Synthesizing High--Value Diffusion-weighted Imaging of the Prostate Using Generative Adversarial Networks.

Purpose: To develop and evaluate a diffusion-weighted imaging (DWI) deep learning framework based on the generative adversarial network (GAN) to generate synthetic high--value ( =1500 sec/mm) DWI (SYN) sets from acquired standard--value ( = 800 sec/mm) DWI (ACQ) and acquired standard--value ( = 1000 sec/mm) DWI (ACQ) sets.

Materials And Methods: This retrospective multicenter study included 395 patients who underwent prostate multiparametric MRI. This cohort was split into internal training (96 patients) and external testing (299 patients) datasets. To create SYN sets from ACQ and ACQ sets, a deep learning model based on GAN (M) was developed by using the internal dataset. M was trained and compared with a conventional model based on the cycle GAN (M). M was further optimized by using denoising and edge-enhancement techniques (optimized version of the M [Opt-M]). The SYN sets were synthesized by using the M and the Opt-M were synthesized by using ACQ and ACQ sets from the external testing dataset. For comparison, traditional calculated ( =1500 sec/mm) DWI (CAL) sets were also obtained. Reader ratings for image quality and prostate cancer detection were performed on the acquired high--value ( = 1500 sec/mm) DWI (ACQ), CAL and SYN sets and the SYN set generated by the Opt-M (Opt-SYN). Wilcoxon signed rank tests were used to compare the readers' scores. A multiple-reader multiple-case receiver operating characteristic curve was used to compare the diagnostic utility of each DWI set.

Results: When compared with the M, the M yielded a lower mean squared difference and higher mean scores for the peak signal-to-noise ratio, structural similarity, and feature similarity ( < .001 for all). Opt-SYN resulted in significantly better image quality ( ≤ .001 for all) and a higher mean area under the curve than ACQ and CAL ( ≤ .042 for all).

Conclusion: A deep learning framework based on GAN is a promising method to synthesize realistic high--value DWI sets with good image quality and accuracy in prostate cancer detection. Prostate Cancer, Abdomen/GI, Diffusion-weighted Imaging, Deep Learning Framework, High Value, Generative Adversarial Networks© RSNA, 2021