The reliability of frozen section diagnosis

Abstract

We undertook a retrospective study of 594 consecutive diagnosis by frozen section. Clinically important discrepancies were found between the frozen section and final diagnoses in 18 (3%), and unimportant discrepancies in 35 (6%). The reasons for the discrepancies were misinterpretation in 22 (41%), the presence of focal lesions in 29 (55%), and technical errors in 2 (4%). There were 4 false positive and 18 false negative diagnoses, giving a sensitivity of 93% and a specificity of 99%. Diagnoses from frozen sections were deferred in 40 cases (7%). These results are in accordance with those of similar studies, and confirm that frozen section is a highly accurate, but not infallible, method of rapid histological diagnosis.