ICon: Improving Inter-Report Consistency in Radiology Report Generation via Lesion-aware Mixup Augmentation

1The Hong Kong Polytechnic University (PolyU), Hong Kong, China
2Southern University of Science and Technology (SUSTech), Shenzhen, China
EMNLP 2024 Findings

*Indicates Equal Contribution

Abstract

Previous research on radiology report generation has made significant progress in terms of increasing the clinical accuracy of generated reports. In this paper, we emphasize another crucial quality that it should possess, i.e., inter-report consistency, which refers to the capability of generating consistent reports for semantically equivalent radiographs. This quality is even of greater significance than the overall report accuracy in terms of ensuring the system's credibility, as a system prone to providing conflicting results would severely erode users' trust. Regrettably, existing approaches struggle to maintain inter-report consistency, exhibiting biases towards common patterns and susceptibility to lesion variants. To address this issue, we propose ICON, which improves the inter-report consistency of radiology report generation. Aiming to enhance the system's ability to capture similarities in semantically equivalent lesions, our approach first involves extracting lesions from input images and examining their characteristics. Then, we introduce a lesion-aware mixup technique to ensure that the representations of the semantically equivalent lesions align with the same attributes, achieved through a linear combination during the training phase. Extensive experiments on three publicly available chest X-ray datasets verify the effectiveness of our approach, both in terms of improving the consistency and accuracy of the generated reports.

Experimental Results

Quantitative Results

Qualitative Results

BibTeX

@inproceedings{hou-etal-2024-icon,
      title = "{ICON}: Improving Inter-Report Consistency in Radiology Report Generation via Lesion-aware Mixup
      Augmentation",
      author = "Hou, Wenjun and Cheng, Yi and Xu, Kaishuai and Hu, Yan and Li, Wenjie and Liu, Jiang",
}