基于机器学习算法模型的卵巢癌化疗患者延迟性化疗所致恶心呕吐风险预测研究-东南大学学报医学版

基于机器学习算法模型的卵巢癌化疗患者延迟性化疗所致恶心呕吐风险预测研究

单位：徐州医科大学附属淮安医院/淮安市第二人民医院, 江苏淮安 223001

分类号：R737.31;R195.1

出版年·卷·期（页码）：2026·45·第三期（379-392）

摘要：

目的: 基于机器学习算法构建卵巢癌化疗患者延迟性化疗所致恶心呕吐(CINV)风险预测模型,筛选最优模型,为临床个体化风险预测与CINV防控提供依据。方法: 便利抽样法选取2023年6月至2025年6月在本院接受化疗的卵巢癌患者为研究对象,按7∶3比例随机划分为训练集和测试集。采用一般资料调查表、常见不良事件评价标准(CTCAE 3.0)、焦虑自评量表(SAS)、Piper疲乏修订量表(RPFS)、中文版患者积极度量表(PAM)13进行调查。根据有无发生延迟性CINV分为延迟性CINV组和非延迟性CINV组。采用LASSO回归分析筛选训练集中的预测变量;基于机器学习构建延迟性CINV的Logistic回归、支持向量机(SVM)、随机森林、极端梯度提升(XGBoost)共4种模型,分析模型性能指标[曲线下面积(AUC)、准确率、精确率、召回率、F1],筛选最优模型。结果: 最终纳入409例卵巢癌化疗患者,发生延迟性CINV有192例(46.94%),其中训练集(286例)中延迟性CINV发生134例(46.85%),测试集(123例)中延迟性CINV发生58例(47.15%)。LASSO回归筛选出8个预测变量,即化疗药物致吐风险、晕动症史、妊娠期剧吐史、ECOG评分、营养风险、焦虑、疲乏、患者积极度。训练集中XGBoost模型的AUC、准确率、精确率、召回率、F1在4种模型中均最高(0.898、0.806、0.825、0.741、0.780),且在测试集上,XGBoost模型的AUC为0.861(95%CI:0.787~0.920),准确率、召回率、F1最高(0.765、0.691、0.733)。结论: 4种预测模型中XGBoost模型的预测性能相对最优,能够为临床早期识别延迟性CINV高风险患者及制定个体化预防干预策略提供依据。

Objective: To construct a risk prediction model for delayed chemotherapy-induced nausea and vomiting(CINV) in ovarian cancer patients undergoing chemotherapy based on machine learning algorithms, and to screen the optimal model, so as to provide evidence for clinical individualized prediction, prevention and control. Methods: A total of ovarian cancer patients who received chemotherapy in our hospital from June 2023 to June 2025 were enrolled by convenience sampling. All subjects were randomly divided into the training set and test set at a ratio of 7∶3. General Information Questionnaire, Common Terminology Criteria for Adverse Events(CTCAE 3.0), Self-Rating Anxiety Scale(SAS), Revised Piper Fatigue Scale(RPFS), and Chinese version of Patient Activation Measure(PAM) 13 were adopted for investigation. Participants were divided into the delayed CINV group and non-delayed CINV group according to the occurrence of delayed CINV. LASSO regression was used to screen predictive variables in the training set. Four machine learning models including Logistic regression, support vector machine(SVM), random forest and extreme gradient boosting(XGBoost) were established. Model performance indicators including the area under the curve(AUC), accuracy, precision, recall rate and F1-score were compared to select the optimal model. Results: A total of 409 ovarian cancer patients undergoing chemotherapy were finally included, among whom 192 cases developed delayed CINV, with an incidence rate of 46.94%. In the training set(286 cases), 134 cases had delayed CINV(46.85%); In the test set(123 cases), 58 cases had delayed CINV(47.15%). Eight predictive variables were screened out by LASSO regression: emetogenic risk of chemotherapeutic drugs, history of motion sickness, history of hyperemesis gravidarum, ECOG performance status score, nutritional risk, anxiety, fatigue, and patient activation level. In the training set, the XGBoost model yielded the highest values of AUC, accuracy, precision, recall rate, and F1-score among the four models(0.898, 0.806, 0.825, 0.741, 0.780). In the test set, the AUC of the XGBoost model was 0.861(95%CI: 0.787-0.920), with the highest accuracy, recall rate, and F1-score of 0.765, 0.691, 0.733, respectively. Conclusion: Among the four prediction models, the XGBoost model presents the relatively optimal predictive performance. It can provide a basis for early clinical identification of patients for high risk of delayed CINV and the formulation of individualized preventive and intervention strategies.

参考文献：

[1] HAN B,ZHENG R,ZENG H,et al.Cancer incidence and mortality in China,2022[J].J Natl Cancer Cent,2024,4(1):47-53.
[2] JIANG B.Analysis of global ovarian cancer disease burden and its changing trend from 1990 to 2021[J].BMC Women's Health,2025,25(1):352.
[3] 中国抗癌协会肿瘤支持治疗专业委员会.延迟性恶心呕吐防治中国专家共识(2022年版)[J].临床肿瘤学杂志,2023,28(5):442-458.
[4] GAO A,GUAN S,SUN Y,et al.Prolonged usage of fosaprepitant for prevention of delayed chemotherapy-induced nausea and vomiting(CINV) in patients receiving highly emetogenic chemotherapy[J].BMC Cancer,2023,23(1):609.
[5] 张晞月.化疗所致恶心呕吐风险因素分析及预测模型构建[D].锦州:锦州医科大学,2024.
[6] ZHANG J,CUI X,YANG C,et al.A deep learning-based interpretable decision tool for predicting high risk of chemotherapy-induced nausea and vomiting in cancer patients prescribed highly emetogenic chemotherapy[J].Cancer Med,2023,12(17):18306-18316.
[7] RILEY R D,ENSOR J,SNELL K I E,et al.Calculating the sample size required for developing a clinical prediction model[J].BMJ,2020,368(1):m441.
[8] 中国抗癌协会妇科肿瘤专业委员会.卵巢恶性肿瘤诊断与治疗指南(2021年版)[J].中国癌症杂志,2021,31(6):490-500.
[9] 中国抗癌协会肿瘤临床化疗专业委员会,中国抗癌协会肿瘤支持治疗专业委员会.中国肿瘤药物治疗相关恶心呕吐防治专家共识(2022年版)[J].中华医学杂志,2022,102(39):3080-3094.
[10] INCI M G,RASCH J,WOOPEN H,et al.ECOG and BMI as preoperative risk factors for severe postoperative complications in ovarian cancer patients:results of a prospective study(RISC-GYN—trial)[J].Arch Gynecol Obstet,2021,304(5):1323-1333.
[11] KONDRUP J.Nutritional risk screening(NRS 2002):a new method based on an analysis of controlled clinical trials[J].Clin Nutr,2003,22(3):321-336.
[12] TROTTI A,COLEVAS A D,SETSER A,et al.CTCAE v3.0:development of a comprehensive grading system for the adverse effects of cancer treatment[J].Semin Radiat Oncol,2003,13(3):176-181.
[13] ZUNG W W K.A rating instrument for anxiety disorders[J].Psychosomatics,1971,12(6):371-379.
[14] PIPER B F,DIBBLE S L,DODD M J,et al.The revised Piper Fatigue Scale:psychometric evaluation in women with breast cancer[J].Oncol Nurs Forum,1998,25(4):677-684.
[15] SO W K W,DODGSON J,TAI J W M.Fatigue and quality of life among Chinese patients with hematologic malignancy after bone marrow transplantation[J].Cancer Nurs,2003,26(3):211-219.
[16] HIBBARD J H,MAHONEY E R,STOCKARD J,et al.Development and testing of a short form of the patient activation measure[J].Health Serv Res,2005,40(6p1):1918-1930.
[17] 洪洋,邹叶芳,吴娟.中文版患者积极度量表的信效度研究[J].中华现代护理杂志,2018,24(25):3009-3015.
[18] ROESEL I,FROEHLICH D,JOOS S,et al.The Patient Activation Measure-13(PAM-13) in an oncology patient population:psychometric properties and dimensionality evaluation[J].Health Qual Life Outcomes,2024,22(1):39.
[19] 楼佳烨,王艳梅,潘欣欣,等.基于机器学习的糖尿病足发病风险预测模型构建[J].护理学杂志,2025,40(9):26-30.
[20] SHIMOKAWA M,HAYASHI T,KOGAWA T,et al.Evaluation of combination antiemetic therapy on CINV in patients with gynecologic cancer receiving TC chemotherapy[J].Anticancer Res,2019,39(1):225-230.
[21] SUN Y,ZHENG Y,YANG X,et al.Incidence of chemotherapy-induced nausea and vomiting among cancer patients receiving moderately to highly emetogenic chemotherapy in cancer centers in Sichuan,China[J].J Cancer Res Clin Oncol,2021,147(9):2701-2708.
[22] PURI S,HYLAND K A,WEISS K C,et al.Prediction of chemotherapy-induced nausea and vomiting from patient-reported and genetic risk factors[J].Support Care Cancer,2018,26(8):2911-2918.
[23] HERRSTEDT J,CLARK-SNOW R,RUHLMANN C H,et al.2023 MASCC and ESMO guideline update for the prevention of chemotherapy-and radiotherapy-induced nausea and vomiting[J].ESMO Open,2024,9(2):102195.
[24] ZHOU X,FAN T.Construction and validation of a nomogram model for predicting CINV in patients with gynecological malignancies[J].Jpn J Clin Oncol,2025,55(7):736-742.
[25] 陈开通,贺园园,刘飞宇,等.临床药师参与化疗所致恶心、呕吐管理的效果评价[J].东南大学学报(医学版),2026,45(1):85-90.
[26] JIN K,LIN X,XIA X,et al.Incidence and risk factors of platinum-based chemotherapy-induced nausea and vomiting:a systematic review and meta-analysis[J].Curr Oncol,2025,32(6):325.
[27] 王鹏程,孟爱凤,智晓旭,等.芳香疗法预防肿瘤患者化疗后恶心及呕吐的系统评价[J].解放军护理杂志,2020,37(4):6-10.
[28] 孙凌雪,胡三莲,刘新义,等.穴位按压缓解癌症病人化疗相关恶心呕吐的最佳证据总结[J].护理研究,2020,34(22):3972-3976.
[29] 王静,胡诗音,龚倩,等.基于Logistic回归和决策树构建卵巢癌病人营养风险预测模型[J].护理研究,2025,39(22):3751-3756.
[30] CHOI J,LEE J,KIM K,et al.Effects of ginger intake on chemotherapy-induced nausea and vomiting:a systematic review of randomized clinical trials[J].Nutrients,2022,14(23):4982.
[31] GALA D,WRIGHT H H,ZIGORI B,et al.Dietary strategies for chemotherapy-induced nausea and vomiting:a systematic review[J].Clin Nutr,2022,41(10):2147-2155.
[32] MOKSHAGUNDAM S,MCGREE M E,TAPIA A L,et al.Physical and psychological distress amongst patients undergoing neoadjuvant chemotherapy for advanced ovarian cancer[J].Gynecol Oncol,2024,190:230-235.
[33] CARNIO S,GALETTA D,SCOTTI V,et al.Chemotherapy-induced nausea and vomiting(CINV) in patients with advanced lung cancer during the first-line treatment:assessment by physicians,nurses,and patients from an Italian multicenter survey[J].Support Care Cancer,2018,26(6):1841-1849.
[34] 李颖颖,蒋晓莲.卵巢癌化疗病人癌因性疲乏现状及影响因素分析[J].护理研究,2021,35(10):1812-1816.
[35] 丁华,李朝凤,程钰,等.正念冥想疗法对慢性心力衰竭患者负性情绪及心功能影响的Meta分析[J].现代医学,2024,52(1):1-9.
[36] NAZARI A M,ZARE-KASEB A,ARBABI Z,et al.The effect of Benson relaxation technique on cancer patients:a systematic review[J].Support Care Cancer,2023,31(12):681.
[37] 邱丽文.妇科恶性肿瘤化疗患者自我管理积极度现状、影响因素及其对策研究[J].现代临床护理,2023,22(6):38-45.
[38] KANU C,BROWN C M,RASCATI K,et al.Are health literacy and patient activation related to health outcomes in breast cancer patients[J].Health Lit Res Pract,2021,5(3):e171-e178.
[39] HIBBARD J H,MAHONEY E,SONET E.Does patient activation level affect the cancer patient journey[J].Patient Educ Couns,2017,100(7):1276-1279.
[40] HüBNER J,WELTER S,CIARLO G,et al.Patient activation,self-efficacy and usage of complementary and alternative medicine in cancer patients[J].Med Oncol,2022,39(12):192.
[41] BONETTI L,TOLOTTI A,ANDERSON G,et al.Nursing interventions to promote patient engagement in cancer care:a systematic review[J].Int J Nurs Stud,2022,133:104289.
[42] LI X,SHANG C,XU C,et al.Development and comparison of machine learning-based models for predicting heart failure after acute myocardial infarction[J].BMC Med Inform Decis Mak,2023,23(1):165.
[43] SCRUTINIO D,AMITRANO F,GUIDA P,et al.Prediction of mortality in heart failure by machine learning.Comparison with statistical modeling[J].Eur J Intern Med,2025,133:106-112.

服务与反馈：

【文章下载】【发表评论】【查看评论】【加入收藏】

提示：您还未登录，请登录！点此登录