Automated tumor segmentation and brain tissue extraction from multiparametric MRI of pediatric brain tumors: A multi-institutional study

Abstract

Background Brain tumors are the most common solid tumors and the leading cause of cancer-related death among all childhood cancers. Tumor segmentation is essential in surgical and treatment planning, and response assessment and monitoring. However, manual segmentation is time-consuming and has high interoperator variability. We present a multi-institutional deep learning-based method for automated brain extraction and segmentation of pediatric brain tumors based on multi-parametric MRI scans. Methods Multi-parametric scans (T1w, T1w-CE, T2, and T2-FLAIR) of 244 pediatric patients (n=215 internal and n=29 external cohorts) with de novo brain tumors, including a variety of tumor subtypes, were preprocessed and manually segmented to identify the brain tissue and tumor subregions into four tumor subregions, i.e., enhancing tumor (ET), non-enhancing tumor (NET), cystic components (CC), and peritumoral edema (ED). The internal cohort was split into training (n=151), validation (n=43), and withheld internal test (n=21) subsets. DeepMedic, a three-dimensional convolutional neural network, was trained and the model parameters were tuned. Finally, the network was evaluated on the withheld internal and external test cohorts. Results Dice similarity score (median±SD) was 0.91±0.10/0.88±0.16 for the whole tumor, 0.73±0.27/0.84±0.29 for ET, 0.79±19/0.74±0.27 for union of all non-enhancing components (i.e., NET, CC, ED), and 0.98±0.02 for brain tissue in both internal/external test sets. Conclusions Our proposed automated brain extraction and tumor subregion segmentation models demonstrated accurate performance on segmentation of the brain tissue and whole tumor regions in pediatric brain tumors and can facilitate detection of abnormal regions for further clinical measurements.

Expert-level pediatric brain tumor segmentation in a limited data scenario with stepwise transfer learning

“…In this study, we developed, validated, and clinically benchmarked a deep learning pipeline using stepwise transfer learning for automated, expert-level pLGG segmentation and volumetric measurement. Accurate tumor auto-segmentation models could be useful for risk-stratification, monitoring tumor progression, assessing treatment response, and surgical approach (6), though have had limited traction in pediatric tumors due to very sparse available training data. We leveraged a novel strategy of in-domain, stepwise transfer learning to demonstrate measurable gains in segmentation accuracy and clinical acceptability that was on par with clinician performance.…”

“…Performance degradation may stem from the distinctive heterogeneous imaging appearance and types of pediatric brain tumors compared to adult brain tumors, as well as the anatomical differences resulting from the ongoing brain development in children. Several studies have proposed various DL solutions to address the segmentation of pediatric brain tumors, achieving Dice similarity coefficients (DSC) ranging between 0.68 and 0.88 (6,16,34,35), however the clinical acceptability of these approaches was not validated. To date, only one study has proposed an algorithm for pLGG segmentation, achieving a DSC of 0.77 (17).…”

“…Given the utility of accurate segmentation and the inherent limitations of manual segmentation, there has been interest in developing auto-segmentation tools for pediatric brain tumors (6–8). The progress in medical imaging techniques and computational methods have led to the development of various approaches for brain tumor segmentation, including machine learning-based methods, deep learning-based methods, and hybrid approaches (8–10).…”

“…Given the utility of accurate segmentation and the inherent limitations of manual segmentation, there has been interest in developing auto-segmentation tools for pediatric brain tumors (6)(7)(8).…”

Noninvasive molecular subtyping of pediatric low-grade glioma with self-supervised transfer learning

“…We report our results in accordance with the Checklist for Artificial Intelligence in Medical Imaging (CLAIM) guidelines (18). A portion of patients from the CBTN dataset (n=140) and an additional subset from the BCH dataset (n=100) had been utilized in two previous studies (10,19). It's worth highlighting that these prior investigations were centered around tumor segmentation, whereas the present study was primarily dedicated to identifying BRAF mutational subtypes.…”