Photo Credit: Rasi Bhadramani

The following is a summary of “Dynamic Evolution of Infarct Volumes at MRI in Ischemic Stroke Due to Large Vessel Occlusion,” published in the May 2024 issue of Neurology by Munsch et al.

Researchers conducted a retrospective study to identify patterns of stroke tissue growth (spatiotemporal evolution) in patients with large vessel occlusion (LVO) and assess whether the patterns can predict patient outcomes.

They collected databases (FRAME, 2017–2019; ETIS, 2015–2022), selecting acute MRI data from anterior patients with LVO stroke with witnessed onset. Splitting the data into training and independent validation datasets, patients were classified in the training set into three growth groups based on infarct growth rate (diffusion volume/time-to-imaging) using Gaussian mixture analysis. Pseudo-longitudinal models of infarct growth were created, generating frequency maps. Growth groups were assigned to patients in the validation dataset, and their 3-month mRS was compared with predicted values from a regression model in the training dataset.

The results included 804 patients (median age 73.0 years [IQR 61.2–82.0 years]; 409 men). In the training dataset, unsupervised clustering revealed 11% (74/703) slow, 62% (437/703) intermediate, and 27% (192/703) fast progressors. Infarct volume evolutions correlated best with linear (r=0.809; P<0.001), cubic (r = 0.471; P<0.001), and power (r=0.63; P<0.001) functions for the slow, intermediate, and fast progressors, respectively. The deep nuclei and insular cortex were rapidly affected in the intermediate and fast groups, with further cortical involvement in the fast group. The growth group significantly predicted the 3-month modified Rankin scale (multivariate OR 0.51; 95% CI 0.37–0.72, P<0.0001) in the training dataset, with a mean area under the receiver operating characteristic curve of 0.78 (95% CI 0.66–0.88) in the independent validation dataset.

Investigators concluded that stroke caused by LVO followed three distinct growth patterns (slow, intermediate, and fast), which can predict patient outcomes and may inform the design of future neuroprotective treatments.

Source: neurology.org/doi/10.1212/WNL.0000000000209427