Does cumulative prostate cancer length in prostate biopsies improve prediction of clinically insignificant cancer at radical prostatectomy in patients eligible for active surveillance? - Abstract

OBJECTIVES: To evaluate if cumulative cancer length on prostate needle biopsy (Bx) divided by the number of biopsy cores (CCL/core) could improve prediction of insignificant cancer (IC) on radical prostatectomy (RP) in patients with prostate cancer (PCA) eligible for active surveillance (AS).

MATERIALS AND METHODS: Patients diagnosed with PCA on extended (≥10 cores) Bx with initial prostate-specific antigen (iPSA) < 15 ng/mL, clinical stage (cT) ≤ 2a, and highest Bx Gleason score (BxGS) 3+3=6 or 3+4=7 with < 3 positive cores who underwent RP were included in the study. CCL/core and presence of IC (organ-confined, volume < 0.5 mL, RPGS ≤ 6) at RP were recorded. pT2 PCA with RPGS ≤ 3+4=7 and volume < 0.5 mL were categorized as low-volume organ-confined disease (LV-OCD).

RESULTS: 221 patients met the inclusion criteria: mean age was 59 years, median iPSA was 4.5 ng/ml. Clinical stage was cT1 in 86% of cases; BxGS was 3+3=6 in 67% (group 1) and 3+4=7 in 33% of patients (group 2). Maximum percent of Bx core involvement was < 50 in 85%; median CCL/core was 0.15 mm. IC was found in 27% and LV-OCD in 44% of patients. Group 2 was associated with higher number of positive cores, maximum percent core involvement, total PCA length, and CCL/core. Group 1 was more likely to have IC (39%) or LV-OCD (54%) compared to group 2 (3% and 23%, respectively). Group 2 had significantly higher RPGS and pathologic stage. Univariate analysis of group 1 demonstrated that iPSA, maximum percent core involvement, PCA length, and CCL/core were all significantly associated with IC and LV-OCD; for group 2, number of positive cores (1 versus 2) was also significantly associated with LV-OCD. On multivariate logistic regression analysis, maximum percent core involvement < 50, and number of positive cores (1 vs. 2) were independent predictors of IC in group 1; BxGS, maximum percent core involvement < 50 and PCA length < 3 mm or CCL/core < 0.2 mm were all independent predictors of LV-OCD in the whole population. Maximum percent of core involvement < 50 and PCA length < 3 mm or CCL/core < 0.2 mm were also independent predictors of LV-OCD in group 1 patients.

CONCLUSION: In patients eligible for AS, CCL/core < 0.20 mm was significantly associated with IC and LV-OCD. However, when parameters of cancer burden were considered, CCL/core did not independently add any additional value for predicting IC in patients with BxGS6. CCL/core was an independent predictor of LV-OCD in the whole population and in group 1 patients, although the model including PCA length showed slightly higher AUC.

Written by:
Chen DJ, Falzarano SM, McKenney JK, Przybycin CG, Reynolds JP, Roma A, Jones JS, Stephenson A, Klein E, Magi-Galluzzi C.   Are you the author?
Robert J. Tomsich Pathology and Laboratory Medicine Institute, Cleveland Clinic, Cleveland, OH.

Reference: BJU Int. 2014 Jul 25. Epub ahead of print.
doi: 10.1111/bju.12880

PubMed Abstract
PMID: 25060664 Prostate Cancer Section