Department of Nuclear Medicine, Beijing Friendship Hospital, Capital Medical University, Beijing 100050, China.
Eur J Radiol. 2022 Sep;154:110444. doi: 10.1016/j.ejrad.2022.110444. Epub 2022 Jul 21.
To develop and validate an F-FDG PET/CT radiomics nomogram for non-invasive differentiation of high-risk and non-high-risk patients of the International Neuroblastoma Risk Group (INRG) Staging System (INRGSS).
One hundred thirty-nine neuroblastoma patients were retrospectively enrolled and classified into a training set (n = 84) and validation set (n = 55). Radiomics features were extracted from F-FDG PET/CT images, a radiomics signature was constructed, and a radiomics score (Rad score) was calculated. Then, univariate and multivariate logistic regression analyses were used to screen out the independent clinical factors and construct the clinical model. A radiomics nomogram was developed based on the Rad score and independent clinical factors. The performance of the clinical model, Rad score, and nomogram was assessed by receiver operating characteristic (ROC) curves, calibration curves, and decision curves analysis (DCA).
Seven radiomics features were selected to build the radiomics signature. The age at diagnosis, the INRG stage, neuron-specific enolase (NSE) and Rad score showed a significant difference between the high-risk and non-high-risk patients. The radiomics nomogram incorporating the Rad score and the above clinical factors demonstrated favorable predictive value for differentiating high-risk from non-high-risk, yielded AUCs of 0.988 and 0.971 in the training and validation sets, respectively. The calibration curves showed that the radiomics nomogram had the goodness of fit, and the DCA demonstrated that the radiomics nomogram was clinically useful.
The radiomics nomogram, which incorporates the Rad score and clinical factors can well predict high-risk and non-high-risk patients of the INRGSS. It may help the disease follow-up and management in clinical practice and assist in personalized and precise treatment of neuroblastoma.
开发和验证一种 F-FDG PET/CT 放射组学列线图，用于对国际神经母细胞瘤风险组（INRG）分期系统（INRGSS）中高危和非高危患者进行非侵入性区分。
回顾性纳入 139 例神经母细胞瘤患者，分为训练集（n=84）和验证集（n=55）。从 F-FDG PET/CT 图像中提取放射组学特征，构建放射组学特征签名，并计算放射组学评分（Rad 评分）。然后，进行单变量和多变量逻辑回归分析，筛选出独立的临床因素并构建临床模型。基于 Rad 评分和独立的临床因素，开发放射组学列线图。通过接受者操作特征（ROC）曲线、校准曲线和决策曲线分析（DCA）评估临床模型、Rad 评分和列线图的性能。
选择 7 个放射组学特征构建放射组学特征签名。诊断时年龄、INRG 分期、神经元特异性烯醇化酶（NSE）和 Rad 评分在高危和非高危患者之间存在显著差异。纳入 Rad 评分和上述临床因素的放射组学列线图在区分高危和非高危方面具有良好的预测价值，在训练集和验证集中的 AUC 分别为 0.988 和 0.971。校准曲线表明放射组学列线图具有良好的拟合度，DCA 表明放射组学列线图具有临床实用性。
纳入 Rad 评分和临床因素的放射组学列线图可以很好地预测 INRGSS 的高危和非高危患者。它可以帮助临床实践中的疾病随访和管理，并有助于神经母细胞瘤的个性化和精确治疗。
