Objective: The Alberta Stroke Program Early CT Score (ASPECTS) is widely used to assess and diagnose acute ischemic stroke (AIS) patients. However, reliability of ASPECTS scoring is poor among physicians with limited expertise. We hypothesize that reliability for ASPECTS scoring can be improved by using algorithm enhanced gray-white matter (AEGWM) NCCT.
Methods: Inter-rater reliability for ASPECTS scoring was assessed between plain and AEGWM NCCT by an expert and a novice reader. 50 AIS patient NCCT images were acquired acutely. NCCT images were then enhanced by skull stripping, image smoothing, histogram equalization, and JET color mapping. ASPECTS scoring was done on AEGWM color mapped NCCT and on standard 5mm NCCT images by a novice reader. ASPECTS scoring was done two days apart on standard NCCT first, followed by AEGWM NCCT. Expert (neuro-radiologist) scores on standard NCCT were then compared with scores from the novice reader on both regular and AEGWM NCCT.
Results: Agreement between novice and expert for trichotomized ASPECTS (0-4, 5-7, 8-10) was best when the novice was reading ASPECTS on AEGWM NCCT (kappa=0.7093) vs. when novice read ASPECTS on standard NCCT (kappa=0.2409). Difference in scoring the full 10-point ASPECTS score was least when the novice read ASPECTS on AEGWM NCCT (mean ASPECTS difference between novice and expert for algorithm-enhanced NCCT 0.68 ± 1.1 vs. 1.48 ± 1.8 for standard NCCT). A Bland Altman plot comparing the difference is attached and the coefficient of variation was found to be 0.14 for AEGWM NCCT scoring.
Conclusion: Algorithm Enhanced Gray-White Matter NCCT allows more accurate/ reliable ASPECTS scoring. Further evaluation on a larger dataset with readers with different levels of expertise is ongoing.