Correlation via synthesis: end-to-end nodule image generation and radiogenomic map learning based on generative adversarial network

TL;DR

This paper introduces an end-to-end GAN-based approach to synthesize lung cancer images conditioned on gene data, enabling simultaneous generation of images and learning of radiogenomic maps, streamlining traditional multi-step processes.

Contribution

It proposes a novel end-to-end GAN framework that fuses gene expression profiles with image features to generate synthetic images and learn radiogenomic relationships simultaneously.

Findings

Generated images are realistic and high-quality.

The method effectively learns gene-image correlations.

It simplifies the process of radiogenomic map creation.

Abstract

Radiogenomic map linking image features and gene expression profiles is useful for noninvasively identifying molecular properties of a particular type of disease. Conventionally, such map is produced in three separate steps: 1) gene-clustering to "metagenes", 2) image feature extraction, and 3) statistical correlation between metagenes and image features. Each step is independently performed and relies on arbitrary measurements. In this work, we investigate the potential of an end-to-end method fusing gene data with image features to generate synthetic image and learn radiogenomic map simultaneously. To achieve this goal, we develop a generative adversarial network (GAN) conditioned on both background images and gene expression profiles, synthesizing the corresponding image. Image and gene features are fused at different scales to ensure the realism and quality of the synthesized image. We tested our method on non-small cell lung cancer (NSCLC) dataset. Results demonstrate that the proposed method produces realistic synthetic images, and provides a promising way to find gene-image relationship in a holistic end-to-end manner.