Diagnosis of malignant tumor: comparison between clinical and autopsy diagnoses

A review was made of the results of 1,036 consecutive autopsies (average patient age, 75 years; 617 male and 419 female patients; autopsy rate, 40.2%) done at the Hospital of Monfalcone from January 1986 through December 1991. In 457 autopsies (44%) one or more malignant neoplasms were found (single tumor in 382, dual tumor in 69, and triple tumor in six) for a total of 538 tumors. In the corresponding death certificates the clinical diagnosis of malignancy was found in 302 autopsies for a total of 310 tumors (single in 294 and dual in eight). The number of clinically unrecognized malignancies was 228 (42%) and the number of patients with undiagnosed tumors was 155 (34%). The neoplasms were subdivided into four stages according to their extension at autopsy: stage 0 (85 tumors), microscopic neoplasm; stage I (146 tumors), neoplasm localized at the site of origin; stage II (43 tumors), local advanced neoplasm; and stage III (264 tumors), neoplasm with metastases. The rate of correct clinical diagnoses increased according to the level of stage; (stage 0, 1%; stage I, 40%; stage II, 58%; and stage III, 66%). The greatest numbers of tumors found at autopsy were located in the gastroenteric apparatus (170), the respiratory apparatus (149), and the urogenital apparatus (130). The number of clinically undiagnosed neoplasms was high with respect to the urogenital apparatus (77% overall and 51% excluding stage 0) and the gastroenteric apparatus (41% and 38% excluding stage 0); in contrast, the number was low for the respiratory apparatus (17%). The rate of correct diagnoses was higher in patients younger than 65 years (65%) than in those older than 65 (54%), but the difference was not significant. The duration of hospitalization in the 12 months before death was statistically higher (P < .001) in patients with a correct diagnosis (36 days) than in undiagnosed patients (14 days). Even taking into account the biases that affect selection of patients for autopsy, the notable discrepancy found between clinical and autopsy diagnoses underlines the fact that autopsy, despite improvements in diagnostic techniques, maintains its fundamental importance in assessing the reliability of clinical diagnoses and furthermore shows the underestimation of the incidence of tumors in epidemiological studies based solely on death certificates.