Generative Adversarial Networks Can Create High Quality Artificial Prostate Cancer Magnetic Resonance Images

Abstract

Purpose: Recent integration of open-source data to machine learning models, especially in the medical field, has opened new doors to study disease progression and/or regression. However, the limitation of using medical data for machine learning approaches is the specificity of data to a particular medical condition. In this context, most recent technologies like generative adversarial networks (GAN) could be used to generate high quality synthetic data that preserves the clinical variability. Materials and Methods: In this study, we used 139 T2-weighted prostate magnetic resonant images (MRI) from various sources as training data for Single Natural Image GAN (SinGAN), to make a generative model. A deep learning semantic segmentation pipeline trained the model to segment the prostate boundary on 2D MRI slices. Synthetic images with a high-level segmentation boundary of the prostate were filtered and used in the quality control assessment by participating scientists with varying degree of experience (more than 10 years, 1 year, or no experience) to work with MRI images. Results: The most experienced participating group correctly identified conventional vs synthetic images with 67% accuracy, the group with 1 year of experience correctly identified the images with 58% accuracy, and group with no prior experience reached 50% accuracy. Nearly half (47%) of the synthetic images were mistakenly evaluated as conventional images. Interestingly, a blinded quality assessment by a board-certified radiologist to differentiate conventional and synthetic images was not significantly different in context of the mean quality of synthetic and conventional images. Conclusions: This study shows promise that high quality synthetic images from MRI can be generated using GAN. Such an AI model may contribute significantly to various clinical applications which involves supervised machine learning approaches.

Picture Perfect: The Status of Image Quality in Prostate MRI

“…In addition, they also found that for younger radiologists, this resulted in an improved ability to detect PCa. However, Xu and colleagues 105 found that a blinded radiologist did not find the quality of their synthetic images to be inferior to those non‐processed.…”

“…Deep learning has also been used in the emerging field of synthetic MRI, where generative adversial models create images based on acquired data 105 . Hu and colleagues 106 compared acquired DWI images to those modified by their model and found improved image quality with reduced distortion and artifacts.…”

Mitigating annotation shift in cancer classification using single-image generative models

“…Given its potential and its successes in medical image synthesis 3,[13][14][15][16] and, particularly, lesion region-of-interest generation, 15 we adopt the SinGAN architecture 17 to generate multiple synthetic images from a single training image. SinGAN is a multi-scale generative adversarial network (GAN) 18 containing a generator-discriminator pair in each scale.…”

Synthetic Histology Images for Training AI Models: A Novel Approach to Improve Prostate Cancer Diagnosis

“…In the context of PCa, these algorithms can be trained on data from tissue samples, medical images, and other clinical information to identify patterns and features associated with the disease [62][63][64][65][66][67][68][69][70][71][72] . Machine learning has proven to be particularly effective in the automated analysis of medical images, including MRI scans and biopsy slides [73][74][75][76][77][78][79][80][81][82] . These algorithms can accurately detect suspicious areas of the prostate and provide a more precise diagnosis than traditional manual analysis.…”