Computed tomography–based acute stroke lesion timing and patient stratification
Minnerup and coworkers proposed a computed tomography (CT)‐based procedure to determine age of acute/hyperacute stroke to aid patient stratification for reperfusion therapy.1 The CT procedure converts Hounsfield units (HU) determined in the stroke lesion into net water uptake due to ischemia. There are issues in the article that deserve attention. First, Minnerup et al1 determined an average increase of brain water by 4.8% in ischemic core within 4.5 hours (mean time = 2.1 hours), rising to 11.5% at 4.5 hours. This value is very high in the light of directly measured water uptake at 4 hours in primate (1.6–2.4%),2 feline (1.2%),3 or rodent (1.0%)4 models of stroke. In feline brain, water uptake exceeding 5% was reported only by 24 to 48 hours of stroke.3 Furthermore, data from primate stroke show that water uptake by white matter (WM) is less severe than that by gray matter (GM).2 Taking into account the water data above, and that the ischemic core is effectively without blood supply, it is difficult to appreciate the water uptake values reported by Minnerup et al.1 One potential reason for high water uptake estimates by Minnerup et al1 could be the use of aqueous HU calibration instead of measurements of HU against water content in the brain tissue.4
Second, the procedure utilizes a mirror‐reference approach, where the ratio of HU in the ischemic core to the contralateral nonischemic mirror reference is used to compute percentage water uptake. The mirror‐reference approach assumes that any deviation from a ratio of 1 is due to pathology (ie, acute ischemia), and that the ratio increases with time. This places strict requirements for accurate positioning of the mirror region5 so that it contains volumes of brain tissue types (GM, WM, and cerebrospinal fluid space) equal to those of the lesion, because inherent water content in these tissue types greatly differs.
The article by Minnerup et al1 underscores the potential value of quantitative imaging data in stroke patient management despite the water uptake calibration issue above. This as such does not affect the potential clinical utility of the CT approach. However, robustness of the presented CT approach for wide clinical use should be demonstrated in cohorts balanced in size with respect to the 4.5‐hour time limit.