Pulse Sequence Dependence of a Simple and Interpretable Deep Learning Method for Detection of Clinically Significant Prostate Cancer Using Multiparametric MRI

Rationale and Objectives: Multiparametric magnetic resonance imaging (mpMRI) is increasingly used for risk stratification and localiza-tion of prostate cancer (PCa). Thanks to the great success of deep learning models in computer vision, the potential application for early detection of PCa using mpMRI is imminent.Materials and Methods: Deep learning analysis of the PROSTATEx dataset.Results: In this study, we show a simple convolutional neural network (CNN) with mpMRI can achieve high performance for detection of clinically significant PCa (csPCa), depending on the pulse sequences used. The mpMRI model with T2-ADC-DWI achieved 0.90 AUC score in the held-out test set, not significantly better than the model using Ktrans instead of DWI (AUC 0.89). Interestingly, the model incor-porating T2-ADC-Ktrans better estimates grade. We also describe a saliency heat map. Our results show that csPCa detection models with mpMRI may be leveraged to guide clinical management strategies.Conclusion: Convolutional neural networks incorporating multiple pulse sequences show high performance for detection of clinically-sig-nificant prostate cancer, and the model including dynamic contrast-enhanced information correlates best with grade.

Automated summary

By utilizing multiparametric magnetic resonance imaging (mpMRI) combined with deep learning models, early detection and localization of prostate cancer can be achieved. The mpMRI model with T2-ADC-DWI sequence achieved a high AUC score of 0.90 in the test set, slightly outperforming the model using Ktrans instead of DWI. The study demonstrates that convolutional neural networks incorporating multiple pulse sequences show high performance for detecting clinically significant prostate cancer.