Cardiovascular Disease Detection from Multi-View Chest X-rays with BI-Mamba

TL;DR

This work tackles cardiovascular disease risk prediction from chest X-ray, aiming to reduce radiation exposure relative to CT while maintaining predictive accuracy. It introduces BI-Mamba, a bidirectional state-space sequence model that processes concatenated multi-view chest X-ray patches with linear time complexity, enabling effective long-range dependency modeling at high image resolutions. On NLST-derived data from 10,395 subjects, BI-Mamba with early input patch concatenation achieves an AUROC of 0.8243—outperforming ViT-S and ResNet-50 baselines with similar parameter counts—and uses at least 30.5% less GPU memory during training; while the CT-based Tri-2D method reaches 0.8710 AUROC, BI-Mamba demonstrates the strong potential of chest X-ray for safe, low-cost CVD risk assessment. These results underscore the practical value of efficient bidirectional sequence modeling for multi-view medical imaging and offer a promising path toward broader adoption of chest X-ray-based risk stratification in clinical settings.

Abstract

Accurate prediction of Cardiovascular disease (CVD) risk in medical imaging is central to effective patient health management. Previous studies have demonstrated that imaging features in computed tomography (CT) can help predict CVD risk. However, CT entails notable radiation exposure, which may result in adverse health effects for patients. In contrast, chest X-ray emits significantly lower levels of radiation, offering a safer option. This rationale motivates our investigation into the feasibility of using chest X-ray for predicting CVD risk. Convolutional Neural Networks (CNNs) and Transformers are two established network architectures for computer-aided diagnosis. However, they struggle to model very high resolution chest X-ray due to the lack of large context modeling power or quadratic time complexity. Inspired by state space sequence models (SSMs), a new class of network architectures with competitive sequence modeling power as Transfomers and linear time complexity, we propose Bidirectional Image Mamba (BI-Mamba) to complement the unidirectional SSMs with opposite directional information. BI-Mamba utilizes parallel forward and backwark blocks to encode longe-range dependencies of multi-view chest X-rays. We conduct extensive experiments on images from 10,395 subjects in National Lung Screening Trail (NLST). Results show that BI-Mamba outperforms ResNet-50 and ViT-S with comparable parameter size, and saves significant amount of GPU memory during training. Besides, BI-Mamba achieves promising performance compared with previous state of the art in CT, unraveling the potential of chest X-ray for CVD risk prediction.

Figures (2)

• Figure 1: Network architecture of the BI-Mamba model.
• Figure 2: Results on different image resolution and comparison of memory footprint at training time.