Annette T. Byrne

We introduce a comprehensive multi-step machine-learning driven pipeline which fuses multi-modal omics datasets and clinical outcomes with survival and treatment response to predict patient outcome following anti-angiogenic therapy in the metastatic colorectal cancer (mCRC) setting. The approach encompasses the following steps: a) employing a sparse Bayesian factor analysis method (PhenMap) to select significant mapping-variables (MVs) and associated biomarkers (features) through joint multivariable modelling of copy number aberrations (CNA) and clinical covariates, including mutations and clinical demographics/outcomes; b) utilizing Cox-proportional hazard analysis on the MVs to select those associated with progression-free survival (PFS); and c) employing elastic net Cox regression analysis on the prognostic features selected by PhenMap to delineate risk groups associated with bevacizumab (BVZ) response. Through this approach, we have identified three putative features (CNA—15q21.1 and 1p36.31 deletions and BRAF mutation) from two prognostically-significant MVs to stratify N = 117 mCRC patients, who received BVZ combination therapy, into 3 (low, medium, high) prognostic risk groups that were significantly associated with PFS and treatment response. Mortality risk was significantly greater in the high-risk group with 100% (n = 12) of patients showing no response to BVZ, compared to the low-risk group where 10 out of 12 patients (88%) showed a response to BVZ. The risk groups were independently negative predictors of survival in BVZ-treated mCRC patients. Overall, Wwe have established a machine learning pipeline integrating multi-modal omics data with response, and have implicated a putative combined CNA/mutation candidate biomarker with associated risk scores that could, in the future, help stratify mCRC patients unlikely to benefit from BVZ combination therapy. Our novel precision medicine approach applies disruptive advancements in artificial intelligence and bioinformatics methodologies to tumour biology datasets.

Objectives Bevacizumab was approved for first-line treatment of metastatic colorectal cancer (mCRC) in 2004. However, adding bevacizumab to treatment consistently fails to be cost-effective owing to modest response rates. Recently, the European Commission (EC) funded ANGIOPREDICT consortium (www.angiopredict.com) identified a link between bevacizumab treatment response and intermediate-to-high chromosomal instability (CIN) in mCRC. Thus, the objective of the current study was to compare the cost-effectiveness of adding bevacizumab with first-line chemotherapy in the bevacizumab responsive CIN subtype across three European countries (Germany, Ireland and Spain) with varying costs of care and reimbursement policies. Methods We developed an open-source health economic model to estimate cost-effectiveness. The ANGIOPREDICT cohort informed progression risks and cause-specific mortality. Health utilities and adverse events probabilities were obtained from the literature. Costs were derived from surveys of collaborating consortium hospitals in Germany, Ireland, and Spain that participated in the recently completed EC funded COLOSSUS translational study (ANGIOPREDICT successor initiative) and the literature. Sensitivity analyses included individual and simultaneous variation of input parameters from a priori defined distributions. Results Bevacizumab was not cost effective even at willingness–to-pay (WTP) thresholds that are appreciably higher than those considered realistic. The highest incremental cost-effectiveness ratio (ICER) was in Germany at €241,188 per quality-adjusted life year (QALY), while the lowest was in Ireland at €180,477 per QALY. All deterministic and probabilistic sensitivity analyses demonstrated that these results were robust. Conclusions Even for patients with mCRC manifesting improved outcomes, adding bevacizumab to first-line chemotherapy is invariably not cost-effective in any of the countries examined. Variability in pricing, healthcare costs and WTP thresholds across countries did not commute this result.

Background: Immunotherapies have improved outcomes in many cancers but show limited efficacy in glioblastoma (GBM). This study aimed to determine whether immunotherapy could be tailored to GBM by functional subtyping of vascular-immune landscapes. Methods: We employed single-cell RNA sequencing, multiplex immunohistochemistry to characterize three distinct TIME subtypes in human and murine GBMs. We evaluated responses to combination of anti-angiogenic immunomodulating therapies (CD40 Ag, anti-PDL1, PI3Kγ/δ inhibition) in orthotopic syngeneic GBM mouse models. Results: We identified three distinct functional TIME subtypes with unique vascular-immune landscapes in human and murine GBM: TIME-low (immune-low/deserted, leaky vasculature), TIME-med (intermediate immune-infiltration, angiogenic), and TIME-high (heavily infiltrated with immunosuppressive myeloid cells and dysfunctional T cells). Representative mouse models of TIME-GBMs responded in subtype specific ways to anti-angiogenic immunomodulating therapies. TIME-low GBMs enhanced T-cell activity but relapsed due to emerging myeloid immunosuppression, concomitant with mesenchymal transition. TIME-med displayed the most immune-activated, yet angiogenic phenotype, and showed overall good responses to various anti-angiogenic immunomodulating therapies. TIME-high GBMs were mostly non-responsive but improved when the myeloid-cell PI3Kγ was targeted. However, CD40 agonist treatment, expected to enhance APC function, unexpectedly worsened survival by promoting angiogenesis and heightening immunosuppression, leading to dysfunctional T cells and reduced NK cell recruitment, and subsequent enhanced tumor propagation. Conclusions: Our study reveals three GBM TIME subtypes with distinct vascular-immune landscapes that require tailored therapies. TIME-med tumors are predicted to respond best to immunotherapies, TIME-low tumors show transient effects with anti-angiogenic immunomodulating therapies, while TIME-high tumors, due to their profound immunosuppression, can even have worse outcomes. Key points: Three TIME subtypes were identified in GBM with distinct vascular-immune landscapes TIME subtypes show divergent immunotherapy responses TIME classification supports personalized treatment strategy for GBM immunotherapy Importance of the Study: This study advances glioblastoma immunotherapy by providing the first comprehensive single-cell characterization of TIME subtypes, moving beyond bulk RNA-sequencing to reveal detailed functional states of immune cells. We establish clinically relevant murine models that recapitulate human TIME subtypes, enabling preclinical testing of TIME-targeted therapies. Our findings identify TIME-low GBM as immune deserted and TIME-med tumors as the most immunotherapy-responsive subtype that should be prioritized for clinical selection. We found that high immune infiltration correlates with non-responsiveness and even unexpected detrimental effects with CD40 agonist treatment in TIME-high tumors—critical information given ongoing clinical trials. Identifying distinct immunosuppressive mechanisms across TIME subtypes and differential treatment responses provides a framework for personalized immunotherapy selection. The immediate translational impact of this work highlights the importance of TIME classification for treatment stratification and the urgent need to consider TIME status in clinical trial design, potentially explaining variable patient responses in previous GBM immunotherapy trials.

To predict outcome to combination bevacizumab (BVZ) therapy, we employed cell-free DNA (cfDNA) to determine chromosomal instability (CIN), nucleosome footprints (NF) and methylation profiles in metastatic colorectal cancer (mCRC) patients. Low-coverage whole-genome sequencing (LC-WGS) was performed on matched tumor and plasma samples, collected from 74 mCRC patients from the AC-ANGIOPREDICT Phase II trial (NCT01822444), and analysed for CIN and NFs. A validation cohort of plasma samples from the University Medical Center Mannheim (UMM) was similarly profiled. 61 AC-ANGIOPREDICT plasma samples collected before and following BVZ treatment were selected for targeted methylation sequencing. Using cfDNA CIN profiles, AC-ANGIOPREDICT samples were subtyped with 92.3% accuracy into low and high CIN clusters, with good concordance observed between matched plasma and tumor. Improved survival was observed in CIN-high patients. Plasma-based CIN clustering was validated in the UMM cohort. Methylation profiling identified differences in CIN-low vs. CIN high (AUC = 0.87). Moreover, significant methylation score decreases following BVZ was associated with improved outcome (p = 0.013). Analysis of CIN, NFs and methylation profiles from cfDNA in plasma samples facilitates stratification into CIN clusters which inform patient response to treatment.

Here, we establish a CT-radiomics based method for application in invasive, orthotopic rodent brain tumour models. Twenty four NOD/SCID mice were implanted with U87R-Luc2 GBM cells and longitudinally imaged via contrast enhanced (CE-CT) imaging. Pyradiomics was employed to extract CT-radiomic features from the tumour-implanted hemisphere and non-tumour-implanted hemisphere of acquired CT-scans. Inter-correlated features were removed (Spearman correlation > 0.85) and remaining features underwent predictive analysis (recursive feature elimination or Boruta algorithm). An area under the curve of the receiver operating characteristic curve was implemented to evaluate radiomic features for their capacity to predict defined outcomes. Firstly, we identified a subset of radiomic features which distinguish the tumour-implanted hemisphere and non- tumour-implanted hemisphere (i.e, tumour presence from normal tissue). Secondly, we successfully translate preclinical CT-radiomic pipelines to GBM patient CT scans (n = 10), identifying similar trends in tumour-specific feature intensities (E.g. ‘glszm Zone Entropy’), thereby suggesting a mouse-to-human species conservation (a conservation of radiomic features across species). Thirdly, comparison of features across timepoints identify features which support preclinical tumour detection earlier than is possible by visual assessment of CT scans. This work establishes robust, preclinical CT-radiomic pipelines and describes the application of CE-CT for in-depth orthotopic brain tumour monitoring. Overall we provide evidence for the role of pre-clinical ‘discovery’ radiomics in the neuro-oncology space.

Chemotherapy using temozolomide is the standard treatment for patients with glioblastoma. Despite treatment, prognosis is still poor largely due to the emergence of temozolomide resistance. This resistance is closely linked to the widely recognized inter- and intra-tumoral heterogeneity in glioblastoma, although the underlying mechanisms are not yet fully understood. To induce temozolomide resistance, we subjected 21 patient-derived glioblastoma cell cultures to Temozolomide treatment for a period of up to 90 days. Prior to treatment, the cells’ molecular characteristics were analyzed using bulk RNA sequencing. Additionally, we performed single-cell RNA sequencing on four of the cell cultures to track the evolution of temozolomide resistance. The induced temozolomide resistance was associated with two distinct phenotypic behaviors, classified as “adaptive” (ADA) or “non-adaptive” (N-ADA) to temozolomide. The ADA phenotype displayed neurodevelopmental and metabolic gene signatures, whereas the N-ADA phenotype expressed genes related to cell cycle regulation, DNA repair, and protein synthesis. Single-cell RNA sequencing revealed that in ADA cell cultures, one or more subpopulations emerged as dominant in the resistant samples, whereas N-ADA cell cultures remained relatively stable. The adaptability and heterogeneity of glioblastoma cells play pivotal roles in temozolomide treatment and contribute to the tumor’s ability to survive. Depending on the tumor’s adaptability potential, subpopulations with acquired resistance mechanisms may arise.