开发机器学习模型预测经口机器人手术(TORS)阳性手术切缘
回顾性收集453例喉、下咽及口咽鳞状细胞癌患者数据,训练(n=316)并验证(n=137)六种二分类监督机器学习模型,使用14项术前变量预测TORS阳性切缘。模型准确度0.67–0.75,AUC 0.68–0.75,特异度高(0.75–0.89)但敏感度低(0.26–0.64),NPV(0.73–0.83)高于PPV(0.45–0.63)。T分类和肿瘤部位是最重要预测因子,模型可识别阳性切缘低风险、适合TORS的患者。
要点
- 六种机器学习模型预测TORS阳性切缘的准确度为0.67–0.75,AUC为0.68–0.75。
- 模型特异度高(0.75–0.89)但敏感度低(0.26–0.64)。
- T分类和肿瘤部位是阳性切缘最重要的预测因子。
核心结论
- 机器学习算法可识别阳性切缘低风险、因此适合TORS的患者。
研究设计与样本
- 回顾性收集三级转诊中心453例喉、下咽及口咽鳞状细胞癌患者数据,训练集n=316、验证集n=137。
- 采用14项术前可获取变量,训练并验证六种二分类监督机器学习模型。
主要发现
- 六种机器学习模型的准确度范围为0.67–0.75,AUC范围为0.68–0.75。
- 检测TORS后阳性切缘时,模型特异度范围为0.75–0.89,敏感度范围为0.26–0.64。
- NPV范围为0.73–0.83,高于PPV范围0.45–0.63。
- T分类和肿瘤部位是阳性手术切缘最重要的预测因子。
临床怎么用
- 作者认为机器学习算法可识别阳性切缘低风险、因此适合TORS的患者。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
PURPOSE: To develop machine learning (ML) models for predicting positive margins in patients undergoing transoral robotic surgery (TORS). METHODS: Data from 453 patients with laryngeal, hypopharyngeal, and oropharyngeal squamous cell carcinoma were retrospectively collected at a tertiary referral center to train (n = 316) and validate (n = 137) six two-class supervised ML models employing 14 variables available pre-operatively. RESULTS: The accuracy of the six ML models ranged between 0.67 and 0.75, while the measured AUC between 0.68 and 0.75. The ML algorithms showed high specificity (range: 0.75-0.89) and low sensitivity (range: 0.26-0.64) in detecting patients with positive margins after TORS. NPV was higher (range: 0.73-0.83) compared to PPV (range: 0.45-0.63). T classification and tumor site were the most important predictors of positive surgical margins. CONCLUSIONS: ML algorithms can identify patients with low risk of positive margins and therefore amenable to TORS.
引用
- Costantino A, Sampieri C, Pirola F, et al. Development of machine learning models for the prediction of positive surgical margins in transoral robotic surgery (TORS). Head Neck. 2022. PMID: 36541686. doi:10.1002/hed.27283
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。