Faint Features Tell: Automatic Vertebrae Fracture Screening Assisted by Contrastive Learning

TL;DR

This paper introduces a supervised contrastive learning approach to improve the detection of faint vertebral fractures in CT scans, significantly enhancing accuracy by emphasizing subtle features.

Contribution

It proposes a novel contrastive learning-based model for vertebral fracture classification and provides a new annotated dataset for this task.

Findings

Achieved 99% specificity and 85% sensitivity in binary classification.

Macro-F1 score of 77% in multi-class fracture grading.

Contrastive learning improves detection of subtle fracture features.

Abstract

Long-term vertebral fractures severely affect the life quality of patients, causing kyphotic, lumbar deformity and even paralysis. Computed tomography (CT) is a common clinical examination to screen for this disease at early stages. However, the faint radiological appearances and unspecific symptoms lead to a high risk of missed diagnosis, especially for the mild vertebral fractures. In this paper, we argue that reinforcing the faint fracture features to encourage the inter-class separability is the key to improving the accuracy. Motivated by this, we propose a supervised contrastive learning based model to estimate Genent's Grade of vertebral fracture with CT scans. The supervised contrastive learning, as an auxiliary task, narrows the distance of features within the same class while pushing others away, enhancing the model's capability of capturing subtle features of vertebral fractures. Our method has a specificity of 99% and a sensitivity of 85% in binary classification, and a macro-F1 of 77% in multi-class classification, indicating that contrastive learning significantly improves the accuracy of vertebrae fracture screening. Considering the lack of datasets in this field, we construct a database including 208 samples annotated by experienced radiologists. Our desensitized data and codes will be made publicly available for the community.