Conditional probabilistic diffusion model driven synthetic radiogenomic applications in breast cancer.

This study addresses the heterogeneity of breast cancer (BC) by employing a Conditional Probabilistic Diffusion Model (CPDM) to synthesize Magnetic Resonance Images (MRIs) based on multi-omic data, including gene expression, copy number variation, and DNA methylation. The lack of paired medical images and genomics data in previous studies presented a challenge, which the CPDM aims to overcome. The well-trained CPDM successfully generated synthetic MRIs for 726 TCGA-BRCA patients, who lacked actual MRIs, using their multi-omic profiles. Evaluation metrics such as Frechet's Inception Distance (FID), Mean Square Error (MSE), and Structural Similarity Index Measure (SSIM) demonstrated the CPDM's effectiveness, with an FID of 2.02, MSE of 0.02, and SSIM of 0.59 based on the 15-fold cross-validation. The synthetic MRIs were used to predict clinical attributes, achieving an Area Under the Receiver-Operating-Characteristic curve (AUROC) of 0.82 and an Area Under the Precision-Recall Curve (AUPRC) of 0.84 for predicting ER+/HER2+ subtypes. Additionally, the MRIs accurately predicted BC patient survival with a Concordance-index (C-index) score of 0.88, outperforming other baseline models. This research demonstrates the potential of CPDMs in generating MRIs based on BC patients' genomic profiles, offering valuable insights for radiogenomic research and advancements in precision medicine. The study provides a novel approach to understanding BC heterogeneity for early detection and personalized treatment. Author summary: Breast cancer (BC) is known for its diverse characteristics, which makes it crucial for early detection and personalized treatment. Combining medical images with genomics provides a fresh approach to studying this diversity, leading to the emergence of a new field called radiogenomics. However, when these two data types (image data and genomic data) are not paired, it becomes a challenge. This study proposes the use of a well-trained Conditional Probabilistic Diffusion Model (CPDM) to address this issue by generating BC medical images based on genomic information. CPDM is a type of advanced Artificial Intelligence (AI)-based generative model, like ChatGPT. CPDM has been very successful in creating images that look real. In this study, we built and trained a CPDM specifically for BC. The well-trained CPDM can generate BC medical images very well and the generated images can accurately predict patients' clinical attributes like gene mutations, receptor statuses, and survival probabilities. This research explores the potential of using CPDMs to generate meaningful medical images from genomic data, aiding in solving crucial clinical problems. These findings have implications for advancing radiogenomic research and the development of personalized medicine approaches using AI. [ABSTRACT FROM AUTHOR]