Frozen section evaluation of margins in radical prostatectomy specimens: a contemporary study and literature review.

The utility of routine frozen section (FS) analysis for margin evaluation during radical prostatectomy (RP) remains controversial. A retrospective search was conducted to identify RPs evaluated by FS over a 5-year period. The potential of FS to discriminate between benign and malignant tissue and to predict final margins was evaluated. During the study period, 71 (12.3%) of 575 cases underwent FS evaluation of margins, generating 192 individual FSs. There were 8 FSs diagnosed as atypical/indeterminate because of significant freezing, crushing, and/or thermal artifacts; 11 as positive for carcinoma; and 173 as benign. Two FSs classified as benign were diagnosed as positive for carcinoma on subsequent permanent section. Frozen sections' sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for diagnosis of prostatic adenocarcinoma were 85%, 100%, 100%, 99%, and 99%, respectively. Overall RP final margin predictive accuracy was 81%. Positive FS was significantly associated with perineural invasion on biopsy and extraprostatic extension and higher stage disease on RP, but not with the overall final margin status. The high FS accuracy supports its use to guide the extent of surgery. However, FS cannot be used to predict the overall final margin status. Recognition of the histological artifacts inherent to the FS procedure is important to ensure appropriate utilization.