Reconfigurable Radiology Labels Without Relabeling

2026-07-06

AI Summary

Existing public chest X-ray datasets are constrained by fixed label taxonomies that fail to capture the rich pathological descriptions found in clinical radiology reports. This work proposes a configurable multi-label construction framework that converts free-text radiology reports into a multi-label matrix in a single pass, enabling dynamic reconfiguration of the label taxonomy at zero inference cost through an edit dictionary and caching mechanism—without requiring re-annotation of the original corpus. The method processes 223K reports from MIMIC-CXR in just 196 seconds with no API overhead, matches state-of-the-art performance on the CheXpert-14 benchmark labels, and achieves an AUROC of 0.78 on expert-reviewed long-tail labels, substantially reducing annotation costs while enhancing label flexibility.

Abstract

Public chest-radiograph (CXR) datasets are typically released with small, fixed label schemas such as CheXpert-14. However, the underlying free-text reports describe far more findings -- and which findings matter depends on the task, site, and reader. We release a pipeline that converts free-text reports into multi-label matrices and then reconfigures the label schema through dictionary edits rather than new inference passes, i.e., without relabeling the corpus. After this one-time pass, reconfiguring MIMIC-CXR (223K reports) from cached annotations takes 196 seconds with no API cost, compared to \$6.6K for an equivalent relabeling pass with Claude Opus 4.7. Using a 58-label taxonomy, we show that 43\% of CXR studies contain at least one finding outside CheXpert-14. Image probes trained on these labels match CheXpert-14 probes on shared targets while also reaching 0.78 AUROC on expert-reviewed long-tail labels that CheXpert-14 cannot represent. These results suggest a different unit of work for radiology labeling: once reports are structured, the label schema becomes a configuration to edit, not a corpus to relabel.

Problem

• reconfigurable labels
• chest radiograph
• label schema
• free-text reports
• long-tail findings

Innovation

• reconfigurable labels
• radiology NLP
• zero-relabeling
• multi-label classification
• long-tail findings