Search

Objectives: Currently, the BRAF status of pediatric low-grade glioma (pLGG) patients is determined through a biopsy. We established a nomogram to predict BRAF status non-invasively using clinical and radiomic factors. Additionally, we assessed an advanced thresholding method to provide only high-confidence predictions for the molecular subtype. Finally, we tested whether radiomic features provide additional predictive information for this classification task, beyond that which is embedded in the location of the tumor. Methods: Random forest (RF) models were trained on radiomic and clinical features both separately and together, to evaluate the utility of each feature set. Instead of using the traditional single threshold technique to convert the model outputs to class predictions, we implemented a double threshold mechanism that accounted for uncertainty. Additionally, a linear model was trained and depicted graphically as a nomogram. Results: The combined RF (AUC: 0.925) outperformed the RFs trained on radiomic (AUC: 0.863) or clinical (AUC: 0.889) features alone. The linear model had a comparable AUC (0.916), despite its lower complexity. Traditional thresholding produced an accuracy of 84.5%, while the double threshold approach yielded 92.2% accuracy on the 80.7% of patients with the highest confidence predictions. Conclusion: Models that included radiomic features outperformed, underscoring their importance for the prediction of BRAF status. A linear model performed similarly to RF but with the added benefit that it can be visualized as a nomogram, improving the explainability of the model. The double threshold technique was able to identify uncertain predictions, enhancing the clinical utility of the model. Clinical relevance statement: Radiomic features and tumor location are both predictive of BRAF status in pLGG patients. We show that they contain complementary information and depict the optimal model as a nomogram, which can be used as a non-invasive alternative to biopsy. Key Points: • Radiomic features provide additional predictive information for the determination of the molecular subtype of pediatric low-grade gliomas patients, beyond what is embedded in the location of the tumor, which has an established relationship with genetic status. • An advanced thresholding method can help to distinguish cases where machine learning models have a high chance of being (in)correct, improving the utility of these models. • A simple linear model performs similarly to a more powerful random forest model at classifying the molecular subtype of pediatric low-grade gliomas but has the added benefit that it can be converted into a nomogram, which may facilitate clinical implementation by improving the explainability of the model. [ABSTRACT FROM AUTHOR]

Low grade gliomas are the most frequent brain tumors in children and encompass a spectrum of histologic entities which are currently assigned World Health Organisation grades I and II. They differ substantially from their adult counterparts in both their underlying genetic alterations and in the infrequency with which they transform to higher grade tumors. Nonetheless, children with low grade glioma are a therapeutic challenge due to the heterogeneity in their clinical behavior – in particular, those with incomplete surgical resection often suffer repeat progressions with resultant morbidity and, in some cases, mortality. The identification of up-regulation of the RAS–mitogen-activated protein kinase (RAS/MAPK) pathway as a near universal feature of these tumors has led to the development of targeted therapeutics aimed at improving responses while mitigating patient morbidity. Here, we review how molecular information can help to further define the entities which fall under the umbrella of pediatric-type low-grade glioma. In doing so we discuss the specific molecular drivers of pediatric low grade glioma and how to effectively test for them, review the newest therapeutic agents and their utility in treating this disease, and propose a risk-based stratification system that considers both clinical and molecular parameters to aid clinicians in making treatment decisions. [ABSTRACT FROM AUTHOR]

Background: The determinants of outcomes for adult survivors of pediatric low-grade glioma (PLGG) are largely unknown.Methods: This study collected population-based follow-up information for all PLGG patients diagnosed in Ontario, Canada from 1985 to 2012 (n = 1202) and determined factors affecting survival. The impact of upfront radiation treatment on overall survival (OS) was determined for a cohort of Ontario patients and an independent reference cohort from the Surveillance, Epidemiology, and End Results database.Results: At a median follow-up of 12.73 years (range, 0.02-33 years), only 93 deaths (7.7%) were recorded, and the 20-year OS rate was 90.1% ± 1.1%. Children with neurofibromatosis type 1 had excellent survival and no tumor-related deaths during adulthood. Adverse risk factors included pleomorphic xanthoastrocytoma (P < .001) and a thalamic location (P < .001). For patients with unresectable tumors surviving more than 5 years after the diagnosis, upfront radiotherapy was associated with an approximately 3-fold increased risk of overall late deaths (hazard ratio [HR], 3.3; 95% confidence interval [CI], 1.6-6.6; P = .001) and an approximately 4-fold increased risk of tumor-related deaths (HR, 4.4; 95% CI, 1.3-14.6; P = .013). In a multivariate analysis, radiotherapy was the most significant factor associated with late all-cause deaths (HR, 3.0; 95% CI, 1.3-7.0; P = .012) and tumor-related deaths (HR, 4.4; 95% CI, 1.3-14.6; P = 0.014). A similar association between radiotherapy and late deaths was observed in the independent reference cohort (P < .001). In contrast to early deaths, late mortality was associated not with PLGG progression but rather with tumor transformation and non-oncological causes.Conclusions: The course of PLGG is associated with excellent long-term survival, but this is hampered by increased delayed mortality in patients receiving upfront radiotherapy. These observations should be considered when treatment options are being weighed for these patients. [ABSTRACT FROM AUTHOR]

Introduction: Diffuse intrinsic pontine glioma (DIPG) and midline high-grade glioma (mHGG) are lethal childhood brain tumors. Spatial genomic heterogeneity has been well-described in adult HGG but has not been comprehensively characterized in pediatric HGG. We performed whole exome sequencing on 38-matched primary, contiguous, and metastatic tumor sites from eight children with DIPG (n = 7) ormHGG (n = 1) collected using a unique MRI-guided autopsy protocol. Validation was performed using Sanger sequencing, Droplet Digital polymerase-chain reaction, immunohistochemistry, and fluorescent in-situ hybridization. Results: Median age at diagnosis was 6.1 years (range: 2.9-23.3 years). Median overall survival was 13.2 months (range: 11.2-32.2 months). Contiguous tumor infiltration and distant metastases were observed in seven and six patients, respectively, including leptomeningeal dissemination in three DIPGs. Histopathological heterogeneity was evident in seven patients, including intra-pontine heterogeneity in two DIPGs, ranging from World Health Organization grade II to IV astrocytoma. We found conservation of heterozygous K27M mutations in H3F3A (n = 4) or HIST1H3B (n = 3) across all primary, contiguous, and metastatic tumor sites in all DIPGs. ACVR1 (n = 2), PIK3CA (n = 2), FGFR1 (n = 2), and MET (n = 1) were also intra-tumorally conserved. ACVR1 was co-mutated with HIST1H3B (n = 2). In contrast, PDGFRA amplification and mutation were spatially heterogeneous, as were mutations in BCOR (n = 1), ATRX (n = 2), and MYC (n = 1). TP53 aberrations (n = 3 patients) varied by type and location between primary and metastatic tumors sites but were intra-tumorally conserved. Conclusion: Spatial conservation of prognostically-relevant and therapeutically-targetable somatic mutations in DIPG and mHGG contrasts the significant heterogeneity of driver mutations seen in adult HGG and supports uniform implementation of diagnostic biopsy in DIPG and mHGG to classify molecular risk groups and guide therapeutic strategy. [ABSTRACT FROM AUTHOR]

Diffuse intrinsic pontine glioma (DIPG) is a devastating pediatric brain cancer with no effective therapy. Histological similarity of DIPG to supratentorial high-grade astrocytomas of adults has led to assumptions that these entities possess similar underlying molecular properties and therefore similar therapeutic responses to standard therapies. The failure of all clinical trials in the last 30years to improve DIPG patient outcome has suggested otherwise. Recent studies employing next-generation sequencing and microarray technologies have provided a breadth of evidence highlighting the unique molecular genetics and epigenetics of this cancer, distinguishing it from both adult and pediatric cerebral high-grade astrocytomas. This review describes the most common molecular genetic and epigenetic signatures of DIPG in the context of molecular subgroups and histopathological diagnosis, including this tumor entity's unique mutational landscape, copy number alterations, and structural variants, as well as epigenetic changes on the global DNA and histone levels. The increased knowledge of DIPG biology and histopathology has opened doors to new diagnostic and therapeutic avenues. [ABSTRACT FROM AUTHOR]

Diffuse intrinsic pontine glioma (DIPG) is the main cause of brain tumour-related death in children. In the majority of cases diagnosis is based on clinical and MRI findings, resulting in the scarcity of pre-treatment specimens available to study. Our group has developed an autopsy-based protocol to investigate the histologic and biologic spectrum of DIPG. This has also allowed us to investigate the terminal pattern of disease and gain a better understanding of what challenges we are facing in treating DIPG. Here, we review 72 DIPG cases with well documented clinical history and molecular data and describe the pathological features of this disease in relation to clinical and genetic features. Fifty-three of the samples were autopsy material (7 pre-treatment) and 19 were pre-treatment biopsy/surgical specimens. Upon histological review, 62 patients had high-grade astrocytomas (18 WHO grade III and 44 WHO grade IV patients), 8 had WHO grade II astrocytomas, and 2 had features of primitive neuroectodermal tumour (PNET). K27M-H3 mutations were exclusively found in tumours with WHO grade II-IV astrocytoma histology. K27M-H3.1 and ACVR1 mutations as well as ALT phenotype were only found in WHO grade III-IV astrocytomas, while PIK3CA mutations and PDGFRA gains/amplifications were found in WHO grade II-IV astrocytomas. Approximately 1/3 of DIPG patients had leptomeningeal spread of their tumour. Further, diffuse invasion of the brainstem, spinal cord and thalamus was common with some cases showing spread as distant as the frontal lobes. These findings suggest that focal radiation may be inadequate for some of these patients. Importantly, we show that clinically classic DIPGs represent a diverse histologic spectrum, including multiple cases which would fit WHO criteria of grade II astrocytoma which nevertheless behave clinically as high-grade astrocytomas and harbour the histone K27M-H3.3 mutation. This suggests that the current WHO astrocytoma grading scheme may not appropriately predict outcome for paediatric brainstem gliomas. [ABSTRACT FROM AUTHOR]