Bruce, Jeffrey N.

PURPOSE Age ≥40 years is regarded as a high-risk feature and an indication for adjuvant chemoradiotherapy for patients with lower-grade glioma in clinical practice guidelines. It is unclear whether age remains a relevant prognostic factor for contemporary definitions of lower-grade gliomas in the molecular era. METHODS The Prospective Gliomas Research (PROGRES) database contains individual patient-level data from 11 prospective clinical trials or observational registries of histologically defined lower-grade 2-3 oligodendroglioma or astrocytoma. We determined the association of age (18-39 years v ≥40 years) with progression-free survival (PFS) stratified by isocitrate dehydrogenase 1 or 2 ( IDH1/2 ) status, using log-rank tests and Cox regression models. We validated our findings in a separate multi-institutional retrospective cohort (Retrospective Glioma Research [REGRES] database). RESULTS We identified 1,619 and 1,292 eligible patients in the PROGRES and REGRES cohorts, respectively. IDH -wildtype tumors were more common in patients 40 years and older (38% v 5%, odds ratio: 11.3 [95% CI, 6.5 to 19.7]). Age was associated with PFS in IDH -wildtype (5-year PFS for ≥40 v 18-39 years: 6% v 24%, hazard ratio [HR], 1.74 [95% CI, 1.21 to 2.50]) but not in IDH -mutant glioma (60% v 59%, HR, 0.89 [95% CI, 0.76 to 1.05], P interaction < .001). In IDH -wildtype tumors, older age predicted aggressive molecular features, including TERT promoter mutation (65% v 28%), EGFR amplification (41% v 15%), and chromosome +7/–10 alteration (57% v 25%). In a pooled analysis of four clinical trials, age was not predictive of a benefit from chemoradiotherapy versus radiotherapy alone for IDH -mutant glioma. CONCLUSION In the absence of additional clinical or molecular risk factors, age alone should not be considered an indication for administration or deferral of adjuvant treatment. Practice guidelines should be revised to reflect contemporary prognostic factors in the molecular era.

Abstract In this study, we investigated tumor intrinsic mechanisms of resistance to cevostamab, a bispecific T-cell engager targeting Fc receptor-like protein 5 (FCRL5) and CD3 that has shown promising clinical activity and durable responses in advanced MM. Longitudinal bone marrow aspirates from nine MM patients treated with cevostamab were analyzed. All patients had previously been treated with both immunomodulatory drugs (IMiDs) and proteasome inhibitors (PI) and eight patients had received daratumumab prior to cevostamab, which was, on average, the 6th line of treatment (mean treatment duration 220±273 days). Five of nine patients responded with ≥PR. Whole genome and transcriptome sequencing was performed on sorted CD138+ MM cells at baseline, and whole exome and transcriptome sequencing was performed on five and three samples, respectively, following cevostamab discontinuation. Baseline FCRL5 expression levels did not correlate with clinical response. However, in one non-responder patient, FCRL5 expression increased four-fold after 28 days of treatment. Given the location of the FCRL5 gene on chromosome 1q—commonly amplified in advanced MM—we evaluated the presence of Gain(1q) in our study cohort. All patients whose best response was ≤PR (n=6) harbored a Gain(1q), whereas it was absent in those who achieved VGPR and sCR (n=3). Crucially, one of the patients that achieved sCR was initially treated with 60 mg of cevostamab for one year and then escalated to a higher dose cohort (90 mg) in the second year. At progression, a clonal FCRL5E115A missense mutation (allelic frequency = 56%) and a subclonal FCRL5G455Efs*19(allelic frequency = 5%) frameshift deletion mutation were detected, both absent at baseline. The missense mutation was located on the extracellular domain of FCRL5 targeted by cevostamab. To validate the functional impact of FCRL5E115Aand FCRL5G455Efs*19 on the sensitivity to cevostamab in vitro, we generated a FCRL5 knockout (KO) model utilizing CRISPR-Cas9 technology in OPM2 MM cells. In vitro cevostamab cytotoxicity assay with PBMCs and pan T cells from healthy donors at various effector to target (E:T) ratios confirmed a complete loss of cevostamab sensitivity in our FCRL5KO models. cDNA encoding patient-derived FCRL5 alterations (FCRL5E115Aand FCRL5G455Efs*19) as well as FCRL5wtwere cloned into mammalian expression vector and reintroduced into FCRL5KO cells. Cevostamab sensitivity was reinstated by FCRL5wt expression in KO cells. In contrast, FCRL5E115A conferred complete resistance to cevostamab (p&lt;0,001) and FCRL5G455Efs*19 depicted a significant decline in cevostamab sensitivity relative to FCRL5wt expressing models. Additional experiment utilizing direct stochastic optical reconstruction microscopy (dSTORM) are ongoing to monitor the membrane localization of FCRL5 mutants. By integrating genomic, transcriptomic and functional data with clinical outcomes, this work is the first to report on acquired FCRL5 mutations causing resistance to cevostamab. We hypothesize that the prolonged treatment duration may have contributed to the acquisition of the target mutations and note a subsequent durable sCR response to BCMA-targeted immunotherapy in the same patient, suggesting immune-related factors such as T cell exhaustion, were not drivers of cevostamab resistance. Due to the limited sample size, these findings require validation in larger cohorts. Nonetheless, our findings begin to elucidate tumor intrinsic mechanisms associated with cevostamab resistance.

Abstract PURPOSE Age over 40 years is commonly regarded as a “high risk” feature and potential indication for adjuvant chemoradiotherapy for patients with lower grade glioma in most clinical practice guidelines. However, these recommendations are based on retrospective observations and their subsequent incorporation into legacy clinical trials. It is unclear whether age remains a relevant prognostic factor in the molecular era. METHODS The Prospective Gliomas Research [PROGRES] Database contains individual patient-level data from 11 prospective clinical trials or observational registries, including seven international trials from the RTOG, EORTC, or Alliance cooperative groups. We determined the association of age with progression-free survival (PFS) in low grade or anaplastic glioma, stratified by isocitrate dehydrogenase (IDH1/2) status. We validated our findings in a separate multi-institutional retrospective cohort (Retrospective Glioma Research [REGRES] Database). RESULTS We identified 1,619 and 1,271 eligible patients in the PROGRES and REGRES cohorts, respectively. IDH-wildtype tumors were more common in patients &gt;40 years (38% vs, 5%, odds ratio: 11.3 [95%CI 6.5 – 19.7]). Age was associated with PFS in IDH-wildtype (5-year PFS: 6% vs, 24%, P =.003) but not in IDH-mutant glioma (60% vs, 59%, P = 0.22, P value of interaction &lt;.001). In IDH-wildtype tumors, older age predicted aggressive molecular features, including TERT promoter mutation (65% vs, 28%), EGFR amplification (41% vs, 15%), and chromosome +7/-10 alteration (57% vs, 25%). Age was not predictive of a benefit from chemoradiotherapy versus radiotherapy alone in IDH-mutant glioma. CONCLUSION Older age is a putative surrogate for IDH-wildtype status, but was not an independent predictor of progression in IDH-mutant lower grade glioma and did not predict benefit from chemoradiotherapy. In the absence of additional clinical or molecular risk factors, age alone should not be considered an indication for administration or deferral adjuvant treatment. Practice guidelines should be revised to reflect contemporary prognostic factors in the molecular era.

Abstract Background IDH-wild type (-wt) status is a prerequisite for the diagnosis of glioblastoma (GBM); however, IDH-wt gliomas with low-grade or anaplastic morphology have historically been excluded from GBM trials and may represent a distinct prognostic entity. While alkylating agent chemotherapy improves overall survival (OS) and progression-free survival (PFS) for IDH-wt GBM and also IDH-mutant gliomas, irrespective of grade, the benefit for IDH-wt diffuse histologic lower-grade gliomas is unclear. Methods We performed a meta-analysis of randomized clinical trials for World Health Organization (WHO) grades 2–3 gliomas (2009 to present) to determine the effect of alkylating chemotherapy on IDH-wt and -mutant gliomas using a random-effects model with inverse-variance pooling. Results We identified 6 trials with 1204 patients (430 IDH-wt, 774 IDH-mutant) that evaluated alkylating chemoradiotherapy versus radiotherapy alone, allowing us to perform an analysis focused on the value of adding alkylating chemotherapy to radiotherapy. For patients with IDH-wt tumors, alkylating chemotherapy added to radiotherapy was associated with improved PFS (HR:0.77 [95% CI: 0.62–0.97], P = .03) but not OS (HR:0.87 [95% CI: 0.64–1.18], P = .17). For patients with IDH-mutant tumors, alkylating chemotherapy added to radiotherapy improved both OS (HR:0.52 [95% CI: 0.42–0.64], P &lt; .001) and PFS (HR = 0.47 [95% CI: 0.39–0.57], P &lt; .001) compared to radiotherapy alone. The magnitude of benefit was similar for IDH-mutant gliomas with or without 1p19q-codeletion. Conclusions Alkylating chemotherapy reduces mortality by 48% and progression by 53% for patients with IDH-mutant gliomas. Optimal management of IDH-wt diffuse histologic lower-grade gliomas remains to be determined, as there is little evidence supporting an OS benefit from alkylating chemotherapy.

Abstract Intratumoral heterogeneity poses a significant challenge to the diagnosis and treatment of glioblastoma (GBM). This heterogeneity is further exacerbated during GBM recurrence, as treatment-induced reactive changes produce additional intratumoral heterogeneity that is ambiguous to differentiate on clinical imaging. There is an urgent need to develop non-invasive approaches to map the heterogeneous landscape of histopathological alterations throughout the entire lesion for each patient. We propose to predictively fuse Magnetic Resonance Imaging (MRI) with the underlying intratumoral heterogeneity in recurrent GBM using machine learning (ML) by leveraging image-localized biopsies with their associated locoregional MRI features. To this end, we develop BioNet, a biologically-informed neural network model, to predict regional distributions of three tissue-specific gene modules: proliferating tumor, reactive/inflammatory cells, and infiltrated brain tissue. BioNet offers valuable insights into the integration of multiple implicit and qualitative biological domain knowledge, which are challenging to describe in mathematical formulations. BioNet performs significantly better than a range of existing methods on cross-validation and blind test datasets. Voxel-level prediction maps of the gene modules by BioNet help reveal intratumoral heterogeneity, which can improve surgical targeting of confirmatory biopsies and evaluation of neuro-oncological treatment effectiveness. The non-invasive nature of the approach can potentially facilitate regular monitoring of the gene modules over time, and making timely therapeutic adjustment. These results also highlight the emerging role of ML in precision medicine.

Abstract Introduction Distinguishing between tumor and treatment effect in post-treatment glioma, although crucial for clinical management, is difficult because contrast-enhancing regions are mixtures of recurrent tumor and reactive tissue, and definitive histopathological criteria do not exist. This study disentangles the marked intra-tumoral heterogeneity in the treatment-recurrent setting by developing an unsupervised framework to algorithmically categorize intraoperative MRI-localized biopsies into three clinically-relevant tissue clusters based on joint analysis of RNA sequencing and histopathological data. Methods A retrospective cohort of 84 MRI-localized biopsies from 37 patients with post-treatment recurrent glioblastoma underwent mRNA extraction and quantification via PLATEseq protocol. For 48 of 84 biopsies, a neighboring piece of tissue underwent quantitative histopathology based on labeling index (LI) for SOX2, CD68, NeuN, Ki67, and H&E. Correlation between LIs and gene expression for these 48 samples was performed. Genes significantly correlated (p&lt;0.05) with ≥1 marker were used for hierarchical clustering of correlation matrix, identifying three mutually-exclusive tissue-specific gene sets. These sets were then used to perform ssGSEA to categorize each of 84 biopsies into one of three tissue types. Results Correlation analysis identified 7779 genes significantly correlated with ≥1 histopathological marker. Clustering revealed three gene sets associated with specific markers: SetA-3688 genes associated with SOX2/Ki67/H&E; SetB-2418 genes associated with CD68; SetC-1673 genes associated with NeuN. ssGSEA using these sets categorized each biopsy into one of three tissue types: 27 biopsies enriched in SetA, 28 in SetB, and 29 in SetC. Conclusions Using MRI-localized biopsies with both RNAseq and histopathological data, this algorithmic approach allows development of three orthogonal tissue-specific gene sets that can be applied to characterize the heterogeneity in post-treatment recurrent glioma: SetA: genes correlated with SOX2/Ki67/H&E, representing recurrent tumor; SetB: genes correlated with CD68 (microglia) representing reactive tissue consistent with treatment effect; SetC: genes correlated with NeuN (neurons), representing infiltrated brain.

Abstract Glioblastoma (GBM) is the one of the most aggressive and common primary brain malignancies, with a poor median overall survival of less than 15 months. While the immune system is activated and brain-resident microglia and blood-derived macrophages combat the tumor, the tumor can convert some microglia and macrophages to instead exhibit an immune-suppressive phenotype. These co-opted immune cells are thereby termed ‘glioma-associated microglia and macrophages’ (GAMMs), as they allow for continued tumor growth. However, limited clinical data has been collected to understand this phenomenon. As a result, we have collected spatially-distributed image-localized biopsies from a cohort of patients and performed RNA sequencing on each sample. Correlations between normalized RNA counts of genetic markers for macrophages (i.e., CD68, CD163), tumor populations (i.e., SOX2, OLIG2), and key cell functions (i.e., KI67, CASP3) were analyzed. To further investigate the temporal effects that GAMMs have on GBM growth, we proposed the Proliferation-Invasion-Macrophage (PIM) model. This system of partial differential equations incorporates the proliferative and invasive behavior of GBM, as well as populations for both ‘healthy’ and ‘glioma-associated’ macrophages. By exploring the parameter space, we classified the various dynamics of tumor progression and how they relate to the immune response. With further insights of the interactions between GBM and macrophage populations, we can begin to parameterize the model on a patient-specific basis and provide insights to personalized immunotherapies and other novel immune-targeted treatments.

Abstract Recurrent glioblastoma is challenging to distinguish from so called “treatment effect” on routine clinical imaging. Further, within tumor heterogeneity reveals that some regions can be histologically dominated by tumor progression whilst others can be dominated by secondary effects of treatment response. Apparent tumor progression on MRI can be very difficult to manage clinically as it is unclear the degree to which the imaging changes are actually tumor progression vs response to treatment (including inflammatory response and necrosis). In this analysis, we study a unique cohort of patients for whom image localized-biopsies reveal heterogeneity in response vs progression. Our dataset included 70 biopsy samples from 32 patients with GBM each histolopatholgically characterized for tumor abundance vs immune infiltrate. Six multiparametric MRI contrasts were available, including T1, T1gd, T2, FLAIR, SWI, and ADC. Images were co-registered. Radiomic (statistical + texture) features were extracted from the region of six image contrasts locally matched with each biopsy sample. Machine learning models were built to predict each biomarker using radiomic features. Leave-one-out cross validation was used to evaluate the prediction accuracy. Radiomic features were found to be informative to the prediction of biomarkers. ANOVA tests show significant improvement of using radiomic features compared with the null model. The prediction accuracy was higher when considering the biomarkers on a binary scale using the median as the cutoff than on a numerical scale. Spatially-informed radiomics models for tumor progression vs treatment effect are possible and can play an instrumental role in navigating confounding imaging changed common during treatment progression.