Contrastive Learning with Counterfactual Explanations for Radiology Report Generation
This work addresses the issue of misdiagnostic reports in radiology AI, which is critical for patient safety, but it is incremental as it builds on existing contrastive learning and counterfactual methods.
The paper tackled the problem of spurious representations in radiology report generation due to data bias, proposing a counterfactual explanations-based framework (CoFE) that improved report generation by learning non-spurious visual representations, resulting in better performance on language generation and clinical efficacy metrics in experiments on two benchmarks.
Due to the common content of anatomy, radiology images with their corresponding reports exhibit high similarity. Such inherent data bias can predispose automatic report generation models to learn entangled and spurious representations resulting in misdiagnostic reports. To tackle these, we propose a novel \textbf{Co}unter\textbf{F}actual \textbf{E}xplanations-based framework (CoFE) for radiology report generation. Counterfactual explanations serve as a potent tool for understanding how decisions made by algorithms can be changed by asking ``what if'' scenarios. By leveraging this concept, CoFE can learn non-spurious visual representations by contrasting the representations between factual and counterfactual images. Specifically, we derive counterfactual images by swapping a patch between positive and negative samples until a predicted diagnosis shift occurs. Here, positive and negative samples are the most semantically similar but have different diagnosis labels. Additionally, CoFE employs a learnable prompt to efficiently fine-tune the pre-trained large language model, encapsulating both factual and counterfactual content to provide a more generalizable prompt representation. Extensive experiments on two benchmarks demonstrate that leveraging the counterfactual explanations enables CoFE to generate semantically coherent and factually complete reports and outperform in terms of language generation and clinical efficacy metrics.