Switching benchmarks in cancer of unknown primary: from autopsy to microarray.

摘要: Abstract Introduction Cancer of unknown primary (CUP) is associated with biology and dismal prognosis. Information on the site origin scant has never been analysed. We systematically reviewed all published evidence CUP identified by two different approaches, either autopsy or microarray gene expression profiling. Methods Published reports identification microarray-based multigene platforms were retrieved analysed for year publication, site, patient age, gender, histology, rate identification, manifestations metastatic deposits, chip technology, training validation sets, mathematical modelling, classification accuracy number classifying genes. Results From 1944 to 2000, a total 884 patients (66% males) underwent in 12 studies after presenting systemic symptoms succumbing their disease. A was 644 (73%) them, mostly lung (27%), pancreas (24%), hepatobiliary tree (8%), kidneys bowel, genital system stomach, as small focus adenocarcinoma poorly differentiated carcinoma. An unpredictable dissemination evident high frequency (46%), nodal (35%), bone (17%), brain (16%) uncommon (18%) deposits. Between 1944–1980 1980–2000 series, female representation increased, ‘undetermined neoplasm’ diagnosis became rarer, pancreatic primaries found less often while colonic ones more frequently. Four using technology profiled than 500 cases classifier set genes (ranging from 10 495) reported strikingly dissimilar frequencies assigned sites (lung 11.5%, 12.5%, bowel 12%, breast 15%, 8%, 6%, 9%, bladder 5%) 75–90% cases. Conclusions Evolution medical imaging diet lifestyle habits probably account changing epidemiology autopsies. Discrepant assignment microarrays may be due presence ‘sanctuary sites’ autopsies, molecular misclassification postulated pro-metastatic genetic signature. In view absence therapeutic prognostic benefit profiling should re-orientated towards unraveling complex pathophysiology metastases.