Automated MRI perfusion-diffusion mismatch estimation may be significantly different in individual patients when...

Area of Science:

• Neuroradiology
• Medical Imaging Analysis
• Stroke Imaging

Background:

• Accurate assessment of brain tissue viability is critical in acute ischemic stroke management.
• Perfusion-diffusion mismatch quantifies the difference between hypoperfused brain tissue and infarct core, guiding treatment decisions.
• Established software applications are used for automated analysis of MRI data in stroke patients.

Purpose of the Study:

• To compare the performance of two established software applications in calculating apparent diffusion coefficient (ADC) lesion volumes.
• To evaluate differences in the calculated volumes of critically hypoperfused brain tissue and perfusion-diffusion mismatch.
• To assess the impact of software variability on clinical decision-making for mechanical thrombectomy.

Main Methods:

• Analysis of brain MRI scans from 81 patients with acute ischemic stroke due to large vessel occlusion.
• Automated calculation of hypoperfused tissue volume, ADC volume, and perfusion-diffusion mismatch using two distinct software packages.
• Quantitative statistical comparison of parameters derived from the different software applications.

Main Results:

• Significant differences were observed in calculated volumes of hypoperfused tissue (median 91.0 vs. 102.2 ml) and ADC lesion volume (median 30.0 vs. 23.9 ml) between the software packages (p < 0.05).
• Perfusion-diffusion mismatch volumes also differed significantly (median 47.0 vs. 67.2 ml; p < 0.05).
• Mean absolute differences per patient were 20.5 ml for hypoperfused tissue, 10.8 ml for ADC volumes, and 27.6 ml for mismatch volumes; applying the DEFUSE 3 threshold could lead to dissenting treatment decisions in 7.4% of cases.

Conclusions:

• Volume segmentation using different software products can yield significantly different results at the individual patient level.
• These discrepancies in infarct core and mismatch volumes may critically influence the decision for or against mechanical thrombectomy.
• Automated calculation of perfusion-diffusion mismatch aids in applying trial-derived inclusion/exclusion criteria, but software variability necessitates careful consideration.