Personalized MR-Informed Diffusion Models for 3D PET Image Reconstruction

TL;DR

This paper introduces a method to generate personalized PET images from MR scans using image registration, which enhances diffusion model-based PET reconstruction accuracy, especially in low-count scenarios, without large training datasets.

Contribution

The authors propose a novel approach to synthesize subject-specific PET images from MR data, improving diffusion model-based reconstruction without extensive training data.

Findings

Improved PET reconstruction accuracy with low-count data.

Enhanced preservation of anatomical features in reconstructed images.

Effective use of synthetic data for personalized medical imaging.

Abstract

Recent work has shown improved lesion detectability and flexibility to reconstruction hyperparameters (e.g. scanner geometry or dose level) when PET images are reconstructed by leveraging pre-trained diffusion models. Such methods train a diffusion model (without sinogram data) on high-quality, but still noisy, PET images. In this work, we propose a simple method for generating subject-specific PET images from a dataset of multi-subject PET-MR scans, synthesizing "pseudo-PET" images by transforming between different patients' anatomy using image registration. The images we synthesize retain information from the subject's MR scan, leading to higher resolution and the retention of anatomical features compared to the original set of PET images. With simulated and real [$^{18}$F]FDG datasets, we show that pre-training a personalized diffusion model with subject-specific "pseudo-PET" images improves reconstruction accuracy with low-count data. In particular, the method shows promise in combining information from a guidance MR scan without overly imposing anatomical features, demonstrating an improved trade-off between reconstructing PET-unique image features versus features present in both PET and MR. We believe this approach for generating and utilizing synthetic data has further applications to medical imaging tasks, particularly because patient-specific PET images can be generated without resorting to generative deep learning or large training datasets.