Dataset Information

MRI-based radiomics for prognosis of pediatric diffuse intrinsic pontine glioma: an international study.

ABSTRACT:

Background

Diffuse intrinsic pontine gliomas (DIPGs) are lethal pediatric brain tumors. Presently, MRI is the mainstay of disease diagnosis and surveillance. We identify clinically significant computational features from MRI and create a prognostic machine learning model.

Methods

We isolated tumor volumes of T1-post-contrast (T1) and T2-weighted (T2) MRIs from 177 treatment-naïve DIPG patients from an international cohort for model training and testing. The Quantitative Image Feature Pipeline and PyRadiomics was used for feature extraction. Ten-fold cross-validation of least absolute shrinkage and selection operator Cox regression selected optimal features to predict overall survival in the training dataset and tested in the independent testing dataset. We analyzed model performance using clinical variables (age at diagnosis and sex) only, radiomics only, and radiomics plus clinical variables.

Results

All selected features were intensity and texture-based on the wavelet-filtered images (3 T1 gray-level co-occurrence matrix (GLCM) texture features, T2 GLCM texture feature, and T2 first-order mean). This multivariable Cox model demonstrated a concordance of 0.68 (95% CI: 0.61-0.74) in the training dataset, significantly outperforming the clinical-only model (C = 0.57 [95% CI: 0.49-0.64]). Adding clinical features to radiomics slightly improved performance (C = 0.70 [95% CI: 0.64-0.77]). The combined radiomics and clinical model was validated in the independent testing dataset (C = 0.59 [95% CI: 0.51-0.67], Noether's test P = .02).

Conclusions

In this international study, we demonstrate the use of radiomic signatures to create a machine learning model for DIPG prognostication. Standardized, quantitative approaches that objectively measure DIPG changes, including computational MRI evaluation, could offer new approaches to assessing tumor phenotype and serve a future role for optimizing clinical trial eligibility and tumor surveillance.

SUBMITTER: Tam LT

PROVIDER: S-EPMC8095337 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Background and purposeThe conventional MR imaging appearance of diffuse intrinsic pontine glioma suggests intralesional histopathologic heterogeneity, and various distinct lesion components, including T2-hypointense foci, have been described. Here we report the prevalence, conventional MR imaging semiology, and advanced MR imaging features of non-necrotic T2-hyperintense foci in diffuse intrinsic pontine glioma.Materials and methodsTwenty-five patients with diffuse intrinsic pontine gliomas were included in this study. MR imaging was performed at 3T by using conventional and advanced MR imaging sequences. Perfusion (CBV), vascular permeability (ve, Ktrans), and diffusion (ADC) metrics were calculated and used to characterize non-necrotic T2-hyperintense foci in comparison with other lesion components, namely necrotic T2-hyperintense foci, T2-hypointense foci, peritumoral edema, and normal brain stem. Statistical analysis was performed by using Kruskal-Wallis and Wilcoxon rank sum tests.ResultsSixteen non-necrotic T2-hyperintense foci were found in 12 tumors. In these foci, ADC values were significantly higher than those in either T2-hypointense foci (P = .002) or normal parenchyma (P = .0002), and relative CBV values were significantly lower than those in either T2-hypointense (P = .0002) or necrotic T2-hyperintense (P = .006) foci. Volume transfer coefficient values in T2-hyperintense foci were lower than those in T2-hypointense (P = .0005) or necrotic T2-hyperintense (P = .0348) foci.ConclusionsNon-necrotic T2-hyperintense foci are common, distinct lesion components within diffuse intrinsic pontine gliomas. Advanced MR imaging data suggest low cellularity and an early stage of angioneogenesis with leaky vessels resulting in expansion of the extracellular space. Because of the lack of biopsy validation, the underlying histoarchitectural and pathophysiologic changes remain unclear; therefore, these foci may correspond to a poorly understood biologic event in tumor evolution.