CT影像组学预测局部进展期直肠癌新辅助治疗的效果

目的探讨CT影像组学在治疗前预测局部进展期直肠癌新辅助治疗效果的价值。资料与方法回顾性分析168例新辅助治疗后行根治术的局部进展期直肠癌患者,收集治疗前临床及CT资料,根据术后病理肿瘤退缩分级分组。采用A.K.软件提取CT影像组学特征并构建影像组学标签。通过多变量Logistic回归筛选疗效预测因子并构建诺莫图模型。利用ROC曲线评价模型诊断效能,并对模型进行内部验证、校准度评价及临床应用价值分析。结果每例患者各提取了396个CT影像组学特征,降维后筛选出6个与局部进展期直肠癌新辅助治疗效果高度相关的特征。联合独立预测因子影像组学标签、癌胚抗原≥3.4 ng/ml和临床T分期(cT4)构建的诺莫图模型ROC曲线下面积(0.881)高于影像组学标签(0.791),且具有较高的校准度、内部验证一致性及临床应用价值(P>0.05)。结论基于治疗前CT及临床资料构建的模型对局部进展期直肠癌新辅助治疗效果预测具有较高的预测效能,且联合预测模型的预测效能优于影像组学标签。

Efficacy of CT Radiomics in Predicting Neoadjuvant Therapy of Locally Advanced Rectal Cancer

Purpose To investigate the value of CT radiomics to predict efficacy of neoadjuvant therapy in locally advanced rectal cancer before treatment.Materials and Methods One hundred and sixty-eight patients with locally advanced rectal cancer who received neoadjuvant therapies were included retrospectively.Clinical and CT material before treatment were collected and the patients were grouped in line with postoperative tumor regression grading.A.K.software was used to extract CT radiomics features and construct radiomics tags.Multivariable Logistic regression was adopted to screen efficacy predictive factors to construct Nomogram.ROC curve was utilized to evaluate diagnostic efficiency of the model.Meanwhile,internal validation,calibration assessment and clinical application value of the model were analyzed.Results Three hundred and ninety-six CT radiomics features of each patient were extracted and 6 features highly related to efficacy of neoadjuvant therapy in locally advanced rectal cancer were screened after dimension reduction.Area under ROC curve(0.881)of Nomogram constructed by tags of combined independent predictive factors radiomics,carcino-embryonic antigen≥3.4 ng/ml and clinical T phasing(cT4)was higher than tags of radiomics(0.791)and was of high calibration,consistency of internal validation and clinical application value(P>0.05).Conclusion Model constructed with CT and clinical materials before treatment is of higher predictive efficiency in predicting efficacy of neoadjuvant therapy in locally advanced rectal cancer and predictive efficiency of combined predictive model is optimal than radiomics tags.