Diffusion-Weighted MRI Stroke Volume Following Recanalization Treatment is Threshold-Dependent.

Abstract

Infarct lesion segmentation has been problematic as there are awide range of relative and absolute diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) thresholds that have been used for this purpose. We examined differences of stroke lesion volume and evolution evaluated by magnetic resonance imaging (MRI) during the immediate post-treatment phase (<5 h) and at 24 h. In this study 33acute ischemic stroke patients were imaged with MRI <5 h and 24 h post-reperfusion treatment. Lesion volumes were segmented on ADC maps and average DWI using literature cited absolute ADC and relative DWI thresholds. The segmented lesion volumes within both time points were compared and the absolute change in lesion volume (infarct growth) between the two time points was calculated and compared using Bland-Altman analysis. Lesion volumes differed significantly when different relative DWI or absolute ADC thresholds were used (p< 0.05), which held true for baseline as well as follow-up lesions. The median absolute changes in lesion volume from baseline to follow-up for ADC thresholds of 550× 10-6 mm2/s, 600× 10-6 mm2/s, 630× 10-6 mm2/s and 650× 10-6 mm2/s were 3.5 ml, 4.2 ml, 4.5 ml, and 6.5 ml, respectively (p< 0.05). Likewise, the median absolute changes in lesion volume from baseline to follow-up for DWI thresholds, k= 0.85, 1.28, 1.64, 1.96, and 2.7 were 10.1ml, 7.3 ml, 5.7 ml, 5.4 ml and 4.2 ml, respectively (p< 0.05). Absolute lesion volumes and changes in lesion volumes (infarct growth) measured after recanalization treatment were dependent on absolute ADC and relative DWI thresholds, which may have clinical significance. Standardization of techniques for measuring DWI lesion volumes requires immediate attention.