Performance Characteristics Of Psa In Patients Undergoing Radical Prostatectomy.

199 Background: The Prostate Cancer Prevention Trial demonstrated that the performance characteristics of PSA are improved for diagnosis of high grade disease. We assessed the performance of PSA for determining total and high grade cancer volume in a radical prostatectomy series.We identified 1307 patients from the Stanford Radical Prostatectomy Database with clinical stage T1c (n=658) and T2 (n=549) disease who had surgery between 1988 and 2002 and detailed morphometric mapping by a single pathologist. Patients with T1c and T2 disease were analyzed separately as the indication for biopsy was different in these two populations (elevated PSA versus abnormal DRE). We generated receiver operating characteristic (ROC) curves for PSA based on total tumor volume and volume of Gleason 4 or 5 disease and compared the areas under the curve (AUC) for different cut points of total and high grade cancer volume.For patients with clinical stage T1c disease, as total and high grade cancer volumes increased, so did the AUC for PSA performance. There was a significant difference in AUC for PSA for total cancer volumes greater than 0.5 cc (AUC=0.62) compared to greater than 2 cc (AUC=0.72, p=0.02), and also between volumes greater than 2 cc (AUC=0.72) and 5 cc (AUC=0.79, p=0.0003, Fig.1). The AUC for PSA performance for volume of high grade disease increased from 0.66, to 0.68, to 0.76 for Gleason 4 or 5 volumes of 0, 0.5, and 3 cc, respectively (Fig. 2). For patients with clinical stage T2 disease, the AUCs for total or high grade cancer volume were not different across cut points, but were overall higher compared to the T1c group.In T1c patients, PSA performance improved in a stepwise fashion with higher total and high grade cancer volumes as evidenced by improved receiver operating characteristics. In T2 patients, no similar trend was observed, although overall PSA performance was improved compared to T1c patients. Prior studies show that PSA performs better for the detection of high grade disease; here, we show that PSA performs better in predicting volume of high grade disease in radical prostatectomy specimens.