Moving Beyond Total Lesion Volume: An Integrative Neuroimaging and Statistical Approach for Evaluating Neuroprotection During Surgical Aortic Valve Replacement (SAVR)

• Activity Number: 25 - Medical Devices and Diagnostics Speed Session
• Type: Contributed
• Date/Time: Sunday, August 8, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Section on Medical Devices and Diagnostics
• Author(s): Lewis E Tomalin* and Mayte Suárez-Fariñas and Guray Erus and Jessica Overbey and Amy Kuceyeski and Steven R Messe and Emilia Bagiella and Alan J Moskowitz and Annetine C Gelijns and Alex Iribarne and Michael A Acker and Michael J Mack
• Companies: Icahn School of Medicine at Mount Sinai and Icahn School of Medicine at Mount Sinai and University of Pennsylvania and Icahn School of Medicine at Mount Sinai and Cornell and University of Pennsylvania and Icahn School of Medicine at Mount Sinai and Icahn School of Medicine at Mount Sinai and Icahn School of Medicine at Mount Sinai and Dartmouth-Hitchcock Medical Center and University of Pennsylvania and Baylor Research Institute
• Keywords: zero-inflated modeling; SAVR; Delirium; DWI-MRI; connectome; cerebellum

Abstract

Incidence of CNS infarcts remains a serious complication of surgical aortic valve replacement (SAVR). Although most CNS infarcts do not lead to clinical stroke, they are associated with neurological complications such as post-operative delirium (PoD). Identifying image-based brain biomarkers associated with PoD is an important step towards identifying individuals with high PoD risk. We analyzed MRI-derived lesionality and connectome data, as well as cognitive outcomes, from SAVR patients (n=308). This data was originally collected as part of the clinical trial ‘Neuroprotection in Patients Undergoing Aortic Valve Replacement’ (NCT02389894, PMID:28787505). Due to the zero-inflated nature of the data, we utilized a zero-inflated gaussian modeling approach (PMID:33166356), originally developed for micro-biome analysis, to identify region-specific peri-operative increases in lesionality/connectivity that were associated with developing PoD, most notably in Cerebellar regions. This analysis paves the way towards identifying MRI-derived image-based biomarkers for peri-operative neurological complications and highlights the utility of zero-inflated modeling approaches.