Quantifying intra-tumoral genetic heterogeneity of glioblastoma toward precision medicine using MRI and a data-inclusive machine learning algorithm.

BACKGROUND AND OBJECTIVE: Glioblastoma (GBM) is one of the most aggressive and lethal human cancers. Intra-tumoral genetic heterogeneity poses a significant challenge for treatment. Biopsy is invasive, which motivates the development of non-invasive, MRI-based machine learning (ML) models to quantify intra-tumoral genetic heterogeneity for each patient. This capability holds great promise for enabling better therapeutic selection to improve patient outcome. METHODS: We proposed a novel Weakly Supervised Ordinal Support Vector Machine (WSO-SVM) to predict regional genetic alteration status within each GBM tumor using MRI. WSO-SVM was applied to a unique dataset of 318 image-localized biopsies with spatially matched multiparametric MRI from 74 GBM patients. The model was trained to predict the regional genetic alteration of three GBM driver genes (EGFR, PDGFRA and PTEN) based on features extracted from the corresponding region of five MRI contrast images. For comparison, a variety of existing ML algorithms were also applied. Classification accuracy of each gene were compared between the different algorithms. The SHapley Additive exPlanations (SHAP) method was further applied to compute contribution scores of different contrast images. Finally, the trained WSO-SVM was used to generate prediction maps within the tumoral area of each patient to help visualize the intra-tumoral genetic heterogeneity. RESULTS: WSO-SVM achieved 0.80 accuracy, 0.79 sensitivity, and 0.81 specificity for classifying EGFR; 0.71 accuracy, 0.70 sensitivity, and 0.72 specificity for classifying PDGFRA; 0.80 accuracy, 0.78 sensitivity, and 0.83 specificity for classifying PTEN; these results significantly outperformed the existing ML algorithms. Using SHAP, we found that the relative contributions of the five contrast images differ between genes, which are consistent with findings in the literature. The prediction maps revealed extensive intra-tumoral region-to-region heterogeneity within each individual tumor in terms of the alteration status of the three genes. CONCLUSIONS: This study demonstrated the feasibility of using MRI and WSO-SVM to enable non-invasive prediction of intra-tumoral regional genetic alteration for each GBM patient, which can inform future adaptive therapies for individualized oncology.

Image-localized biopsy mapping of brain tumor heterogeneity: A single-center study protocol.

Brain cancers pose a novel set of difficulties due to the limited accessibility of human brain tumor tissue. For this reason, clinical decision-making relies heavily on MR imaging interpretation, yet the mapping between MRI features and underlying biology remains ambiguous. Standard (clinical) tissue sampling fails to capture the full heterogeneity of the disease. Biopsies are required to obtain a pathological diagnosis and are predominantly taken from the tumor core, which often has different traits to the surrounding invasive tumor that typically leads to recurrent disease. One approach to solving this issue is to characterize the spatial heterogeneity of molecular, genetic, and cellular features of glioma through the intraoperative collection of multiple image-localized biopsy samples paired with multi-parametric MRIs. We have adopted this approach and are currently actively enrolling patients for our 'Image-Based Mapping of Brain Tumors' study. Patients are eligible for this research study (IRB #16-002424) if they are 18 years or older and undergoing surgical intervention for a brain lesion. Once identified, candidate patients receive dynamic susceptibility contrast (DSC) perfusion MRI and diffusion tensor imaging (DTI), in addition to standard sequences (T1, T1Gd, T2, T2-FLAIR) at their presurgical scan. During surgery, sample anatomical locations are tracked using neuronavigation. The collected specimens from this research study are used to capture the intra-tumoral heterogeneity across brain tumors including quantification of genetic aberrations through whole-exome and RNA sequencing as well as other tissue analysis techniques. To date, these data (made available through a public portal) have been used to generate, test, and validate predictive regional maps of the spatial distribution of tumor cell density and/or treatment-related key genetic marker status to identify biopsy and/or treatment targets based on insight from the entire tumor makeup. This type of methodology, when delivered within clinically feasible time frames, has the potential to further inform medical decision-making by improving surgical intervention, radiation, and targeted drug therapy for patients with glioma.

Integrated molecular and multiparametric MRI mapping of high-grade glioma identifies regional biologic signatures.

Sampling restrictions have hindered the comprehensive study of invasive non-enhancing (NE) high-grade glioma (HGG) cell populations driving tumor progression. Here, we present an integrated multi-omic analysis of spatially matched molecular and multi-parametric magnetic resonance imaging (MRI) profiling across 313 multi-regional tumor biopsies, including 111 from the NE, across 68 HGG patients. Whole exome and RNA sequencing uncover unique genomic alterations to unresectable invasive NE tumor, including subclonal events, which inform genomic models predictive of geographic evolution. Infiltrative NE tumor is alternatively enriched with tumor cells exhibiting neuronal or glycolytic/plurimetabolic cellular states, two principal transcriptomic pathway-based glioma subtypes, which respectively demonstrate abundant private mutations or enrichment in immune cell signatures. These NE phenotypes are non-invasively identified through normalized K2 imaging signatures, which discern cell size heterogeneity on dynamic susceptibility contrast (DSC)-MRI. NE tumor populations predicted to display increased cellular proliferation by mean diffusivity (MD) MRI metrics are uniquely associated with EGFR amplification and CDKN2A homozygous deletion. The biophysical mapping of infiltrative HGG potentially enables the clinical recognition of tumor subpopulations with aggressive molecular signatures driving tumor progression, thereby informing precision medicine targeting.

NCCN Guidelines® Insights: Central Nervous System Cancers, Version 2.2022.

The NCCN Guidelines for Central Nervous System (CNS) Cancers focus on management of the following adult CNS cancers: glioma (WHO grade 1, WHO grade 2-3 oligodendroglioma [1p19q codeleted, IDH-mutant], WHO grade 2-4 IDH-mutant astrocytoma, WHO grade 4 glioblastoma), intracranial and spinal ependymomas, medulloblastoma, limited and extensive brain metastases, leptomeningeal metastases, non-AIDS-related primary CNS lymphomas, metastatic spine tumors, meningiomas, and primary spinal cord tumors. The information contained in the algorithms and principles of management sections in the NCCN Guidelines for CNS Cancers are designed to help clinicians navigate through the complex management of patients with CNS tumors. Several important principles guide surgical management and treatment with radiotherapy and systemic therapy for adults with brain tumors. The NCCN CNS Cancers Panel meets at least annually to review comments from reviewers within their institutions, examine relevant new data from publications and abstracts, and reevaluate and update their recommendations. These NCCN Guidelines Insights summarize the panel's most recent recommendations regarding molecular profiling of gliomas.

Evaluating the Patient with Neurotoxicity after Chimeric Antigen Receptor T-cell Therapy.

OPINION STATEMENT: Chimeric antigen receptor (CAR) T-cells are now a well-established treatment for hematologic malignancies. Their use in clinical practice has expanded quite rapidly and hospitals have developed CAR T-cell protocols to evaluate patients for associated toxicities, and particularly for neurotoxicity. There are many variables that influence the risk for developing this complication, many of which are not fully understood. The severity can be related to a particular product. Clinical vigilance is critical to facilitate early recognition of neurotoxicity, hence the importance of pre-CAR T-cell neurological evaluation of each patient. While details of such an evaluation may slightly differ between institutions, generally a comprehensive neurological evaluation including assessment of cognitive abilities along with magnetic resonance imaging (MRI) of the brain is a gold standard. Management of neurotoxicity requires a well-orchestrated team approach with specialists from oncology, neurology, oftentimes neurosurgery and neuro-intensive care. Diagnostic work-up frequently includes detailed neurologic evaluation with comparison to the baseline assessment, imaging of the brain, electroencephalogram, and lumbar puncture. While steroids are uniformly used for treatment, many patients also receive tocilizumab for an underlying and frequently concomitant cytokine release syndrome (CRS) in addition to symptom-driven supportive care. Novel CAR T-cell constructs and other agents allowing for potentially lower risk of toxicity are being explored. While neurotoxicity is predominantly an early, and reversible, event, a growing body of literature suggests that late neurotoxicity with variable clinical presentation can also occur.

Challenging Cases in Neuro-Oncology.

Neuro-oncology encompasses a broad field focusing on an array of neoplasms, many of which can mimic several diseases. Neurologists will often be involved in the initial diagnostic evaluation and management of these patients. Their insight is central to optimizing the diagnostic yield and providing high-level clinical care. Several neuro-oncologic cases are reviewed with a goal of increasing the understanding of these diseases in a clinically relevant manner and providing updates on the contemporary thinking in the subspecialty.

Lung cancer with brain metastases remaining in continuous complete remission due to pembrolizumab and temozolomide: a case report.

Background: Immunotherapy has been shown to improve the overall survival (OS) in patients with advanced or metastatic non-small cell lung cancer (NSCLC) without driver gene mutations. However, monotherapy with immunotherapy alone or combined with chemotherapy in NSCLC patients with untreated brain metastases (BM) is still under debate. Data regarding treatment of BM with immunotherapy and temozolomide (TMZ) in patients with NSCLC is rare. Case Presentation: A 60-year-old male due to cough and expectoration presented in our hospital. Chest computed tomography (CT), brain magnetic resonance imaging (MRI) and immunohistochemistry of a mediastinal lymph node biopsy were administered, he was diagnosed with stage IIIB lung adenocarcinoma. Without driver gene mutations, he was treated with platinum-based chemotherapy because he refused to accept concurrent radiation therapy (RT). Heavy cough companied with hemoptysis and chest CT scan both revealed progressive disease (PD) after 6 cycles of chemotherapy. Immunotherapy was consequently considered, while two metastatic lesions in the brain were confirmed after combined treatment of pembrolizumab with docetaxel. TMZ was administered in combination with pembrolizumab (200 mg, day 1). A new metastasis in the right occipital lobe was detected on a scan 1 month later, though the other 2 lesions continued to shrink. The treatment was continued, MRI and CT scans suggested complete response (CR) was achieved for both the BM and lung lesions after 3 cycles. Consolidation therapy with TMZ and pembrolizumab (100 mg) per month was considered for another 7 months. Maintenance monotherapy with pembrolizumab (100 mg) was selected because of his stable CR status. At 59 months since diagnosis, the patient remains alive, with CR for both the primary lesions and BM. The patient experienced slight numbness on each side of his feet. There was no occurrence of adverse effects greater than grade 3. Conclusions: The data indicates that immunotherapy combined with TMZ for untreated BM in NSCLC patients maybe an efficient and safe decision making therapeutic choice. Despite the encouraging efficacy of the combination, it is an isolated case and the speculation of synergism need to be proved in further pharmacokinetic/pharmacodynamic studies even in large randomized controlled trials.

In Patients With Melanoma Brain Metastases, Is Combination Immune Checkpoint Inhibition a Safe and Effective First-Line Treatment? A Critically Appraised Topic.

BACKGROUND: Combined PD-1/PD-L1 and CTLA-4 immune checkpoint inhibition for the has been shown to produce superior results in the treatment of malignant melanoma when compared to monotherapy. However, patients with intracranial disease were excluded from these studies given their poor prognosis. OBJECTIVE: The objective of this study was to critically assess current evidence supporting the co-administration of PD-1/PD-L1 and CTLA-4 inhibitors in the treatment of melanoma brain metastases. METHODS: The objective was addressed through the development of a critically appraised topic that included a clinical scenario, structured question, literature search strategy, critical appraisal, assessment of results, evidence summary, commentary, and bottom-line conclusions. Participants included consultant and resident neurologists, a medical librarian, clinical epidemiologists, and a content expert in the field of neuro-oncology. RESULTS: A recent, open-label, non-comparative randomized phase II trial was selected for critical appraisal. This trial evaluated the efficacy and safety of nivolumab alone or in combination with ipilimumab in 79 adult patients with untreated, asymptomatic melanoma brain metastases. The rates of the primary outcome (intracranial response at ≥12 wk) in the primary endpoint cohort were 46% for cohort A (combination therapy) and 20% for cohort B (nivolumab monotherapy). No treatment related deaths were observed in the study. Grade 4 adverse events occurred in 9% of patients in cohort A and none in cohort B. CONCLUSIONS: Co-administration of ipilimumab and nivolumab as first-line therapy is effective in the treatment of asymptomatic melanoma brain metastases, with an acceptable safety profile.

Comprehensive Commissioning and Clinical Implementation of GammaTiles STaRT for Intracranial Brain Cancer.

Purpose: To validate the dose calculation accuracy and dose distribution of GammaTiles for brain tumors, and to suggest a surgically targeted radiation therapy (STaRT) workflow for planning, delivery, radiation safety documentation, and posttreatment validation. Methods and Materials: Novel surgically targeted radiation therapy, GammaTiles, uses Cs-131 radiation isotopes embedded in collagen-based tiles that can be resorbed after surgery. GammaTile target delineation and dose calculation were performed on MIM Symphony software. Point-based and complex seed distribution calculations in MIM Symphony were verified with hand calculations and BrachyVision calculations. Vendor-provided 2-dimensional dose distribution calculation accuracy was validated using gafchromic EBT3 film measurements at various depths. A workflow was established for safe and effective GammaTile implants. Results: Good agreement was observed between different calculations. Calculation accuracy of less than 0.5% was achieved for all points except one, which had rounding issues for very low doses and resulted in just below 5% difference. Differences in anisotropy and geometry positioning were noticed in the delineation of Cs-131 IsoRay seeds in the compared systems, resulting in minor discrepancies in the calculated dosimetry distributions. Film measurements showed profiles with relatively good agreement of 0% to 5% in nongradient regions with higher differences between 5% to 10% in the sharp dose fall-off regions. Conclusions: A comprehensive evaluation of GammaTile geometry, dose distribution, and clinical workflow was conducted. Safe intro-operative implantation of GammaTiles requires extensive preplanning and interdisciplinary collaboration. A STaRT workflow was outlined to provide a guideline for an accurate treatment planning and safe implant process at other institutions.

Practical guidance for telemedicine use in neuro-oncology.

While the COVID-19 pandemic has catalyzed the expansion of telemedicine into nearly every specialty of medicine, few articles have summarized current practices and recommendations for integrating virtual care in the practice of neuro-oncology. This article identifies current telemedicine practice, provides practical guidance for conducting telemedicine visits, and generates recommendations for integrating virtual care into neuro-oncology practice. Practical aspects of telemedicine are summarized including when to use and not use telemedicine, how to conduct a virtual visit, who to include in the virtual encounter, unique aspects of telehealth in neuro-oncology, and emerging innovations.

Central Nervous System Lymphoma: Novel Therapies.

OPINION STATEMENT: Primary central nervous system lymphomas (PCNSLs) are very rare neoplasms and continue to be challenging to treat. While high-dose methotrexate (HD-MTX)-based regimens are the currently accepted standard first-line therapy for newly diagnosed patients, the optimal induction therapies are still unknown. The role of consolidation therapies continues to evolve with a variety of chemotherapy regimens, including high-dose chemotherapy with stem cell rescue and reduced or deferred whole brain radiotherapy being used. Importantly, several recent advances have been made in the treatment of PCNSL. The incorporation of targeted therapy and immune therapy remain promising strategies. Several agents, successfully used in treatment of systemic lymphomas, have shown activity in PCNSL, frequently leading to durable responses in the relapsed/refractory patients. Many ongoing studies will likely lead to a better understanding of the roles of these treatments, especially as the first line and potentially also as maintenance. In addition, the use of molecular profiling to predict disease response to targeted agents and understand relapse patterns will become increasingly important. Clinical trials in PCNSL are critical yet frequently challenging to conduct given the rarity of the condition and lack of suitable subjects. Therefore, multi-institutional and international collaboration is of utmost importance to accelerate progress in understanding the biology and design better treatments for this disease. It is critical to consider patients of all demographics in the design and study of future treatment algorithms to have the largest impact on patient care and outcomes.

Shape matters: morphological metrics of glioblastoma imaging abnormalities as biomarkers of prognosis.

Lacunarity, a quantitative morphological measure of how shapes fill space, and fractal dimension, a morphological measure of the complexity of pixel arrangement, have shown relationships with outcome across a variety of cancers. However, the application of these metrics to glioblastoma (GBM), a very aggressive primary brain tumor, has not been fully explored. In this project, we computed lacunarity and fractal dimension values for GBM-induced abnormalities on clinically standard magnetic resonance imaging (MRI). In our patient cohort (n = 402), we connect these morphological metrics calculated on pretreatment MRI with the survival of patients with GBM. We calculated lacunarity and fractal dimension on necrotic regions (n = 390), all abnormalities present on T1Gd MRI (n = 402), and abnormalities present on T2/FLAIR MRI (n = 257). We also explored the relationship between these metrics and age at diagnosis, as well as abnormality volume. We found statistically significant relationships to outcome for all three imaging regions that we tested, with the shape of T2/FLAIR abnormalities that are typically associated with edema showing the strongest relationship with overall survival. This link between morphological and survival metrics could be driven by underlying biological phenomena, tumor location or microenvironmental factors that should be further explored.

Glioblastoma as an age-related neurological disorder in adults.

BACKGROUND: Advanced age is a major risk factor for the development of many diseases including those affecting the central nervous system. Wild-type isocitrate dehydrogenase glioblastoma (IDHwt GBM) is the most common primary malignant brain cancer and accounts for ≥90% of all adult GBM diagnoses. Patients with IDHwt GBM have a median age of diagnosis at 68-70 years of age, and increasing age is associated with an increasingly worse prognosis for patients with this type of GBM. METHODS: The Surveillance, Epidemiology, and End Results, The Cancer Genome Atlas, and the Chinese Glioma Genome Atlas databases were analyzed for mortality indices. Meta-analysis of 80 clinical trials was evaluated for log hazard ratio for aging to tumor survivorship. RESULTS: Despite significant advances in the understanding of intratumoral genetic alterations, molecular characteristics of tumor microenvironments, and relationships between tumor molecular characteristics and the use of targeted therapeutics, life expectancy for older adults with GBM has yet to improve. CONCLUSIONS: Based upon the results of our analysis, we propose that age-dependent factors that are yet to be fully elucidated, contribute to IDHwt GBM patient outcomes.

Is Autologous Stem Cell Transplantation a Safe and Effective Alternative to Whole Brain Radiation as Consolidation Therapy in Patients With Primary Central Nervous System Lymphoma?: A Critically Appraised Topic.

BACKGROUND: High-dose chemotherapy followed by autologous stem cell transplantation (HD-ASCT) is a promising alternative to whole brain radiation therapy (WBRT) in the treatment of primary central nervous system lymphoma (PCNSL). The objective of this study was to critically assess current evidence supporting the use of HD-ASCT as first-line consolidative therapy in PCNSL. METHODS: The objective was addressed through the development of a critically appraised topic that included a clinical scenario, structured question, literature search strategy, critical appraisal, assessment of results, evidence summary, commentary, and bottom-line conclusions. Participants included consultant and resident neurologists, a medical librarian, clinical epidemiologists, and a content expert in the field of neuro-oncology. RESULTS: A recent, open-label, noncomparative randomized phase II trial was selected for critical appraisal. This trial evaluated the efficacy and toxicity of consolidative therapy with HD-ASCT and WBRT in PCNSL in 2 separate treatment arms. A total of 140 patients with newly diagnosed PCNSL between the ages of 18 and 60 years were included. The primary endpoint of 2-year progression-free survival was met in 63% of patients in the WBRT arm and 87% in the HD-ASCT arm. Notably, an overall improvement in neurocognitive scores was observed following HD-ASCT, while WBRT was associated with worsened cognitive outcomes. CONCLUSIONS: In young patients with newly diagnosed PCNSL, consolidative therapy with HD-ASCT appears to be associated with less neurocognitive toxicity and may be more effective than WBRT at preventing relapses, however, at the cost of a higher treatment-related mortality.

Supportive care for patients with brain metastases from lung cancer.

Lung cancer is the most common cause of intracranial metastases (ICM). Metastases in the brain can result in a broad range of uncomfortable symptoms and significant morbidity secondary to neurological disability. Treatment options can range from surgical resection of solitary metastases to radiotherapy and more recently systemic targeted therapies and immunotherapy. Patient survival continues to improve with innovations made in treatments for this condition, but each of these treatments carry their own adverse effects that must be appropriately managed. These patients can benefit greatly from multidisciplinary care throughout the course of their disease. Clinicians involved in their care must be equipped with the ability to communicate skillfully and compassionately and set expectations for the road ahead, including symptoms, treatment plans, and prognosis. Involvement of a palliative care team can be very helpful, especially for patients who are nearing the terminal stages of the disease. Palliative care skills may be invaluable in the management of symptoms and can ease suffering for patients and their caregivers, thus allowing for maximum quality of life for as long as possible. End of life may bring its own complications and challenges; and opinion of an experienced and knowledgeable clinician can alleviate the pain and distress of the patient and also bring peace to the caregivers and loved ones.

Assessing the utility and attitudes toward molecular testing in neuro-oncology: a survey of the Society for Neuro-Oncology members.

BACKGROUND: Molecular testing (MT) is utilized in neuro-oncology with increasing frequency. The aim of this study was to determine clinical practice patterns to acquire this information, interpret and utilize MT for patient care, and identify unmet needs in the practical clinical application of MT. METHODS: We conducted a voluntary online survey of providers within the Society for Neuro-Oncology (SNO) membership database between March and April 2019. RESULTS: We received 152 responses out of 2022 SNO members (7.5% of membership). 88.8% of respondents routinely order MT for newly diagnosed gliomas. Of those who do not, testing is preferentially performed in younger patients or those with midline tumors. 82.8% use MT in recurrent gliomas. Other common indications included: metastatic tumors, meningioma, and medulloblastoma. Many providers utilize more than one resource (36.0%), most frequently using in-house (41.8%) over commercially available panels. 78.1% used the results for clinical decision-making, with BRAF, EGFR, ALK, and H3K27 mutations most commonly directing treatment decisions. Approximately, half (48.5%) of respondents have molecular tumor boards at their institutions. Respondents would like to see SNO-endorsed guidelines on MT, organized lists of targeted agents available for specific mutations, a database of targetable mutations and clinical trials, and more educational programs on MT. CONCLUSION: This survey was marked by several limitations including response rate and interpretation of MT. Among respondents, there is routine use of MT in Neuro-Oncology, however, there remains a need for increased guidance for providers to effectively incorporate the expanding genomic data resulting from MT into daily Neuro-Oncology practice.

The state of neuro-oncology during the COVID-19 pandemic: a worldwide assessment.

BACKGROUND: It remains unknown how the COVID-19 pandemic has changed neuro-oncology clinical practice, training, and research efforts. METHODS: We performed an international survey of practitioners, scientists, and trainees from 21 neuro-oncology organizations across 6 continents, April 24-May 17, 2020. We assessed clinical practice and research environments, institutional preparedness and support, and perceived impact on patients. RESULTS: Of 582 respondents, 258 (45%) were US-based and 314 (55%) international. Ninety-four percent of participants reported changes in their clinical practice. Ninety-five percent of respondents converted at least some practice to telemedicine. Ten percent of practitioners felt the need to see patients in person, specifically because of billing concerns and pressure from their institutions. Sixty-seven percent of practitioners suspended enrollment for at least one clinical trial, including 62% suspending phase III trial enrollments. More than 50% believed neuro-oncology patients were at increased risk for COVID-19. Seventy-one percent of clinicians feared for their own personal safety or that of their families, specifically because of their clinical duties; 20% had inadequate personal protective equipment. While 69% reported increased stress, 44% received no psychosocial support from their institutions. Thirty-seven percent had salary reductions and 63% of researchers temporarily closed their laboratories. However, the pandemic created positive changes in perceived patient satisfaction, communication quality, and technology use to deliver care and mediate interactions with other practitioners. CONCLUSIONS: The pandemic has changed treatment schedules and limited investigational treatment options. Institutional lack of support created clinician and researcher anxiety. Communication with patients was satisfactory. We make recommendations to guide clinical and scientific infrastructure moving forward and address the personal challenges of providers and researchers.

Uncertainty quantification in the radiogenomics modeling of EGFR amplification in glioblastoma.

Radiogenomics uses machine-learning (ML) to directly connect the morphologic and physiological appearance of tumors on clinical imaging with underlying genomic features. Despite extensive growth in the area of radiogenomics across many cancers, and its potential role in advancing clinical decision making, no published studies have directly addressed uncertainty in these model predictions. We developed a radiogenomics ML model to quantify uncertainty using transductive Gaussian Processes (GP) and a unique dataset of 95 image-localized biopsies with spatially matched MRI from 25 untreated Glioblastoma (GBM) patients. The model generated predictions for regional EGFR amplification status (a common and important target in GBM) to resolve the intratumoral genetic heterogeneity across each individual tumor-a key factor for future personalized therapeutic paradigms. The model used probability distributions for each sample prediction to quantify uncertainty, and used transductive learning to reduce the overall uncertainty. We compared predictive accuracy and uncertainty of the transductive learning GP model against a standard GP model using leave-one-patient-out cross validation. Additionally, we used a separate dataset containing 24 image-localized biopsies from 7 high-grade glioma patients to validate the model. Predictive uncertainty informed the likelihood of achieving an accurate sample prediction. When stratifying predictions based on uncertainty, we observed substantially higher performance in the group cohort (75% accuracy, n = 95) and amongst sample predictions with the lowest uncertainty (83% accuracy, n = 72) compared to predictions with higher uncertainty (48% accuracy, n = 23), due largely to data interpolation (rather than extrapolation). On the separate validation set, our model achieved 78% accuracy amongst the sample predictions with lowest uncertainty. We present a novel approach to quantify radiogenomics uncertainty to enhance model performance and clinical interpretability. This should help integrate more reliable radiogenomics models for improved medical decision-making.