南方医科大学学报

南方医科大学学报 ›› 2026, Vol. 46 ›› Issue (7): 1509-1519.doi: 10.12122/j.issn.1673-4254.2026.07.05

• • 上一篇    

可解释性机器学习模型对宫颈癌周围神经浸润的术前精准预测及临床决策支持：一项多中心回顾性研究

朱梦欣1,4(), 刘晓2, 赵珊3, 张新1,4, 刘超4, 刘献孔4, 綦浩辰4, 韩铁生1, 马冬1,4()   

  1. 1.华北理工大学公共卫生学院，河北 唐山 063210
    2.河北医科大学第四医院感染科，河北 石家庄 050010
    3.河北医科大学第二医院肿瘤科，河北 石家庄 050000
    4.河北医科大学生物化学与分子生物学研究室，河北 石家庄 050011
  • 收稿日期:2026-01-07 出版日期:2026-07-20 发布日期:2026-07-20
  • 通讯作者: 马冬 E-mail:zmx13375521993@163.com;madong119@ hebmu.edu.cn
  • 作者简介:朱梦欣，在读硕士研究生，E-mail: zmx13375521993@163.com
  • 基金资助:
    国家自然科学基金(81541149);河北省青年科学基金(H2016209095)

Interpretable machine learning models for preoperative precision prediction of perineural invasion in cervical cancer to support treatment decision: a multicenter retrospective study

Mengxin ZHU1,4(), Xiao LIU2, Shan ZHAO3, Xin ZHANG1,4, Chao LIU4, Xiankong LIU4, Haochen QI4, Tiesheng HAN1, Dong MA1,4()   

  1. 1.School of Public Health, North China University of Science and Technology, Tangshan 063210, China
    2.Department of Infection Control, Fourth Hospital of Hebei Medical University, Shijiazhuang 051010, China
    3.Department of Cancer Second Division, Second Hospital of Hebei Medical University, Shijiazhuang 050000, China
    4.Laboratory of Biochemistry and Molecular Biology, Hebei Medical University, Shijiazhuang 050011, China
  • Received:2026-01-07 Online:2026-07-20 Published:2026-07-20
  • Contact: Dong MA E-mail:zmx13375521993@163.com;madong119@ hebmu.edu.cn
  • Supported by:
    National Natural Science Foundation of China(81541149)

RichHTML

16

PDF

121

摘要：

目的 建立术前预测宫颈癌患者术后周围神经浸润（PNI）发生风险的机器学习模型和线上预测器，实现临床诊疗决策制定的人工智能应用。 方法 回顾性纳入845例于河北医科大学第四医院接受根治性手术治疗的宫颈癌患者病理资料，以7∶3的比例分为训练集（n=593）和测试集（n=252）；同期纳入河北医科大学第二医院收治的223例宫颈癌患者组成外部验证集。LASSO筛选宫颈癌患者术后PNI发生相关术前特征变量，通过7种机器学习算法构建预测模型并通过ROC曲线下面积（AUC）和临床决策曲线（DCA）评价其预测效能及其在不同数据集的一致性。确定最优预测模型后进行SHAP解释，最终构建宫颈癌患者术后PNI发生风险的预测网站。 结果 共纳入1068例宫颈癌患者，其中192例（17.98%）患者术后病理结果呈现为PNI。LASSO回归分析共筛选：脉管间隙侵犯、间质浸润深度、淋巴结转移、阴道镜引导下穿刺活检、肿瘤最大直径、癌胚抗原、SCC-Ag、血小板/淋巴细胞比值、中性粒细胞/淋巴细胞比值、淋巴细胞*白蛋白/中性粒细胞比值、绝经情况、年龄和组织学类型等13个术前特征变量，7种算法对比，XGBoost模型在训练集和测试集中效能最佳，AUC分别为0.962和0.923，95% CI：0.942~0.979和0.874~0.960，敏感度为0.873、0.767，特异度为0.939、0.942。DCA分析结果同样显示XGBoost模型在较宽阈值范围内具有正向更大的净效益。基于SHAP可解释模型构建预测网站并进行外部验证，结果显示，该模型具有良好的泛化能力和预测准确性（AUC为0.924，准确率为0.933）。 结论 本研究开发的可解释性SHAP-XGBoost模型能够较好的预测宫颈癌患者术后周围神经浸润结局，基于该模型开发的预测网站可为宫颈癌诊疗决策制定提供辅助工具。

关键词: 宫颈癌, 周围神经浸润, 机器学习

Abstract:

Objective To develop an interpretable machine learning model and web-based prediction tool for preoperative risk assessment of perineural invasion (PNI) in cervical cancer. Methods A total of 845 cervical cancer patients undergoing radical surgery at Fourth Hospital of Hebei Medical University were retrospectively enrolled and divided into training and testing sets in a 7:3 ratio, with another 223 cervical cancer patients at Hebei Medical University Second Hospital during the same period serving as the external validation cohort. LASSO regression identified 13 preoperative predictors, which were incorporated into 7 machine learning algorithms. Model performance was evaluated using AUC and decision curve analysis. The optimal model was interpreted using SHAP values and deployed as a web-based prediction tool. Results Of the total of 1068 patients enrolled, 192 (17.98%) were diagnosed to have PNI. Thirteen preoperative features, namely lymphovascular space invasion (LVSI), depth of stromal invasion, lymph node metastasis (LNM), colposcopy-directed biopsy (CDB), tumor maximum diameter, carcinoembryonic antigen, SCC-Ag, platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), lymphocyte-albumin-neutrophil ratio (LANR), menopausal status, age, and histological type were selected. Comparison of model performance revealed that the Extreme Gradient Boosting (XGBoost) model resulted in the best efficacy in both the training and testing datasets with AUC of 0.962 and 0.923 (95% CI: 0.942-0.979 and 0.874-0.960), sensitivity of 0.873 and 0.767, and specificity of 0.939 and 0.942, respectively. Decision curve analysis demonstrated greater net benefit of the XGBoost model across a broader threshold range. The SHAP-XGBoost model showed excellent performance in external validation with an AUC of 0.924 and an accuracy of 0.933. Conclusion The interpretable SHAP-XGBoost model effectively predicts PNI risk preoperatively. The predictive website derived from this model provides an useful tool to facilitate clinical decision-making in cervical cancer treatment.

Key words: cervical cancer, perineural invasion, machine learning