一项前瞻性研究:利用光学与红外图像的放射组学特征实现放疗皮肤毒性的早期预测
该前瞻性研究纳入20例乳腺癌放疗患者,每日采集乳腺及胸壁光学与红外图像并每周评估皮肤毒性,从108个特征中筛选出9个与皮肤毒性显著相关的颜色和温度特征,构建机器学习模型预测皮肤毒性分级;当输入图像仅限生物有效剂量不超过30 Gy的照射次数时,立方支持向量机的交叉验证准确率仍大于85%,受试者工作特征曲线下面积仍大于0.75。
要点
- 从光学与红外图像提取的9个颜色和温度特征与皮肤毒性显著相关。
- 生物有效剂量不超过30 Gy时,立方支持向量机交叉验证准确率仍大于85%。
- 同一条件下受试者工作特征曲线下面积仍大于0.75。
核心结论
- 皮肤光学与红外图像的放射组学特征定量分析在预测皮肤毒性方面显示出前景。
研究设计与样本
- 前瞻性研究,纳入20例接受放疗的乳腺癌患者,放疗期间每日采集乳腺及胸壁光学与红外图像,放疗结束后3周内每周采集。
- 皮肤毒性评估每周进行直至患者末次访视。
主要发现
- 从108个特征中识别出9个与皮肤毒性显著相关的独立颜色和温度特征。
- 当输入图像数据仅限生物有效剂量不超过30 Gy(约占总治疗剂量的前三分之一至前一半)的照射次数时,立方支持向量机的交叉验证准确率仍大于85%,受试者工作特征曲线下面积仍大于0.75。
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: Approximately 90% of patients undergoing breast cancer radiation therapy experience skin toxicities that are difficult to classify and predict ahead of time. A prediction of toxicity at the early stages of the treatment would provide clinicians with a prompt to intervene. The objectives of this study were to evaluate the correlation between skin toxicity and radiomic features extracted from optical and infrared (thermal) images of skin, and to develop a model for predicting a patient's skin response to radiation. METHODS AND MATERIALS: Optical and infrared breast and chest-wall images were acquired daily during the course of radiation therapy, as well as weekly for 3 weeks after the end of treatment for 20 patients with breast cancer. Skin-toxicity assessments were conducted weekly until the patients' final visit. Skin color and temperature trends from histogram-based and texture-based radiomic features, extracted from the treatment area, were analyzed, reduced, and used in a cross-validation machine learning model to predict the patients' skin toxicity grades. RESULTS: A set of 9 independent color and temperature features with significant correlation to skin toxicity were identified from 108 features. The cross-validation accuracy of a cubic Support Vector Machine remained >85% and area under the receiver operating characteristic curve remained >0.75, when reducing the input imaging data to include only the sessions with a biologically effective dose not exceeding 30 Gy (approximately the first third to first half of the total treatment dose). CONCLUSIONS: The quantitative analysis of radiomic features extracted from optical and infrared (thermal) images of skin was shown to be promising for predicting skin toxicities.
引用
- Yashayaeva A, Dahn H, Svatos M, et al. A Prospective Study Demonstrating Early Prediction of Skin Toxicity From Radiation Therapy Using Radiomic Features From Optical and Infrared Images. Int J Radiat Oncol Biol Phys. 2023. PMID: 37778424. doi:10.1016/j.ijrobp.2023.09.043
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。