Letter to the Editor on “Comparative performance of fully-automated and semi-automated artificial intelligence methods for the detection of clinically significant prostate cancer on MRI: a systematic review”

Excerpt

We have read the article entitled “Comparative performance of fully-automated and semi-automated artificial intelligence methods for the detection of clinically significant prostate cancer on MRI: a systematic review” by Sushentsev et al. [1], recently published in Insights into Imaging, which also mentions our recent publication entitled “The primacy of high B-value 3 T-DWI radiomics in the prediction of clinically significant prostate cancer” [2]. In their comparative review, the Authors address several state-of-art research studies employing Magnetic Resonance Imaging (MRI) and exploiting deep learning and machine learning methods for predicting clinically significant prostate cancer (csPCa). Accordingly, our work is cited because we compare the predictive performance achieved with b2000 Diffusion-Weighted Imaging (DWI_b2000) and Apparent Diffusion Coefficient (ADC) MRI sequences to classify csPCa and non-csPCa (ncsPCa), finally stating the primacy of DWI_b2000, that provides by far the best results. …