[828] Cumulative Prostate Cancer Length in Prostate Biopsy Cores Improves Prediction of Clinically Insignificant Cancer at Radical Prostatectomy in Patients Eligible for Active Surveillance

Derrick Chen, Sara M Falzarano, Jesse McKenney, Chris Przybycin, Jordan Reynolds, Andres Roma, Eric Klein, Cristina Magi-Galluzzi. Cleveland Clinic, Cleveland, OH

Background: Although the rationale for active surveillance (AS) in patients with prostate cancer (PCA) is well established, eligibility criteria vary significantly across institutions. We evaluated if cumulative cancer length divided by number of biopsy cores (CCL/core) could improve the prediction of insignificant cancer (IC) at radical prostatectomy (RP) in a population eligible for AS.
Design: We searched our database for consecutive patients diagnosed with PCA on extended (≥10 cores) biopsies (Bx) performed between 2010 and 2011. Patients with prostate-specific antigen (PSA) <20 ng/mL, clinical (cT) stage ≤2a, and highest Bx Gleason score (BxGS) ≤7 with <3 positive cores who underwent RP were included in the study. Preoperative features {age, PSA, cT stage, BxGS, D'Amico risk class (low [L] vs. intermediate [I]), # of positive cores, and max % of Bx core involvement} were recorded in an IRB approved database. CCL/core was calculated and presence or absence of IC (organ-confined, volume <0.5 mL, RP GS≤6) at RP was recorded.
Results: 188 patients met the inclusion criteria. GS was 6 in 61% and 7 in 39% of cases. IC was found in 46 (24%) patients, 43 BxGS6 and 3 BxGS7. Preoperative characteristics are summarized in table I. On univariate analysis BxGS (p<0.0001), D'Amico risk class (p<0.0001), 1 vs. 2 positive Bx cores (p=0.002), max % Bx involvement <50 (p<0.0001) and CCL/core (p<0.0001) were significantly associated to IC. All IC had max % Bx involvement <50. Significant variables were combined in multivariate models to predict IC at RP. The best resulting model [BxGS (6 vs. 7), D'Amico risk (L vs. I), # of positive cores (1 vs. 2), max % Bx involvement ≤10%, and CCL/core ≤0.10] could predict IC with a p value of <0.0001 and AUC = 0.82. The only independently significant variable within the model was CCL/core ≤0.10 mm (p<0.03).
Conclusions: CCL/core ≤0.10 mm improves prediction of IC at RP in patients eligible for AS and is a useful parameter for candidate selection in addition to BxGS, D'Amico risk class, number of positive Bx cores, and max % Bx involvement.

Table 1
Age, mean (±SD), years60±6
PSA, median (range), ng/mL4.7 (0.6-17.0)
cT (%)
1c89%
2a11%
GS (%)
661%
739%
D Amico (%)
L58%
I42%
#Bx cores taken, mean (±SD)14 (±3)
#Bx cores + (%)
154%
246%
Max % Bx involvement (%)
Median (range)15 (2.5-95)
<5084%
≥5016%
CCL/core (mm)
Median (range)0.17 (0.02-1.14)
<0.1 (%)35%
≥0.1 (%)65%



Category: Genitourinary (including renal tumors)

Monday, March 4, 2013 1:00 PM

Proffered Papers: Section A, Monday Afternoon

 

Close Window