Synthesis of Positron Emission Tomography (PET) Images via Multi-channel Generative Adversarial Networks (GANs)

TL;DR

This paper introduces a multi-channel GAN approach for synthesizing high-resolution PET images, leveraging high-level semantic features and multi-modal inputs to improve realism and clinical relevance.

Contribution

The proposed M-GAN method effectively synthesizes PET images using semantic labels and CT data, addressing low-resolution issues of prior methods with high-level feature representation.

Findings

M-GAN produces PET images closer to real scans than existing methods.

The method effectively uses annotations and CT images to guide synthesis.

Results on 50 lung cancer studies demonstrate improved accuracy.

Abstract

Positron emission tomography (PET) image synthesis plays an important role, which can be used to boost the training data for computer aided diagnosis systems. However, existing image synthesis methods have problems in synthesizing the low resolution PET images. To address these limitations, we propose multi-channel generative adversarial networks (M-GAN) based PET image synthesis method. Different to the existing methods which rely on using low-level features, the proposed M-GAN is capable to represent the features in a high-level of semantic based on the adversarial learning concept. In addition, M-GAN enables to take the input from the annotation (label) to synthesize the high uptake regions e.g., tumors and from the computed tomography (CT) images to constrain the appearance consistency and output the synthetic PET images directly. Our results on 50 lung cancer PET-CT studies indicate that our method was much closer to the real PET images when compared with the existing methods.