Most solid tumors are diagnosed in their organ of origin; however, approximately 20% of patients will present with a tumor identified in one or more metastatic sites. In the overwhelming majority of cases, a clinical history, physical examination, laboratory tests, functional and radiographic imaging (positron emission tomography/computed tomography), and histologic assessment will disclose the primary site, enabling site-directed chemotherapy. Yet, again, in approximately 20% of cases, the primary site eludes determination, even after examination of broad panels of immunohistochemical assays. These cancers of unknown primary organ (CUP), defined as metastatic cancers whose anatomic origin is clinically not detectable even after a thorough diagnostic evaluation, represent a heterogeneous group of malignancies and account for approximately 4% of cancer diagnoses (1,2). Interestingly, in our experience, even after postmortem examination, 20% of CUPs, or about 1% of all cancers, are never anatomically defined. Although these cancers present as metastases and represent a spectrum of biological behavior, oncologists have stratified them into favorable (approximately 20%) and poor (approximately 80%) prognostic groups based on such factors as clinical presentation, host factors, tumor histology, number and location of metastatic sites, and their sensitivity to chemoradiation treatment (1–3). In general, patients with CUP have an overall survival of 6 to 9 months, although the favorable prognostic group may have a median survival of nearly The pathologic diagnoses of CUPs in metastatic sites tend to be carcinomas, of which the majority are adenocarcinomas. The initial diagnostic approach seeks to exclude atypical but benign reactive process and then classify the malignancy as a carcinoma or other malignancies such as sarcomas, lymphomas, and melanomas. An immunohistochemical (IHC) panel can separate the majority of these tumor types. Histopathologic features combined with histochemical mucin stains usually permit distinction between adenocarcinomas and other cancers, namely, squamous cell carcinomas, poorly differentiated carcinomas, germ cell cancers, neuroendocrine carcinomas, and the occasional mesotheliomas that mimic a sarcoma or adenocarcinoma. The increasing versatility of IHC panels based on the pairwise findings of CK7 and CK20, and the more refined organ-specific panels, have increased the pathologist’s ability to narrow the field of primary organ sites of cancer or in some cases to readily define it (1,3–5). Tumors with unique immunohistochemical signatures can provide high probability of primary sites such as the panel of CK7–, CK20+, CDX2+ for colorectal adenocarcinomas, and CK7+, CK20–, TTF-1+, Napsin-A+ for pulmonary adenocarcinomas. As more organ-specific and therapy-directed IHC markers (such as human epidermal growth factor receptor-2 and estrogen receptor) are introduced and validated, they should increasingly enable CUPs to be better defined and managed. Despite these advances, as a result of tumor heterogeneity, IHC only supports a differential diagnosis, and additional and alternative modalities for primary site designation are clearly needed. In this regard, several attempts at molecular profiling have been initiated that have focused on multiple gene expression profiles or microRNA signatures using sequencing, reverse-transcription polymerase chain reaction (RT-PCR), or microarray platform technologies. Some reports have claimed an 80% overall accuracy at organ-specific identification based on a focused group of tumor types (6–11).Commercial assays and reference laboratories advertise to help the oncologist in this endeavor and support the hypothesis that identification of the primary site will focus therapy and improve clinical outcome. In this issue of the Journal, Greco and colleagues at the Sarah Cannon Research Institute in Nashville, Tennessee, have contributed to the clinical laboratory science of organ site prediction of CUPs based on a molecular profiling approach that complements the results of diagnostic pathology (12). Much of the work and the thrust of molecular tumor profiling (MTP) technology have already been published with the demonstration
Read full abstract
Jun 27, 2017
Laura Macías-García + 5
Open Access
