Accuracy of frozen-section diagnosis in a teaching hospital.

A retrospective quality assurance study of frozen-section diagnoses was performed to determine the source and nature of inaccuracies associated with this procedure over an 18-month period in a large teaching hospital. Of 30,278 surgical pathology specimens accessioned, 1414 (4.7%) had frozen-section examination. Of these, there were five false-positive diagnoses of malignancy (0.4%), 16 false-negative diagnoses of malignancy (1.1%), and 53 deferrals of diagnosis (3.7%). Soft-tissue, breast, and lymph node sites accounted for 12 errors (57%), while central nervous system, breast, and soft-tissue sites were associated with 30 deferrals (57%). Erroneous frozen-section diagnoses were attributed to interpretation (57%), microscopic sampling (24%), gross sampling (9.5%), and lack of communication between pathologist and surgeon (9.5%). Some of these diagnostic errors might have been avoided by changes in procedure or technique.