Citations

Abstract Glioblastoma (GBM), the most aggressive primary brain tumor in adults, presents significant challenges due to its universal recurrence and limited survival rates, exacerbated when proximal to the lateral ventricles (LV). The subpopulation of brain tumor-initiating cells (BTICs) plays a pivotal role in tumor initiation and invasiveness, interacting with the microenvironment, particularly the cellular components in the subventricular zone (SVZ), like Neural Progenitor Cells (NPCs). Mechanisms of intercellular communication present in the SVZ include paracrine, autocrine, direct cell contact, and more recently extracellular vesicles (EVs). EVs are membrane-bound particles released by cells that are central to intercellular communication. They carry bioactive proteins, RNA, and other molecules that influence recipient cell behavior. This study investigates the role of EVs in GBM’s communication with non-cancer cells. Our findings indicate that BTIC-derived EVs promote migration and proliferation (*p < 0.02, ****p <0.001) in NPCs in vitro, highlighting the EVs’ role in altering the microenvironment. Additionally, we leverage a multi-omic approach to profile the small RNA and protein cargo of BTIC-EVs while also evaluating changes in the transcriptome of NPCs after BTIC-EV treatment. To identify the protein cargo contained in GBM-derived EVs, we utilized the innovative proximity-labeling system TurboID. We successfully induced TurboID expression in primary-cultured human BTICs from GBM patients, enabling the unbiased biotinylation of proteins in EVs. This method marks the first implementation of TurboID for unbiased global labeling of EV protein cargo in primary GBM cells. We identified unique miRNAs to be regulators of NPC’s transcriptome, and ontogenic analysis revealed upregulation of pathways involved in maintaining stem-like state, and upregulating PI3k-AKT, mTOR, and WNT signaling. This approach facilitates the investigation of EV-mediated communication and potential therapeutic targets, contributing to the understanding of GBM’s complex interactions with the brain’s microenvironment and the identification of biomarkers for improved diagnosis and treatment response.

Abstract Glioblastoma, IDH-wildtype (GBM), and primary central nervous system lymphoma (PCNSL) are two aggressive brain tumors with distinct treatment strategies but overlapping imaging features on MRI. Accurate preoperative differentiation remains a clinical challenge, due to their overlapping imaging features, even when advanced MRI techniques are employed. This study explored deep learning approaches leveraging multi-sequence MRI to distinguish GBM from PCNSL. Models were developed utilizing 146 PCNSL cases, which were matched to146 GBM by age, sex and year of diagnosis. A 3D-DenseNet121 convolutional neural network was trained using two analytical strategies: (i) loss-based model selection and (ii) ensemble prediction using five-fold cross-validation. Model performance was evaluated on an independent validation set of 256 GBM and 37 PCNSL from the same institution as well as 36 PCNSL cases from an external institution. All MRIs were preprocessed using the FeTS pipeline. Using the independent validation cohort, area under the receiver operating characteristic curve (AUC) and 95% confidence intervals (CI) were computed overall and stratified by demographic characteristics. Using both T1 post-contrast (T1Gd) and T2-weighted images, the loss-based model achieved AUC=0.83 (95% CI: 0.77-0.88) and the ensemble model yielded an AUC=0.82 (95% CI: 0.76-0.89). Model performance was similar across males and females; however, model performance was better for subjects older than 50 years of age versus younger than 50. When using only T1Gd images, the loss-based model dropped to AUC=0.55 (95% CI: 0.48-0.63), whereas the ensemble model had AUC=0.79 (95% CI: 0.72-0.85). Models trained with only T2 images achieved AUCs of 0.67 (95% CI: 0.60-0.73) (loss-based) and 0.75 (95% CI: 0.69-0.82) (ensemble). These results highlight the added value of the multi-sequence input. Our results confirm the feasibility of using automated deep learning models to differentiate GBM from PCNSL. Notably, the model generalized well to external MRI data, emphasizing its potential utility as a diagnostic aid in real-world clinical workflows.

Abstract Glioblastoma (GBM) diagnosis integrates morphology and tumor molecular alterations. Patients with classic morphology are designated as histologic GBM (hGBM). Patients without classic morphology harboring TERT promoter mutation (TERTp), or epidermal growth factor receptor amplification (EGFRamp), or combined gain of chromosome 7 and loss of 10 (+7/-10) designate molecular GBM (mGBM). We compared radiographic features, genomic alterations, and clinical outcome between hGBM and mGBM. 109 mGBM Mayo Clinic cases diagnosed between 1993-2021 were matched to 101 cGBM by age, sex, year of diagnosis, and TERT promoter (TERTp) status. Clinical and radiographic data abstraction was blinded. Progression-free and overall survival were compared across groups in the subset of patients who received standard of care. Contrast enhancement was present on index MRI (MRI at diagnosis) for 98% (96/98) hGBM and 50% (54/107) mGBM (p<0.0001) patients. mGBM tumors harbored fewer total copy number alterations (median of 16.5 vs 20, p=0.029) and were less likely to have CDKN2A/B homozygous deletion (56% vs 77%, p=0.005). Among the 87 mGBM and 82 hGBM receiving standard of care, median overall survival was longer for mGBM versus hGBM (2.24 vs 1.62 years, p=0.016). mGBM patients with +7/-10 (HR=2.60, p=0.035) or CDKN2A/B (HR=3.61, p=0.002) had poorer overall survival compared to mGBM patients without the corresponding alteration. On index MRI mGBM patients without contrast enhancement had better overall survival compared to hGBM (HR=0.50; p=0.003) whereas there was no survival difference between mGBM with enhancement and hGBM. mGBM with enhancement on index MRI likely represent undersampled, or early / evolving hGBM. mGBM patients with non-enhancing MRI, younger than 60, with long symptom profile including seizures, and some combination of non-anaplastic morphology, TERTp only, no homozygous deletion of CDKN2A/B, and low copy number complexity may represent other IDHwt glioma entities. Caution is advised in counseling, management, discussing prognosis, and inclusion in clinical trials.

Abstract BACKGROUND CNS ependymal tumors rarely demonstrate mixed subependymal and ependymal pathology. The clinical course of these tumors is difficult to predict, but prior studies have identified TERT mutations and loss of chromosome 6 as poor prognostic factors in these tumors when located in the posterior fossa. METHODS In this retrospective single-institution study, patients with a diagnosis of mixed subependymoma/ependymoma were identified and chart reviewed to describe demographics, pathological findings, treatment courses and clinical outcomes. RESULTS Eleven patients (median age 57, 2 female, 9 male) were identified. Ten tumors were in the posterior fossa and 1 in the cervical spine. At diagnosis, 2 underwent gross total resection, 8 underwent subtotal resection, and 1 was only biopsied. Radiotherapy was administered to 5 patients at initial diagnosis (54Gy in 30 fractions). Systemic therapy (temozolomide and bevacizumab) was administered at recurrence to 2 patients. Loss of chromosome 6 was present in 5 /11 patients and TERT mutations were present in 4/11 patients, with both present in 3/11 patients. Median progression-free survival (mPFS) was 90.1 months (95% CI: 90.1–NA), with 3 progression events observed at 22, 90, and 107 months. Median follow-up duration was 27.9 months. Receiving radiotherapy had a longer mPFS (98 vs 22 months, p=0.15), although non-significant. At the time of data cut-off none of the patients were deceased. CONCLUSIONS Mixed subependymoma/ependymoma is a rare diagnosis with varying clinical outcomes. Our case series demonstrated that the majority of patients did not exhibit the previously identified poor prognostic factors. Although non-significant due to limited case numbers, radiotherapy was suggestive of longer mPFS.

Abstract Computer vision now enables deep learning models to link histopathologic features to molecular features in a quantitative manner not previously achievable by human pathologists. However, human interpretability is often limited. We trained a deep neural network to predict gene expression from digital whole slide images (WSIs) using 848 central nervous system (CNS) tumors with paired WSI and RNA-sequencing data. We used inferred RNA expression levels as a surrogate for protein expression of the CNS tumor lineage markers GFAP, OLIG2, and SSTR2, as well as the proliferation marker MKI67 (Ki67). We established thresholds for categorizing the inferred expression levels as positive or negative based on levels observed in cross-validation testing in select tumor types to approximate routine clinical interpretation of immunohistochemical (IHC) staining. We tested the sensitivity and specificity of our ‘virtual’ IHC on an independent, multi-institutional cohort of over 2,000 CNS tumors with objective diagnostic labels derived from DNA methylation-based tumor classification. As a dichotomous variable, inferred GFAP expression showed a sensitivity of 74% and a specificity of 99% as a glial marker in a cohort of gliomas and meningiomas. In the same cohort, OLIG2 showed a sensitivity of 75% and a specificity of 97% as a glial marker, while SSTR2 showed a sensitivity of 83% and a specificity of 97% as a marker of meningioma. In a cohort of ependymomas and gliomas, virtual IHC for OLIG2 showed a sensitivity of 75% and a specificity of 82% in distinguishing ependymomas from gliomas. Finally, in a cohort of gliomas, inferred expression of MKI67 demonstrated a trend consistent with what would be expected by actual IHC, with increasing MKI67 expression with increasing tumor grade. Our model provides the basis for a human interpretable and clinically applicable deep neural network to aid human pathologists in the diagnosis and grading of CNS tumors.

Abstract The 2021 WHO classification redefined glioblastoma IDH-wildtype (GBM), integrating histopathological and tumor molecular features. Diffuse astrocytic glioma lacking histologic features of necrosis and microvascular proliferation are classified as (molecular) GBM if the tumor is IDH- and H3-wildtype and has one or more additional tumor molecular alterations: TERT promoter mutation, EGFR amplification, or +7/-10. Due to the absence of histopathological features of classic GBM and presentation without enhancement on MRI in approximately 50% of cases, molecular GBM frequently resemble low-grade IDH-mutant gliomas. We hypothesized that MRI-based deep learning models could distinguish IDH-mutant glioma from molecular GBM. We retrospectively identified 97 individuals with molecular GBM, who were matched to 97 individuals with IDH-mutant glioma by age, sex, year of diagnosis, and enhancement on MRI. Of the 97 IDH-mutant gliomas, 51% were 1p/19q codeleted, 42% grade 2, 47% grade 3, and 10% grade 4. Preoperative postcontrast T1-weighted (T1C) and T2-weighted (T2) MRI sequences were used to train a 3D-DenseNet121 model. The model was validated on an independent prospective set of 150 IDH-mutant gliomas and 71 molecular GBM. The MRI-based deep learning model achieved an area under the receiver operator curve (AUC) of 0.75 (95% CI: 0.68-0.82) in the prospective validation set. Sex-stratified AUCs were 0.72 (95% CI: 0.59-0.85) for females and 0.76 (95% CI: 0.68-0.85) for males. Age stratified AUCs were 0.71 (95% CI: 0.60-0.83) for patients under 55 years of age and 0.64 (95% CI: 0.50-0.79) for patients older than 55. These results suggest that discriminative MRI features of IDH-mutant glioma and molecular GBM exist and can be identified from deep learning models. However, radiographic ambiguity exists and underscores the need for further optimization of the deep learning architecture or by integrating additional non-invasive modalities. For example, future work entails integrating germline variants into the model, including rs55705857, which is a causal variant for IDH-mutant glioma. With further validation and integration of multi-modal data, a noninvasive tool may support more personalized decision-making.

Abstract Molecular testing has become critical components in the diagnostic classification of the central nervous system (CNS) tumors. However, these methods require substantial resources, limiting accessibility for many patients. Recent advances in artificial intelligence (AI) and computer vision empower deep learning models to infer molecular features from histopathology images, which may often be more readily available than molecular testing, to classify CNS tumors. We trained pan-CNS tumor models to predict DNA methylation and gene expression from whole-slide images (WSIs). These molecular predictions were then used in a hierarchical machine learning framework, termed Neuropath-AI, to predict nine broad tumor families and 52 tumor types using a large diverse cohort of 5,715 histopathology samples, of which 2,988 had paired DNA methylation profiling and 848 had paired RNA-sequencing. We evaluated Neuropath-AI on another large independent multi-institutional cohort of 5,516 CNS tumors, for which Neuropath-AI predicted tumor class with an associated confidence score. Neuropath-AI designated 46% of test samples as predictable with high-confidence, which it then accuracy classified in 97% of samples. It made moderate-confidence (and above) predictions for 87% of test samples, for which it achieved a diagnostic accuracy of 80% for the top-1 prediction and 86% accuracy for a top-2 predictions. A comparison with human neuropathologists showed that Neuropath-AI is comparable at classifying tumors. A second test showed that human neuropathologists were more accurate when the Neuropath-AI classifier results were provided to them, testifying to the potential translational benefit of integrating AI classification tools into the pathologist’s workflow. Our model provides the basis for a clinically applicable deep learning assistant to improve human efficiency and accuracy of CNS tumor diagnosis. The model will be made publicly available and can be readily implemented to assist human pathologists in clinical workflows, in further expanded prospective studies.

Abstract IDH-wildtype diffuse astrocytic gliomas with lower-grade histological features alongside specific genomic alterations (e.g., TERT promoter mutation, EGFR amplification, gain Chr7/loss Chr10) are presently diagnosed as glioblastoma, IDH-wildtype (GBM) under WHO 2021 criteria. Previous analyses categorized these into “early/evolving” or “surgically undersampled” GBMs. However, it remains unclear whether these tumors share identical clinical and prognostic features with histologic GBM. Data from IDH-wt molecular GBMs across six institutions were aggregated. Astrocytic IDH-wt tumors of any grade that lacked necrosis or microvascular proliferation on initial histopathology yet exhibited molecular alterations consistent with WHO 2021 GBM diagnosis, were included. Recursive partitioning analysis (RPA) was used to categorize patients into risk groups based on overall survival (OS). Among the 264 patients included, the median age was 60 years (IQR: 51-67), with 162 (61%) being male. Additionally, 210 (80%) were diagnosed since 2016, and 216 (82%) received chemoradiation plus adjuvant temozolomide (GBM-SOC). The median follow-up period was 3.9 years (95% CI: 3.4, 4.6). Identified univariate risk factors for worse survival included older age, biopsy-only status, EGFR amplification, absence of initial adjuvant treatment, low post-operative KPS, and tumor histologic grade of 3. RPA divided patients into four distinct survival groups (p<0.0001). The “best” risk group comprised lower grade (<=2) tumors (n=58, median OS (mOS): 40 months (95% CI: 33, NA)). The “second best” group included Grade 3 patients receiving GBM-SOC without steroids during initial RT (n=139, mOS: 25 months (95% CI: 22, 28)). The “second worst” group encompassed Grade 3 patients who were on steroids during initial RT while receiving GBM-SOC (n=38, mOS: 15 months (95% CI: 13, 20)). The “worst” group contained Grade 3 tumors that had not undergone SOC (n=29, mOS: 9 months (95% CI: 4, 26)). These data show that histologic features and treatment are strong indicators of patient outcomes.

Abstract BACKGROUND The majority of pituitary neuroendocrine tumors (PitNETs) are effectively managed with medical therapy, surgery, and/or radiation therapy; however, a subset of pituitary adenomas exhibit aggressive behavior. Temozolomide is thought to be an effective treatment option for patients with PitNETs. METHODS We identified and retrospectively reviewed 11 patients with pituitary tumors treated with temozolomide. We then conducted a review of the literature, combining 114 previously published cases with our cohort. RESULTS 125 cases were identified with diagnoses of pituitary adenomas (n=45), aggressive pituitary tumors (n=61), and pituitary carcinoma (n=19) treated with temozolomide. 79 patients were male (68.1%) with a median age of 47.5 (range 9-72). A variety of tumor types were represented: 50 corticotroph adenomas (40.7%), 37 prolactinomas (30.1%), 22 hormone nonfunctioning adenomas (17.9%), 8 somatotroph adenomas (6.5%), 6 other functioning adenomas (4.9%), and 2 cases where the functional status was not mentioned. Median treatment duration was 9 cycles (n=91, 150-200 mg/m2 days 1-5, 28-day cycles; n=2, 75 mg/m2, days 1-21, 28-day cycles; n=1, 75 mg/m2 daily). 27 patients received concurrent therapy with a dopamine agonist/somatostatin analogue (n=10), other chemotherapy agent (n=9) or radiotherapy (n=8). Radiographic responses included stable disease (n=40), partial response (n=35), complete response (n=8), and progressive disease (n=30). Hormonal responses included complete response (n=13), partial response (n=32), and no response (n=16). 22 patients (47.8%; out of 46 patients where progression/recurrence was assessed) experienced progression/recurrence following temozolomide initiation with a median progression free survival (PFS) of 24 months. There was no significant association between MGMT methylation status and radiological or hormonal response. No unexpected toxicities were observed. CONCLUSION In this cohort, temozolomide was well-tolerated with a median PFS of 24 months. MGMT methylation status did not predict temozolomide response. Temozolomide is an option with a survival benefit for patients with PitNETs.

Abstract BACKGROUND Patients with primary brain tumors often have very complex histories, but an accurate understanding of their care timeline is critical to provide the best care possible. Extracting pertinent facts from the medical record can be time consuming and frustrating. We sought to apply generative AI large language models (LLM) to this problem in order to supply providers with concise, accurate, and focused data including graphs and tables. METHODS After evaluating multiple LLMs, GEMINI 2.5 Pro was selected. Prompt engineering was iteratively refined across a range of patients to generate concise outputs using tables, graphs, and prose to describe timing of chemoradiation, tumor markers, procedures, and trends in MRI results. The model inferred BT-RADS scores from the MRI report prose. Using a 5-point Likert scale, neuroradiologists assessed the AI-output both prospectively and retrospectively comparing it to the medical record. LLM’s compared AI summaries to source material. RESULTS Six neuroradiologists each ranked 25 prospective patient outputs as accurate (82%/16%/1%), useful (94%/6%/0%), trustworthy (86%/13%/1%), and satisfactory (86%/10%/1%) for Likert ratings of highly agree/agree/neutral, respectively. No responses were disagree or highly disagree. Interobserver agreement was high (Gwent AC2 > 0.95). The LLM evaluation of referenced portions of the record determined > 85% of summary statements were supported by source documents. While users estimated time savings of a 1-10 minutes per case, they felt the greater benefit was having more organized data presented efficiently. CONCLUSIONS AI summaries of the medical record of complex patients with primary brain tumors give providers an accurate, well-organized display of the pertinent facts necessary for high quality care and save time in chart review. These summaries, which are generated early each morning, have been found to be helpful and are now being incorporated into our clinical practice and multidisciplinary tumor board.