OBJECTIVE. The objective of this study was to assess the practical usefulne
ss of the Bosniak classification system for separating surgical from nonsur
gical cystic renal masses in a large number of patients examined with prope
rly performed renal CT. The study included only patients whose scans were t
echnically adequate to allow proper assignment of the lesion to a category.
MATERIALS AND METHODS. The scans of 109 patients were gathered from two lar
ge teaching institutions both prospectively and retrospectively, yielding a
total of 116 analyzable renal cystic lesions. Eighty-two masses were resec
ted from 77 of these patients, retrospectively categorized by two experienc
ed uroradiologists using the Bosniak classification system, and correlated
with pathology reports, A second group of 34 lesions in 32 patients with at
ypical cysts was followed up prospectively for periods ranging from 3 month
s to 10 years.
RESULTS. The results were similar for the two institutions: 15 resected cat
egories I and II lesions were correctly identified as benign, and all 18 ca
tegory IV lesions were malignant, Twenty-nine (59%) of 49 pooled category I
II masses were malignant. No malignancies have been identified in the prosp
ectively monitored group of patients.
CONCLUSION, Our results are compared with earlier, smaller series and suppo
rt those that show that the Bosniak classification system is useful in sepa
rating lesions requiring surgery from those that can be safely followed up,
provided proper CT techniques are used.