Improving nonalcoholic fatty liver disease classification performance with latent diffusion models

Synthetic Breast Ultrasound Images: A Study to Overcome Medical Data Sharing Barriers

The vast potential of medical big data to enhance healthcare outcomes remains underutilized due to privacy concerns, which restrict cross-center data sharing and the construction of diverse, large-scale datasets. To address this challenge, we developed a deep generative model aimed at synthesizing medical data to overcome data sharing barriers, with a focus on breast ultrasound (US) image synthesis. Specifically, we introduce CoLDiT, a conditional latent diffusion model with a transformer backbone, to generate US images of breast lesions across various Breast Imaging Reporting and Data System (BI-RADS) categories. Using a training dataset of 9,705 US images from 5,243 patients across 202 hospitals with diverse US systems, CoLDiT generated breast US images without duplicating private information, as confirmed through nearest-neighbor analysis. Blinded reader studies further validated the realism of these images, with area under the receiver operating characteristic curve (AUC) scores ranging from 0.53 to 0.77. Additionally, synthetic breast US images effectively augmented the training dataset for BI-RADS classification, achieving performance comparable to that using an equal-sized training set comprising solely real images (P = 0.81 for AUC). Our findings suggest that synthetic data, such as CoLDiT-generated images, offer a viable, privacy-preserving solution to facilitate secure medical data sharing and advance the utilization of medical big data.

Addressing Class Imbalance with Latent Diffusion-based Data Augmentation for Improving Disease Classification in Pediatric Chest X-rays

Deep learning (DL) has transformed medical image classification; however, its efficacy is often limited by significant data imbalance due to far fewer cases (minority class) compared to controls (majority class). It has been shown that synthetic image augmentation techniques can simulate clinical variability, leading to enhanced model performance. We hypothesize that they could also mitigate the challenge of data imbalance, thereby addressing overfitting to the majority class and enhancing generalization. Recently, latent diffusion models (LDMs) have shown promise in synthesizing high-quality medical images. This study evaluates the effectiveness of a text-guided image-to-image LDM in synthesizing disease-positive chest X-rays (CXRs) and augmenting a pediatric CXR dataset to improve classification performance. We first establish baseline performance by fine-tuning an ImageNet-pretrained Inception-V3 model on class-imbalanced data for two tasks-normal vs. pneumonia and normal vs. bronchopneumonia. Next, we fine-tune individual text-guided image-to-image LDMs to generate CXRs showing signs of pneumonia and bronchopneumonia. The Inception-V3 model is retrained on an updated data set that includes these synthesized images as part of augmented training and validation sets. Classification performance is compared using balanced accuracy, sensitivity, specificity, F-score, Matthews correlation coefficient (MCC), Kappa, and Youden's index against the baseline performance. Results show that the augmentation significantly improves Youden's index (p<0.05) and markedly enhances other metrics, indicating that data augmentation using LDM-synthesized images is an effective strategy for addressing class imbalance in medical image classification.

Improving nonalcoholic fatty liver disease classification performance with latent diffusion models

Integrating deep learning with clinical expertise holds great potential for addressing healthcare challenges and empowering medical professionals with improved diagnostic tools. However, the need for annotated medical images is often an obstacle to leveraging the full power of machine learning models. Our research demonstrates that by combining synthetic images, generated using diffusion models, with real images, we can enhance nonalcoholic fatty liver disease (NAFLD) classification performance even in low-data regime settings. We evaluate the quality of the synthetic images by comparing two metrics: Inception Score (IS) and Fréchet Inception Distance (FID), computed on diffusion- and generative adversarial network (GAN)-generated images. Our results show superior performance for the diffusion-generated images, with a maximum IS score of 1.90 compared to 1.67 for GANs, and a minimum FID score of 69.45 compared to 100.05 for GANs. Utilizing a partially frozen CNN backbone (EfficientNet v1), our synthetic augmentation method achieves a maximum image-level ROC AUC of 0.904 on a NAFLD prediction task.