专有和微调的大型语言模型的比较,用于从放射学报告中多标签分类的计费代码
Kamyar Arzideh1,2, Henning Schäfer3,4, Ahmad Idrissi-Yaghir3,5
1Central IT Department, Data Integration Center, University Hospital Essen, Essen, Germany. kamyar.arzideh@uk-essen.de.
European radiology
|March 14, 2026
概括
一个微调的大型语言模型 (LLM) 有效地自动化了从放射学报告中提取德国医疗计费代码 (GOÄ),超过了一些商业LLM. 这种方法提高了医疗保健中的计费效率和准确性.
科学领域:
- 医疗信息学 医疗信息学
- 医疗保健中的人工智能
- 医疗编码自动化 医疗编码自动化
背景情况:
- 大型语言模型 (LLM) 显示了医学文本分析的潜力.
- 自动化医疗计费代码提取,特别是对于德国GOÄ系统来说,还没有得到充分的探索.
- 目前的LLM可能没有针对专业医疗计费任务进行优化.
研究的目的:
- 微调一个大型语言模型 (LLM) 来自动化多标签分类德国医疗费用表 (GOÄ) 代码从放射学报告.
- 将微调的LLM与最先进的商业和开源LLM的性能进行比较.
- 评估微调的LLM在提高医疗保健计费效率和准确性方面的潜力.
主要方法:
- 分析了来自124,497名患者的499,601份放射学报告,其中1,799,971名患者手动识别了GOÄ代码.
- 使用五倍交叉验证微调MediPhi-Instruct 4B模型.
- 在持久测试集上的性能评估和与多个LLM (GPT-5,Gemini 2.5等) 的比较. 在报告子集上使用零射击和少数射击提示.
主要成果:
- 这种微调的模型在持久测试套件上实现了77.15%的准确性和87.79%的微平均F1得分.
- 在现实世界样本中,微调的模型表现优于双子座2.5闪 (70.32%对58.22%F1得分).
- 在清理数据上,GPT-5获得了最高的F1得分 (89.51%),超过了微调模型.
结论:
- 精心调整的LLM可以有效地自动化从放射学报告中对GOÄ代码的分类.
- 这种自动化方法显示出可能超越商业LLM的性能,提高了计费效率和准确性.
- 仍然建议手动验证,精细调整的模型为临床编码辅助提供了有效的替代方案.
相关概念视频
Multiple Comparison Tests
Multiple comparison test, abbreviated as MCT, is a post hoc analysis generally performed after comparing multiple samples with one or more tests. An MCT will help identify a significantly different sample among multiple samples or a factor among multiple factors.
It would be easy to compare two samples using a significance alpha level of 0.05. In other words, there is only one sample pair to be compared. However, it would be difficult to identify a significantly different sample if the number...
It would be easy to compare two samples using a significance alpha level of 0.05. In other words, there is only one sample pair to be compared. However, it would be difficult to identify a significantly different sample if the number...
Mass Analyzers: Common Types
The quadrupole mass analyzer consists of four cylindrical metal rods arranged in a diamond carrying a DC voltage and a radio-frequency AC voltage. The motion of ions through the quadrupole depends on the field strength, causing only ions of a certain m/z to resonate successfully and strike the detector at a given field strength. Though the transmission rate for these analyzers is high, the exact elemental composition of the sample is not determined because of low resolution; however, they are...
Classification of Systems-II
Continuous-time systems have continuous input and output signals, with time measured continuously. These systems are generally defined by differential or algebraic equations. For instance, in an RC circuit, the relationship between input and output voltage is expressed through a differential equation derived from Ohm's law and the capacitor relation,
Aggregates Classification
Aggregate classification is generally based on its size, petrographic characteristics, weight, and source. Size classification ranges from coarse to fine aggregates, defined by the size of the particles. Coarse aggregates are particles that do not pass through ASTM sieve No. 4, and aggregates that pass through the sieve are fine aggregates.
Petrographic classification groups aggregates based on common mineralogical characteristics. Some of the common mineral groups found in aggregates are...
Petrographic classification groups aggregates based on common mineralogical characteristics. Some of the common mineral groups found in aggregates are...
Modern Molecular Taxonomy
Advancements in molecular biology have revolutionized the identification and characterization of bacteria, with multiple methods leveraging DNA sequencing for enhanced precision. As sequencing technologies improve and costs decline, these approaches are increasingly used in clinical, environmental, and evolutionary studies.Multilocus Sequence Typing (MLST) examines several housekeeping genes, essential chromosomal genes encoding cellular functions, to distinguish strains. Approximately...
Methods of Medium Optimization
Optimizing growth media enhances microbial proliferation and maximizes product yield. Statistical experimental design methodologies provide structured and reproducible approaches, offering progressively higher levels of robustness and efficiency.The One-Factor-at-a-Time (OFAT) MethodThe One-Factor-at-a-Time (OFAT) method involves adjusting a single variable while keeping all others constant. However, it cannot detect interactions between variables, often leading to suboptimal outcomes when...

