The purpose of this study was to develop an automatic method for the identification and segmentation of clinically significant prostate cancer in low-risk patients and evaluate this performance in a routine clinical setting. The authors discovered that the proposed deep learning computer-aided method showed promising results in the previously-mentioned identification and segmentation of clinically significant prostate cancer in patients on active surveillance.
- Clinically significant prostate cancer identification and segmentation on multi-parametric MRI is feasible in low-risk patients using a deep neural network.
- The deep neural network for significant prostate cancer localization performs better for lesions with larger volumes sizes (> 0.5 cc) as compared to small lesions (> 0.03 cc).
- For the evaluation of automatic prostate cancer segmentation methods in the active surveillance cohort, the large discordance group (MRI positive, targeted biopsy negative) should be included.
Authors: Muhammad Arif, Ivo G. Schoots, Jose Castillo Tovar, Chris H. Bangma, Gabriel P. Krestin, Monique J. Roobol, Wiro Niessen & Jifke F. Veenland