Vol. 181, No. 4, Supplement, Monday, April 27, 2009 290 THE JOURNAL OF UROLOGY ® intermediate and high risk groups, respectively; p=0.09). ROC coordinate plot indicated that removal of 28 nodes yields 90% ability to detect LNI in each risk category. Assessment of 10 or fewer nodes is associated with virtually zero probability of finding LNI in each risk group. At MVA, the number of removed nodes predicted LNI in each risk group (p 0.04). CONCLUSIONS: We demonstrated that in all patient categories limited PLND is associated with virtually null staging accuracy. This is confirmed by the evidence that the number of lymph nodes removed was significantly associated with the presence of LNI in patients with low, intermediate and high risk patients. Therefore, if a staging PLND is contemplated, this should be extended, even in the low risk group, where a limited PLND has been often wrongly considered sufficient for staging purposes. Source of Funding: None 812 PATIENTS WITH ORGAN CONFINED PROSTATE CANCER AND POSTITIVE SURGICAL MARGINS HAVE SIMILAR RECURRENCE RATES COMPARED TO PATIENTS WITH EXTRA-CAPSULAR EXTENSION AND NEGATIVE SURGICAL MARGINS. A PLEA FOR STAGE RE-CLASSIFICATION. Alberto Briganti*, Andrea Gallina, Nazareno Suardi, Giuseppe Zanni, Marco Bianchi, Manuela Tutolo, Niccolò Passoni, Umberto Capitanio, Milano, Italy; Pierre I Karakiewicz, Montreal, QCCanada; Patrizio Rigatti, Francesco Montorsi, Milano, Italy INTRODUCTION AND OBJECTIVE: Several studies have shown a negative impact of positive surgical margins (+SM) on biochemical recurrence (BCR) free survival of patients with organ confined (pT2) prostate cancer (PCa). We hypothesized that patients with pT2 and +SM have the same outcome of patients with pT3a PC with negative SM (-SM). METHODS: Between January 1988 and July 2008, 4574 consecutive patients were treated with radical prostatectomy (RP). Only patients with either pT2 (+/-SM ) or pT3a -SM and negative lymph nodes were included. These resulted in 1415 patients with available follow-up data. No patients received any adjuvant therapy. Patients were divided into three groups: pT2 -SM (group 1; n=1114 [78.7%]), pT2 +SM (group 2; n=215 [15.2%]), pT3a R0 (group 3; n=86 [6.1%]). The Kaplan Meier method was used to explore BCR free survival rates at 5,8 and 10 years after surgery. The log-rank test was used to compare the clinical outcome of patients among the groups. Univariable and multivariable Cox regression models tested the association between a new proposed classification (pT2 -SM vs pT2 +SM plus pT3a -SM) and BCR. Finally, the accuracy of the current TNM (pT2a-c vs pT3a) classification in predicting BCR was compared to that of our proposed re-classification (pT2 -SM vs pT2 +SM plus pT3a -SM). RESULTS: Mean follow-up was 52 months (median 41). Overall BCR-free survival rates at 5,8 and 10 years were 90,83 and 80%, respectively. Patients with pT2 + SM PCa had similar BCR free survival rates compared to patients with pT3a -SM PCa (80, 75 and 60% vs 89, 70 and 59%, respectively; p=0.9). Conversely, patients with pT2 SM- disease had significantly higher BCR free survival rates (90, 86 and 80%; p<0.001). At multivariable analyses the new classification (pT2 -SM vs pT2 +SM plus pT3a -SM) was an independent predictor of BCR (p<0.001) while the current TNM was not (p=0.4). Moreover, the accuracy of a multivariable model including pathological Gleason score, PSA and our new staging classification was 74.8% vs 68.9% for a model including PSA, pathological Gleason sum and the current TNM (pT2 vs pT3a; gain:5.9%;p<0.001). CONCLUSIONS: Patients with pT2 -SM have the same outcome of patients with pT3a -SM. Our proposed new staging classification is significantly more accurate than the current TNM in predicting post- operative BCR (5.9% gain; p<0.001). This should be taken into account in prostate cancer prediction models. Source of Funding: None 813 NOT ALL HIGH RISK PROSTATE CANCER PATIENTS HAVE THE SAME RISK OF BIOCHEMICAL RECURRENCE AFTER SURGERY. A MULTI-INSTITUTIONAL ANALYSIS OF PREDICTORS OF GOOD OUTCOME. Alberto Briganti*, Milan, Italy; Patrick J Bastian, Munich, Germany; Felix K h Chun, Hamburg, Germany; Andrea Gallina, Milan, Italy; Sascha A Ahyai, Hamburg, Germany; Michael Seitz, Munich, Germany; Nazareno Suardi, Milan, Italy; Mario Zacharias, Hamburg, Germany; Alexander Buchner, Munich, Germany; Luigi F Da Pozzo, Milan, Italy; Markus Graefen, Hamburg, Germany; Patrizio Rigatti, Milan, Italy; Christian G Stief, Munich, Germany; Hartwig Huland, Hamburg, Germany; Francesco Montorsi, Milan, Italy INTRODUCTION AND OBJECTIVE: High risk patients, defined as clinical T3 stage, PSA values >20 ng/ml or biopsy Gleason sum 8-10, are associated with poor long term outcome. We hypothesized that not all high risk patients have per se a bad prognosis. We thus investigated factors associated with better outcomes among patients with pre- operatively defined high risk prostate cancer. METHODS: Between 1992 and 2008, 1087 patients with high risk prostate cancer were treated with radical prostatectomy (RP) at 3 tertiary referral institutions (n=559, n=410 and n=118, respectibely). Clinical and pathological characteristics were evaluated. Kaplan-Meier analyses targeted time to biochemical recurrence (BCR), according to clinical and pathological characteristics. Univariable and multivariable Cox regression analyses were used to identify clinical and pathological predictors of BCR. RESULTS: Mean pre-operative PSA was 27.2 ng/ml (median 18.9 ng/ml). Clinical stage was T1c in 305 (28.1%), T2 in 296 (27.2%) and T3 in 293 (27.0%) patients, respectively. Biopsy Gleason sum was 2-6 in 315 (29.0%), 7 in 267 (24.6%) and 8-10 in 305 (46.5%) patients, respectively. Final pathology showed pT2 in 377 (34.7%), pT3a in 302 (27.8%), pT3b in 361 (33.2%), pT4 disease in 44 (4.0%) patients. LNI was found in 271 (24.9%) patients. Pathological Gleason sum was 2-6 in 159 (14.6%), 7 in 607 (55.8%) and 8-10 in 321 (29.5%) patients. Interestingly, of 293 clinical T3 prostate cancers, 89 patients (30.5%) showed organ confined disease (pT2) at final pathology. Mean follow up was 46.9 months (median 35). BCR free survival rates at 5,8 and 10 years were 72, 69 and 68%, respectively. At univariable and multivariable analyses biopsy Gleason sum and PSA were the only predictors of BCR free survival (all p 0.04). BCR free survival rates at 5 years were 62%, 74% and 72% in patients with Gleason 8-10, PSA > 20 and clinical stage T3, respectively (p<0.001). Patients with Gleason score 8-10 had lower BCR free survival rates as compared to patients with Gleason 7 and either clinical stage T3 or PSA>20 (71% vs 81 and 80%, respectively; p=0.05). CONCLUSIONS: Not all patients with pre-operative high risk prostate cancer are at the same risk of developing BCR. A sub-cohort of high risk patients have good outcome after surgery. This is represented by patients with biopsy Gleason 7 even in presence of very high PSA levels or clinical T3 prostate cancer. This reinforces the need for a sub- stratification of patients with high risk prostate cancer. Source of Funding: None 814 ISOLATED BLADDER NECK INVOLVEMENT FROM PROSTATE CANCER SHOULD NOT BE CONSIDERED AS PT4 DISEASE Alberto Briganti*, Andrea Gallina, Nazareno Suardi, Marco Bianchi, Niccolò Passoni, Andrea Salonia, Firas Abdollah, Massimo Freschi, Renzo Colombo, Patrizio Rigatti, Francesco Montorsi, Milan, Italy INTRODUCTION AND OBJECTIVE: Currently, bladder neck involvement (+BN) after radical prostatectomy (RP) is considered as pT4 disease. However, no study tested the cancer specific survival (CSS) rates of patients with positive bladder margins compared to patients with truly pT4 disease (namely, invasion of adjacent structures other than seminal vesicle and bladder neck). METHODS: The study included 201 consecutive patients with