Evaluating Large Language Models for Zero-Shot Disease Labeling in CT Radiology Reports Across Organ Systems

Abstract

Purpose: This study aims to evaluate the effectiveness of large language models (LLMs) in automating disease annotation of CT radiology reports. We compare a rule-based algorithm (RBA), RadBERT, and three lightweight open-weight LLMs for multi-disease labeling of chest, abdomen, and pelvis (CAP) CT reports. Materials and Methods: This retrospective study analyzed 40,833 chest-abdomen-pelvis (CAP) CT reports from 29,540 patients, with 1,789 reports manually annotated across three organ systems. External validation was conducted using the CT RATE dataset. Three open-weight LLMs were tested with zero-shot prompting. Performance was evaluated using Cohen's Kappa ($κ$) and micro/macro-averaged F1 scores. Results: In the internal test set of 12,197 CAP reports from 8,854 patients, Llama-3.1 8B and Gemma-3 27B showed the highest agreement ($κ$ median: 0.87). On the manually annotated set, Gemma-3 27B achieved the top macro-F1 (0.82), followed by Llama-3.1 8B (0.79), while the RBA scored lowest (0.64). On the CT RATE dataset (lungs/pleura labels only), Llama-3.1 8B performed best (0.91), with Gemma-3 27B close behind (0.89). Performance differences were mainly due to differing labeling practices, especially for labels with high subjectivity such as atelectasis. Conclusion: Lightweight LLMs outperform rule-based methods for CT report annotation and generalize across organ systems with zero-shot prompting. However, binary labels alone cannot capture the full nuance of report language. LLMs can provide a flexible, efficient solution aligned with clinical judgment and user needs.