Date of Award

January 2024

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)



First Advisor

Alexandra Lansky


Stroke is a feared complication of transcatheter aortic valve replacement (TAVR), affecting 2–8% of patients at 30 days. Beyond clinically evident stroke, covert brain injury (CBI) defined as clinically silent evidence of brain injury on imaging is strikingly prevalent: 70–100% of patients have evidence of new infarcts on diffusion-weighted magnetic resonance imaging (DW-MRI). The clinical significance of these lesions is unknown. Recent guidance has highlighted the importance of imaging in the assessment of cerebral embolic protection (CEP) devices that aim to counter the debris generated during TAVR. This work aims to a) determine the clinical significance of new ischemic lesions, b) propose metrics for ischemic lesion burden, and c) identify factors associated with stroke and CBI, including their association with surgical risk as defined by the Society of Thoracic Surgeons Predicted Risk of Mortality STS score.Patient-level data were pooled from 4 prospective multicenter TAVR studies (DEFLECT III [N=87], NeuroTAVR [N=44], REFLECT I [N=258], and REFLECT II [N=214]). All studies shared a common independent imaging core laboratory and clinical event adjudication committee. DW-MRI were assessed for total lesion number (TLN) per subject, individual lesion volumes (ILV), and total lesion volume (TLV). Receiver operating characteristic (ROC) analysis was performed to identify the optimal DW-MRI measure and thresholds to discriminate ischemic stroke at 30 days. A total of 495 of 603 patients undergoing TAVR completed DW-MRI, with 97% clinical follow-up at 30 days. At 30 days, the rates of death, ischemic stroke, and disabling stroke were 0.8%, 6.9%, and 3.1%. New ischemic lesions were observed in 85% of patients, with a mean TLN of 5.5±7.3 per patient, a mean ILV of 78.2±257.1 mm3, and a mean TLV of 555±1039 mm3. The area under the ROC curve was 0.84 for TLV with an optimal cut point of 440 mm3 (Youden criteria) to 547 mm3 (distance 0,1 criteria), 0.82 for maximum ILV (cut point 216 mm3 by both criteria), and 0.81 for TLN (cut point 4–5 lesions). Compared with patients with a TLV ≤500 mm3, patients with TLV >500mm3 had more ischemic stroke at 30 days (18.2% vs 2.3%, p<0.001), more disabling strokes (8.8 vs 0.9%, p<0.001), and less complete stroke recovery (44 vs 62.5%, p=0.001). Stroke was independently associated with older age, self-expanding valve use, and worse baseline MoCA and mRS scores. While low (<4%), intermediate (4–8%), and high (>8%) STS PROM predicted mortality at 30 days, stroke rates did not differ across the STS risk groups. This patient-level pooled analysis is the first to demonstrate that acute brain injury measures on DW-MRI can discriminate clinical ischemic stroke and worse recovery in patients undergoing TAVR. A TLV threshold of >500 mm3 had excellent discrimination when categorizing patients with ischemic and disabling stroke. Thresholds of TLN >5 and maximum ILV of >200 mm3 also performed strongly. Our study provides new measures to better predict clinical outcomes of patients undergoing TAVR that will enable better initial evaluation of the efficacy of preventive CEP devices in future trials.


This is an Open Access Thesis.

Open Access

This Article is Open Access