Prospective analysis of intraoperative frozen needle biopsy of solid renal masses in adults

We prospectively determined the accuracy of intraoperative needle biopsy of solid renal masses.

A total of 103 patients diagnosed with a solid renal mass and scheduled for surgery were prospectively evaluated. Radical or partial nephrectomy was performed, and biopsy of the surgical specimen was done twice through the tumor using an 18 gauge biopsy gun. Biopsy specimens of 106 tumors were sent for frozen sectioning, stained with hematoxylin and eosin, and reviewed by 2 independent pathologists blinded to each other and whole tissue specimens. Biopsy results were compared to whole tissue specimens.

Specimens were obtained from 60 radical and 46 partial nephrectomy cases. Malignant neoplasms were present in 91 cases (86%). Overall, 15 cases (14%) were benign, of which 11 were oncocytomas. If lesions 4 cm. or less only were included in analysis, the incidence of benign lesions increased to 22%. Overall accuracy of the 2 pathologists was 76 and 80%. Nondiagnostic rates were 11 and 17%. Both observers incorrectly diagnosed 4 malignant lesions (5%) as benign, and incorrectly diagnosed 3 and 5 benign lesions (21 and 36%), respectively, as malignant. Analysis of values for both observers yielded a sensitivity of 77 and 84%, specificity 60 and 73%, positive predictive value 94 and 96%, and negative predictive value 69 and 73%.

Overall frozen needle biopsy was accurate in more than 75% of cases and showed an excellent positive predictive value for carcinoma of more than 94%. Unfortunately, there was a large degree of inaccuracy for benign lesions and we do not recommend the routine use of intraoperative frozen needle biopsy to guide surgical decision making.