ABSTRACT
The use of targeted agents in the treatment of pediatric low-grade gliomas (pLGGs) relies on the determination of molecular status. It has been shown that genetic alterations in pLGG can be identified non-invasively using MRI-based radiomic features or convolutional neural networks (CNNs). We aimed to build and assess a combined radiomics and CNN non-invasive pLGG molecular status identification model. This retrospective study used the tumor regions, manually segmented from T2-FLAIR MR images, of 336 patients treated for pLGG between 1999 and 2018. We designed a CNN and Random Forest radiomics model, along with a model relying on a combination of CNN and radiomic features, to predict the genetic status of pLGG. Additionally, we investigated whether CNNs could predict radiomic feature values from MR images. The combined model (mean AUC: 0.824) outperformed the radiomics model (0.802) and CNN (0.764). The differences in model performance were statistically significant (p-values < 0.05). The CNN was able to learn predictive radiomic features such as surface-to-volume ratio (average correlation: 0.864), and difference matrix dependence non-uniformity normalized (0.924) well but was unable to learn others such as run-length matrix variance (- 0.017) and non-uniformity normalized (- 0.042). Our results show that a model relying on both CNN and radiomic-based features performs better than either approach separately in differentiating the genetic status of pLGGs, and that CNNs are unable to express all handcrafted features.
Subject(s)
Brain Neoplasms , Glioma , Magnetic Resonance Imaging , Neural Networks, Computer , Humans , Glioma/genetics , Glioma/diagnostic imaging , Glioma/pathology , Child , Female , Retrospective Studies , Male , Magnetic Resonance Imaging/methods , Brain Neoplasms/genetics , Brain Neoplasms/diagnostic imaging , Brain Neoplasms/pathology , Adolescent , Child, Preschool , Neoplasm Grading , InfantABSTRACT
Individuals with 21 trisomy or Down syndrome (DS) are known to have an increased risk of acute leukemia, while they rarely develop solid or central nervous system (CNS) tumors. Atypical teratoid rhabdoid tumor (ATRT) is a highly aggressive CNS-WHO grade 4 neoplasm, which has never been reported in association with Down syndrome. We present a case study of a 14-year-old female with Down syndrome, diagnosed with intradural-extramedullary spinal ATRT. The chief complaints included bilateral lower limb weakness, constipation, and urinary incontinence for 2 weeks. Surgery was scheduled, and a biopsy was taken. The histopathology, immunohistochemistry, and molecular analysis confirmed the diagnosis of the ATRT-MYC/group 2B subgroup. This report highlights the challenges of managing a patient with complex medical conditions. Moreover, it adds to the existing literature on CNS tumors in patients with Down syndrome.
Subject(s)
Down Syndrome , Rhabdoid Tumor , Teratoma , Humans , Down Syndrome/complications , Rhabdoid Tumor/complications , Rhabdoid Tumor/pathology , Female , Adolescent , Teratoma/pathology , Teratoma/complications , Teratoma/diagnosis , Spinal Cord Neoplasms/pathology , Spinal Cord Neoplasms/complicationsABSTRACT
With the success of immunotherapy in cancer, understanding the tumor immune microenvironment (TIME) has become increasingly important; however in pediatric brain tumors this remains poorly characterized. Accordingly, we developed a clinical immune-oncology gene expression assay and used it to profile a diverse range of 1382 samples with detailed clinical and molecular annotation. In low-grade gliomas we identify distinct patterns of immune activation with prognostic significance in BRAF V600E-mutant tumors. In high-grade gliomas, we observe immune activation and T-cell infiltrates in tumors that have historically been considered immune cold, as well as genomic correlates of inflammation levels. In mismatch repair deficient high-grade gliomas, we find that high tumor inflammation signature is a significant predictor of response to immune checkpoint inhibition, and demonstrate the potential for multimodal biomarkers to improve treatment stratification. Importantly, while overall patterns of immune activation are observed for histologically and genetically defined tumor types, there is significant variability within each entity, indicating that the TIME must be evaluated as an independent feature from diagnosis. In sum, in addition to the histology and molecular profile, this work underscores the importance of reporting on the TIME as an essential axis of cancer diagnosis in the era of personalized medicine.
Subject(s)
Brain Neoplasms , Glioma , Tumor Microenvironment , Humans , Tumor Microenvironment/immunology , Tumor Microenvironment/genetics , Brain Neoplasms/immunology , Brain Neoplasms/genetics , Brain Neoplasms/pathology , Child , Glioma/immunology , Glioma/genetics , Glioma/pathology , Biomarkers, Tumor/genetics , Biomarkers, Tumor/immunology , Female , Male , Adolescent , Gene Expression Regulation, Neoplastic , Prognosis , Proto-Oncogene Proteins B-raf/genetics , Child, Preschool , Gene Expression Profiling , Immunotherapy/methods , Immune Checkpoint Inhibitors/therapeutic use , Immune Checkpoint Inhibitors/pharmacology , Mutation , T-Lymphocytes/immunology , Precision Medicine/methods , Lymphocytes, Tumor-Infiltrating/immunology , Clinical RelevanceABSTRACT
PURPOSE: Molecular characterization is key to optimally diagnose and manage cancer. The complexity and cost of routine genomic analysis have unfortunately limited its use and denied many patients access to precision medicine. A possible solution is to rationalize use-creating a tiered approach to testing which uses inexpensive techniques for most patients and limits expensive testing to patients with the highest needs. Here, we tested the utility of this approach to molecularly characterize pediatric glioma in a cost- and time-sensitive manner. METHODS: We used a tiered testing pipeline of immunohistochemistry (IHC), customized fusion panels or fluorescence in situ hybridization (FISH), and targeted RNA sequencing in pediatric gliomas. Two distinct diagnostic algorithms were used for low- and high-grade gliomas (LGGs and HGGs). The percentage of driver alterations identified, associated testing costs, and turnaround time (TAT) are reported. RESULTS: The tiered approach successfully characterized 96% (95 of 99) of gliomas. For 82 LGGs, IHC, targeted fusion panel or FISH, and targeted RNA sequencing solved 35% (29 of 82), 29% (24 of 82), and 30% (25 of 82) of cases, respectively. A total of 64% (53 of 82) of samples were characterized without targeted RNA sequencing. Of 17 HGG samples, 13 were characterized by IHC and four were characterized by targeted RNA sequencing. The average cost per sample was more affordable when using the tiered approach as compared with up-front targeted RNA sequencing in LGG ($405 US dollars [USD] v $745 USD) and HGGs ($282 USD v $745 USD). The average TAT per sample was also shorter using the tiered approach (10 days for LGG, 5 days for HGG v 14 days for targeted RNA sequencing). CONCLUSION: Our tiered approach molecularly characterized 96% of samples in a cost- and time-sensitive manner. Such an approach may be feasible in neuro-oncology centers worldwide, particularly in resource-limited settings.
Subject(s)
Glioma , Humans , Glioma/genetics , Glioma/diagnosis , Glioma/pathology , Child , Male , Child, Preschool , Female , Adolescent , Brain Neoplasms/genetics , Brain Neoplasms/pathology , Brain Neoplasms/economics , Brain Neoplasms/diagnosis , In Situ Hybridization, Fluorescence/economics , Infant , Immunohistochemistry/economics , Health Resources/economics , Sequence Analysis, RNA/economics , Resource-Limited SettingsABSTRACT
Pediatric-type low-grade glioma (PLGG) encompasses a heterogeneous group of WHO grade 1 or 2 tumors and is the most common central nervous system tumor found in children. PLGG extends beyond pediatrics, into adolescents and young adults (AYA, ages 15-40). PLGG represents 25% of all gliomas diagnosed in AYA with differences in tumor location and molecular alterations compared to children, resulting in improved outcome for AYAs. Long-term outcome is excellent, though patients may suffer significant morbidity depending on tumor location. There are differences in treatment practices with radiation used to treat PLGG in AYAs more often than in children. Most PLGG in AYA harbor an alteration in the RAS/MAPK pathway, with limited insight into response to targeted therapy in this age group. This review discusses the epidemiology, current therapeutic approaches, and challenges in the management of PLGG in AYA.
ABSTRACT
Background: Accurate and precise diagnosis is central to treating central nervous system (CNS) tumors, yet tissue diagnosis is often a neglected focus in low- and middle-income countries (LMICs). Since 2016, the WHO classification of CNS tumors has increasingly incorporated molecular biomarkers into the diagnosis of CNS tumors. While this shift to precision diagnostics promises a high degree of diagnostic accuracy and prognostic precision, it has also resulted in increasing divergence in diagnostic and management practices between LMICs and high-income countries (HICs). Pathologists and laboratory professionals in LMICs lack the proper training and tools to join the molecular diagnostic revolution. We describe the impact of a 7-year long twinning program between Canada and Pakistan on pathology services. Methods: During the study period, 141 challenging cases of pediatric CNS tumors initially diagnosed at Aga Khan University Hospital (AKUH), Karachi, were sent to the Hospital for Sick Children in Toronto, Canada (SickKids), for a second opinion. Each case received histologic review and often immunohistochemical staining and relevant molecular testing. A monthly multidisciplinary online tumor board (MDTB) was conducted to discuss the results with pathologists from both institutions in attendance. Results: Diagnostic discordance was seen in 30 cases. Expert review provided subclassification for 53 cases most notably for diffuse gliomas and medulloblastoma. Poorly differentiated tumors benefited the most from second review, mainly because of the resolving power of specialized immunohistochemical stains, NanoString, and targeted gene panel next-generation sequencing. Collaboration with expert neuropathologists led to validation of over half a dozen immunostains at AKUH facilitating diagnosis of CNS tumors. Conclusions: LMIC-HIC Institutional twinning provides much-needed training and mentorship to pathologists and can help in infrastructure development by adopting and validating new immunohistochemical stains. Persistent unresolved cases indicate that molecular techniques are indispensable in for diagnosis in a minority of cases. The development of affordable alternative molecular techniques may help with these histologically unresolved cases.
ABSTRACT
BACKGROUND: Hippocampal sclerosis (HS) is a common surgical substrate in adult epilepsy surgery cohorts but variably reported in various pediatric cohorts. OBJECTIVE: We aimed to study the epilepsy phenotype, radiological and pathological variability, seizure and neurocognitive outcomes in children with drug-resistant epilepsy and hippocampal sclerosis (HS) with or without additional subtle signal changes in anterior temporal lobe who underwent surgery. METHODS: This retrospective study enrolled children with drug-resistant focal epilepsy and hippocampal sclerosis with or without additional subtle T2-Fluid Attenuated Inversion Recovery (FLAR)/Proton Density (PD) signal changes in anterior temporal lobe who underwent anterior temporal lobectomy with amygdalohippocampectomy. Their clinical, EEG, neuropsychological, radiological and pathological data were reviewed and summarized. RESULTS: Thirty-six eligible patients were identified. The mean age at seizure onset was 3.7 years; 25% had daily seizures at time of surgery. Isolated HS was noted in 22 (61.1%) cases and additional subtle signal changes in ipsilateral temporal lobe in 14 (38.9%) cases. Compared to the normative population, the group mean performance in intellectual functioning and most auditory and visual memory tasks were significantly lower than the normative sample. The mean age at surgery was 12.3 years; 22 patients (61.1%) had left hemispheric surgeries. ILAE class 1 outcomes was seen in 28 (77.8%) patients after a mean follow up duration of 2.3 years. Hippocampal sclerosis was noted pathologically in 32 (88.9%) cases; type 2 (54.5%) was predominant subtype where further classification was possible. Additional pathological abnormalities were seen in 11 cases (30.6%); these had had similar rates of seizure freedom as compared to children with isolated hippocampal sclerosis/gliosis (63.6% vs 84%, p=0.21). Significant reliable changes were observed across auditory and visual memory tasks at an individual level post surgery. CONCLUSIONS: Favourable seizure outcomes were seen in most children with isolated radiological hippocampal sclerosis. Patients with additional pathological abnormalities had similar rates of seizure freedom as compared to children with isolated hippocampal sclerosis/gliosis.
Subject(s)
Drug Resistant Epilepsy , Hippocampus , Sclerosis , Humans , Hippocampus/pathology , Hippocampus/surgery , Sclerosis/surgery , Male , Female , Child , Drug Resistant Epilepsy/surgery , Drug Resistant Epilepsy/pathology , Adolescent , Retrospective Studies , Treatment Outcome , Child, Preschool , Magnetic Resonance Imaging , Electroencephalography/methods , Neuropsychological Tests , Anterior Temporal Lobectomy/methods , Hippocampal SclerosisABSTRACT
INTODUCTION: Diffuse leptomeningeal glioneuronal tumors (DLGNTs) pose a rare and challenging entity within pediatric central nervous system neoplasms. Despite their rarity, DLGNTs exhibit complex clinical presentations and unique molecular characteristics, necessitating a deeper understanding of their diagnostic and therapeutic nuances. METHODS: This review synthesizes contemporary literature on DLGNT, encompassing epidemiology, clinical manifestations, pathological features, treatment strategies, prognostic markers, and future research directions. To compile the existing body of knowledge on DLGNT, a comprehensive search of relevant databases was conducted. RESULTS: DLGNT primarily affects pediatric populations but can manifest across all age groups. Its diagnosis is confounded by nonspecific clinical presentations and overlapping radiological features with other CNS neoplasms. Magnetic resonance imaging (MRI) serves as a cornerstone for DLGNT diagnosis, revealing characteristic leptomeningeal enhancement and intraparenchymal involvement. Histologically, DLGNT presents with low to moderate cellularity and exhibits molecular alterations in the MAPK/ERK signalling pathway. Optimal management of DLGNT necessitates a multidisciplinary approach encompassing surgical resection, chemotherapy, radiotherapy, and emerging targeted therapies directed against specific genetic alterations. Prognostication remains challenging, with factors such as age at diagnosis, histological subtypes, and genetic alterations influencing disease progression and treatment response. Long-term survival data are limited, underscoring the need for collaborative research efforts. CONCLUSION: Advancements in molecular profiling, targeted therapies, and international collaborations hold promise for improving DLGNT outcomes. Harnessing the collective expertise of clinicians, researchers, and patient advocates, can advance the field of DLGNT research and optimize patient care paradigms.
Subject(s)
Meningeal Neoplasms , Humans , Meningeal Neoplasms/therapy , Meningeal Neoplasms/pathology , Meningeal Neoplasms/diagnostic imaging , Meningeal Neoplasms/genetics , ChildABSTRACT
Diffuse intrinsic pontine gliomas are lethal tumors with a prognosis generally less than 1 year. Few cases of survivors of 5 years or more have been reported. This case report highlights the journey of a 9.5-year survivor who underwent 3 rounds of focal radiotherapy; she experienced 6 years of progression-free survival following the first round but ultimately succumbed to her disease. An autopsy revealed a favorable IDH1 mutation and the absence of H3K27M. This case reiterates the importance of extensive molecular analyses in diffuse intrinsic pontine gliomas and explores the potential benefit of re-irradiation in patients with positive responses and long periods of remission.
Subject(s)
Brain Stem Neoplasms , Diffuse Intrinsic Pontine Glioma , Humans , Female , Brain Stem Neoplasms/pathology , Brain Stem Neoplasms/therapy , Brain Stem Neoplasms/mortality , Diffuse Intrinsic Pontine Glioma/pathology , Diffuse Intrinsic Pontine Glioma/therapy , Diffuse Intrinsic Pontine Glioma/genetics , Child , Survivorship , Cancer Survivors , Fatal Outcome , Isocitrate Dehydrogenase/genetics , Prognosis , MutationABSTRACT
BACKGROUND AND PURPOSE: Molecular biomarker identification increasingly influences the treatment planning of pediatric low-grade neuroepithelial tumors (PLGNTs). We aimed to develop and validate a radiomics-based ADC signature predictive of the molecular status of PLGNTs. MATERIALS AND METHODS: In this retrospective bi-institutional study, we searched the PACS for baseline brain MRIs from children with PLGNTs. Semiautomated tumor segmentation on ADC maps was performed using the semiautomated level tracing effect tool with 3D Slicer. Clinical variables, including age, sex, and tumor location, were collected from chart review. The molecular status of tumors was derived from biopsy. Multiclass random forests were used to predict the molecular status and fine-tuned using a grid search on the validation sets. Models were evaluated using independent and unseen test sets based on the combined data, and the area under the receiver operating characteristic curve (AUC) was calculated for the prediction of 3 classes: KIAA1549-BRAF fusion, BRAF V600E mutation, and non-BRAF cohorts. Experiments were repeated 100 times using different random data splits and model initializations to ensure reproducible results. RESULTS: Two hundred ninety-nine children from the first institution and 23 children from the second institution were included (53.6% male; mean, age 8.01 years; 51.8% supratentorial; 52.2% with KIAA1549-BRAF fusion). For the 3-class prediction using radiomics features only, the average test AUC was 0.74 (95% CI, 0.73-0.75), and using clinical features only, the average test AUC was 0.67 (95% CI, 0.66-0.68). The combination of both radiomics and clinical features improved the AUC to 0.77 (95% CI, 0.75-0.77). The diagnostic performance of the per-class test AUC was higher in identifying KIAA1549-BRAF fusion tumors among the other subgroups (AUC = 0.81 for the combined radiomics and clinical features versus 0.75 and 0.74 for BRAF V600E mutation and non-BRAF, respectively). CONCLUSIONS: ADC values of tumor segmentations have differentiative signals that can be used for training machine learning classifiers for molecular biomarker identification of PLGNTs. ADC-based pretherapeutic differentiation of the BRAF status of PLGNTs has the potential to avoid invasive tumor biopsy and enable earlier initiation of targeted therapy.
Subject(s)
Brain Neoplasms , Diffusion Magnetic Resonance Imaging , Machine Learning , Neoplasms, Neuroepithelial , Humans , Child , Female , Male , Retrospective Studies , Neoplasms, Neuroepithelial/diagnostic imaging , Neoplasms, Neuroepithelial/genetics , Brain Neoplasms/diagnostic imaging , Brain Neoplasms/genetics , Brain Neoplasms/pathology , Child, Preschool , Adolescent , Diffusion Magnetic Resonance Imaging/methods , Proto-Oncogene Proteins B-raf/genetics , Infant , Neoplasm Grading , Biomarkers, Tumor/geneticsABSTRACT
PURPOSE: The use of trametinib in the treatment of pediatric low-grade gliomas (PLGG) and plexiform neurofibroma (PN) is being investigated in an ongoing multicenter phase II trial (NCT03363217). Preliminary data shows potential benefits with significant response in the majority of PLGG and PN and an overall good tolerance. Moreover, possible benefits of MEK inhibitor therapy on cognitive functioning in neurofibromatosis type 1 (NF1) were recently shown which supports the need for further evaluation. METHODS: Thirty-six patients with NF1 (age range 3-19 years) enrolled in the phase II study of trametinib underwent a neurocognitive assessment at inclusion and at completion of the 72-week treatment. Age-appropriate Wechsler Intelligence Scales and the Trail Making Test (for children over 8 years old) were administered at each assessment. Paired t-tests and Reliable Change Index (RCI) analyses were performed to investigate change in neurocognitive outcomes. Regression analyses were used to investigate the contribution of age and baseline score in the prediction of change. RESULTS: Stable performance on neurocognitive tests was revealed at a group-level using paired t-tests. Clinically significant improvements were however found on specific indexes of the Wechsler intelligence scales and Trail Making Test, using RCI analyses. No significant impact of age on cognitive change was evidenced. However, lower initial cognitive performance was associated with increased odds of presenting clinically significant improvements on neurocognitive outcomes. CONCLUSION: These preliminary results show a potential positive effect of trametinib on cognition in patients with NF1. We observed significant improvements in processing speed, visuo-motor and verbal abilities. This study demonstrates the importance of including neuropsychological evaluations into clinical trial when using MEK inhibitors for patients with NF1.
Subject(s)
Neurofibromatosis 1 , Neuropsychological Tests , Pyridones , Pyrimidinones , Humans , Pyridones/therapeutic use , Pyrimidinones/therapeutic use , Pyrimidinones/pharmacology , Pyrimidinones/administration & dosage , Male , Female , Adolescent , Child , Neurofibromatosis 1/drug therapy , Neurofibromatosis 1/complications , Neurofibromatosis 1/psychology , Young Adult , Child, Preschool , Glioma/drug therapy , Glioma/psychology , Glioma/complications , Brain Neoplasms/drug therapy , Brain Neoplasms/psychology , Brain Neoplasms/complications , Adult , Protein Kinase Inhibitors/therapeutic use , Antineoplastic Agents/adverse effectsABSTRACT
Introduction: Advances in molecular diagnostics led to improved targeted interventions in the treatment of pediatric CNS tumors. However, the capacity to test for these is limited in LMICs, and thus their value needs exploration. Methods: We reviewed our experience with NGS testing (TruSight RNA Pan-Cancer-seq panel) for pediatric CNS tumors at KHCC/Jordan (March/2022-April/2023). Paraffin blocks' scrolls were shipped to the SickKids laboratory based on the multidisciplinary clinic (MDC) recommendations. We reviewed the patients' characteristics, the tumor types, and the NGS results' impact on treatment. Results: Of 237 patients discussed during the MDC meetings, 32 patients (14%) were included. They were 16 boys and 16 girls; the median age at time of testing was 9.5 years (range, 0.9-21.9 years). There were 21 samples sent at diagnosis and 11 upon tumor progression. The main diagnoses were low-grade-glioma (15), high-grade-glioma (10), and other histologies (7). Reasons to request NGS included searching for a targetable alteration (20) and to better characterize the tumor behavior (12). The median turnaround time from samples' shipment to receiving the results was 23.5 days (range, 15-49 days) with a median laboratory processing time of 16 days (range, 8-39 days) at a cost of US$1,000/sample. There were 19 (59%) tumors that had targetable alterations (FGFR/MAPK pathway inhibitors (14), checkpoint inhibitors (2), NTRK inhibitors (2), and one with PI3K inhibitor or IDH1 inhibitor). Two rare BRAF mutations were identified (BRAFp.G469A, BRAFp.K601E). One tumor diagnosed initially as undifferentiated round cell sarcoma harbored NAB2::STAT6 fusion and was reclassified as an aggressive metastatic solitary fibrous tumor. Another tumor initially diagnosed as grade 2 astroblastoma grade 2 was reclassified as low-grade-glioma in the absence of MN1 alteration. NGS failed to help characterize a tumor that was diagnosed histologically as small round blue cell tumor. Nine patients received targeted therapy; dabrafenib/trametinib (6), pembrolizumab (2), and entrectinib (1), mostly upon tumor progression (7). Conclusion: In this highly selective cohort, a high percentage of targetable mutations was identified facilitating targeted therapies. Outsourcing of NGS testing was feasible; however, criteria for case selection are needed. In addition, local capacity-building in conducting the test, interpretation of the results, and access to "new drugs" continue to be a challenge in LMICs.
ABSTRACT
Published in 2021, the fifth edition of the World Health Organization (WHO) classification of tumors of the central nervous system (CNS) introduced new molecular criteria for tumor types that commonly occur in either pediatric or adult age groups. Adolescents and young adults (AYAs) are at the intersection of adult and pediatric care, and both pediatric-type and adult-type CNS tumors occur at that age. Mortality rates for AYAs with CNS tumors have increased by 0.6% per year for males and 1% per year for females from 2007 to 2016. To best serve patients, it is crucial that both pediatric and adult radiologists who interpret neuroimages are familiar with the various pediatric- and adult-type brain tumors and their typical imaging morphologic characteristics. Gliomas account for approximately 80% of all malignant CNS tumors in the AYA age group, with the most common types observed being diffuse astrocytic and glioneuronal tumors. Ependymomas and medulloblastomas also occur in the AYA population but are seen less frequently. Importantly, biologic behavior and progression of distinct molecular subgroups of brain tumors differ across ages. This review discusses newly added or revised gliomas in the fifth edition of the CNS WHO classification, as well as other CNS tumor types common in the AYA population.
Subject(s)
Brain Neoplasms , Cerebellar Neoplasms , Glioma , Medulloblastoma , Female , Male , Humans , Adolescent , Young Adult , Child , Brain Neoplasms/diagnostic imaging , Glioma/diagnostic imaging , World Health OrganizationABSTRACT
Pediatric high-grade gliomas (pHGG) are devastating and incurable brain tumors with recurrent mutations in histone H3.3. These mutations promote oncogenesis by dysregulating gene expression through alterations of histone modifications. We identify aberrant DNA repair as an independent mechanism, which fosters genome instability in H3.3 mutant pHGG, and opens new therapeutic options. The two most frequent H3.3 mutations in pHGG, K27M and G34R, drive aberrant repair of replication-associated damage by non-homologous end joining (NHEJ). Aberrant NHEJ is mediated by the DNA repair enzyme polynucleotide kinase 3'-phosphatase (PNKP), which shows increased association with mutant H3.3 at damaged replication forks. PNKP sustains the proliferation of cells bearing H3.3 mutations, thus conferring a molecular vulnerability, specific to mutant cells, with potential for therapeutic targeting.
Subject(s)
Brain Neoplasms , Glioma , Histones , Child , Humans , Brain Neoplasms/pathology , DNA Repair/genetics , DNA Repair Enzymes/metabolism , Glioma/pathology , Histones/genetics , Histones/metabolism , Mutation , Phosphotransferases (Alcohol Group Acceptor)/geneticsABSTRACT
Background: Despite genomic simplicity, recent studies have reported at least 3 major atypical teratoid rhabdoid tumor (ATRT) subgroups with distinct molecular and clinical features. Reliable ATRT subgrouping in clinical settings remains challenging due to a lack of suitable biological markers, sample rarity, and the relatively high cost of conventional subgrouping methods. This study aimed to develop a reliable ATRT molecular stratification method to implement in clinical settings. Methods: We have developed an ATRT subgroup predictor assay using a custom genes panel for the NanoString nCounter System and a flexible machine learning classifier package. Seventy-one ATRT primary tumors with matching gene expression array and NanoString data were used to construct a multi-algorithms ensemble classifier. Additional validation was performed using an independent gene expression array against the independently generated dataset. We also analyzed 11 extra-cranial rhabdoid tumors with our classifier and compared our approach against DNA methylation classification to evaluate the result consistency with existing methods. Results: We have demonstrated that our novel ensemble classifier has an overall average of 93.6% accuracy in the validation dataset, and a striking 98.9% accuracy was achieved with the high-prediction score samples. Using our classifier, all analyzed extra-cranial rhabdoid tumors are classified as MYC subgroups. Compared with the DNA methylation classification, the results show high agreement, with 84.5% concordance and up to 95.8% concordance for high-confidence predictions. Conclusions: Here we present a rapid, cost-effective, and accurate ATRT subgrouping assay applicable for clinical use.
ABSTRACT
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.
Subject(s)
Brain Neoplasms , Glioma , Humans , Child , Proto-Oncogene Proteins B-raf/genetics , Brain Neoplasms/pathology , Radiomics , Retrospective Studies , Glioma/pathologyABSTRACT
Recent clinical trials for H3K27-altered diffuse midline gliomas (DMGs) have shown much promise. We present a consensus roadmap and identify three major barriers: (1) refinement of experimental models to include immune and brain-specific components; (2) collaboration among researchers, clinicians, and industry to integrate patient-derived data through sharing, transparency, and regulatory considerations; and (3) streamlining clinical efforts including biopsy, CNS-drug delivery, endpoint determination, and response monitoring. We highlight the importance of comprehensive collaboration to advance the understanding, diagnostics, and therapeutics for DMGs.
Subject(s)
Brain Neoplasms , Glioma , Humans , Child , Brain Neoplasms/genetics , Brain Neoplasms/therapy , Glioma/diagnosis , Glioma/genetics , Glioma/therapy , Mutation , Brain/pathology , BiopsyABSTRACT
Purpose: MRI-based radiomics models can predict genetic markers in pediatric low-grade glioma (pLGG). These models usually require tumour segmentation, which is tedious and time consuming if done manually. We propose a deep learning (DL) model to automate tumour segmentation and build an end-to-end radiomics-based pipeline for pLGG classification. Methods: The proposed architecture is a 2-step U-Net based DL network. The first U-Net is trained on downsampled images to locate the tumour. The second U-Net is trained using image patches centred around the located tumour to produce more refined segmentations. The segmented tumour is then fed into a radiomics-based model to predict the genetic marker of the tumour. Results: Our segmentation model achieved a correlation value of over 80% for all volume-related radiomic features and an average Dice score of .795 in test cases. Feeding the auto-segmentation results into a radiomics model resulted in a mean area under the ROC curve (AUC) of .843, with 95% confidence interval (CI) [.78-.906] and .730, with 95% CI [.671-.789] on the test set for 2-class (BRAF V600E mutation BRAF fusion) and 3-class (BRAF V600E mutation BRAF fusion and Other) classification, respectively. This result was comparable to the AUC of .874, 95% CI [.829-.919] and .758, 95% CI [.724-.792] for the radiomics model trained and tested on the manual segmentations in 2-class and 3-class classification scenarios, respectively. Conclusion: The proposed end-to-end pipeline for pLGG segmentation and classification produced results comparable to manual segmentation when it was used for a radiomics-based genetic marker prediction model.
Subject(s)
Glioma , Proto-Oncogene Proteins B-raf , Humans , Child , Genetic Markers , Glioma/pathology , Magnetic Resonance Imaging/methods , Area Under CurveABSTRACT
Pediatric high-grade glioma (pHGG) is highly malignant central nervous system tumor and constitute 10% of the pediatric gliomas. Effective treatment needs a functioning multi-disciplinary team including pediatric neuro oncologist, neurosurgeon, neuroradiologist, neuropathologist and radiation oncologist. Despite surgical resection, radiotherapy and chemotherapy, most HGG will recur resulting in early death. A significant proportion of HGG occurs in context of cancer predisposition syndromes like Constitutional Mismatch Repair Deficiency (CMMRD) also known as Biallelic Mismatch Repair Deficiency (bMMRD) characterized by high mutational burden. The incidence of HGG with CMMRD is one per million patients. bMMRD is caused by homozygous germline mutations in one of the four Mis Match Repair (MMR) genes (PMS2, MLH1, MSH2, and MSH6). The use of TMZ is now avoided in CMMRD related HGG due to its limited response and known ability to increase the accumulation of somatic mutations in these patients, increasing the risk of secondary tumors. HGG should be managed under the care of multidisciplinary team to receive optimum treatment. This is particularly important for low middle-income countries (LMIC) with limited resources like Pakistan.
ABSTRACT
Pediatric low-grade gliomas (pLGGs) are the most common brain tumor in young children. While they are typically associated with good overall survival, children with these central nervous system tumors often experience chronic tumor- and therapy-related morbidities. Moreover, individuals with unresectable tumors frequently have multiple recurrences and persistent neurological symptoms. Deep molecular analyses of pLGGs reveal that they are caused by genetic alterations that converge on a single mitogenic pathway (MEK/ERK), but their growth is heavily influenced by nonneoplastic cells (neurons, T cells, microglia) in their local microenvironment. The interplay between neoplastic cell MEK/ERK pathway activation and stromal cell support necessitates the use of predictive preclinical models to identify the most promising drug candidates for clinical evaluation. As part of a series of white papers focused on pLGGs, we discuss the current status of preclinical pLGG modeling, with the goal of improving clinical translation for children with these common brain tumors.