Background: Radiologic volumetric evaluation of Wilms’ tumor (WT) is an important indicator to guide treatment decisions. However, due to the heterogeneity of the tumors, radiologists have main-guard differences in diagnosis that can lead to misdiagnosis and poor treatment. The aim of this study was to explore whether CT-based outlining of WT foci can be automated using deep learning. Methods: We included CT intravenous phase images of 105 patients with WT and double-blind outlining of lesions by two radiologists. Then, we trained an automatic segmentation model using nnUnet. The Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (HD₉₅) were used to assess the performance. Next, we optimized the automatic segmentation results based on the ratio of the three-dimensional diameter of the lesion to improve the performance of volumetric assessment. Results: The DSC and HD₉₅ was 0.83 ± 0.22 and 10.50 ± 8.98 mm. The absolute difference and percentage difference in tumor size was 72.27 ± 134.84 cm³ and 21.08% ± 30.46%. After optimization according to our method, it decreased to 40.22 ± 96.06 cm³ and 10.16% ± 9.70%. Conclusion: We introduce a novel method that enhances the accuracy of predicting WT volume by integrating AI automated outlining and 3D tumor diameters. This approach surpasses the accuracy of using AI outcomes alone and has the potential to enhance the clinical evaluation of pediatric patients with WT. By intertwining AI outcomes with clinical data, this method becomes more interpretive and offers promising applications beyond Wilms tumor, extending to other pediatric diseases.

Background: Medulloblastoma (MB) is classified into different molecular subgroups: wingless (WNT), sonic hedgehog (SHH), group 3, and group 4. This study aims to investigate the potential of the molecular-structural mismatch signs for distinguishing the MB molecular subgroups in children non-invasively. Methods: In this prospective study, 60 newly diagnosed pediatric MB patients underwent conventional structural MRI and molecular amide proton transfer (APT) MRI from February 2018 to April 2022. Molecular-structural mismatch images were generated based on the APTw, T2-weighted (T2w), and fluid-attenuated inversion recovery (FLAIR) images. The mean values were calculated from the region of interest defined on APTw and mismatch images. The classification performance of mean values was assessed using the receiver operating characteristic (ROC) analysis. The statistical significance of differences between the area under the curve (AUC) values was evaluated using the DeLong test. Results: In total, 46 participants (mean age: 75.1±43.4 months; 33 males) were included in the final analysis: 5 WNT, 13 SHH, 9 group 3, and 19 group 4 patients. The AUC values of the mean values of APTw images for discriminating the WNT, SHH, group 3, and group 4 subgroups were 0.87, 0.66, 0.56, and 0.56, respectively. The AUC values for distinguishing the WNT, SHH, group 3, and group 4 subgroups with the mean values of molecular-structural mismatch images were improved to 0.90, 0.67, 0.68, and 0.77, respectively. Conclusions: Molecular-structural mismatch signs have the potential for distinguishing the molecular subgroups of MB in children non-invasively.

Medulloblastoma (MB), a malignant intracranial tumor, is classified into four molecular subgroups (WNT, SHH, Group 3, Group 4) with distinct risk profiles. This study aimed to establish APT MRI-based radiomic signature for distinguishing risk subgroups of pediatric MB. From February 2018 to May 2023, 72 diagnosed pediatric MB patients were enrolled, all undergoing brain MRI examinations. Radiomic features were extracted from regions of interest in APT-related metric maps and structural MR images, and were subject to a feature selection process. The performance of retained individual features and radiomic signatures generated via the least absolute shrinkage and selection operator (LASSO) was assessed using receiver operating characteristic (ROC) analysis. Final analysis included 56 subjects (mean age: 76.5 ± 41.9 months; 37 males), consisting of 7 WNT, 17 SHH, 9 Group 3, and 23 Group 4 cases. The optimal individual radiomic feature exhibited an area under the ROC curve (AUC) of 0.83 for risk stratification, while the APT-based radiomic signature achieved a higher AUC of 0.86. Notably, structural images and clinical factors failed to provide additional value to the APT-based signature for risk stratification. In conclusion, APT MRI-based radiomic signature demonstrated favorable performance in distinguishing risk groups of pediatric MB, offering valuable insights for clinical diagnosis of MB.

Neuroblastoma (NB) is most commonly diagnosed in young children, with high-risk NB indicating a poor prognosis and requiring aggressive treatment. To inves