NEOPLASMS OF UNKNOWN PRIMARY SITE - A CLINICOPATHOLOGICAL STUDY OF AUTOPSIED PATIENTS

Background: Malignant neoplasms of an unknown primary site (NUPS) rema in a diagnostic and therapeutic challenge in clinical practice. With t his in mind, we have reviewed all autopsies performed in patients with NUPS in a single institution. Patients and methods: By reviewing 1656 autopsies performed on adults in our institution (1974-1990), 43 case s of NUPS were found. (NUPS was defined as histologically proven malig n;ant tumor for which a primary site could not be found after anamnesi s, complete physical examination, chest X-ray and routine chemistries. ) Results: There were 24 men (56 %) and 19 women. Mean age was 62 year s (76 % of patients were aged 40-75). Clinical presentation included g eneral deterioration (73 %), digestive symptoms (58 %), liver enlargem ent (58 %) abdominal pain (56 %), respiratory symptoms (45 %), ascites (26 %) and node enlargement (16 %). Abnormalities in analysis and ima ge tests were frequent but nonspecific. Median time from admission to death was 42 days (range, 4-135). Pathologic diagnoses at autopsy were : 23 adenocarcinomas (53 %), arising from pancreas (6), biliary tree ( 6), lung (3), prostate (2), stomach (1), kidney (1) and unknown (4); 3 squamous carcinomas (5 %) (1 renal pelvis, 1 biliary tree, 1 stomach) ; 5 undifferentiated carcinomas (1 lung, 4 unknown); and 12 miscellane ous tumors (including 3 lymphomas, 3 neuroendocrine tumors, 3 hepatoca rcinomas, 2 mesotheliomas and 1 melanoma). There was a tendency toward s a metastatic pattern different from that expected from the primary t umor. Image tests were of little usefulness in the search for the prim ary tumor. Conclusions: 1) Adenocarciomas were the most frequent tumor presenting as NUPS, especially from the pancreas and biliary tree. 2) In this series, at least 11 % of patients were amenable to standard s ystemic therapies (3 lymphomas and 2 prostatic adenocarcinomas) if a c orrect pathologic diagnosis could have been established when alive. 3) Presenting symptoms and metastatic pattern differed from those expect ed for the primary neoplasm eventually found. 4) Image tests were ofte n misleading as regards the primary site, although they were useful to quantify the dissemination of the tumor.