Gene Expression Profiling to Identify the Histogenetic Origin of Metastatic Adenocarcinomas of Unknown Primary

  1. Daphne de Jong
  1. From the Divisions of Pathology and Medical Oncology, the Netherlands Cancer Institute; Agendia BV, Amsterdam Science Park, Amsterdam; and Department of Medical Oncology, Medical Centre Alkmaar, Alkmaar, the Netherlands
  1. Corresponding author: Daphne de Jong, MD, PhD, Department of Pathology, Plesmanlaan 121, 1066 CX Amsterdam, the Netherlands; e-mail: d.d.jong{at}nki.nl
 
Next Section

Abstract

Purpose Patients with adenocarcinoma of unknown primary origin (ACUP) constitute approximately 4% of all malignancies. For effective treatment of these patients, it is considered optimal to identify the primary tumor origins. Currently, the success rate of the diagnostic work-up is only 20% to 30%. Our goal was to evaluate the contribution of gene expression profiling for routine clinical practice in patients with ACUP.

Patients and Methods Formalin-fixed, paraffin-embedded (FFPE) samples were obtained from 84 patients with a known primary adenocarcinoma and from 38 patients with ACUP. An extensive immunohistochemical panel classified 16 of the patients with ACUP, whereas 22 patients remained unclassified for their histogenetic origin. Information about staging procedures and clinical follow-up were available in all patient cases. The expression data were analyzed in relation to clinicopathologic variables and immunohistochemical results.

Results The gene expression–based assay classified the primary site correctly in 70 (83%) of 84 patient cases of primary and metastatic tumors of known origin, with good sensitivity for the majority of the tumor classes and relatively poor sensitivity for primary lung adenocarcinoma. Gene expression profiling identified 15 (94%) of 16 patients with initial ACUP who were classified by immunohistochemistry, and it made a valuable contribution to a potential site of origin in 14 of the 22 patients with ACUP.

Conclusion The gene expression platform can classify correctly from FFPE samples the majority of tumors classes both in patients with known primary and in patients with ACUP. Therefore, gene expression profiling represents an additional analytic approach to assist with the histogenetic diagnosis of patients with ACUP.

Previous SectionNext Section

INTRODUCTION

Patients with metastatic carcinoma of unknown primary origin (CUP), for which no primary site of malignancy can be identified despite standardized and extensive investigation, constitute approximately 4% of all malignancies. For optimal treatment decisions, it is important to identify the true origin of the disease. This need is enhanced by the increasing availability of specific and more effective therapy regimens, often chemotherapy or therapy designed to target biologic characteristics of specific malignancies, such as trastuzamab in breast cancer1; erlotinib in lung cancer2; sorafenib and sunitinib in renal cell carcinoma3; and bevacizumab in renal cell carcinoma,4 breast cancer,5 and colorectal cancer.6 The diagnosis of metastatic disease in patients who have no history of earlier malignancy can be made on the basis of a representative biopsy sample of a metastatic lesion. Generally, 80% of malignancies are adenocarcinoma (50% well to moderately differentiated and 30% poorly differentiated or undifferentiated), 15% are squamous cell carcinoma, and 5% are undifferentiated neoplasms.7-13

Standardized work-up for patients with CUP includes a thorough physical examination (including head and neck, rectal, pelvic, and breast examination), basic blood and biochemistry survey, urinalysis, fecal occult blood test, chest x-ray, and computed tomography scan of the abdomen and pelvis.14 Mammography has been recommended for female patients with metastatic adenocarcinoma, and magnetic resonance imaging has been sensitive for the detection of mammographically occult breast cancer.15 [18F]fluoro-2-deoxy-D-glucose positron emission tomography scans may be a valuable modern imaging technique for patients with CUP, particularly for patients with squamous cell carcinoma that involves the cervical lymph nodes,16,17 but also for those with noncervical metastases of unknown primary origin.18 Additional evaluation and endoscopies are recommended and should be sign- or symptom-guided.14

Immunohistochemical analysis is a crucial component of the clinical work-up to identify the histogenetic origin of a malignancy.19-22 Nevertheless, the current success rate of the diagnostic work-up, including clinical, radiologic, and extensive pathologic methods, varies between 20% to 30%.23-25 Therefore, new methods are required to increase the diagnostic accuracy.

Adenocarcinomas are most frequently seen entity in CUP (ACUP). The morphologic features of these tumors are usually uncharacteristic and rarely contribute to the identification of the primary site. However, each type of glandular epithelium has a different biologic function and, therefore, expresses specific genes associated with this differentiation. It is likely that this histogenetic information is maintained during the metastasis process.26 Therefore, the gene expression pattern of a metastasis may reflect the histogenetic make-up of the primary tumor, which would open the door to the use of gene expression–based analysis to assist in identification of the primary site. To date, several gene expression–based tests have demonstrated the potential value of this approach.19,27-31 Although these studies used different microarray platforms, classification algorithms, and sample selection criteria, the overall accuracy of the confirmation of the primary site of origin was 78% to 88%. In most studies, fresh frozen tumor samples, which are often not available in daily practice, were used. Therefore, a diagnostic tool that uses formalin-fixed paraffin-embedded (FFPE) biopsy samples and analytic methods optimized for clinical application use is more suitable for routine practice.

Tothill et al32 translated an expression-based classifier from microarray gene expression data to a reverse transcriptase quantitative polymerase chain reaction (RT-qPCR). They selected a 79-gene panel to design an RT-qPCR assay. Five tumor types can be analyzed with this assay, which might limit its usefulness in daily practice. Talentov et al33 developed an RT-PCR assay that included 10 tissue-specific markers for lung, colon, pancreatic, breast, prostate, and ovarian carcinomas. An independent validation of this assay on 48 metatastic tumors (37 that were of known origin or were subsequently solved patient cases) was able to predict its origin in 76% of patient cases.33

Ma et al29 developed a gene expression database of 466 frozen and 112 FFPE samples that measured the expression of 22,000 genes for the classification of 32 tumor classes, and it had an accuracy of 87% in an independent set of 119 FFPE tumor samples.

By using this extensive database, we developed a diagnostic gene expression–based classifier (CupPrint, Agendia, Amsterdam, the Netherlands). In this study, we evaluated the potential usefulness of this assay to diagnose the histogenetic background of ACUP in clinical practice, with an emphasis on frequently encountered potential sites of origin in the differential diagnosis of ACUP.

Previous SectionNext Section

PATIENTS AND METHODS

One hundred twenty-two patients (81 women and 41 men) who had metastatic adenocarcinoma and who were treated at the Netherlands Cancer Institute, Antoni van Leeuwenhoek Hospital, Amsterdam, the Netherlands, or at the University Medical Center, Groningen, the Netherlands, between 1996 and 2006 were included in this study. Full medical histories were retrieved for all patients, and representative FFPE biopsy samples were obtained. Tissue was collected and studied under an anonymous tissue protocol approved by the ethical committee of both institutions.

Histogenesis of Known Primary Sites

We selected 84 patients who had a known primary tumor, 17 who had primary tumors, and 67 who had metastases. These 84 known tumors spanned 10 different tumor sites. These 10 classes are recognized as frequent sites of origin in the differential diagnosis of ACUP: breast (n = 21), lung (n = 18), colon (n = 22), kidney (clear cell; n = 10), stomach (n = 5), pancreas (n = 3), prostate (n = 5), ovary (n = 8), thyroid (n = 5), bladder (n = 1), and dysgerminoma (n = 1). For all patient cases, the histologic type, tumor grade, and biopsy site were extracted from the pathology reports and were reviewed by one pathologist. Patient and tumor characteristics are listed in Table 1.

View this table:
Table 1.

Patient Characteristics With Known Tumor Localization

Histogenesis of Unknown Primary Sites

We selected 38 patients who had an unknown primary, 16 patients who initially presented with a carcinoma of unknown primary but who were subsequently resolved by immunohistochemistry, and 22 patients whose final diagnosis had been ACUP. All patients received a thorough diagnostic work-up according to the European Society of Medical Oncology guidelines.14 We divided the 22 patients who had an unknown primary into two subgroups. The first group consisted of 12 patients in which clinical and radiologic work-up could not define the primary tumor. In these patient cases, immunohistochemistry reduced the differential diagnoses to two or three possibilities. The second group consisted of 10 patients in which clinical, radiologic, or immunohistochemistry information could not define a tumor origin. Details of the clinical evaluation of these patients are listed in Table 2.

View this table:
Table 2.

Patients Characteristics With Unknown Primary

Tissue Microarray and Immunohistochemistry Analysis

To identify the primary site in patients who had metastatic disease of unknown origin, standard pathologic approaches were used, and additional tissue microarrays (TMAs) that contained 0.6-mm cores from all 122 FFPE samples were constructed. A panel of 10 markers that included primary antibodies to estrogen receptor (ER), progesterone receptor (PR), carcinoembryonic antigen (CEA), clustered antibody 10 (CD10), cytokeratin (CK) 7, CK20, octamer-4 (OCT3/4), thyroid transcription factor 1 (TTF-1), renal cell carcinoma marker (RCC), caclitonin, prostate-specific antigen (PSA), and cancer antigen 125 (CA-125) was selected to identify the primary tumor, according to standard methods (Table 3). Staining was performed with the Lab Vision Immunohistochemical Autostainer (Lab Vision Corporation, Fremont, CA), according to standard protocol that included antigen retrieval.

View this table:
Table 3.

Panel of Antibodies to Detect the Origin of Primary Adenocarcinoma Tumor

RNA Isolation, cRNA Synthesis, and Gene Expression Profiling

To isolate RNA for gene expression profiling, areas of sufficient relative tumor cell content (at least 50%) were selected by using a hematoxylin and eosine–stained section. Three 5-μm sections for each FFPE sample were used for RNA extraction. RNA was isolated, amplified, and labeled, as previously described in Ma et al.29 All samples were hybridized to a customized eight-pack oligonucleotide microarray (CupPrint) that contained 1,900 features (60-mer probes) and allowed eight hybridizations simultaneously, according to standard procedures.34 The CupPrint customized eight-pack microarray contains 495 genes that were selected as highly differentially expressed between 48 tumor types. The array also contains 255 positive and 100 negative controls and an additional 1,050 features, which are used in combination with the 495 predictive genes for the purpose of normalization.34 Reference RNA consisted of a pool of general human cell-lines (Stratagene, La Jolla, CA). As the platform is suitable for archived FFPE biopsy samples, we tested if gene expression profiling was possible for samples that were archived for up to 14 years (1992 to 2006). We selected archived samples from 1992 (n = 8), 1996 (n = 6), 2000 (n = 8), and 2004 (n = 8).

Bioinformatic Approach for Tumor Origin Prediction

To predict the origin of a tumor, a weighted five-nearest neighbor (5NN) algorithm was used to determine the five most molecularly similar tumors in the database, according to the expression of 495-gene set. The k-nearest neighbor (kNN) method was selected as the method to predict the origin of test specimens on the basis of their molecular similarity to tumors in our classification database. This method of classification allows adjustment of the number (k) of samples in the classification database that are used to infer the identity of a given test sample. The optimal value of k-five was identified by cross-validating the complete gene expression database (LOOCV) by using values of k from one to 15. To obtain a single prediction of tumor origin even when multiple tumor types are present in the k-nearest neighbors, the tumor type with the largest summed weight (ie, ∑[1÷vector angle θ]) was selected. In routine application of the classifier, the identity of the top five nearest neighbors was reported. For each sample, tumor origin—as predicted by gene expression profiling—was compared with clinical and pathologic data and with the additional immunological panel.

Previous SectionNext Section

RESULTS

Archival FFPE Biopsies Up to 10 Years Old Are Suitable for Gene Expression Analysis

RNA extraction, amplification, labeling, and hybridization from FFPE tumor tissue was possible for all samples archived in 2004, for seven of eight that were archived in 2000, for five of six that were archived in 1996, and for three of eight samples from 1992. These results indicate that RNA extraction from FFPE material is feasible from recent material but is increasingly difficult from older material.

Gene Expression Profiling Correctly Classifies 83% of Adenocarcinomas From Known Primaries

The microarray gene expression database used in this study has been shown to identify up to 84% of adenocarcinomas of known histogenetic origin.29 In this study, we evaluated an independent series of 84 patient samples to validate the claimed gene expression profiling accuracy of this system on CupPrint, customized, eight-pack microarrays. Clinical characteristics of these patients are listed in Table 1. Correct classifications were obtained for 70 (83%) of 84 tumors with a known origin. These 84 tumors consisted of 17 primary tumors and 67 metastases. Sixteen (94%) of 17 primary tumors were classified correctly, and one primary renal cell carcinoma (TUO-0018) was misclassified as a mesothelioma. Of the 67 metastases from a known primary tumor, 54 (81%) were classified correctly. All patients who had a known primary of the breast (n = 16), colon (n = 9), kidney (n = 7), prostate (n = 5), and thyroid (n = 5) were classified according to their origin. Seven of 11 metastases from primary lung, one of six from primary ovary, two of five from primary stomach, and three of three from primary pancreatic adenocarcinomas were misclassified. A summary of the gene expression profiling results for all known primary tumors is listed in Table 4.

View this table:
Table 4.

Comparison of Microarray Classification Results With Histology and Clinical Information for Patients With a Known Primary Tumor

When the classification performance was examined according to tumor differentiation, gene expression profiling accuracies were 79% (17 of 21), 96% (24 of 25), and 82% (44 of 54) for well-, moderately, and poorly, or undifferentiated tumors, respectively, which indicated no influence of dedifferentiation on gene expression profiling results in our series (χ2 P > .05). Also, sex of the patient did not influence the gene expression profiling (t test > .05).

Gene Expression Profiling Can Provide Additional Information About Localization of the Primary Tumor

For the 16 patients who initially presented with an ACUP and whose diagnosis was subsequently solved by immunohistochemistry, the gene expression profiling identified the correct primary tumor in 15 (93.8%) patient cases (Table 5). Additionally, for eight of the 12 patient cases in which the immunohistochemistry report suggested a differential diagnosis of two or three anatomic localizations, gene expression profiling predicted a single primary origin that was consistent with the clinicopathologic information (Table 6). Gene expression profiling could additionally specify a likely primary origin site in six of 10 patient cases that had an unclear combination of the immunohistochemistry panel and no suspected site of primary tumor. The gene expression profiling result coincided with clinical suspicion (Table 6).

View this table:
Table 5.

Comparison of Microarray Classification Results With Clinicopathologic Features for Patients With Initial Adenocarcinoma of Unknown Primary Origin (N = 16)

View this table:
Table 6.

Detailed Patient Characteristics for Patients With an Adenocarcinoma of Unknown Primary Origin (N = 22)

Using Gene Expression Analyses to Guide Therapy Choice: A Case History

A 42-year-old female (TUO-0148) who had a history of ER-positive, PR-positive, human epidermal growth factor receptor 2–negative, p53–negative breast carcinoma and who was a carrier of a BRCA1 mutation was treated with radiotherapy, neoadjuvant chemotherapy, mastectomy, axillary lymph node dissection for locally advanced breast cancer, and prophylactic bilateral ovariectomy and prophylactic contralateral mastectomy. She presented 7 years later with a progressive skull mass. Imaging techniques showed brain, leptomenigial, lung, and liver metastases. A biopsy sample from the skull showed adenocarcinoma that was positive for cytokeratin 8 (CAM5.2), tumor protein 53 (TP53), S-100 protein (S100), and vimentin. In addition, ER, PR, clustered antibody 56 (CD56), melanoma antigen recognized by T-cells (MART-1), CK7, CK20, CEA, epidermal growth factor receptor (EGFR), and TTF-1 were negative. The second manifestation of malignancy was not clinically and histologically concordant with the primary breast cancer. A second primary tumor of origin was not detected, and the patient received radiotherapy for the intracerebral manifestations of disseminated disease. According to a protocol for ACUP, she received a cisplatin-based regimen. However, the gene expression profile indicated that the metastases were more likely to have originated from a primary breast carcinoma. It was decided to change the chemotherapy to a taxane-based regimen, which resulted in good partial remission. Two months after the end of this treatment, she had rapid progression of liver metastases, which did not respond to vinorelbine. Cisplatin and gemcitabine resulted in a persisting partial remission. She is in good physical and mental condition and has only small metastases visible in the liver.

Previous SectionNext Section

DISCUSSION

In this study, we evaluated the potential usefulness of a gene expression profiling–based microarray test (CupPrint) to diagnose the histogenetic background of ACUP by using FFPE samples. We tested this in diagnostically challenging tumors to determine the utility of the assay in daily practice. First, the microarray test showed accuracy of 83% for samples with known primary tumors and of 94% for ACUPs that could also be classified by immunohistochemistry. In total, the assay showed accuracy of 85%. The microarray test used here did not show a systematic problem in the identification of poorly differentiated tumors, as observed in other studies.30,32 Systematic errors were noted, however, in the classes of lung and pancreatic tumors (63% and 100% misclassified, respectively). Six of the seven misclassified lung adenocarcinomas were poorly differentiated tumors. Although we have shown that dedifferentiation of a tumor has generally no significant influence on the accuracy of the microarray test, it could be argued that some tumor types are more susceptible to deterioration of gene expression quality with increasing dedifferentiation. It is known that lung adenocarcinomas are heterogeneous on the basis of gene expression profiling studies35,36 and that they can be divided into distinctive groups (ie, bronchioid, squamoid, and magnoid) according to their respective correlations with gene expression patterns from histologically defined bronchioalveolar carcinoma, squamous cell carcinoma, and large-cell carcinoma.35,36 Although all lung adenocarcinomas used to develop the classifier were poorly differentiated tumors, the lung adenocarcinoma in this study might belong to different lung adenocarcinoma subtypes. However, the number of patient cases was too small to investigate this hypothesis.

The accuracy of the microarray test used here is quite similar to results with RT-PCR methods in FFPE samples from patients with ACUP.29,32,33 However, the CupPrint is able to recognize more classes than most RT-PCR methods, which recognize only five to six tumor origins.32,33 The high number of 48 classes recognized with CupPrint is obviously an advantage for the usefulness of a diagnostic ACUP test in daily practice.

Currently, the success rate of the diagnostic work-up, including clinical, radiologic, and extensive pathologic methods, is between 20% and 30% for these patients with ACUP.23-25 This is partially caused by relatively small biopsy samples in patients who have tumor of unknown primary, in which the use of various standard techniques is limited.

Molecular analyses, such as gene expression–based tumor classifiers, do not require large samples. Additionally, the microarray test used in this study can utilize RNA from FFPE tissue of up to 10-year-old archived material, which allows retrospective evaluation of old sample sets.

Although a microarray-based gene expression profiling test will not obviate the need for thorough clinical investigation, it may facilitate more focused testing, which would result in reduced cost, lower patient morbidity, and improved outcome. Of the patients with ACUP who were analyzed in this study, the microarray test gave confident and clinically valuable results, which could be strongly supported by their detailed clinical histories, in 14 of 22 tumors with an unknown primary. Specifically, these were patients who had a differential diagnosis in upper abdominal localization (CK7 positive and CEA positive), and who had metastases of the female genital tract (ER positive, CA-125 positive). The CupPrint test has already been useful in the determination of the origin of a CK7-positive and CEA-positive unknown primary tumor.37

With the availability of increasingly more specific and effective chemotherapy agents and of targeted therapies for metastatic adenocarcinomas, it is important to define the localization and the histogenetic background of a primary tumor. Gastrointestinal adenocarcinomas are one example for the specification of treatment: colon tumors are commonly treated with a fluorouracil-based combination, whereas gastric cancer is treated with fluorouracil and a platinum agent, and pancreatic tumors are treated with gemcitabine alone. Gemcitabine alone is less effective for other tumors.19

The combination of gene expression profiling results with a panel of immunohistochemistry markers should be considered to increase the diagnostic accuracy in these gastrointestinal tumors and in metastases in the female genital tract without a known primary.

Furthermore, the diagnostic microarray testing may provide reason for a more targeted therapy in individual patients. In addition to the CupPrint test, the evaluation of a panel of selected markers that are associated with the choice of targeted therapy could additionally direct treatment planning, possibly guided by specific predictive tests. The more accurate diagnosis may lead to consequent improvement in survival, which could not be evaluated in this retrospective study and which remains to be tested in future clinical studies. A clinical diagnostic trial and a cost-effectiveness study may also give valid information (ie, to what extent the gene expression profiling can be adjunct or can replace laboratory testing, pathology, immunologic phenotyping, and or radiologic imaging, such as computed tomography or positron emission tomography scans).

In conclusion, our study demonstrates that microarray-based gene expression profiling tests hold a promise to be a useful additional tool for the determination of the origin of ACUPs.

Previous SectionNext Section

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: Ryan K. van Laar, Agendia BV (U); Arno Floore, Agendia BV (U); Anke Witteveen, Agendia BV (U); Jaana Lathi-Domenici, Agendia BV (U); Annuska M. Glas, Agendia BV (U); Laura J. Van't Veer, Agendia BV (U) Consultant or Advisory Role: None Stock Ownership: Laura J. Van't Veer, Agendia BV Honoraria: None Research Funding: None Expert Testimony: None Other Remuneration: None

Previous SectionNext Section

AUTHOR CONTRIBUTIONS

Conception and design: Hugo M. Horlings, Ryan K. van Laar, Jan-Martijn Kerst, Jelle Wesseling, Laura J. Van't Veer, Daphne de Jong

Administrative support: Hugo M. Horlings, Jan-Martijn Kerst, Helgi H. Helgason, Jelle Wesseling, Jacobus J.M. van der Hoeven, Arno Floore, Daphne de Jong

Provision of study materials or patients: Hugo M. Horlings, Jan-Martijn Kerst, Helgi H. Helgason, Jelle Wesseling, Jacobus J.M. van der Hoeven, Daphne de Jong

Collection and assembly of data: Hugo M. Horlings, Jan-Martijn Kerst, Helgi H. Helgason, Jelle Wesseling, Jacobus J.M. van der Hoeven, Marc O. Warmoes, Arno Floore, Anke Witteveen, Jaana Lathi-Domenici, Annuska M. Glas, Laura J. Van't Veer, Daphne de Jong

Data analysis and interpretation: Hugo M. Horlings, Ryan K. van Laar, Jan-Martijn Kerst, Jacobus J.M. van der Hoeven, Marc O. Warmoes, Arno Floore, Annuska M. Glas, Laura J. Van't Veer, Daphne de Jong

Manuscript writing: Hugo M. Horlings, Ryan K. van Laar, Daphne de Jong

Final approval of manuscript: Hugo M. Horlings, Ryan K. van Laar, Jan-Martijn Kerst, Helgi H. Helgason, Jelle Wesseling, Jacobus J.M. van der Hoeven, Arno Floore, Anke Witteveen, Jaana Lathi-Domenici, Annuska M. Glas, Laura J. Van't Veer, Daphne de Jong

Previous SectionNext Section

Acknowledgments

We thank Iris Simon for helpful discussions and research support and the technicians at Agendia BV for carrying out experiments.

Previous SectionNext Section

Footnotes

  • Supported by Agendia BV.

  • Presented at the 31st European Society for Medical Oncology Congress, September 29-October 3, 2006, Istanbul, Turkey.

  • Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.

  • Received October 3, 2007.
  • Accepted April 8, 2008.
Previous Section
 

REFERENCES

  1. ↵
    Hudis CA: Trastuzumab–mechanism of action and use in clinical practice. N Engl J Med 357:39-51, 2007
    CrossRefMedline
  2. ↵
    Tsao MS, Sakurada A, Cutz JC, et al: Erlotinib in lung cancer: Molecular and clinical predictors of outcome. N Engl J Med 353:133-144, 2005
    CrossRefMedline
  3. ↵
    Brugarolas J: Renal-cell carcinoma: Molecular pathways and therapies. N Engl J Med 356:185-187, 2007
    CrossRefMedline
  4. ↵
    Escudier B, Pluzanska A, Koralewski P, et al: Bevacizumab plus interferon alfa-2a for treatment of metastatic renal cell carcinoma: A randomised, double-blind phase III trial. Lancet 370:2103-2111, 2007
    CrossRefMedline
  5. ↵
    Miller K, Wang M, Gralow J, et al: Paclitaxel plus bevacizumab versus paclitaxel alone for metastatic breast cancer. N Engl J Med 357:2666-2676, 2007
    CrossRefMedline
  6. ↵
    Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004
    CrossRefMedline
  7. ↵
    Lembersky BC, Thomas LC: Metastases of unknown primary site. Med Clin North Am 80:153-171, 1996
    CrossRefMedline
  8. Daugaard G: Unknown primary tumours. Cancer Treat Rev 20:119-147, 1994
    CrossRefMedline
  9. van de Wouw AJ, Janssen-Heijnen ML, Coebergh JW, et al: Epidemiology of unknown primary tumours; incidence and population-based survival of 1,285 patients in Southeast Netherlands, 1984-1992. Eur J Cancer 38:409-413, 2002
    CrossRefMedline
  10. Altman E, Cadman E: An analysis of 1,539 patients with cancer of unknown primary site. Cancer 57:120-124, 1986
    CrossRefMedline
  11. Abbruzzese JL, Abbruzzese MC, Hess KR, et al: Unknown primary carcinoma: Natural history and prognostic factors in 657 consecutive patients. J Clin Oncol 12:1272-1280, 1994
    Abstract/FREE Full Text
  12. Hess KR, Abbruzzese MC, Lenzi R, et al: Classification and regression tree analysis of 1000 consecutive patients with unknown primary carcinoma. Clin Cancer Res 5:3403-3410, 1999
    Abstract/FREE Full Text
  13. ↵
    Le Chevalier T, Cvitkovic E, Caille P, et al: Early metastatic cancer of unknown primary origin at presentation: A clinical study of 302 consecutive autopsied patients. Arch Intern Med 148:2035-2039, 1988
    Abstract/FREE Full Text
  14. ↵
    Briasoulis E, Tolis C, Bergh J, et al: ESMO Minimum Clinical Recommendations for diagnosis, treatment and follow-up of cancers of unknown primary site (CUP). Ann Oncol 16:i75–i76, 2005 (suppl 1)
    FREE Full Text
  15. ↵
    Orel SG, Weinstein SP, Schnall MD, et al: Breast MR imaging in patients with axillary node metastases and unknown primary malignancy. Radiology 212:543-549, 1999
    Abstract/FREE Full Text
  16. ↵
    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 35:1076-1082, 1999
    CrossRefMedline
  17. ↵
    Rades D, Kuhnel G, Wildfang I, et al: Localised disease in cancer of unknown primary (CUP): The value of positron emission tomography (PET) for individual therapeutic management. Ann Oncol 12:1605-1609, 2001
    Abstract/FREE Full Text
  18. ↵
    Joshi U, van der Hoeven JJM, Comans EFI, et al: In search of an unknown primary tumour presenting with extracervical metastases: The diagnostic performance of FDG-PET. Br J Radiol 77:1000-1006, 2004
    Abstract/FREE Full Text
  19. ↵
    Dennis JL, Hvidsten TR, Wit EC, et al: Markers of adenocarcinoma characteristic of the site of origin: Development of a diagnostic algorithm. Clin Cancer Res 11:3766-3772, 2005
    Abstract/FREE Full Text
  20. DeYoung BR, Wick MR: Immunohistologic evaluation of metastatic carcinomas of unknown origin: An algorithmic approach. Semin Diagn Pathol 17:184-193, 2000
    Medline
  21. Tot T: Cytokeratins 20 and 7 as biomarkers: Usefulness in discriminating primary from metastatic adenocarcinoma. Eur J Cancer 38:758-763, 2002
    CrossRefMedline
  22. ↵
    Varadhachary GR, Abbruzzese JL, Lenzi R: Diagnostic strategies for unknown primary cancer. Cancer 100:1776-1785, 2004
    CrossRefMedline
  23. ↵
    Abbruzzese JL, Abbruzzese MC, Lenzi R, et al: Analysis of a diagnostic strategy for patients with suspected tumors of unknown origin. J Clin Oncol 13:2094-2103, 1995
    Abstract/FREE Full Text
  24. Ayoub JP, Hess KR, Abbruzzese MC, et al: Unknown primary tumors metastatic to liver. J Clin Oncol 16:2105-2112, 1998
    Abstract
  25. ↵
    van de Wouw AJ, Jansen RL, Griffioen AW, et al: Clinical and immunohistochemical analysis of patients with unknown primary tumour: A search for prognostic factors in UPT. Anticancer Res 24:297-301, 2004
    Abstract/FREE Full Text
  26. ↵
    Weigelt B, Glas AM, Wessels LF, et al: Gene expression profiles of primary breast tumors maintained in distant metastases. Proc Natl Acad Sci U S A 100:15901-15905, 2003
    Abstract/FREE Full Text
  27. ↵
    Bloom G, Yang IV, Boulware D, et al: Multi-platform, multi-site, microarray-based human tumor classification. Am J Pathol 164:9-16, 2004
    Medline
  28. Dennis JL, Vass JK, Wit EC, et al: Identification from public data of molecular markers of adenocarcinoma characteristic of the site of origin. Cancer Res 62:5999-6005, 2002
    Abstract/FREE Full Text
  29. ↵
    Ma XJ, Patel R, Wang X, et al: Molecular classification of human cancers using a 92-gene real-time quantitative polymerase chain reaction assay. Arch Pathol Lab Med 130:465-473, 2006
    Medline
  30. ↵
    Ramaswamy S, Tamayo P, Rifkin R, et al: Multiclass cancer diagnosis using tumor gene expression signatures. Proc Natl Acad Sci U S A 98:15149-15154, 2001
    Abstract/FREE Full Text
  31. ↵
    Su AI, Welsh JB, Sapinoso LM, et al: Molecular classification of human carcinomas by use of gene expression signatures. Cancer Res 61:7388-7393, 2001
    Abstract/FREE Full Text
  32. ↵
    Tothill RW, Kowalczyk A, Rischin D, et al: An expression-based site of origin diagnostic method designed for clinical application to cancer of unknown origin. Cancer Res 65:4031-4040, 2005
    Abstract/FREE Full Text
  33. ↵
    Talantov D, Baden J, Jatkoe T, et al: A quantitative reverse transcriptase-polymerase chain reaction assay to identify metastatic carcinoma tissue of origin. J Mol Diagn 8:320-329, 2006
    CrossRefMedline
  34. ↵
    Glas AM, Floore A, Delahaye LJ, et al: Converting a breast cancer microarray signature into a high-throughput diagnostic test. BMC Genomics 7:278, 2006
    CrossRefMedline
  35. ↵
    Garber ME, Troyanskaya OG, Schluens K, et al: Diversity of gene expression in adenocarcinoma of the lung. Proc Natl Acad Sci U S A 98:13784-13789, 2001
    Abstract/FREE Full Text
  36. ↵
    Hayes DN, Monti S, Parmigiani G, et al: Gene expression profiling reveals reproducible human lung adenocarcinoma subtypes in multiple independent patient cohorts. J Clin Oncol 24:5079-5090, 2006
    Abstract/FREE Full Text
  37. ↵
    Ismael G, de Azambuja E, Awada A: Molecular profiling of a tumor of unknown origin. N Engl J Med 355:1071-1072, 2006
    CrossRefMedline

Related Article

  • Unknown Primary Site Cancer: Molecular Profiling of Carcinoma of Unknown Primary and Correlation With Clinical Evaluation
    • Gauri R. Varadhachary,
    • Dmitri Talantov,
    • Martin N. Raber,
    • Christina Meng,
    • Kenneth R. Hess,
    • Tim Jatkoe,
    • Renato Lenzi,
    • David R. Spigel,
    • Yixin Wang,
    • F. Anthony Greco,
    • James L. Abbruzzese,
    • and John D. Hainsworth
    JCO September 20, 2008:4442-4448
| Table of Contents

This Article

  1. doi: 10.1200/JCO.2007.14.6969 JCO vol. 26 no. 27 4435-4441
  1. Abstract
  2. » Full Text
  3. PDF

Classifications

Navigate This Article

Current Issue

  1. July 20, 2013, 31 (21)
  1. Alert me to new issues of JCO
    • Advertisement
    • Advertisement
    • Advertisement