In-silico trial of intracranial flow diverters replicates and expands insights from conventional clinical trials

Abstract

Although the cost of clinical trials is ever-increasing, in-silico trials, which rely on virtual populations and interventions simulated using patient-specificc models, may offer a solution to contain these costs. However, in-silico trial endpoints need to be compared to those available from conventional clinical trials to ensure that the predictions of safety or effcacy from the in-silico approach are valid. Here, we present the flow diverter performance assessment (FDPASS) in-silico trial, which modelled the treatment of intracranial aneurysms in 82 virtual patients with a flow-diverting stent, using computational fluid dynamics (CFD) to quantify post-treatment flow reduction in the aneurysm sac. The predicted FD-PASS flow-diversion success rate replicated the values previously reported in three reference clinical trials. The in-silico approach allowed broader investigation of factors associated with insuficient flow reduction and increased stroke risk after flow diversion than would be feasible in a conventional trial. These ndings demonstrate for the rst time that in-silico trials of medical devices can (i) replicate ndings of conventional clinical trials and (ii) incorporate virtual experiments that are impossible in conventional trials.

Hemodynamics of thrombus formation in intracranial aneurysms: an in-silico observational study

“…These four groups were very similar in terms of patient age and sex, but the size and AR of the thrombosed groups were larger than that of the non-thrombosed groups. 62.5% (15/24) of the thrombosed group were high AR ( >1.6) 36 IAs. In contrast, high AR cases only accounted for 24.7% (21/85) of the non-thrombosed group.…”

“…However, in-vivo or image-based population-based analysis of thrombosis haemodynamics in realistic anatomies and physiologies is currently very difficult, if not impossible. Computational modeling has proved a powerful tool in predicting thrombosis in aneurysms before and after treatment, and thus in patient-specific treatment planning or in-silico trials 35,36 . But such approach remains dependent on hemodynamic thresholds usually chosen based on ranges derived from limited highly-controlled in-vitro experiments, not accurate representatives of real anatomy and physiology.…”

“…Hypertension is a well-known risk factor of unruptured IAs 42 , and more than 75% of saccular IA patients have hypertension 17 . The flow diverter performance assessment (FD-PASS) in-silico trial showed that hypertension may cause less effective flow diversion 36 , while the IntrePED study reported an association between hypertension and ischaemic stroke in flow-diverted aneurysms 5 . We modeled hypertension as boundary conditions from a cerebral autoregulation system (CARS) 19 .…”

Comparison of stent retriever thrombectomy using 3‐dimensional patient‐specific models of intracranial circulation with actual middle cerebral artery occlusion thrombectomy cases

“…Contrary to our findings, the study by Bernava et al 21 tual experiments in subgroup that are particularly challenging for conventional clinical trial design. 22 Stroke thrombectomy trials could be another promising target for the in vitro testing approach because these trails are particularity challenging to conduct given the acute nature of the disease, large number of variables that could affect procedural outcomes, and the consequences of suboptimal device performance on clinical outcomes. 23 Our findings support this overarching hypothesis but also outline a number of limitations that the current models present.…”

“…Recent work by Sarrami‐Foroushani et al. demonstrated the feasibility of the in vitro approach in replicating findings of clinical trials and conducting virtual experiments in subgroup that are particularly challenging for conventional clinical trial design 22 . Stroke thrombectomy trials could be another promising target for the in vitro testing approach because these trails are particularity challenging to conduct given the acute nature of the disease, large number of variables that could affect procedural outcomes, and the consequences of suboptimal device performance on clinical outcomes 23 .…”