This Article Abstract Full Text (PDF) eLetters: Submit a response to this article Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services E-mail this article link to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pentheroudakis, G. Articles by Pavlidis, N. Search for Related Content PubMed PubMed Citation Articles by Pentheroudakis, G. Articles by Pavlidis, N.

Cancer Medicine: Case Discussions

Cancer of Unknown Primary Site: Missing Primary or Missing Biology?

Department of Medical Oncology, Ioannina University Hospital, Ioannina, Greece

Key Words. Cancer • Unknown primary • Molecular biology

Correspondence: Nicholas Pavlidis, M.D., Ph.D., Department of Medical Oncology, Ioannina University Hospital, Niarxou Avenue, 45500 Ioannina, Greece. Telephone and Fax: 30-26510 99394; e-mail: npavlid{at}cc.uoi.gr

Received May 30, 2006; accepted for publication February 1, 2007.

	ABSTRACT

Top Abstract Introduction Distinct Clinical Features Molecular Biology Conclusions Disclosure of Potential... References

 
Cancer of unknown primary site (CUP) ranks as the fourth most common cause of cancer deaths and represents both a diagnostic and a management challenge. In CUP, the regression or dormancy of the primary tumor, the development of early, uncommon, systemic metastases, and the resistance to therapy are hallmarks of this heterogeneous clinical entity. Still, no consensus exists on whether CUP is simply a group of metastatic tumors with unidentified primaries or a distinct entity with specific genetic/phenotypic aberrations that define it as "primary metastatic disease." In this review, we present karyotypic analyses as well as the single-gene, single-protein studies done on the expression of oncogenes, tumor- or metastasis-suppressor genes, as well as angiogenesis effectors. These studies show frequent expression of oncoproteins, lack of activating epidermal growth factor receptor/c-Kit mutations or amplification, uncommon presence of tumor- or metastasis-suppressor gene mutations and highly active angiogenesis in CUP. Informative as they may be, these data have been observed in several solid tumors of known primary and failed to identify a CUP-specific molecular signature. The latter, if it exists, probably consists of a multigene expression pattern not captured by single-gene studies. Gene and protein microarray technologies offer promise for the unraveling of complex genetic programs that would either identify each CUP's primary tissue of origin or instead define the CUP-specific molecular signature. Confirmation of one of the two hypotheses would either improve primary disease–oriented therapy or develop CUP-oriented treatments targeting molecular aberrations that drive neoplastic growth/dissemination.

Disclosure of potential conflicts of interest is found at the end of this article.

	INTRODUCTION

Top Abstract Introduction Distinct Clinical Features Molecular Biology Conclusions Disclosure of Potential... References

 
Cancer of unknown primary site (CUP) is defined as histologically confirmed metastases in the absence of identifiable primary tumor despite a standardized diagnostic approach [1]. CUP accounts for 5% of all cancers, with a reported annual incidence per 100,000 population of 18 cases in Australia, 7–12 cases in the U.S., six cases in the Netherlands, ranking as the fourth commonest cause of cancer death in both sexes [2]. Because the management of patients with malignant diseases relies heavily on identification of the primary tumor, its absence poses significant diagnostic and therapeutic problems. Although it is widely accepted that CUP is a heterogeneous cohort of metastatic malignancies, no consensus exists yet on the true nature of this clinical entity [3]. CUP might prove to be an artificial classification group made up from malignant metastatic deposits from primaries that have simply not been found, in which case an intensive diagnostic approach for identification of the primary is justified so as to enable definition of prognosis and administration of disease-oriented therapy. An opposing hypothesis could maintain that, whatever the primary site, CUP represents a unique clinical entity harboring distinct genetic and epigenetic aberrations, obviating the need to pinpoint the primary site. Instead, the unraveling of the molecular mechanisms would be a priority and would enable the identification of biochemical-molecular targets and administration of CUP-specific therapy. The research effort to prove or disprove either of the two theories has been fragmented, a fact that reflects the heterogeneity of the disease, differing investigator attitudes, as well as the lack of cooperative group clinical trials and translational research.

	DISTINCT CLINICAL FEATURES

Top Abstract Introduction Distinct Clinical Features Molecular Biology Conclusions Disclosure of Potential... References

 
A review of available data on the clinical presentation and natural history of CUP provides insights into some common clinicopathological characteristics. CUP has a more or less differentiated glandular epithelial phenotype in 80% of cases. Patients present with nonspecific systemic complaints of short duration (weight loss, anorexia, malaise, fatigue), not contributing to organ-site localization. The primary tumor remains unidentified in >80% of antemortem series, despite an aggressive endoscopic/positron emission tomography–based workup, and in 30%–70% of postmortem autopsy series, highlighting the regression of the primary as one of the hallmarks of this clinical entity [4–6]. When it is found, commonly in the lung or pancreas, the primary tumor is a small, asymptomatic nodule: this phenomenon is not compatible with the usual natural history of pancreatic or lung carcinomas and makes the dormancy of the primary a second CUP hallmark. Patients with CUP have early, systemic metastases: frequently three or more organs are involved, with metastases being commonly localized in unexpected sites such as the kidneys, skin, scalp, heart, and distant lymph nodes [7, 8]. With the exception of a minority of treatable subsets, CUP is characterized by resistance to combination chemotherapy: responses occur in less than one fourth of patients and are short-lived, with a median survival duration in the range of 3–6 months in unfavorable and 10–16 months in favorable subsets [9]. Although not a proof of principle, the clinical hallmarks of regression/dormancy of the primary, early systemic uncommon metastases, and resistance provide hints for the presence of distinct genetic alterations that characterize CUP.

	MOLECULAR BIOLOGY

Top Abstract Introduction Distinct Clinical Features Molecular Biology Conclusions Disclosure of Potential... References

 
Attempts at elucidating the molecular biology of CUP initially focused on chromosomal abnormalities, immunohistochemical (IHC) protein expression studies, and mutational screening of single or a few genes. Underlying these efforts rests the belief that dysregulation of one or a few genes and encoded proteins drives systemic dissemination and regression of the primary. Most such simplistic hypotheses have not been confirmed, because the research efforts yielded neither consistent nor specific abnormalities of genes or proteins "pivotal" for the development and survival of a metastasis-prone malignant clone.

Chromosomal Abnormalities
Aneuploidy is seen in 70%–90% of solid tumors and probably reflects derangements of chromosomal replication during cell division [10]. Hedley et al. [11] established aneuploidy in 70% of 152 CUP, without any relationship to patterns of metastatic involvement or survival. Abnormalities of the short arm of chromosome 1 (1p) were reported in 12 of 13 CUP, a finding consistent with the common occurrence of 1p structural alterations in advanced solid tumors [12]. As the presence of an additional copy of the short arm of chromosome 12 specifically characterizes germ-cell tumors, Motzer et al. [13] analyzed 40 CUP and reported it in 12 of these. The presence of i(12p) predicted for response to cisplatin-based chemotherapy, making it a germ cell cancer–specific marker for patients presenting with midline nodal metastases of unknown primary. Comparative genomic hybridization may provide a tool more sensitive than karyotypic analysis for the allocation of CUP cases to various classes of primary tumor [14].

Oncogenes and Proteins
Oncogenes are overexpressed or amplified in many solid tumors. The encoded proteins favor malignant transformation and survival by activating cellular proliferation, inducing cell migration, inhibiting apoptosis, and promoting neoangiogenesis. Data on oncogene and oncoprotein expression in CUP are summarized inTable 1.

We studied the presence of c-Myc, Ras, and human epidermal growth factor receptor (HER)-2 proteins by IHC in 26 cases of CUP and observed almost universal staining for the first two (96%, 92%, and 65%, respectively), although overexpression was seen in less than one third of cases for any of the studied proteins [15]. Hainsworth et al. [16] examined 100 CUP and similarly observed HER-2 overexpression in 11% of specimens, with no differences in overall survival between patients with overexpressing and nonoverexpressing HER-2 tumors. Fizazi et al. [17] found HER-2 IHC overexpression in only 2 of 56 CUP cases (4%); no association of HER-2 status with histological differentiation, treatment activity, or clinical outcome was seen. In keeping with the initial results published by our group, Rashid et al. [18] reported HER-2 IHC expression in 68% and overexpression in 24% of 76 cases of CUP. Van de Wouw et al. [3] also observed IHC HER-2 overexpression in 16 of 45 CUP cases (35%). Interestingly, coexpression of HER-2 with epidermal growth factor receptor (EGFR) and vascular endothelial growth factor (VEGF) or cyclooxygenase (COX)-2 was observed in more than half of all tumors, while patients with combined EGFR, HER-2, and COX-2–positive tumors had a survival that was twice that of the rest. That study suggests that expression profiling of multiple genes may lead to a better understanding of the molecular pathophysiology and reliable selection of combination targeted therapies. Overall, these studies agree in depicting overexpression of c-Myc, Ras, and HER-2 oncoproteins in one fourth to one third of the patients with CUP and create optimism that such Ras/HER-2 profiling will help identify patients more or less likely to benefit from therapies targeting these molecules.

EGFR recently came under scrutiny after cumulating knowledge of its role in malignant transformation. The predictive value of activating mutations in exons 18–21 for lung cancer patient benefit from EGFR tyrosine kinase inhibitors (TKIs) increased interest in the role of the receptor, despite contradictory data on the prognostic significance of EGFR tissue protein expression [19]. Fizazi et al. [17] reported IHC EGFR protein overexpression in 2 of 56 CUP specimens (4%), whereas Rashid et al. [18] reported it in 46 of 76 CUP samples (61%). These discrepant results serve to emphasize the heterogeneity of this group of tumors. Both studies, however, failed to establish any correlation between EGFR protein staining and clinicopathological parameters, response to chemotherapy, or patient survival. We recently completed a global EGFR profiling screen in 50 patients with CUP [20]. Seventy-four percent of the tumors stained positively for EGFR protein, though only 12% overexpressed it. EGFR protein expression or overexpression had no prognostic significance for patient survival. Using polymerase chain reaction (PCR) techniques, we found neither activating mutations nor amplification of exons 18, 19, or 21 in any of the 50 tumors. Our data highlighting the absence of markers of active EGFR signaling in CUP constitute proof of principle for a foreseeable lack of benefit from EGFR modulation.

The c-Kit and platelet-derived growth factor receptor (PDGFR) proteins are transmembrane receptors signaling toward cell survival, inhibition of apoptosis, and angiogenesis upon binding of their specific ligands. Activating receptor mutations drive neoplastic proliferation of gastrointestinal stromal tumors (GISTs) and render them exquisitely sensitive to inhibition of c-Kit/PDGFR tyrosine kinase activity by small molecule TKIs [21]. In studies using IHC methodologies, Fizazi et al. [17] reported c-Kit protein overexpression in 6 of 56 CUP samples (11%), while Rashid et al. [18] reported it in 3 of 76 cases (4%). Our group screened 50 CUP specimens and found c-Kit protein overexpression in 13% of tumors. Moreover, we screened all samples for mutations in exon 11 of the c-kit and exons 12 and 18 of the PDGFR genes by means of PCR–single strand conformational polymorphism (SSCP) and found none [20]. In all three studies, c-Kit protein expression failed to demonstrate prognostic utility for response to therapy or survival. Overall, these data show an absence of contribution of active c-Kit signaling in the pathogenesis of CUP, at least analogous to that seen in GISTs. Consequently, c-Kit tyrosine kinase inhibition is unlikely to drastically improve the dismal prognosis of patients with CUP.

The IHC expression of the antiapoptotic protein Bcl-2 was analyzed by our group in 40 patients with CUP. Briasoulis et al. [22] reported Bcl-2 staining in >5% of malignant cells in 65% of the tumors. Forty percent of the CUP overexpressed Bcl-2, though no prognostic value was demonstrated. These data imply that the antiapoptotic effect of Bcl-2 may retain biological significance in a subset of CUP, in contrast to other advanced malignancies that lose Bcl-2 expression early in the course of clonal evolution [3].

Tumor Suppressor Genes and Proteins
Tumor suppressor genes encode proteins that suppress malignant transformation, survival, and metastatic dissemination by maintaining the integrity of cellular DNA and by controlling vital cell cycle processes, with the "guardian of the genome," p53, being the most investigated one. IHC studies showed that 55% of advanced malignancies express p53 protein [23]. Data on tumor suppressor gene and protein expression in CUP are summarized in Table 2.

Briasoulis et al. [22] studied IHC p53 expression in 47 patients with CUP and found positive staining in 33 of them (70%). Approximately half of the tumors (53%) exhibited a strong immunoreactivity index for the protein. The expression of p53 protein by itself had no prognostic value for treatment benefit or survival, though co-overexpression of p53 and Bcl-2 (20% of tumors) predicted for a higher response rate to cisplatin-based combination chemotherapy. Van de Wouw et al. [24] also found no prognostic value of IHC p53 detection (present in 23 of 48 tumors, 48%) in a relatively large series of CUP patients from the Netherlands. Bar-Eli et al. [25] investigated the frequency of p53 exon 5–9 mutations in a series of 15 CUP cases and eight cell lines derived from CUP by means of PCR-SSCP. The gene was found mutated in six tumors (26%), suggesting a relatively low frequency of p53 mutations in this group of metastasis-prone, frequently aneuploid tumors. The discordance of IHC and mutational analysis p53 data may be due to the differing specificities of antibodies for wild-type and mutated p53 gene products, the occurrence of metastases outside the exon 5–9 region, and the differential impact of p53-regulating factors such as murine double minute (MDM)-2, p14 alternate reading frame (p14ARF), and p21 [3, 26].

Metastasis-suppressor genes have been recently recognized. They are postulated to modulate the capability of the malignant clone for systemic dissemination, for homing and growth at distant tissues without affecting the malignant transformation process. Up to 14 such genes have been identified to date, with the KiSS-1 gene being the most extensively studied [27]. Loss-of-function of KiSS-1 was seen in several human malignancies and was correlated with advanced stage, high tumor burden, and poor prognosis [28]. We screened 50 cases of CUP for KiSS-1 gene mutations by PCR-SSCP and direct sequencing and found only one mutated sample (242C>G resulting in P81R) [29]. Accordingly, the metastasis-prone behavior of CUP is probably defined by functional deficiency of additional metastasis-suppressor or tumor-suppressor genes other than KiSS-1.

Angiogenesis
Angiogenesis is defined as the process of new vessel formation, and tumors depend on it for growth, survival, and invasion. Emergence of an angiogenic clone that readily disseminates systemically with concurrent involution of the primary site tumor population is an elegant hypothesis for the development of CUP [30]. However, Hillen et al. [31] found no difference in microvessel density between 39 liver metastases from CUP and 30 liver metastases from colon and breast cancer, as both groups showed a highly angiogenic profile. Microvessel density was an adverse prognostic factor for patient outcome in both univariate and multivariate analyses. In a more recent study, van de Wouw et al. [24] found no prognostic value of IHC CD34 (a microvessel density marker) and VEGF-A detection for the outcome of patients with CUP. VEGF was overexpressed in 12 of 46 tumors (26%). Our group studied the expression of CD34, VEGF, and thrombospondin-1 (TSP-1) in paraffin-embedded tissue from 81 patients with CUP [32]. Strong expression of VEGF and TSP-1 was seen, respectively, in 83% and 20% of tumors. Neither of the two proteins was associated with any clinicopathological parameters. Tumor microvessel density was higher in the unfavorable group CUP, compared with the more favorable group, showing a positive correlation with VEGF and a negative one with TSP-1. The same investigators found IHC overexpression of matrix metalloproteinases (MMP)-2 and -9, key effectors of extracellular matrix degradation, in 49% and 36% of 75 CUP cases, respectively [33]. The expression of MMP-2 correlated positively with that of MMP-9, while tissue expression of their inhibitor, tissue inhibitor of metalloproteinases (TIMP)-1, was significantly higher in the unfavorable group than in the favorable group CUP and was associated with a shorter patient survival. Rashid et al. [18] recently observed VEGF expression by IHC in half of 75 CUP, 29% of which overexpressed it. Forty-seven percent of tumors coexpressed VEGF, EGFR, and HER-2 or COX-2, a common expression pattern of genes involved in cellular proliferation and angiogenesis.

The accumulating data (Table 2) draw the picture of a highly active angiogenic profile of CUP metastases, with universal or frequent expression of key regulators such as VEGF, MMPs, and TIMP-1. Although these features are not different from those seen in several advanced malignancies of known primary site, they provide a sound basis for implementation of antiangiogenic therapies coupled with antiproliferative approaches in the clinical setting.

Multigene Expression Profiling
Technological advances during the 1990s enabled the transition from investigating individual genes to the simultaneous assessment of thousands of genes or gene products in a tissue. Complementary DNA (cDNA) or oligonucleotide microarrays, serial analysis of gene expression (SAGE) platforms, and tissue protein microarrays are powerful tools in the quest to explore unidentified subgroups of CUP, to attempt to assign them to primary site groups or to search for an as yet elusive common molecular signature.

Two different concepts underlie the research efforts based on nucleic acid or protein microarray methodology. According to the first, each type of epithelium has a different biological function and therefore expresses specific genes associated with this function. Conservation of this tissue-specific gene-expression profile during carcinogenesis may allow class definition of a CUP case according to primary site. Dennis et al. [34] used 15 publicly available SAGE data libraries and performed hierarchical clustering to demonstrate that the common adenocarcinomas clustered according to their site of origin. Moreover, gene-expression patterns of metastatic tumors more closely resembled those of primary tumors of the same origin than adenocarcinomas from alternative primary sites. These findings suggest that many cancers retain their "tissue of origin" genetic identity throughout metastatic evolution. Su et al. [35] used oligonucleotide arrays to profile the expression of 12,533 genes from 11 tumor types in a sample of 100 primary tumors. Hierarchical clustering showed that some, but not all, common carcinomas grouped according to anatomical sites. A set of 110 genes (10 per tumor type) differentially expressed according to anatomical site succeeded in accurately classifying a validation set of 75 tumors according to tissue and site of origin in 75%–87% of cases. Bloom et al. [36] used both oligonucleotide and spotted cDNA microarrays to profile 78 adenocarcinomas from seven common sites. A set of 400 genes was required for 50 metastatic tumors with known primaries to be classified according to primary site with 84% accuracy. Tothill et al. [37] used a single cDNA microarray platform and managed to profile 229 primary and metastatic tumors representing 14 tumor types with 89% accuracy in training and validation sets. Use of the microarray classified 11 of 13 CUP to a probable anatomic site of origin. Dennis et al. [38] used tissue microarrays to study the expression of 27 proteins in 352 primary adenocarcinomas from seven main sites. A simplified diagnostic panel of 10 peptides and a decision tree were devised, able to provide correct classification of 88% of primary and metastatic tumors of an independent, validation set.

The alternative concept holds that CUP do not have tissue-specific gene-expression profiles according to the site of origin but instead share common multigene expression patterns unique for the development of primary metastatic disease. Ramaswamy et al. [39] used oligonucleotide microarrays and hierarchical clustering methodology to study the expression of 16,063 genes in 144 primary tumors spanning 14 tumor classes. Although leukemias, lymphomas, mesotheliomas, and central nervous system tumors readily grouped according to histology and site of origin, adenocarcinomas did not cluster according to their primary site. Moreover, high-grade tumors could not be accurately classified by multigene expression profiling according to their tissues of origin, indicating that poorly differentiated malignancies have fundamentally distinct gene-expression patterns, instead of simply lacking a few key markers of differentiation. Finally, when Tothill et al. [37] attempted to reduce the number of classifier genes from thousands to 79 in order to incorporate them in a more easily used quantitative reverse transcription (RT)-PCR assay, they were forced to reduce the number of tumor types from 14 to 5, limiting its usefulness.

Ma et al. [40] recently used data from a 22,000-gene microarray and devised a 92-gene RT-PCR assay that classified a validation set of 119 tumors to 32 primary site/histological types with 87% accuracy. In view of the cost, technical, and administrative issues that hinder the routine diagnostic use of microarrays, such simplified easy-to-use assays may provide a useful compromise between what is scientifically available and what is realistically obtainable. Emerging data on nucleic acid and protein microarrays are shown in Table 3.

	CONCLUSIONS

Top Abstract Introduction Distinct Clinical Features Molecular Biology Conclusions Disclosure of Potential... References

 
The identification of pivotal molecular aberrations that characterize CUP is slowly progressing. Accordingly, it is not clear whether CUP is a clinicopathological entity with distinct genetic/phenotypic characteristics or a "by convention" grouping of tumors with unidentified primaries. Profiling of the expression of multiple genes and gene products instead of individual ones is likely to shed more light on the pathophysiology of CUP, as multiclass classification or "hidden signatures" are encoded in complex gene-expression patterns. It is still obscure whether cancer biology is driven by primary tissue-specific gene-expression patterns or by tissue-independent gene-expression profiles that transcend tissue of origin distinctions and are defined by cancer stem cells. If the latter proves right, the term "primary metastatic disease" we coined would be better suited to describe the true nature of CUP. Multigene and tissue protein microarray methodology should be used to compare expression profiles between CUP and known primary tumors, as well as for assigning CUP cases to sites of origin. A comparison of molecular signatures between CUP oriented to a specific primary site by molecular criteria and primary tumors of the relevant sites may disclose important differences. Moreover, microarray technology should be used not solely to assign multigene expression profiles to tissue-of-origin clusters, but also to seek out a tissue of origin–independent, "primary metastatic disease" genetic cluster. Our deepest hope is that the clinical research implementation of genome-wide profiling will elucidate what is really missing in CUP: the primary or the biology?

	DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

Top Abstract Introduction Distinct Clinical Features Molecular Biology Conclusions Disclosure of Potential... References

 
The authors indicate no potential conflicts of interest.

	REFERENCES

Top Abstract Introduction Distinct Clinical Features Molecular Biology Conclusions Disclosure of Potential... References

1. Pavlidis N, Fizazi K. Cancer of unknown primary (CUP). Crit Rev Oncol Hematol 2005;54:243–250.[Medline]
2. Levi F, Te VC, Erler G et al. Epidemiology of unknown primary tumours. Eur J Cancer 2002;38:1810–1812.[CrossRef][Medline]
3. Van de Wouw AJ, Jansen RL, Speel EJ et al. The unknown biology of the unknown primary tumour: A literature review. Ann Oncol 2003;14:191–196.[Abstract/Free Full Text]
4. Pavlidis N, Briasoulis E, Hainsworth J et al. Diagnostic and therapeutic management of cancer of an unknown primary. Eur J Cancer 2003;39:1990–2005.[CrossRef][Medline]
5. Mousseau M, Schaerer R, Lutz JM et al. [Hepatic metastasis of unknown primary site.]. Bull Cancer 1991;78:725–736; French.[Medline]
6. Kolesnikov-Gauthier H, Levy E, Merlet P et al. FDG PET in patients with cancer of an unknown primary. Nucl Med Commun 2005;26:1059–1066.[CrossRef][Medline]
7. Nystrom JS, Weiner JM, Heffelfinger-Juttner J et al. Metastatic and histologic presentations in unknown primary cancer. Semin Oncol 1977;4:53–58.[Medline]
8. 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 1988;148:2035–2039.[Abstract/Free Full Text]
9. Pentheroudakis G, Briasoulis E, Karavasilis V et al. Chemotherapy for patients with two favourable subsets of unknown primary carcinoma: Active, but how effective? Acta Oncol 2005;44:155–160.[CrossRef][Medline]
10. Jallepalli PV, Lengauer C. Chromosome segregation and cancer: Cutting through the mystery. Nat Rev Cancer 2001;1:109–117.[CrossRef][Medline]
11. Hedley DW, Leary JA, Kirsten F. Metastatic adenocarcinoma of unknown primary site: Abnormalities of cellular DNA content and survival. Eur J Cancer Clin Oncol 1985;21:185–189.[CrossRef][Medline]
12. Abbruzzese JL, Lenzi R, Raber MN et al. The biology of unknown primary tumors. Semin Oncol 1993;20:238–243.[Medline]
13. Motzer RJ, Rodriguez E, Reuter VE et al. Genetic analysis as an aid in diagnosis for patients with midline carcinomas of uncertain histologies. J Natl Cancer Inst 1991;83:341–346.[Abstract/Free Full Text]
14. Wessendorf S, Fritz B, Wrobel G et al. Automated screening for genomic imbalances using matrix-based comparative genomic hybridization. Lab Invest 2002;82:47–60.[Medline]
15. Pavlidis N, Briasoulis E, Bai M et al. Overexpression of C-myc, Ras and C-erbB-2 oncoproteins in carcinoma of unknown primary origin. Anticancer Res 1995;15:2563–2567.[Medline]
16. 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;18:632–635.[Abstract/Free Full Text]
17. Fizazi K, Voigt JJ, Lesimple T et al. Carcinoma of unknown primary (CUP): Are the tyrosine kinase receptors HER-2, EGF-R, and c-Kit suitable targets for therapy? Proc Am Soc Clin Oncol 2003;22:3549.
18. Rashid A, Hess KR, Lenzi R et al. Overexpression and prevalence of molecular markers in patients with cancer of unknown primary (CUP). Proc Am Soc Clin Oncol 2005;24:9683.
19. Baselga J, Arteaga CL. Critical update and emerging trends in epidermal growth factor receptor targeting in cancer. J Clin Oncol 2005;23:2445–2459.[Abstract/Free Full Text]
20. Dova L, Georgiou I, Vartholomatos G et al. EGFR and C-KIT/CD117 gene mutational screening and oncoprotein expression in patients with cancer of unknown primary: Implications for molecular pathophysiology and therapy. Eur J Cancer 2005;3:424.
21. Debiec-Rychter M, Dumez H, Judson I et al. Use of c-KIT/PDGFRA mutational analysis to predict the clinical response to imatinib in patients with advanced gastrointestinal stromal tumours entered on phase I and II studies of the EORTC Soft Tissue and Bone Sarcoma Group. Eur J Cancer 2004;40:689–695.[CrossRef][Medline]
22. 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 Cooperative Oncology Group study. Anticancer Res 1998;18:1907–1914.[Medline]
23. Hollstein M, Sidransky D, Vogelstein B et al. p53 mutations in human cancers. Science 1991;253:49–53.[Abstract/Free Full Text]
24. 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 2004;24:297–301.[Abstract/Free Full Text]
25. Bar-Eli M, Abbruzzese JL, Lee-Jackson D et al. p53 gene mutation spectrum in human unknown primary tumors. Anticancer Res 1993;13:1619–1623.[Medline]
26. Vousden KH, Woude GF. The ins and outs of p53. Nat Cell Biol 2000;2:E178–E180.[CrossRef][Medline]
27. Yoshida BA, Sokoloff MM, Welch DR et al. Metastasis-suppressor genes: A review and perspective on an emerging field. J Natl Cancer Inst 2000;92:1717–1730.[Abstract/Free Full Text]
28. Navenot JM, Wang Z, Chopin M et al. Kisspeptin-10-induced signaling of GPR54 negatively regulates chemotactic responses mediated by CXCR4: A potential mechanism for the metastasis suppressor activity of kisspeptins. Cancer Res 2005;65:10450–10456.[Abstract/Free Full Text]
29. Dova L, Golfinopoulos V, Pentheroudakis G et al. The metastatic potential of cancer of unknown primary is independent of the KiSS-1 metastasis-suppressor gene: A study of 50 patients. Forum of Clinical Oncology 2006;1:Abstract EA23.
30. Naresh KN. Do metastatic tumours from an unknown primary reflect angiogenic incompetence of the tumour at the primary site?—a hypothesis. Med Hypotheses 2002;59:357–360.[CrossRef][Medline]
31. Hillen HF, Hak LE, Joosten-Achjanie SR et al. Microvessel density in unknown primary tumors. Int J Cancer 1997;74:81–85.[CrossRef][Medline]
32. Karavasilis V, Malamou-Mitsi V, Briasoulis E et al. Angiogenesis in cancer of unknown primary: Clinicopathological study of CD34, VEGF and TSP-1. BMC Cancer 2005;5:25–32.[CrossRef][Medline]
33. Karavasilis V, Malamou-Mitsi V, Briasoulis E et al. Matrix metalloproteinases in carcinoma of unknown primary. Cancer 2005;104:2282–2287.[CrossRef][Medline]
34. 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 2002;62:5999–6005.[Abstract/Free Full Text]
35. Su AI, Welsh JB, Sapinoso LM et al. Molecular classification of human carcinomas by use of gene expression signatures. Cancer Res 2001;61:7388–7393.[Abstract/Free Full Text]
36. Bloom G, Yang IV, Boulware D et al. Multi-platform, multi-site, microarray-based human tumor classification. Am J Pathol 2004;164:9–16.[Abstract/Free Full Text]
37. 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 2005;65:4031–4040.[Abstract/Free Full Text]
38. 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 2005;11:3766–3772.[Abstract/Free Full Text]
39. Ramaswamy S, Tamayo P, Rifkin R et al. Multiclass cancer diagnosis using tumor gene expression signatures. Proc Natl Acad Sci U S A 2001;98:15149–15154.[Abstract/Free Full Text]
40. 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 2006;130:465–473.[Medline]

This article has been cited by other articles:

				G. Pentheroudakis and N. Pavlidis Probing the unknown in cancer of unknown primary: which way is the right way? Ann. Onc., January 20, 2010; (2010) mdp609v1. [Full Text] [PDF]

This Article Abstract Full Text (PDF) eLetters: Submit a response to this article Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services E-mail this article link to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pentheroudakis, G. Articles by Pavlidis, N. Search for Related Content PubMed PubMed Citation Articles by Pentheroudakis, G. Articles by Pavlidis, N.