primary site. J Clin Oncol 1995;13:1720 –5. 108. Farrugia DC, Norman AR, Nicolson MC, et al. Unknown primary carcinoma: randomised studies are needed to identify optimal treatments and their benefits. Eur J Cancer 1996;32A:2256 – 61. 109. Woods RL, Fox RM, Tattersall MH, et al. Metastatic adenocarcinomas of unknown primary site: a randomized

46

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

study of two combination-chemotherapy regimens. N Engl J Med 1980;303:87–9. Walach N, Horn Y. Combination chemotherapy in the treatment of adenocarcinoma of unknown primary origin. Cancer Treat Rep 1987;71:605–7. Sporn JR, Greenberg BR. Empirical chemotherapy for adenocarcinoma of unknown primary tumor site. Semin Oncol 1993;20:261–7. Sporn JR, Greenberg BR. Empiric chemotherapy in patients with carcinoma of unknown primary site. Am J Med 1990;88:49 –55. Goldberg RM, Smith FP, Ueno W, et al. 5-fluorouracil, adriamycin, and mitomycin in the treatment of adenocarcinoma of unknown primary. J Clin Oncol 1986;4:395–9. Bedikian AY, Bodey GP, Valdivieso M, et al. Sequential chemotherapy for adenocarcinoma of unknown primary. Am J Clin Oncol 1983;6:219 –24. Briasoulis E, Kalofonos H, Bafaloukos D, et al. Carboplatin plus paclitaxel in unknown primary carcinoma: a phase II Hellenic Cooperative Oncology Group Study. J Clin Oncol 2000;18:3101–7. Dowell JE, Garrett AM, Shyr Y, et al. A randomized Phase II trial in patients with carcinoma of an unknown primary site. Cancer 2001;91:592–7. Greco FA, Erland JB, Morrissey LH, et al. Carcinoma of unknown primary site: phase II trials with docetaxel plus cisplatin or carboplatin. Ann Oncol 2000;11:211–5. Hainsworth JD, Burris HA 3rd, Calvert SW, et al. Gemcitabine in the second-line therapy of patients with carcinoma of unknown primary site: a phase II trial of the Minnie Pearl Cancer Research Network. Cancer Invest 2001;19:335–9. Rigg A, Cunningham D, Gore M, et al. A phase I/II study of leucovorin, carboplatin and 5-fluorouracil (LCF) in patients with carcinoma of unknown primary site or advanced oesophagogastric/pancreatic adenocarcinomas. Br J Cancer 1997;75:101–5. Voog E, Merrouche Y, Trillet-Lenoir V, et al. Multicentric phase II study of cisplatin and etoposide in patients with metastatic carcinoma of unknown primary. Am J Clin Oncol 2000;23:614 – 6. Warner E, Goel R, Chang J, et al. A multicentre phase II study of carboplatin and prolonged oral etoposide in the treatment of cancer of unknown primary site (CUPS). Br J Cancer 1998;77:2376 – 80. Hainsworth JD, Erland JB, Kalman LA, et al. Carcinoma of unknown primary site: treatment with 1-hour paclitaxel, carboplatin, and extended-schedule etoposide. J Clin Oncol 1997;15:2385–93. Greco FA, Burris HA 3rd, Litchy S, et al. Gemcitabine, carboplatin, and paclitaxel for patients with carcinoma of unknown primary site: a Minnie Pearl Cancer Research Network study. J Clin Oncol 2002;20:1651– 6.

July 2003 Volume 326 Number 1