Lachance, Daniel Honore

Abstract Clinical management of IDH wildtype glioblastoma (GBM) and tumefactive multiple sclerosis (tMS) is drastically different. GBM requires maximal safe resection followed by chemoradiation, while tMS outcome is worsened by surgery and radiotherapy. Misdiagnosis of brain lesions may expose patients to unnecessary anxiety, surgery, or radiotherapy. Noninvasive methods to help with accurate diagnosis of tumor and non-tumor etiologies are needed. tMS subjects were identified by the Mayo Clinic Center for Multiple Sclerosis and Autoimmune Neurology (CMSAN). GBM subjects were matched to tMS by age at diagnosis, index MRI date, and sex. Inclusion criteria included one cm minimal lesion size and pre-operative post-contrast T1 and T2 images available for analysis. To identify potential experimental design bias, MRI technology and clinical characteristics were compared across groups. A 3D-DenseNet121 was used to develop a classification model using prespecified parameters: 600 epochs, batch size 16, learning rate 10-3, cross entropy loss, and AdamW optimizer. The stopping rule was defined as three sequential epoch cross entropy loss < 0.02. Five-fold cross validation (CV) was utilized to obtain an unbiased estimate of predictive performance (overfitting) and sensitivity of the 3D-DenseNet121 algorithm was evaluated. 224 subjects (114 GBM, 110 tMS) were analyzed: 40% male, median age 46. After one round of CV, the average unbiased estimate of sensitivity and specificity for predicting GBM from the final classification model was 0.86 and 0.73, respectively. To assess the reproducibility of the deep learning algorithm (3D-DenseNet121), the exact same CV folds were rerun resulting in average sensitivity and specificity of 0.82 and 0.75, respectively. Obtaining unbiased estimates of model performance using internal validation is critical before evaluating a final model on an external dataset. Understanding sensitivity of a deep learning algorithm is important for model validity. Differentiation of GBM and tMS appears reasonable with MRI; external validation is in progress.

Abstract INTRODUCTION Prior to 2021, (classic) glioblastoma (GBM) required absence of IDH mutation and histologic features of hypercellularity, nuclear atypia, microvascular proliferation (MVP), and necrosis. Absence of histologic features can be a result of undersampling, with molecular, radiologic, and clinical features consistent with GBM. Accordingly, WHO 2021 defined molecular glioblastoma as diffuse glioma IDHwt, H3wt, with one or more of MVP, necrosis, TERT promoter mutation, EGFR gene amplification (EGFRamp), and/or +7/−10. We hypothesized that there is clinical and radiographic heterogeneity among tumors defined as molecular GBM. METHODS We identified 121 patients that met the criteria for molecular GBM. 121 Classic GBM, defined as IDHwt with hypercellularity, MVP or necrosis, were matched to molecular GBM cases by age, sex, and diagnosis date. All cases were reviewed for relevant histologic features and molecular features. Examined clinical variables included: first symptom, time from first symptom to date of pathological diagnosis, date of first MRI, tumor location, radiographic features of enhancement, diffusion or mass effect, time to progression and overall survival. Clinical, radiologic, and molecular features were compared between classic and molecular GBM. RESULTS Currently,160 cases have been reviewed. Preliminary results show that there is a longer time from first symptom to diagnosis for molecular GBM. Molecular GBM is significantly more likely to be discovered coincidentally, and without enhancement or radiographic necrosis. Classic GBM is significantly more likely to present with focal neurologic deficit as first symptom, regional mass effect on index imaging and have monosomy 10. CONCLUSION While pathologic undersampling accounts for a significant proportion of molecular GBM, our data suggest that there are significant clinical, radiographic, and outcome differences between classical and molecular GBM. Final analysis of the entire cohort will help better define this clinical spectrum.

Abstract BACKGROUND The diagnosis and management of patients with brain tumors currently uses WHO defined morphologic and molecular features. However, neuro-oncology practice in low resource settings relies on histomorphology. This study aims to reclassify glioma cases diagnosed in the Department of Anatomic and Molecular Pathology Lagos University Teaching Hospital using the 2021 WHO classification. METHODS Fifty-six cases of glioma diagnosed between 2013 to 2021 were evaluated. Morphologic diagnosis was reassessed. Five-micron sections were obtained for immunohistochemistry (IHC), and genetic testing (DNA and RNA); after manual macrodissection for tumor enrichment using 10μm sections. IHC for IDH1-R132H, ATRX, BRAFV600E, P53,Ki67, and H3K27M and OncoScan chromosomal microarray analysis were performed. NGS mutation/fusion analysis and EPIC methylation array profiling are in progress. RESULTS Of 56 cases evaluated, median age was 22 years (range=1-60 years); 71% were males. The initial morphologic diagnosis include 21 pilocytic astrocytoma, 11 glioblastoma, 9 ependymoma, 7 diffuse astrocytoma, 3 anaplastic astrocytoma, 2 high-grade gliomas, 1 oligodendroglioma, 1 diffuse midline glioma, and 1 pleomorphic xanthoastrocytoma (PXA). Histomorphologic re-evaluation revealed discordant diagnosis in 29% (16/56) of cases. Of the 56 cases, 73% (41/56) had usable DNA yield and 38% (21/56) had DNA yield of ≥ 750ng. Chromosomal microarray analysis was completed for 50% (28/56) cases. The revised 2021 WHO diagnosis were: 5 glioblastoma, IDH-wildtype, 5 pilocytic astrocytoma, 4 glioma NEC, 3 posterior fossa ependymoma, group B, 3 PXAs, 3 astrocytoma, IDH-mutant, 2 pediatric low-grade glioma, 2 supratentorial ependymoma, ZFTA fusion-positive, 1 posterior fossa ependymoma, group A. In 71% (20/28) of cases, diagnosis was revised using the 2021 WHO criteria. CONCLUSION Most cases had usable DNA yield. Overall, hsitologic re-evaluation and molecular testing based on the 2021 WHO classification guidelines refined the diagnosis in 48% (27/56) of cases.

Abstract INTRODUCTION Brain tumors constitute significant morbidity and mortality in sub-Saharan Africa (SSA). Improvements in neuro-oncology practice and research have seen increased diagnosis of brain tumors. However, to our knowledge, there have been no previous projects on brain tumor genetics in SSA. To understand the feasibility of developing brain tumor biobanking and registries, this project aims to start by mapping out the current neuro-oncology capacity in terms of workforce and available services. METHODOLOGY We will conduct a descriptive questionnaire-based online survey among a cross section of neuro-oncology practitioners spread across SSA countries. The dissemination of the survey questionnaire will be facilitated by the Society for Neuro-oncology in SSA (SNOSSA). The Google designed questionnaire will capture information on the distribution, and number of available neuro-oncology care practitioners, capacity for brain tumor diagnosis and management, cancer registration, tumor biobanking, and research. CLINICAL IMPACT AND CONCLUSION This study will generate robust data from SSA, which can be used to improve future research and promote advocacy towards brain tumor diagnosis and management. This work will also serve as a pilot study for subsequent research that will focus on the development of brain tumor-specific registries and biorepositories in SSA.

Abstract BACKGROUND The diagnosis and management of patients with brain tumors currently uses WHO defined morphologic and molecular features. However, neuro-oncology practice in low resource settings relies solely on histomorphology. This study aims to reclassify glioma cases diagnosed in the Department of Anatomic and Molecular Pathology Lagos University Teaching Hospital using the WHO 2021 classification. METHODS Fifty-six cases of glioma diagnosed between 2013 to 2021 were evaluated. Morphologic diagnosis was reassessed. Five-micron sections were obtained for immunohistochemistry (IHC), and genetic testing (DNA and RNA) after manual macrodissection for tumor enrichment used 10-µm sections. IHC for IDH1-R132H, ATRX, BRAF-V600E, p53, Ki67 and H3-K27M and Oncoscan chromosomal microarray analysis were performed. NGS mutation/fusion analysis and EPIC methylation array profiling are in progress. RESULTS Of the 56 cases evaluated, median age was 22 years (range, 1-60); 71% were male. The initial morphologic diagnoses included 21 pilocytic astrocytoma, 11 glioblastoma, 9 ependymoma, 7 diffuse astrocytoma, 3 anaplastic astrocytoma, 2 high-grade gliomas, 1 oligodendroglioma, 1 diffuse midline glioma, and 1 pleomorphic xanthoastrocytoma (PXA). Histomorphologic re-evaluation revealed discordant diagnosis in 29% (16/56) of cases. Of the 56 cases, 73% (41/56) had usable DNA yield and 38% (21/56) had DNA yield of ≥750ng. Chromosomal microarray analysis was completed for 50% (28/56) cases. The revised WHO 2021 diagnosis was: 5 glioblastoma IDH-wildtype, 5 pilocytic astrocytoma, 4 glioma NEC, 3 posterior fossa ependymoma, group B, 3 PXA, 3 astrocytoma IDH-mutant, 2 pediatric low-grade gliomas, 2 supratentorial ependymoma, ZFTA fusion-positive, 1 posterior fossa ependymoma, group A. In 71% (20/28) of cases with molecular testing, diagnosis was revised using WHO 2021 criteria. CONCLUSION Most cases had usable DNA. Overall, histologic re-evaluation and molecular based on the WHO 2021 criteria refined the diagnosis in 48% (27/56) of cases.

e24089 Background: Little is known about risk and protective factors for dementia in cancer survivors. With increasing longevity due to improved cancer therapy, it’s prudent to study factors contributing to cognitive impairment in this patient population. Methods: We performed multivariable cox regression analyses to assess established risk factors for all-cause dementia (age, education, APOE4-status, social and physical activity engagement) in breast and prostate cancer survivors. Patients 50+ years old with pathology-confirmed cancer diagnosis at Mayo Clinic were identified from a longitudinal, population-based cohort – Mayo Clinic Study of Aging. Analyses were controlled for age, APOE4-status, education, and cancer treatment. Results: We included a total of 209 patients with breast cancer (mean age 75.4 years, 24% APOE4-positive, median follow-up 6.8 years) and 352 with prostate cancer (mean age 77 years, 26% APOE4-positive, median follow-up 6.4 years). 31 patients with breast cancer (15%) and 66 patients with prostate cancer (20%) were diagnosed with dementia at last follow-up. Older age was associated with increased risk for dementia (5-year increase – breast: hazard ratio (HR) 1.9 [95%CI 1.3-2.7, p = 0.001]; prostate: HR 2.4 [95%CI 1.8-3.2, p < 0.0001]). APOE4-positivity was associated with increased dementia risk in breast (HR 3.6 [95%CI 1.6-7.8, p < 0.002]), but not prostate cancer (HR 1.2 [95%CI 0.7-2.0, p = 0.5]). More education was associated with lower dementia risk in prostate cancer (education year: HR 0.91 [95%CI 0.84-0.98, p < 0.02]; education years > 12: HR 0.26 [95%CI 0.1-0.6, p < 0.001]). Physical activity engagement in mid-life was associated with lower dementia risk (prostate: HR 0.56 [95%CI 0.4-0.8, p < 0.005]; breast: HR 0.48 [95%CI 0.3-0.9, p < 0.02]). Conclusions: In this longitudinal, population-based study we found that some risk and protective factors for dementia in cancer survivors vary by cancer type. While advanced age is an expected dementia risk factor, there were mixed results for APOE4-positivity (increased risk only in breast cancer) and education (only protective in prostate cancer). Our results support further study into factors influencing dementia risk by cancer type in hopes of identifying preventative strategies.

AbstractBackgroundMeningiomas are the most common primary central nervous system (CNS) tumor in adults and CNS World Health Organization grade 2 (atypical) meningiomas show an intermediate risk of recurrence/progression. Molecular parameters are needed to better inform management following gross total resection (GTR).MethodsWe performed comprehensive genomic analysis of tumor tissue from 63 patients who underwent radiologically confirmed GTR of a primary grade 2 meningioma, including a CLIA-certified target next-generation sequencing panel (n = 61), chromosomal microarray (n = 63), genome-wide methylation profiling (n = 62), H3K27me3 immunohistochemistry (n = 62), and RNA-sequencing (n = 19). Genomic features were correlated with long-term clinical outcomes (median follow-up: 10 years) using Cox proportional hazards regression modeling and published molecular prognostic signatures were evaluated.ResultsThe presence of specific copy number variants (CNVs), including -1p, -10q, -7p, and -4p, was the strongest predictor of decreased recurrence-free survival (RFS) within our cohort (P &lt; .05). NF2 mutations were frequent (51%) but did not show a significant association with RFS. DNA methylation-based classification assigned tumors to DKFZ Heidelberg benign (52%) or intermediate (47%) meningioma subclasses and was not associated with RFS. H3K27 trimethylation (H3K27me3) was unequivocally lost in 4 tumors, insufficient for RFS analysis. Application of published integrated histologic/molecular grading systems did not improve prediction of recurrence risk over the presence of -1p or -10q alone.ConclusionsCNVs are strong predictors of RFS in grade 2 meningiomas following GTR. Our study supports incorporation of CNV profiling into clinical evaluation to better guide postoperative patient management, which can be readily implemented using existing, clinically validated technologies.

Abstract BACKGROUND The topologically associated domain (TAD) on 8q24 surrounding MYC is important for many types of cancer. The region is also important in the formation of IDH-mutant glioma; the risk allele rs55705857 being associated with the development of these tumors. The rs55705857 risk allele significantly increased the formation of tumors in an IDH-mutant mouse model. Cerebral organoids have been found to recapitulate the early development of the brain. We hypothesized that the rs55705857 risk allele may alter the phenotype of cerebral organoids. METHODS Isogenic induced pluripotent stem cells (iPSCs) were developed with (n=3) and without (n=2) the rs55705857 risk allele from the parental cell line PGP-1 (GM23338). These cells were cultured and differentiated into unpatterned cerebral organoids and observed throughout differentiation. Organoids were fixed, sectioned, and stained by H&E and immunofluorescence. RESULTS Organoids containing the risk allele were smaller throughout development, starting at the embryoid body stage and continuing up to over 75 days in culture, compared to two isogenic non-risk lines. At 4 weeks the risk allele containing organoids also contained a larger number of small rosette-like neuroepithelia, instead of the more continuous neuroepithelia observed in the non-risk allele organoids. At 75 days, the organoids with the risk allele maintained a larger proportion of Sox2 positive cells, which are enriched in the tight clusters of small cells with round nuclei. CONCLUSIONS The risk allele rs55705857 in the MYC TAD at 8q24 is altering the development of unpatterned cerebral organoids. This difference is visible as early as embryoid body development and persists through neuroepithelial development. Further studies of these differences may help our understanding of how rs55705857 accelerates IDH-mutant glioma development.

Abstract Missing MRI sequences are common in radiology due to differences in protocols, time constraints, and image artifacts. Missing sequences are a nuisance for subsequent algorithms, such as automated segmentation algorithms requiring multiple MRI sequences (e.g., HD-GLIO, which requires T2, T1, post-contrast T1 (T1c), and FLAIR). Accurate tumor segmentation is a fundamental step in extracting radiomics analysis to use, for example, to predict the IDH status of glioma or their overall survival. We validated a conditional (pix2pix) and super-resolution GAN (SOUP-GAN) for imputing MRI sequences on 285 adult glioma with complete MRI sequences available. We simulated missing T1 sequences and subsequently imputed them from T1c using pix2pix and from sagittal T1 using SOUP-GAN. Using HD-GLIO, we obtained the true segmentations from the complete MRI sequences. HD-GLIO was also used after replacing the original T1 sequence with the one imputed with pix2pix and SOUP-GAN. Thus, there were three sets of segmentations for each subject. We compared T2 and T1c radiomics features extracted with Pyradiomics using segmentations obtained with complete data with features extracted from segmentation obtained with the imputed T1 sequences using intraclass correlation coefficient (ICC) and Bland-Altman (BA) analysis. Median (interquartile range) ICC for shape features extracted from T2 and T1c were 0.82(0.76-0.92) using pix2pix and 0.89(0.86-0.94) using SOUP-GAN. Median ICC for first-order-statistics features were 0.90(0.83-0.93) for T2 and 0.90(0.88-0.92) for T1c using pix2pix, and 0.61(0.22-0.82) for T2 and 0.42(0.34-0.57) for T1c using SOUP-GAN. Median ICC for texture features were 0.86(0.82-0.87) for T2 and 0.87(0.79-0.90) for T1c using pix2pix, and 0.69(0.47-0.81) for T2 and 0.76(0.65-0.83) for T1c using SOUP-GAN. We show that T2 and T1c shape features (e.g., volumetric analysis) can be adequately extracted from segmentations obtained with an algorithm requiring multiple MRI sequences using T1 sequences imputed with GANs; however, the effect of imputation differed for first-order-statistics and texture features.

Abstract Atypical meningiomas (CNS WHO grade 2) show an intermediate risk of recurrence/progression, and molecular parameters informing management following gross total resection (GTR) remain to be established. We performed comprehensive genomic analysis in 63 patients who underwent GTR of an atypical meningioma. Median age was 59.8 years, 41 patients (65.1%) were female, and 8/62 patients (12.9%) underwent radiation therapy following resection. Follow-up was available for all patients (median 10 years), with recurrence/progression in 17 (5-year PFS: 84.2%; 10-year PFS: 71.0%). Tumors were evaluated by next generation sequencing (n=61), chromosomal microarray (n=63), DNA methylation profiling (n=62), and H3K27me3 immunohistochemistry (n=62). Next generation sequencing using a CLIA-certified 50-gene Neuro-Oncology specific panel identified pathogenic alterations in NF2 (n=31; 50.8%), SMARCB1 (n=5; 8.2%), AKT1 (n=2; 3.3%), TERT (n=1, 1.6%), SMO (n=1), SUFU (n=1), TP53 (n=1), and PTEN (n=1). There was trend toward decreased PFS in patients with NF2 alterations (p=0.078) which was not statistically significant. Copy number variants (CNVs) were identified using OncoScan (ThermoFisher). The most frequent CNVs were loss -22 (n=45; 71.4%), -1p (n=30, 47.6%), -14 (n=21, 33.3%), -6q (n=18, 28.6%), -10q (n=15, 23.8%), -X (n=13, 20.6%), and -18 (n=11, 17.4%). CNVs associated with decreased PFS included -1p (p=0.015), -7p (p=0.007), -10q (p=0.0007), and -18 (p=0.0376). DNA methylation profiling was performed using the Illumina MethylationEPIC array (n=62) and evaluated using the DKFZ classifier (v12.3). Based on the highest score, tumors matched to benign (n=32; 51.6%) or intermediate (n=29; 46.7%) meningioma subclasses or showed non-contributory profiles (n=1; 1.6%). Methylation class (benign vs. intermediate) showed no association with PFS (p=0.96). H3K27 trimethylation (H3K27me3) was unequivocally lost in 4 (6.5%) tumors, which was insufficient for PFS analysis. Overall, our study supports CNV profiling of atypical meningiomas following GTR, which can be implemented using existing, clinically validated technologies. Molecular predictors could guide use of radiotherapy in these patients.