ProstAttention-Net: A deep attention model for prostate cancer segmentation by aggressiveness in MRI scans

TL;DR

ProstAttention-Net is a novel deep learning model that jointly segments prostate and cancer lesions in MRI scans while grading their aggressiveness, aiding non-invasive prostate cancer assessment.

Contribution

The paper introduces a multi-class attention-based network that combines segmentation and grading of prostate cancer in MRI, using zonal priors to improve accuracy.

Findings

Achieved 69% sensitivity at 2.9 false positives per patient for significant lesions.

Demonstrated effective grading of prostate cancer aggressiveness in MRI scans.

Validated on a diverse dataset with cross-validation across multiple scanners.

Abstract

Multiparametric magnetic resonance imaging (mp-MRI) has shown excellent results in the detection of prostate cancer (PCa). However, characterizing prostate lesions aggressiveness in mp-MRI sequences is impossible in clinical practice, and biopsy remains the reference to determine the Gleason score (GS). In this work, we propose a novel end-to-end multi-class network that jointly segments the prostate gland and cancer lesions with GS group grading. After encoding the information on a latent space, the network is separated in two branches: 1) the first branch performs prostate segmentation 2) the second branch uses this zonal prior as an attention gate for the detection and grading of prostate lesions. The model was trained and validated with a 5-fold cross-validation on an heterogeneous series of 219 MRI exams acquired on three different scanners prior prostatectomy. In the free-response receiver operating characteristics (FROC) analysis for clinically significant lesions (defined as GS > 6) detection, our model achieves 69.0% $\pm$14.5% sensitivity at 2.9 false positive per patient on the whole prostate and 70.8% $\pm$14.4% sensitivity at 1.5 false positive when considering the peripheral zone (PZ) only. Regarding the automatic GS group