Multiparametric analysis of amide proton transfer imaging for preoperative risk stratification of pediatric neuroblastoma

Neuroblastoma (NB) is most commonly diagnosed in young children, with high-risk NB indicating a poor prognosis and requiring aggressive treatment. To investigate the feasibility of multi-parametric amide proton transfer (APT) imaging for preoperative risk stratification in children with abdominal NB. This prospective study enrolled 121 consecutive pediatric volunteers with suspected NB, and all subjects underwent abdominal APT imaging on a 3T MRI scanner. Four APT-related metrics (CESTR, CESTR nr , MTR Rex , and AREX) were measured and a semi-automatic tumor segmentation method was applied after two experienced radiologists delineated initial tumor regions. Single-metric and multi-metric models were constructed from the four APT metrics using seven commonly used machine learning classifiers. The performance of the models in predicting the risk groups of NB patients was evaluated using the area under the receiver operating characteristic curve (AUC) and compared with the mean indices. Fifty-eight cases (mean age, 41.68 ± 29.32 months; 24 non-high-risk and 34 high-risk ones) were included in the final analysis. APT metrics identified NB risk groups more effectively than conventional MR images. The single-metric models (max AUC = 0.81) achieved better performance than the mean APT values (max AUC = 0.60). Moreover, the single-metric models based on the MTR Rex metric yielded higher AUCs (AUC = 0.67 ~ 0.81) than the models based on the other three APT metrics (AUC = 0.62 ~ 0.72). The multi-metric model integrating MTR Rex with CESTR or AREX, using the least absolute shrinkage and selection operator classifier, further improved the stratification performance, achieving an AUC of 0.83. Models established APT metrics, particularly MTR Rex , showed potential clinical aid for identifying high-risk abdominal NB in children. Integrating multiple APT metrics can further enhance the risk stratification ability of pre-treatment NB, providing a novel tool for personalized management of NB.