可解释人工智能识别局部晚期非小细胞肺癌患者毒性剂量学预测因素:RTOG 0617二次分析
本研究为RTOG 0617前瞻性临床试验的二次分析,使用机器学习和可解释人工智能预测局部晚期非小细胞肺癌根治性放化疗患者≥3级肺、心脏和食管毒性。肺毒性最佳模型为极端梯度提升(AUC 0.739),心脏毒性为随机森林(AUC 0.706),食管毒性为朴素贝叶斯(AUC 0.721),均优于逻辑回归。识别出肺毒性平均剂量>18 Gy(OR 2.467)和V20>37%(OR 2.722)阈值,食管毒性平均剂量>34 Gy(OR 4.006)和V20>37%(OR 3.725)阈值,心脏毒性未发现显著阈值。
要点
- 肺毒性阈值:平均剂量>18 Gy(OR 2.467,95%CI 1.049-5.800,P=0.038)和V20>37%(OR 2.722,95%CI 1.034-7.163,P=0.043)。
- 食管毒性阈值:平均剂量>34 Gy(OR 4.006,95%CI 2.183-7.354,P<0.001)和V20>37%(OR 3.725,95%CI 1.308-10.603,P=0.014)。
- 心脏毒性未识别出显著剂量学阈值。
核心结论
- 在该数据集中,机器学习方法验证了已知的剂量学阈值,且在预测毒性方面优于逻辑回归。
- 使用可解释人工智能可能识别出临床有用的剂量学阈值,并随后进行外部验证。
研究设计与样本
- 本研究为RTOG 0617前瞻性临床试验的二次分析。
- 训练多个机器学习模型,使用临床和剂量学特征预测≥3级肺、心脏和食管毒性。
- 模型性能通过曲线下面积(AUC)评估,最佳模型使用Shapley加法解释(SHAP)框架解释,SHAP值转换为比值比(OR)及自举法生成的置信区间(CI),阈值通过逻辑回归验证。
主要发现
- 肺毒性最佳模型为极端梯度提升(AUC 0.739),心脏毒性最佳模型为随机森林(AUC 0.706),食管毒性最佳模型为朴素贝叶斯(AUC 0.721),均优于逻辑回归。
- 肺毒性阈值:平均剂量>18 Gy(OR 2.467;95%CI 1.049-5.800;P=0.038)和V20>37%(OR 2.722;95%CI 1.034-7.163;P=0.043)。
- 食管毒性阈值:平均剂量>34 Gy(OR 4.006;95%CI 2.183-7.354;P<0.001)和V20>37%(OR 3.725;95%CI 1.308-10.603;P=0.014)。
- 心脏毒性未识别出显著阈值。
关键数值与推荐
- 肺毒性:平均剂量>18 Gy(OR 2.467;95%CI 1.049-5.800;P=0.038)和V20>37%(OR 2.722;95%CI 1.034-7.163;P=0.043)。
- 食管毒性:平均剂量>34 Gy(OR 4.006;95%CI 2.183-7.354;P<0.001)和V20>37%(OR 3.725;95%CI 1.308-10.603;P=0.014)。
- 心脏毒性:未识别出显著阈值。
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: Dosimetric predictors of toxicity in patients treated with definitive chemoradiation for locally advanced non-small cell lung cancer are often identified through trial and error. This study used machine learning (ML) and explainable artificial intelligence to empirically characterize dosimetric predictors of toxicity in patients treated as part of a prospective clinical trial. METHODS AND MATERIALS: A secondary analysis of the Radiation Therapy Oncology Group (RTOG) 0617 trial was performed. Multiple ML models were trained to predict grade ≥3 pulmonary, cardiac, and esophageal toxicities using clinical and dosimetric features. Model performance was evaluated using the area under the curve (AUC). The best performing model for each toxicity was explained using the Shapley Additive Explanation (SHAP) framework; SHAP values were used to identify relevant dosimetric thresholds and were converted to odds ratios (ORs) with confidence intervals (CIs) generated using bootstrapping to obtain quantitative measures of risk. Thresholds were validated using logistic regression. RESULTS: The best-performing models for pulmonary, cardiac, and esophageal toxicities, outperforming logistic regression, were extreme gradient boosting (AUC, 0.739), random forest (AUC, 0.706), and naive Bayes (AUC, 0.721), respectively. For pulmonary toxicity, thresholds of a mean dose >18 Gy (OR, 2.467; 95% CI, 1.049-5.800; P = .038) and lung volume receiving ≥20 Gy (V20) > 37% (OR, 2.722; 95% CI, 1.034-7.163; P = .043) were identified. For esophageal toxicity, thresholds of a mean dose >34 Gy (OR, 4.006; 95% CI, 2.183-7.354; P < .001) and V20 > 37% (OR, 3.725; 95% CI, 1.308-10.603; P = .014) were identified. No significant thresholds were identified for cardiac toxicity. CONCLUSIONS: In this data set, ML approaches validated known dosimetric thresholds and outperformed logistic regression at predicting toxicity. Furthermore, using explainable artificial intelligence, clinically useful dosimetric thresholds might be identified and subsequently externally validated.
引用
- Ladbury C, Li R, Danesharasteh A, et al. Explainable Artificial Intelligence to Identify Dosimetric Predictors of Toxicity in Patients with Locally Advanced Non-Small Cell Lung Cancer: A Secondary Analysis of RTOG 0617. Int J Radiat Oncol Biol Phys. 2023. PMID: 37406826. doi:10.1016/j.ijrobp.2023.06.019
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。