In this retrospective cohort of 162 patients with acute isolated pontine infarction, BAD and LAA imaging phenotypes (relative to SAD) were independently associated with early neurological deterioration, TyG was not, and the imaging model's discrimination advantage over the metabolic/coagulation model was not statistically established; the full model and nomogram need external validation.
Early neurological deterioration occurred in about 30% of patients with acute isolated pontine infarction.
The BAD imaging phenotype was very strongly associated with END relative to SAD.
The LAA imaging phenotype was independently associated with END relative to SAD.
Thrombin time was associated with END, but this association is described as exploratory.
TyG was not independently associated with END, and diabetes showed only borderline evidence of association.
Adding imaging phenotype to a base model improved discrimination more than adding metabolic/coagulation markers did, but the difference between the two was not statistically significant.
The full model showed good discrimination after bootstrap validation, with reasonable calibration.
The authors conclude that the prediction model and nomogram require external validation before clinical use.
This study looked back at the records of 162 patients who had a stroke in one specific part of the brainstem, the pons. About 3 in 10 of them got worse neurologically within the first week in hospital. The researchers wanted to know whether the way the stroke looked on brain imaging, or blood markers related to metabolism and clotting, were more closely linked to this early worsening. Patients whose scans showed a particular pattern involving blockage of a branch of the basilar artery (BAD) or large-artery disease (LAA) were much more likely to get worse than those with small-vessel disease (SAD). The link was especially strong for BAD. A longer thrombin time, a clotting test, was also linked to worsening, though the authors consider this finding exploratory. A blood sugar and fat measure called TyG was not independently linked to worsening, and diabetes showed only a borderline link. A prediction model that combined all the factors performed fairly well, even after internal testing. However, the study could not show statistically that imaging predicted worsening better than the metabolic and clotting markers did. The study was retrospective, small, and from a single group of patients. This research suggests that imaging pattern may help identify patients at higher risk, but the model needs testing in other groups of patients before it could be used in clinical practice.
Chen Y, Li X, Wang H, Zhou G. (2026). Imaging phenotype is more strongly associated with early neurological deterioration after acute isolated pontine infarction: a retrospective cohort study.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1913826