Improving Reproducibility and Performance of Radiomics in Low Dose CT using Cycle GANs

TL;DR

This study demonstrates that cycle GANs can effectively denoise low dose CT scans, enhancing radiomics reproducibility and survival prediction performance, even when trained on unpaired datasets, thus addressing data collection challenges.

Contribution

It introduces a novel approach using cycle GANs trained on unpaired data to improve radiomics in low dose CT, outperforming traditional models in reproducibility and predictive accuracy.

Findings

Cycle GANs improved radiomic feature concordance from 0.87 to 0.93.

Survival prediction AUC increased from 0.52 to 0.59.

Unpaired training achieved comparable results to paired data models.

Abstract

As a means to extract biomarkers from medical imaging, radiomics has attracted increased attention from researchers. However, reproducibility and performance of radiomics in low dose CT scans are still poor, mostly due to noise. Deep learning generative models can be used to denoise these images and in turn improve radiomics' reproducibility and performance. However, most generative models are trained on paired data, which can be difficult or impossible to collect. In this article, we investigate the possibility of denoising low dose CTs using cycle generative adversarial networks (GANs) to improve radiomics reproducibility and performance based on unpaired datasets. Two cycle GANs were trained: 1) from paired data, by simulating low dose CTs (i.e., introducing noise) from high dose CTs; and 2) from unpaired real low dose CTs. To accelerate convergence, during GAN training, a slice-paired training strategy was introduced. The trained GANs were applied to three scenarios: 1) improving radiomics reproducibility in simulated low dose CT images and 2) same-day repeat low dose CTs (RIDER dataset) and 3) improving radiomics performance in survival prediction. Cycle GAN results were compared with a conditional GAN (CGAN) and an encoder-decoder network (EDN) trained on simulated paired data.The cycle GAN trained on simulated data improved concordance correlation coefficients (CCC) of radiomic features from 0.87 to 0.93 on simulated noise CT and from 0.89 to 0.92 on RIDER dataset, as well improving the AUC of survival prediction from 0.52 to 0.59. The cycle GAN trained on real data increased the CCCs of features in RIDER to 0.95 and the AUC of survival prediction to 0.58. The results show that cycle GANs trained on both simulated and real data can improve radiomics' reproducibility and performance in low dose CT and achieve similar results compared to CGANs and EDNs.