文献速递头颈外科咽喉与嗓音

机器学习与统计模型预测口腔癌游离皮瓣术后长期胃造瘘管依赖

Evaluation of Machine Learning and Statistical Models for Predicting Long Term Gastrostomy Tube Dependency in Patients Undergoing Major Oral Cavity Cancer Surgery With Free Flap Reconstruction.
PubMed 文献 · Head Neck · 2026-07-12

本研究为多中心回顾性 pooled 队列,纳入500例接受游离皮瓣重建的非转移性口腔癌患者,比较三种模型预测12个月胃造瘘管依赖。随机森林预测最准确(AUC 0.77),LASSO和Firth惩罚逻辑回归相当(AUC 0.71和0.70)。关键预测因素包括IV期、T4病变、单侧颈清扫、重建部位及未行舌切除。

要点

  • 随机森林模型预测12个月胃造瘘管依赖AUC为0.77,优于LASSO(0.71)和Firth回归(0.70)。
  • 500例中78例(16%)在12个月时依赖胃造瘘管。
  • 关键预测因素包括IV期、T4病变、单侧颈清扫、重建部位及未行舌切除。

核心结论

  • 三种模型对口腔癌游离皮瓣术后12个月胃造瘘管依赖具有中等准确度预测能力,随机森林表现最佳。
  • 新预测因素(单侧颈清扫、重建部位、未行舌切除)值得关注,更大样本可能提升预测性能。

研究设计与样本

  • 多中心回顾性 pooled 队列研究,纳入500例接受游离皮瓣重建的非转移性口腔癌患者。
  • 使用10折交叉验证训练Firth惩罚逻辑回归、LASSO和随机森林模型,以AUC评估预测性能。

主要发现

  • 500例患者平均年龄61.7岁,78例(16%)在12个月时依赖胃造瘘管。
  • 随机森林预测12个月胃造瘘管依赖的AUC为0.77,LASSO为0.71,Firth惩罚逻辑回归为0.70。
  • 关键预测因素包括TNM IV期、T4病变、单侧颈清扫、重建部位及未行舌切除。

关键数值与推荐

  • 随机森林AUC 0.77,LASSO AUC 0.71,Firth惩罚逻辑回归AUC 0.70。
  • 胃造瘘管依赖定义为12个月时仍依赖胃造瘘管。

临床怎么用

  • 术前可利用随机森林模型评估口腔癌游离皮瓣术后长期胃造瘘管依赖风险,辅助决策。
  • 关注IV期、T4病变、单侧颈清扫、特定重建部位及未行舌切除等高危因素。

局限性

  • 回顾性研究设计可能引入选择偏倚。
  • 样本量有限(500例),可能影响模型性能,需更大 multicenter 数据验证。

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

BACKGROUND: Swallowing dysfunction and long-term gastrostomy tube dependence are common morbidities after oral cavity cancer resection with free flap reconstruction, highlighting the need for preoperative risk prediction. METHODS: We performed a multicenter retrospective pooled cohort study of patients with non-metastatic oral cavity cancer undergoing free flap reconstruction. Firth's penalized logistic regression, LASSO, and random forest models were trained using 10-fold cross-validation and predictive performance was assessed via area under the receiver operator curve (AUC). RESULTS: Among 500 patients (mean age 61.7 years), 78 (16%) were gastrostomy tube dependent at 12 months. The random forest model was most accurate in predicting 12-month gastrostomy tube dependency (AUC; 0.77), with comparable performance from LASSO (AUC; 0.71) and Firth's penalized logistic regression (AUC; 0.70). Key predictors included TNM stage IV, T4 disease, unilateral neck dissection, reconstruction site, and no tongue resection. CONCLUSIONS: The three models predicted 12-month gastrostomy tube dependency after oral cavity free-flap surgery with fair accuracy and novel predictors; larger multicenter datasets may further improve predictive performance.

来源:Head Neck,查看原文。版权归原作者所有。

引用

  1. Namavarian A, Alenazi A, Hueniken K, et al. Evaluation of Machine Learning and Statistical Models for Predicting Long Term Gastrostomy Tube Dependency in Patients Undergoing Major Oral Cavity Cancer Surgery With Free Flap Reconstruction. Head Neck. 2026. PMID: 42438145. doi:10.1002/hed.70353

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。