Answers are grounded in this paper and cite their sources.
Randomized trial demonstrates AI-enabled MRI assessments in neuro-oncology, indicating improved accessibility for care.
Background Quantitative measures of brain tumor burden are critical for neuro-oncology care and clinical trials. However, MRI-based volumetric assessments remain inaccessible in most settings due to limited subspecialty expertise, narrow disease application, and lack of local computational infrastructure. We present HOPE4KIDS, a tumor- and age-agnostic, AI-enabled web platform that delivers standardized, automated MRI tumor segmentation and volumetric analysis, with performance validated on international benchmarks. Methods HOPE4KIDS performs fully automated multi-class segmentation on standard multi-sequence MRI(T1, contrast-enhanced T1, T2, T2-FLAIR) using an ensemble of state-of-the-art deep learning architectures(nnU-Net, MedNeXt, SwinUNETR)[1]. Radiomic feature-guided post-processing improves robustness across heterogeneous scanners, institutions, and tumor phenotypes(see Table 1). Segmentation performance was assessed using the Dice similarity coefficient(DSC). All trained models are deployed through a secure web application that performs computations entirely offsite. Results The different models of the HOPE4KIDS platform were benchmarked on multiple international datasets, and performance is summarized in Table 2. Best performance is denoted by a score 1.0 and worst perfromance by a score 0.0. Conclusions HOPE4KIDS offers drag-and-drop, tumor-and-age-agnostic MRI segmentation with internationally benchmarked performance, supporting free, accessible volumetric analysis for sites without computational infrastructure. The HIPAA-compliant platform and open-source models are ready to use at https://segmenter.hope4kids.io. 1. Capellán-Martín D, Jiang Z, Parida A, Liu X, Lam V, Nisar H, Tapp A, Elsharkawi S, Ledesma-Carbayo MJ, Anwar SM, Linguraru MG. Model ensemble for brain tumor segmentation in magnetic resonance imaging. InInternational Challenge on Cross-Modality Domain Adaptation for Medical Image Segmentation 2023 Oct 8 (pp. 221-232). Cham: Springer Nature Switzerland. 2. Zapaishchykova, A.,Vajapeyam, S., Liu, K.X., Poussaint, T.Y., Kann, B.H.. (2024) MR imaging of pediatric subjects with high-grade gliomas (DFCI-BCH-BWH-PEDs-HGG) [Dataset] (Version 1). The Cancer Imaging Archive. https://doi.org/10.7937/v8h6-bg25
No takes yet. Share an insight, caveat, or question.
Parida et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: