Segmentation-Free Pretherapeutic Assessment of BRAF-Status in Pediatric Low-Grade Gliomas

Abstract Background BRAF status is crucial for treating pediatric low-grade gliomas (pLGG) and can be assessed non-invasively from segmented tumor regions on MRI using machine learning (ML). However, there are inherent limitations to manual and automated tumor segmentations. Purpose To assess the performance of automated segmentation algorithms and to develop and assess a segmentation-free ML classification pipeline that identifies BRAF status from whole-brain FLAIR MRI sequences. Materials and Methods In this REB-approved retrospective study, molecularly-characterized tumors and whole-brain FLAIR MR images were collected from 455 patients with pLGG treated between 1999 and 2023 at a single tertiary care children’s hospital. We trained and evaluated three medical segmentation models, TransBTS, MedNeXt, and MedicalNet. Next, we developed a model to identify BRAF status from whole-brain FLAIR MRI, without any reliance on manual or automated segmentations. We then implemented a novel pretraining regimen that embedded segmentation knowledge into the whole-brain FLAIR MRI classification model. Finally, we trained and evaluated a baseline model that used manual segmentations as inputs. All ML models were trained and evaluated under a nested-cross validation scheme, and mean performance across all test folds was compared using the corresponding t-test. Results The MedNeXt segmentation model (mean Dice score: 0.555) outperformed both the convolutional neural network (CNN) based MedicalNet (0.516) and the CNN-transformer hybrid TransBTS (0.449) (p <0.05 for all comparisons). The MedNeXt style classification model achieved a one-vs-rest area under the ROC curve of 0.741 using the whole brain FLAIR sequence as an input, without any segmentation knowledge. This was improved to 0.772 through pretraining on the segmentation task, which was not significantly different from the baseline manual segmentation-based model (0.756, p-value: 0.141). Conclusion BRAF status can be assessed non-invasively using ML models based on whole-brain FLAIR sequences. Dependence on inconsistent manual or automated segmentations can be reduced by integrating tumor region information into the model through pretraining.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.