A retrospective review of 470 consecutive frozen section consultations from two Venetian hospitals (O.C.R. and O.A.M.) was performed in order to assess the accuracy of the procedure. The overall accuracy was 89.4% with a cumulative error rate of 3.6% (false positive + false negative + different diagnosis). Such values are in accord with that reported in literature with the exception of deferred diagnosis whose rate was 7.0% of total diagnosis. It was examined with criticism the latter aspect, focusing on problems and limitations of the method and suggesting a system of quality control by means of periodic review of cases.