arXiv cs.AI· Stefan Lenz, Lakisha Ortiz Rosario, Georg Vollmar, Arsenij Ustjanzew, Fatma Alickovic, Thomas Kindler, Torsten Panholzer·· 4 小时前AI 评分44
用医学分类数据对大语言模型做指令微调提升德国肿瘤诊断的 ICD 编码准确率
Instruction-Tuning Large Language Models with Medical Classification Data Improves ICD Coding of German Tumor Diagnoses
AI 导读
一项研究基于 ICD-10-GM、ICD-O-3 和 OPS 目录构建超 50 万条问答对，对 Qwen、Llama 和 Mistral 系列 8 个 7-70B 参数开源模型做指令微调，提升德语肿瘤诊断的 ICD 编码准确率。ICD-10-GM 精确准确率从 1.4-24% 升至 41-58%，畸形编码输出降至 0%，肿瘤诊断识别率达 99%。训练数据与最优模型权重已公开。
正文 · 原文
Abstract:Accurate coding of tumor diagnoses with ICD-10-GM and ICD-O-3 is essential for structured cancer documentation in Germany. Smaller open-weight LLMs are appealing for privacy-preserving automation but often struggle with coding accuracy in German-language contexts. This study investigates whether instruction-based fine-tuning on public datasets improves the coding accuracy of open-weight LLMs for German tumor diagnosis texts. The evaluation uses coded diagnoses from the local tumor documentation system as test data. In a systematic data quality assessment, the upper limit for ICD-10 coding performance was estimated at 60-79% for exact and 81-94% for partial (three-character codes only) derivation. As training data, over 500,000 question-answer pairs were created based on the ICD-10-GM, ICD-O-3, and OPS catalogues. Eight open-weight models from the Qwen, Llama, and Mistral families (7-70 B parameters) were fine-tuned. ICD-10-GM accuracy rose from 1.4-24% to 41-58%, and partial accuracy from 31-74% to 73-83%. The accuracy of ICD-O-3 topography coding also improved but started and remained considerably lower with an exact accuracy of 22-40% and a partial accuracy of 56-67% after fine-tuning. Malformed code outputs dropped to 0% for all models. Tumor-diagnosis recognition reached 99%. Accuracy correlated positively with model size, but gaps between small and large models narrowed after fine-tuning. The reasoning mode in Qwen3 generally yielded a lower performance than fine-tuning and was over 100 times slower. Our findings highlight the potential of leveraging public catalogues to build instruction datasets that improve LLMs in medical documentation tasks. The complete training dataset and the best-performing checkpoints of the fine-tuned models are available from this https URL.
| Comments: | 43 pages, 4 figures |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2510.13624 [cs.CL] |
| (or arXiv:2510.13624v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2510.13624 arXiv-issued DOI via DataCite |
Submission history
From: Stefan Lenz [view email]
[v1]
Wed, 15 Oct 2025 14:51:28 UTC (965 KB)
[v2]
Thu, 5 Nov 2026 14:00:59 UTC (1,121 KB)
来源：arXiv cs.AI · arxiv.org