多组学机器学习预测宫颈癌放疗急性直肠毒性及生存结局
该回顾性研究纳入370例接受放疗的宫颈癌患者,构建多组学机器学习模型预测放疗诱导的急性直肠毒性(RI-ARTs)。联合模型下的XGBoost模型在测试集中AUC为0.80(95%CI 0.68–0.90),直肠D0.1cc与RI-ARTs发生显著相关,D0.1cc≤93 Gy时风险显著降低;低模型评分与总生存(OS)相关(HR 0.31,95%CI 0.11–0.88,p=0.028)。
要点
- 联合多组学XGBoost模型预测RI-ARTs的测试集AUC为0.80(95%CI 0.68–0.90)。
- 直肠D0.1cc与RI-ARTs发生显著相关,D0.1cc≤93 Gy时风险显著降低。
- 低XGBoost多组学模型评分与OS相关(HR 0.31,95%CI 0.11–0.88,p=0.028)。
核心结论
- 整合多组学模型对宫颈癌放疗诱导的急性直肠毒性(RI-ARTs)具有无创预测潜力。
- 探索性分析提示RI-ARTs发生与生存结局存在关联。
- 尽管存在单中心回顾性局限,研究结果提示该模型在临床决策中可能具有应用价值,尚待进一步前瞻性验证。
研究设计与样本
- 回顾性分析2016至2023年接受放疗的370例宫颈癌患者,其中231例发生RI-ARTs,129例未发生。
- 获取每例患者的CT影像特征、直肠放疗剂量相关信息和临床特征。
- 队列首先按80%训练集和20%完全独立测试集划分。
- 通过单变量(Mann-Whitney U检验/卡方检验)和多变量(LASSO)方法筛选特征,并基于Pearson相关分析剔除冗余特征。
- 从六种机器学习算法和深度学习模型中选出最合适的预测模型,采用AUC和生存分析评估模型有效性和临床适用性。
主要发现
- 共筛选出12个显著特征。
- 预测RI-ARTs的最佳性能模型为联合模型(剂量组学、影像组学和临床模型)下的XGBoost模型,测试队列AUC为0.80(95%CI 0.68–0.90)。
- 直肠D0.1cc特征与RI-ARTs发生率显著相关,D0.1cc≤93 Gy时RI-ARTs风险显著降低。
- 在结局队列中,相对于中位数的低XGBoost多组学模型评分与OS相关(风险比0.31;95%CI 0.11–0.88;p=0.028)。
关键数值与推荐
- 联合模型下XGBoost模型预测RI-ARTs的测试集AUC为0.80(95%CI 0.68–0.90)。
- 直肠D0.1cc≤93 Gy时RI-ARTs风险显著降低。
- 低XGBoost多组学模型评分(相对于中位数)与OS相关:HR 0.31(95%CI 0.11–0.88),p=0.028。
临床怎么用
- 该多组学XGBoost模型可用于无创预测宫颈癌放疗患者发生RI-ARTs的风险。
- 直肠D0.1cc可作为与RI-ARTs发生相关的剂量学特征参考,D0.1cc≤93 Gy时风险显著降低。
- 模型评分可能辅助判断患者生存预后,但需进一步前瞻性验证后方可用于临床决策。
局限性
- 本研究为单中心回顾性分析。
- 研究结论尚待进一步前瞻性验证。
- 模型需在多中心队列和更大样本量中进一步验证。
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: Radiotherapy-based regimens are the main treatment for cervical cancer, which has good curative effect and some toxicity. However, radiation-induced toxicities are not easily detected and may be associated with patient survival. OBJECTIVE: Machine-learning with multiomics model was constructed and evaluated to predict key features, treatment response, and survival outcomes of radiotherapy-induced acute rectal toxicities (RI-ARTs) in cervical cancer. METHODS: A retrospective analysis was conducted on 370 patients with cervical cancer who received radiotherapy from 2016 to 2023, of whom 231 had RI-ARTs, and 129 had no RI-ARTs. Corresponding CT imaging features, rectal radiotherapy dose-related information, and clinical features were also obtained for individual patients. The cohort was first partitioned into a training set (80%) and a completely independent test set (20%). Screening was then performed through univariate (Mann-Whitney U test/Chi-square test) and multivariate (LASSO) approaches, in which redundant features were eliminated based on Pearson correlation analyses, leading to the final selection of the most appropriate predictive model from among six analyzed machine learning algorithms and deep learning models. Area under the curve (AUC) and survival analyses were then utilized to assess model validity and clinical applicability. RESULTS: Twelve salient features were screened. The best performance model for predicting RI-ARTs was the XGBoost model under the combined model (dosiomics, radiomics, and clinical models), with an AUC of 0.80 (95% CI, 0.68–0.90) in the test cohort. Rectal D0·1cc feature was significantly associated with the incidence of RI-ARTs, and the risk of RI-ARTs was significantly reduced when D0·1cc ≤ 93 Gy. Low XGBoost-based multi-omics model scores relative to the median in outcome cohort was associated with OS (Hazard ratio, 0.31; 95% CI, 0.11–0.88; p = 0.028). CONCLUSION: The integrated multi-omics model demonstrates potential for non-invasive RI-ART prediction. Exploratory analysis further links RI-ART incidence to survival outcomes. Despite the single-center retrospective constraints, these findings suggest potential utility in clinical decision-making, pending further prospective validation. What is already known on this topic What this study adds How this study might affect research, practice or policy Consider that the majority of pelvic tumors treated with radiation, especially cervical cancer patients, more than 75% will develop radiotherapy-induced acute rectal toxicities (RI-ARTs), which in severe cases can affect the patient’s prognosis. However, there is currently no systematic model to integrate these risk factors and accurately predict various prognoses. We searched the PubMed database for peer-reviewed articles published up to October 10, 2024, using the search terms “(radiotherapy-induced acute rectal toxicities OR RI-ART), (cervical cancer), (Machine learning), and (models)”, with no language restriction. We could not find any effective models and clinical application tools for predicting the outcomes of patients with RI-ART in cervical cancer. To our knowledge, this is the first time an applicable model based on limit gradient enhancement (XGBoost) has been established to predict the risk of RI-ART in cervical cancer patients and its performance has been validated in an internal validation and test cohort. These models can facilitate clinician utility. In addition, a significant correlation between model score and OS and treatment response was first proposed. The XGBoost model under our combined model was first established to classify the RI-ART, treatment response, and long-term clinical survival of cervical cancer patients receiving radiotherapy, which is worthy of further verification in a multicenter cohort and a larger sample size.
引用
- Zhu Y, Liu K, Yang Z, et al. Machine-learning with multiomics for radiotherapy-induced acute rectal toxicities and survival outcome in patients with cervical cancer. Radiat Oncol. 2026. PMID: 41787421. doi:10.1186/s13014-026-02814-z
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。