Study of diagnostic accuracy of Fagan's two-step nomogram in increasing the value of predictive tools for prostate cancer: application of specific spatial distribution of positive/negative bioptic cores to predict extracapsular extension.

BACKGROUND: Prostate cancer (PC) represents the second most frequent cancer in the male population worldwide. It is mandatory to have a very accurate staging to choice the best possible treatment.

AIMS: To test the possibility of improving the performance of Partin's tables in predicting the pathological staging of PC by introducing bioptic parameters through an innovative statistic tool (Fagan's two-step nomogram).

METHODS: We prospectivelly collected data of all 1048 consecutive patients undergoing saturation 24-core transrectal prostate biopsy. Then, in eligible 94 patients, we compared the prediction of presence/absence of extracapsular extension of neoplasm (EPE+/-), with pathological assessment of invasion through (pseudo)capsule in the prostatectomy specimens. Starting from the probability of EPE- (pre-test probability, calculated with formula "100%-risk of EPE+"), we used Fagan's nomogram to examine the diagnostic sensitivity (DSe) and specificity (DSp) of negative "lateral" bioptic cores.

RESULTS: We specifically analyzed the status of "lateral" cores in each side (94 patients × 2 sides = 188 sides). "Lateral" cores were negative in 42.5% of sides (80/188) with a DSe and DSp of 91.7 and 45.4%, respectively. In these sides, the mean probability of EPE+ according to Partin's tables was 21.6%. With Fagan's nomogram, the post-test probability of EPE+ when all "lateral" cores were negative was 14.1%, with a substantial gain of 7.5%.

DISCUSSION: The spatial distribution of bioptic positive cores allowed us to demonstrate the role Fagan's nomogram in increasing the accuracy of already existing, predictive tools for PC.

CONCLUSIONS: This pioneering study may justify the use of the above nomogram in testing "local" predictive parameters in combination with pre-existing nomograms.