Advancing Lung Disease Diagnosis in 3D CT Scans

TL;DR

This paper introduces a new 3D CT scan analysis model that improves lung disease diagnosis accuracy by focusing on lung regions, using ResNeSt50 for feature extraction, and addressing class imbalance, achieving high F1 scores.

Contribution

The study presents a simple yet effective model that enhances lung disease classification in 3D CT scans through targeted region analysis and advanced feature extraction techniques.

Findings

Achieved a Macro F1 Score of 0.80 on validation data.

Effectively reduces computational cost by removing non-lung regions.

Addresses class imbalance with weighted loss function.

Abstract

To enable more accurate diagnosis of lung disease in chest CT scans, we propose a straightforward yet effective model. Firstly, we analyze the characteristics of 3D CT scans and remove non-lung regions, which helps the model focus on lesion-related areas and reduces computational cost. We adopt ResNeSt50 as a strong feature extractor, and use a weighted cross-entropy loss to mitigate class imbalance, especially for the underrepresented squamous cell carcinoma category. Our model achieves a Macro F1 Score of 0.80 on the validation set of the Fair Disease Diagnosis Challenge, demonstrating its strong performance in distinguishing between different lung conditions.