前列腺癌诊断及预后评估中MRI影像组学的应用分析

Application of MRI in Diagnosis and Prognosis Evaluation of Prostate Cancer

目的:分析应用MRI影像组学诊断和预后评估前列腺癌的技术要点和临床价值。方法:选取2021年1月~2022年12月疑似前列腺癌患者120例,采用MRI影像组学方法和病理检查,提取影像纹理特征和构建模型,分析病理检查结果、主要纹理特征,评价模型应用价值。结果:病理检查显示,前列腺癌61例,前列腺增生59例;影像纹理特征显示,共6类1129个;模型分析显示,前列腺病灶预测模型应用效能良好,训练组AUC0.87,95%可信任区间(0.72~1.00),敏感度81.0%,特异度90.0%,模型准确率100.0%,阳性预测值100.0%,阴性预测值85.00%,验证组AUC0.98,95%可信任区间(0.93~1.00),敏感度85.0%,特异度85.0%,模型准确率85.0%,阳性预测值85.0%,阴性预测值85.0%。结论:MRI影像组学可辅助医师有效检出前列腺病变,促进病灶识别,以ADC图像为基础确定影像组学特征搭建模型,可比较准确地判别前列腺癌与前列腺增生,指导临床医师科学治疗疾病。

Objective:To analyze the technical points and clinical value of MRI imaging in diagnosis and prognosis evaluation of prostate cancer.Methods:A total of 120 patients with prostate cancer from January 2021 to December 2022 were selected.MRI imaging omics method and pathological examination were used to extract image texture features and construct models,and pathological examination results and main texture features were analyzed to evaluate the application value of the models.Results:Pathological examination showed 61 cases of prostate cancer and 59 cases of prostate hyperplasia.The image texture features showed 1129 images in 6 categories.Model analysis showed that the prostatic lesion prediction model had good application efficacy,with AUC 0.87,95%confidence interval(0.72-1.00),sensitivity 81.0%,specificity 90.0%,model accuracy 100.0%,positive predictive value 100.0%and negative predictive value 85.00%in the training group.In the verification group,AUC 0.98,95%confidence interval(0.93-1.00),sensitivity 85.0%,specificity 85.0%,model accuracy 85.0%,positive predictive value 85.0%,negative predictive value 85.0%.Conclusions:MRI imaging omics can assist physicians to effectively detect prostate lesions and promote lesion identification.Based on ADC images,imaging features can be determined and models built,which can more accurately distinguish prostate cancer and prostate hyperplasia and guide clinicians to treat diseases scientifically.