Development of a Molecular Diagnostic to Identify Dangerous Intracranial Aneurysms

Project Summary Intracranial aneurysm (IA) rupture is the primary cause of non-traumatic subarachnoid hemorrhage, a catastrophic event that carries high rates of mortality (>50%) and permanent disability (>50% among survivors). When an unruptured IA is discovered, clinicians face the difficult decision of whether or not it should be treated. Since rupture rates are low (risk of rupture ~ 0.5-1% per year), while complication risks of intervention can be significant, it is critical to stratify rupture risk such that the dangerous IAs are treated right away while less dangerous ones can be periodically monitored. In clinical practice, aneurysm size has been the principal criterion for treatment, but it is not always reliable. Recently, other clinical metrics, such as the PHASES (Population, Hypertension, Age, Size of IA, Earlier subarachnoid hemorrhage, and Site of IA) score, which use IA characteristics and patient demographics, have been developed and validated in attempts to better stratify risk. However, like IA size, such risk metrics can only be accurately assessed after the patient has received digital subtraction angiography (DSA), which is invasive, expensive ($5,000-8,500 per scan), and risky for transient or permanent complications. Furthermore, these metrics lack information about biological differences underlying IAs that may better discriminate high- vs low-risk cases. A blood-based diagnostic of dangerous IAs would enable more informed IA management and could offer a low-cost, non-invasive way to monitor patients during watchful waiting or after treatment (between follow-up imaging). Over the past several years, Neurovascular Diagnostics has been developing a blood-based IA diagnostic called AneuScreenTM. During development, we found that expression differences of certain genes also stratified IA cases by IA size in a “dose-dependent” manner, leading us to hypothesize that patients with high-risk IAs have distinguishable gene expression patterns in their blood. In a successfully completed Phase I SBIR project, we used whole blood transcriptomes from a modestly-sized dataset of IA patients to develop and test machine learning classifiers of high-risk aneurysm cases (delineated by PHASES score). This model had 88% testing accuracy, 78% sensitivity, and 95% specificity. Despite these exciting results, further work is needed to translate this biomarker into a diagnostic test, which is the focus of this Phase II SBIR project. In Aim 1, we will validate the genes within the IA risk biomarker in a large dataset of previously-collected blood samples that will be subjected to RNA sequencing. In Aim 2, we will standardize the assessment of the biomarker on an established clinical platform that utilizes qPCR and capillary electrophoresis for expression readout. The models will be re- trained to perform well using this new output data-type. Lastly, in Aim 3, we will test the developed assay, which we call AneuScreenTM+(or AneuScreenTM Plus), in a prospective cohort of n=400 patients with IA and n=200 controls. Successful completion of this Phase II will determine if our prototype IA diagnostic can expand its capabilities and increase its value by stratifying IA risk, in addition to identifying IA.