What question did this study set out to answer?

To develop an automated system for lung cancer TNM classification from radiology reports using large language models.

April 1, 2026Open Access

Automated Lung Cancer Staging from Radiological Reports: A Large Language Model Approach for the NTCIR-18 RadNLP Task

Key Points

To develop an automated system for lung cancer TNM classification from radiology reports using large language models.
Utilized large language models with supervised fine-tuning and specialized prompting.
Evaluated system on the NTCIR-18 RadNLP 2024 Task dataset.
Achieved classification accuracy for both Japanese and English radiology reports.
Achieved 72.69% accuracy for Japanese and 55.56% for English radiology reports.
Demonstrated >93.98% accuracy in N-factor classification.
Ranked 5th among 15 teams in the task, highlighting the system's competitiveness.

Abstract

Lung cancer TNM classification from narrative radiology reports presents challenges due to expression variability and complex relationships between findings. This study develops an automated TNM classification system utilizing large language models (LLMs) with supervised fine-tuning (SFT) and specialized prompting (SP) approaches. We evaluated our system on the NTCIR-18 RadNLP 2024 Task dataset, achieving 72.69\% (Japanese) and 55.56\% (English) fine-grained accuracy, ranking 5th among 15 teams. Our system demonstrated particularly high performance in N-factor classification (>93.98\% accuracy) and in the subtask of textual analysis (ranking 1st in Japanese and 3rd in English tracks). Error analysis revealed challenges in interpreting complex expressions and implicit information. This system shows potential for clinical workflow optimization, standardization of TNM classification, and educational support, with implications for improving cancer staging practices.

AI에게 질문

Bookmark

View Full Paper