Improved $\alpha$-GAN architecture for generating 3D connected volumes with an application to radiosurgery treatment planning

TL;DR

This paper introduces an enhanced 3D $oldsymbol{ extalpha}$-GAN architecture capable of generating realistic, connected 3D medical volumes, including tumor shapes, which can aid in medical data augmentation and treatment planning.

Contribution

The paper presents an improved 3D $oldsymbol{ extalpha}$-GAN with architectural enhancements for generating connected 3D volumes, addressing the challenge of high-dimensional medical data synthesis.

Findings

Successfully generates connected 3D shapes with similar geometry to training data

Produces high-quality 3D tumor volumes and treatment specifications

Implicitly learns data distribution to generate realistic samples

Abstract

Generative Adversarial Networks (GANs) have gained significant attention in several computer vision tasks for generating high-quality synthetic data. Various medical applications including diagnostic imaging and radiation therapy can benefit greatly from synthetic data generation due to data scarcity in the domain. However, medical image data is typically kept in 3D space, and generative models suffer from the curse of dimensionality issues in generating such synthetic data. In this paper, we investigate the potential of GANs for generating connected 3D volumes. We propose an improved version of 3D $\alpha$-GAN by incorporating various architectural enhancements. On a synthetic dataset of connected 3D spheres and ellipsoids, our model can generate fully connected 3D shapes with similar geometrical characteristics to that of training data. We also show that our 3D GAN model can successfully generate high-quality 3D tumor volumes and associated treatment specifications (e.g., isocenter locations). Similar moment invariants to the training data as well as fully connected 3D shapes confirm that improved 3D $\alpha$-GAN implicitly learns the training data distribution, and generates realistic-looking samples. The capability of improved 3D $\alpha$-GAN makes it a valuable source for generating synthetic medical image data that can help future research in this domain.