Predicting clinically significant prostate cancer with a deep learning approach: a multicentre retrospective study

Abstract

Purpose This study aimed to develop deep learning (DL) models based on multicentre biparametric magnetic resonance imaging (bpMRI) for the diagnosis of clinically significant prostate cancer (csPCa) and compare the performance of these models with that of the Prostate Imaging and Reporting and Data System (PI-RADS) assessment by expert radiologists based on multiparametric MRI (mpMRI). Methods We included 1861 consecutive male patients who underwent radical prostatectomy or biopsy at seven hospitals with mpMRI. These patients were divided into the training (1216 patients in three hospitals) and external validation cohorts (645 patients in four hospitals). PI-RADS assessment was performed by expert radiologists. We developed DL models for the classification between benign and malignant lesions (DL-BM) and that between csPCa and non-csPCa (DL-CS). An integrated model combining PI-RADS and the DL-CS model, abbreviated as PIDL-CS, was developed. The performances of the DL models and PIDL-CS were compared with that of PI-RADS. Results In each external validation cohort, the area under the receiver operating characteristic curve (AUC) values of the DL-BM and DL-CS models were not significantly different from that of PI-RADS (P > 0.05), whereas the AUC of PIDL-CS was superior to that of PI-RADS (P < 0.05), except for one external validation cohort (P > 0.05). The specificity of PIDL-CS for the detection of csPCa was much higher than that of PI-RADS (P < 0.05). Conclusion Our proposed DL models can be a potential non-invasive auxiliary tool for predicting csPCa. Furthermore, PIDL-CS greatly increased the specificity of csPCa detection compared with PI-RADS assessment by expert radiologists, greatly reducing unnecessary biopsies and helping radiologists achieve a precise diagnosis of csPCa.

Detecting Adverse Pathology of Prostate Cancer With a Deep Learning Approach Based on a 3D Swin‐Transformer Model and Biparametric MRI: A Multicenter Retrospective Study

“…For the diagnosis based on medical images, compared to the radiologists' interpretation, the main advantage of deep learning models is that they can not only mine the subtle and even invisible image features, 17,33 but also analyze the features concurrently using a multivariable method, therefore achieving higher sensitivity and specificity. 17 The higher performance of TransCL relative to radiologists' interpretation in the current study was consistent with previous studies on the diagnosis of PCa 16,23,34 and other tumors. [35][36][37][38] Thus, the proposed deep learning models, particularly TransCL which combined TransNet signature and the clinical characteristics, have potential to aid radiologists in the pre-operation diagnosis of AP presence.…”

“…Furthermore, some subtle, and even invisible features (eg, textural features, advanced features), that are also associated with PCa aggressiveness and progression may be missed. [14][15][16] Deep learning is an artificial intelligence (AI) methodology that can convert medical images into high-dimensional, mineable, and quantitative features that are difficult to detect through visual assessment. 14,17 Using the high-throughput and deeply mined features as input information, a deep learning model can output a quantitative score indicating the risk of adverse outcome.…”

“…Deep learning has been widely used to assist in the diagnosis and treatment decision-making of PCa. 16,[19][20][21][22] However, deep learning methods for detecting AP presence are lacking, other than one recent study that focused on the detection of ECE. 23 Thus, the aim of this study was to develop and evaluate a deep learning model to detect the presence of AP (i.e., ECE, SVI, or PSM) based on biparametric MRI (bpMRI), which included T2WI, DWI, and apparent diffusion coefficient (ADC) images.…”

Effectiveness and Accuracy of MRI‐Ultrasound Fusion Targeted Biopsy Based on PI‐RADS v2.1 Category in Transition/Peripheral Zone of the Prostate

“…Additionally, for the diagnosis in TZ of this study, MRI‐TBx detected 23% cases missed by TRUS‐SBx in which 68% were csPCa, and in 22% cases Gleason score was upgraded by MRI‐TBx. Previous research has demonstrated that MRI‐TBx could detect PCa with a higher Gleason score compared to TRUS‐SBx due to the most suspicious lesions precisely identified by MRI 23–25 . It was shown that prebiopsy MRI scan could guide a more effective biopsy for csPCa and thus was recommended by many guidelines 10,26 .…”

“…Previous research has demonstrated that MRI-TBx could detect PCa with a higher Gleason score compared to TRUS-SBx due to the most suspicious lesions precisely identified by MRI. [23][24][25] It was shown that prebiopsy MRI scan could guide a more effective biopsy for csPCa and thus was recommended by many guidelines. 10,26 However, 16% (13/83) prostate cancers were only detected by TRUS-SBx in which 69% (9/13) were clinically significant.…”