高召回率生物医学语言模型用于放射肿瘤学证据合成：鼻咽癌放疗并发症预测模型的整合系统评价
High-Recall Biomedical Language Models for Radiation Oncology Evidence Synthesis: An Integrated Systematic Review of Prediction Models for Radiotherapy-Induced Complications in Nasopharyngeal Carcinoma.
文献信息
| PMID | 42795430 |
|---|---|
| 原文 | 在 PubMed 查看原文 ↗ |
| 发表日期 | 2026 |
| 作者 | Guang-Zhi Lin |
| 作者单位 | Medical Physics and Informatics Laboratory of Electronic Engineering, National Kaohsiung University of Science and Technology, Kaohsiung 80778, Taiwan. |
| 期刊 | Life (Basel, Switzerland) |
| SCI 分区 | Q1 |
| IF | 3.9 |
| 研究类型 | AI/ML · 临床 |
| 所属专科 | 鼻咽癌 |
中文摘要
背景： 放疗相关并发症可能损害鼻咽癌（NPC）的长期结局，而高通量文献筛选仍是证据合成的瓶颈。
方法： 检索了四个数据库，截止至2026年1月。去重后，4916条记录经过两阶段人工筛选，最终纳入38项研究。使用分层五折交叉验证对三种双向编码器表示变换器（BERT）模型（BERT-base、BioBERT和PubMedBERT）进行微调，并使用受试者工作特征曲线下面积（ROC-AUC）、精确率-召回率曲线下面积（PR-AUC）以及在100%和85%召回阈值下的工作节省率（WSS@100%和WSS@85%）评估筛选性能。临床证据使用预测模型偏倚风险评估工具（PROBAST）进行评估，并根据终点、区分度指标、时间范围、分析单位、验证级别和不确定性报告进行叙述性综合。
结果： 在完全召回时，BioBERT和PubMedBERT的平均WSS值分别为97.8%和97.5%，而BERT-base为89.7%。另外，临床证据评价确定了39个区分度估计值：33个常规二元ROC-AUC、3个Harrell's C指数和3个时间依赖性AUC。只有两项研究提供了外部验证估计，一项研究提供了时间验证估计。16个估计值缺乏可用的不确定性度量，PROBAST将所有研究评为具有高总体偏倚风险，主要由分析领域驱动。
结论： 在检索并人工标注的语料库中，生物医学预训练语言模型显示出在保持完全召回的同时减少标题和摘要筛选负担的潜力。独立地，临床证据图识别出有前景的区分度，但证据的稳健性和可移植性不足以支持常规临床实施。
英文摘要
BACKGROUND: Radiotherapy-related complications can impair long-term outcomes in nasopharyngeal carcinoma (NPC), while high-volume literature screening remains a bottleneck in evidence synthesis.
METHODS: Four databases were searched through January 2026. After deduplication, 4916 records underwent two-stage manual screening, yielding 38 studies. Three Bidirectional Encoder Representations from Transformers (BERT) models, BERT-base, BioBERT, and PubMedBERT, were fine-tuned using stratified five-fold cross-validation, and screening performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC), the area under the precision-recall curve (PR-AUC), and work saved over sampling (WSS) at the 100% and 85% recall thresholds (WSS@100% and WSS@85%).The clinical evidence was evaluated using the Prediction Model Risk of Bias Assessment Tool (PROBAST) and synthesized narratively according to endpoint, discrimination metric, time horizon, unit of analysis, validation level, and uncertainty reporting.
RESULTS: At complete recall, BioBERT and PubMedBERT achieved mean WSS values of 97.8% and 97.5%, respectively, compared with 89.7% for BERT-base. Separately, the clinical evidence review identified 39 discrimination estimates: 33 conventional binary ROC-AUCs, three Harrell's C-indices, and three time-dependent AUCs. Only two studies contributed external-validation estimates and one contributed a temporal-validation estimate. Sixteen estimates lacked a usable measure of uncertainty, and PROBAST rated all studies as having a high overall risk of bias, driven by the analysis domain.
CONCLUSIONS: Within the retrieved and manually annotated corpus, biomedical pretrained language models showed the potential to reduce the title-and-abstract screening burden while maintaining complete recall. Independently, the clinical evidence map identified promising discrimination but insufficiently robust and transportable evidence for routine clinical implementation.