Department of Urology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, People's Republic of China.
Guangdong Provincial Key Laboratory of Malignant Tumor Epigenetics and Gene Regulation, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, People's Republic of China.
Cancer. 2019 Dec 15;125(24):4388-4398. doi: 10.1002/cncr.32490. Epub 2019 Aug 30.
Bladder cancer (BCa) can be divided into muscle-invasive BCa (MIBC) and non-muscle-invasive BCa (NMIBC). Whether the tumor infiltrates the detrusor muscle is a critical determinant of disease management in patients with BCa. However, the current preoperative diagnostic accuracy of muscular invasiveness is less than satisfactory. The authors report a radiomic-clinical nomogram for the individualized preoperative differentiation of MIBC from NMIBC.
In total, 2602 radiomics features were extracted from whole bladder tumors and the basal part of the lesions on T2-weighted magnetic resonance imaging. Then, a radiomics signature was constructed using the least absolute shrinkage and selection operator algorithm in the training set (n = 130). Furthermore, a radiomic-clinical nomogram was developed incorporating the radiomics signature and selected clinical predictors based on a multivariable logistic regression analysis. The performance of the nomogram (discrimination, calibration, and clinical usefulness) was assessed and validated in an independent validation set (n = 69).
The radiomics signature, consisting of 23 selected features, showed good discrimination in the training and validation sets (area under the curve [AUC], 0.913 and 0.874, respectively). Incorporating the radiomics signature and magnetic resonance imaging-determined tumor size, the radiomic-clinical nomogram showed favorable calibration and discrimination in the training set with an AUC of 0.922, which was confirmed in the validation set (AUC, 0.876). Decision curve analysis and net reclassification improvement and integrated discrimination improvement indices (net reclassification improvement, 0.338, integrated discrimination improvement, 0.385) demonstrated the clinical usefulness of the nomogram.
The proposed noninvasive radiomic-clinical nomogram can increase the accuracy of preoperatively discriminating MIBC from NMIBC, which may aid in clinical decision making and improve patient prognosis.
膀胱癌（BCa）可分为肌层浸润性膀胱癌（MIBC）和非肌层浸润性膀胱癌（NMIBC）。肿瘤是否浸润逼尿肌是决定 BCa 患者治疗方案的关键因素。然而，目前术前判断肌层浸润的准确性仍不理想。作者报告了一种基于影像组学和临床的列线图模型，用于个体化术前鉴别 MIBC 和 NMIBC。
共从 2602 例膀胱癌患者的全膀胱肿瘤和病变基底部分 T2 加权磁共振成像上提取了 2602 个影像组学特征。然后，使用最小绝对收缩和选择算子算法在训练集（n=130）中构建影像组学特征。此外，还基于多变量逻辑回归分析，结合影像组学特征和选定的临床预测因子，开发了一个影像组学-临床列线图。通过在独立验证集（n=69）中评估和验证该列线图的性能（区分度、校准度和临床实用性）。
由 23 个选定特征组成的影像组学特征在训练集和验证集中均具有良好的区分度（曲线下面积[AUC]分别为 0.913 和 0.874）。结合影像组学特征和磁共振成像确定的肿瘤大小，影像组学-临床列线图在训练集上具有较好的校准度和区分度（AUC 为 0.922），在验证集上得到了验证（AUC 为 0.876）。决策曲线分析和净重新分类改善和综合判别改善指数（净重新分类改善 0.338，综合判别改善 0.385）表明该列线图具有临床实用性。
本研究提出的非侵入性影像组学-临床列线图可提高术前鉴别 MIBC 和 NMIBC 的准确性，有助于临床决策并改善患者预后。
