Abstract WP247: Enhanced Detection of Acute Ischemic Stroke with Low-Field MRI

Introduction: Portable, low-field (LF) magnetic resonance imaging (MRI) has the potential to expand access to neuroimaging in environments where conventional MRI is limited. However, diffusion-weighted imaging (DWI) at LF is challenged by a low signal-to-noise ratio and gradient strength, which may limit diagnostic confidence in acute ischemic stroke (AIS) evaluation, particularly for very small strokes. To address this gap, we evaluated a combination of novel LF-MRI hardware and pulse sequences to enhance lesion detection. Methods: Patients with a suspected diagnosis of AIS were prospectively enrolled at three US centers and underwent single-direction DWI and apparent diffusion coefficient (ADC) imaging on a 0.064 T MRI (v1, Hyperfine Inc.). A multi-direction DWI sequence (comprising 3 orthogonal diffusion direcions) was developed for comparison. Additionally, a next-generation 0.064 T scanner with optimized gradient amplifiers, form factor, and cooling system was evaluated (v2, Hyperfine Inc.). Diagnostic accuracy and the lower limit of lesion detection was calculated for both single-direction and multi-direction images compared to ground-truth MRI (1.5-3 T). Results: A total of 95 patients ( n =62 confirmed ischemic stroke; n =33 stroke mimics) were included. On single-direction images, agreement between assessors regarding lesion detection was κ=0.72, and κ=0.84 on multi-direction images. The positive predictive value for differentiating AIS from stroke mimics was 78.2% on single-direction images and 95% for multi-direction trace maps. On single-direction images, a lesion volume cut-point of 0.6 mL yielded a sensitivity of 89% and specificity of 88%. For multi-direction trace maps, the lesion volume cut-point was 0.4 mL, with a corresponding sensitivity of 86% and specificity of 83%. Multi-direction imaging on the next generation v2 hardware improved image uniformity (p