Conditional Diffusion Model for Longitudinal Medical Image Generation

TL;DR

This paper introduces a diffusion-based model for generating 3D longitudinal medical images from MRI data, effectively handling missing data and irregular follow-ups to improve image quality.

Contribution

The proposed model uniquely incorporates conditioning MRI and time-visit encoding for controlled longitudinal image generation, addressing common issues in medical data.

Findings

Generated images are of higher quality than competing methods.

The model effectively manages missing data and irregular intervals.

It enables controlled change between source and target images.

Abstract

Alzheimers disease progresses slowly and involves complex interaction between various biological factors. Longitudinal medical imaging data can capture this progression over time. However, longitudinal data frequently encounter issues such as missing data due to patient dropouts, irregular follow-up intervals, and varying lengths of observation periods. To address these issues, we designed a diffusion-based model for 3D longitudinal medical imaging generation using single magnetic resonance imaging (MRI). This involves the injection of a conditioning MRI and time-visit encoding to the model, enabling control in change between source and target images. The experimental results indicate that the proposed method generates higher-quality images compared to other competing methods.