FD-AA: A Lightweight Focal-Diffuse And Attenuation-Aware Head for Incidental Abdominal Abnormality Detection in Chest CT

2026-09-28

AI Summary

This study addresses the challenge of detecting incidental abdominal abnormalities in chest CT scans, which is complicated by substantial spatial variability and heterogeneous attenuation patterns. To this end, we propose FD-AA, a lightweight organ-aware classification head designed to interface with frozen 3D encoders such as Pillar-0 and MedicalNet. This module incorporates an attenuation-aware mechanism to preserve sparse focal evidence and employs masked generalized mean pooling to capture diffuse abnormality patterns. Furthermore, it explicitly integrates Hounsfield Unit values as physical priors to achieve focal-diffuse feature aggregation. Extensive evaluations on the CT-RATE and RAD-ChestCT datasets demonstrate that the proposed approach achieves state-of-the-art performance, improving the Macro AUC to 0.798 while exhibiting significant generalization gains across different encoder architectures.

Abstract

Routine chest CT captures upper-abdominal structures that may contain clinically relevant incidental abnormalities. Detecting these findings requires feature extraction from organs with different spatial extents and attenuation patterns. We propose FD-AA, a lightweight organ-aware classification head adaptable for frozen 3-D CT encoders. Within each organ, an attenuation-aware module preserves sparse focal evidence, while masked generalized-mean pooling captures diffuse anomaly patterns. In seven abdominal organs, FD-AA with Pillar-0 achieved state-of-the-art (SOTA) performance in both the CT-RATE test set (AUC = 0.798) and the external RAD-ChestCT dataset (AUC = 0.713). More specifically, FD-AA improved macro AUC/AP from 0.763/0.346 to 0.798/0.405 over direct classification using frozen Pillar-0 only (p = 0.034/0.016). Such performance gain generalizes across multiple frozen encoders (AUC improvement on MedicalNet +9.8%, CT-CLIP +14.7%, ResNet +3.7%), demonstrating the effectiveness of FD-AA across different feature representations. These results support the effectiveness of integrating focal-diffuse aggregation with explicit HU evidence for incidental abdominal abnormality detection.

Problem

• Incidental abdominal abnormality detection
• Chest CT
• Upper-abdominal structures
• Attenuation patterns
• Feature extraction

Innovation

• Focal-Diffuse Aggregation
• Attenuation-Aware Module
• Lightweight Classification Head
• Incidental Abnormality Detection
• Masked Generalized-Mean Pooling