From the Departments of Radiology (X.Y.Z., H.T.Z., M.Y., X.T.L., Y.J.S., H.C.Z., Y.S.S.), Gastrointestinal Surgery (L.W.), and Pathology (Z.W.L.), Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Peking University Cancer Hospital & Institute, No. 52 Fu Cheng Rd, Hai Dian District, Beijing 100142, China.
Radiology. 2020 Jul;296(1):56-64. doi: 10.1148/radiol.2020190936. Epub 2020 Apr 21.
Background Preoperative response evaluation with neoadjuvant chemoradiotherapy remains a challenge in the setting of locally advanced rectal cancer. Recently, deep learning (DL) has been widely used in tumor diagnosis and treatment and has produced exciting results. Purpose To develop and validate a DL method to predict response of rectal cancer to neoadjuvant therapy based on diffusion kurtosis and T2-weighted MRI. Materials and Methods In this prospective study, participants with locally advanced rectal adenocarcinoma (≥cT3 or N+) proved at histopathology and baseline MRI who were scheduled to undergo preoperative chemoradiotherapy were enrolled from October 2015 to December 2017 and were chronologically divided into 308 training samples and 104 test samples. DL models were constructed primarily to predict pathologic complete response (pCR) and secondarily to assess tumor regression grade (TRG) (TRG0 and TRG1 vs TRG2 and TRG3) and T downstaging. Other analysis included comparisons of diffusion kurtosis MRI parameters and subjective evaluation by radiologists. Results A total of 383 participants (mean age, 57 years ± 10 [standard deviation]; 229 men) were evaluated (290 in the training cohort, 93 in the test cohort). The area under the receiver operating characteristic curve (AUC) was 0.99 for the pCR model in the test cohort, which was higher than the AUC for raters 1 and 2 (0.66 and 0.72, respectively; < .001 for both). AUC for the DL model was 0.70 for TRG and 0.79 for T downstaging. AUC for pCR with the DL model was better than AUC for the best-performing diffusion kurtosis MRI parameters alone (diffusion coefficient in normal diffusion after correcting the non-Gaussian effect [ value] before neoadjuvant therapy, AUC = 0.76). Subjective evaluation by radiologists yielded a higher error rate (1 - accuracy) (25 of 93 [26.9%] and 23 of 93 [24.8%] for raters 1 and 2, respectively) in predicting pCR than did evaluation with the DL model (two of 93 [2.2%]); the radiologists achieved a lower error rate (12 of 93 [12.9%] and 13 of 93 [14.0%] for raters 1 and 2, respectively) when assisted by the DL model. Conclusion A deep learning model based on diffusion kurtosis MRI showed good performance for predicting pathologic complete response and aided the radiologist in assessing response of locally advanced rectal cancer after neoadjuvant chemoradiotherapy. © RSNA, 2020 See also the editorial by Koh in this issue.
背景 局部晚期直肠癌新辅助放化疗的术前疗效评估仍然是一个挑战。最近，深度学习（DL）已广泛应用于肿瘤的诊断和治疗，并取得了令人兴奋的结果。目的 开发和验证一种基于扩散峰度和 T2 加权 MRI 预测直肠癌对新辅助治疗反应的 DL 方法。材料与方法 本前瞻性研究纳入了 2015 年 10 月至 2017 年 12 月在组织病理学和基线 MRI 证实为局部晚期直肠腺癌（≥cT3 或 N+）并计划接受术前放化疗的患者，根据入组时间先后将患者分为 308 例训练样本和 104 例测试样本。主要构建 DL 模型预测病理完全缓解（pCR），其次评估肿瘤消退分级（TRG）（TRG0 和 TRG1 与 TRG2 和 TRG3）和 T 分期降级。其他分析包括比较扩散峰度 MRI 参数和放射科医生的主观评估。结果 共评估了 383 例患者（平均年龄 57 岁±10[标准差]；229 例男性）（训练队列 290 例，测试队列 93 例）。测试队列中 pCR 模型的受试者工作特征曲线下面积（AUC）为 0.99，高于评分者 1 和 2（分别为 0.66 和 0.72；均<0.001）。DL 模型的 AUC 用于 TRG 和 T 分期降级分别为 0.70 和 0.79。DL 模型预测 pCR 的 AUC 优于最佳扩散峰度 MRI 参数的 AUC（新辅助治疗前校正非高斯效应后的扩散系数[ 值]，AUC=0.76）。放射科医生的主观评估在预测 pCR 方面的准确率（1-准确性）（评分者 1 为 25/93[26.9%]，评分者 2 为 23/93[24.8%]）高于 DL 模型（93 例中有 2 例[2.2%]）；当放射科医生使用 DL 模型辅助时，其准确率（评分者 1 为 12/93[12.9%]，评分者 2 为 13/93[14.0%]）较低。结论 基于扩散峰度 MRI 的深度学习模型在预测病理完全缓解方面表现良好，并有助于放射科医生评估局部晚期直肠癌新辅助放化疗后的反应。©RSNA，2020 本期杂志还刊登了 Koh 等人的述评。
