Introduction

Although prostate biopsies targeted only to MRI detected lesions allow for the detection of clinically significant diseases, they would not result into a complete sampling of the entire prostate. Therefore, lower grade tumors (i.e., non-index lesions) in other areas would not be detected. We hypothesized that the presence of non-index lesions might impact on the risk of adverse outcomes at radical prostatectomy (RP)

Methods

761 PCa patients treated with RP between 2012 and 2016 were identified. All biopsy specimens were re-reviewed by two high-volume dedicated uro-pathologists. The index lesion was defined as the highest-grade core at biopsy. When multiple positive cores were present, the index lesion was defined as higher-grade disease or higher number of positive cores with higher-grade disease from the same location. Non-index lesions were defined as lower grade or lower number of positive cores in other areas. Multivariable logistic regression (MVA) analyses tested the impact of the non-index lesions and of the number of positive non-index lesion cores on the risk of extracapsular extension (ECE), seminal vesicle involvement (SVI), and positive surgical margins (PSM). AUC of the models without information on the presence of non-index lesions were compared with full models using the DeLong method.

Results

Overall, 284 (37.5%), 83 (10.9%), and 145 (19.1%) patients had ECE, SVI, and PSM at final pathology. At MVA, the presence of non-index lesions was a predictor of ECE (Odds ratio [OR]: 2.12; P=0.001), SVI (OR: 2.75; P=0.02), and PSM (OR: 2.16; P=0.01). Similarly, the number of positive cores in the non-index lesion was associated with the risk of ECE (OR: 1.09; P=0.02), SVI (OR: 1.13; P<0.001), and PSM (OR: 1.07; P=0.01). The inclusion of information on non-index lesions improved the accuracy of the model predicting PSM (AUC: 67.0 vs. 69.4%; P=0.04). No differences in the AUCs of the base model and of the model including the presence of non-index lesions were observed for ECE (78.8 vs. 78.6%; P=0.7) and SVI (81.5 vs. 82.1%; P=0.3).

Conclusions

The presence of non-index lesions and the number of positive cores in the non-index lesion represent predictors of ECE, SVI, and PSM. The inclusion of these parameters improves our ability to identify patients at higher risk of PSM. A systematic sample of the prostate provides useful preoperative information on the risk of adverse pathologic outcomes and should be always considered in association with targeted biopsies.

Funding

none