Abstract #0584

Quantitative MRI-Driven Deep Learning for Detection of Clinical Significant Prostate Cancer

Shiwen Shen^1,2, Xinran Zhong^1,3, Willam Hsu¹, Alex Bui¹, Holden Wu¹, Michael Kuo¹, Steven Raman¹, Daniel Margolis¹, and Kyunghyun Sung¹

¹Department of Radiological Sciences, University of California, Los Angeles, Los Angeles, CA, United States, ²Department of Bioengineering, University of California, Los Angeles, Los Angeles, CA, United States, ³Physics and Biology in Medicine IDP, University of California, Los Angeles, Los Angeles, CA, United States

We present a novel automatic classification method to distinguish between indolent and clinically significant prostatic carcinoma using multi-parametric MRI (mp-MRI). The main contributions are 1) utilizing state-of-art deep learning method to characterize the lesion in mp-MRI through a pre-trained convolutional neural network model, OverFeat, 2) building a hybrid two-order classification model that combines deep learning and conventional statistical features, and 3) avoiding annotation of the lesion boundaries and anatomical-location-specific training. The proposed method was evaluated using 102 lesions of prostate cancer and achieved significantly higher accuracy than the method with traditional statistical features.

This abstract and the presentation materials are available to members only; a login is required.

Join Here