Predictors of congruency between clinical and radiographic Oxfordshire Community Stroke Project Classification subtypes.

BACKGROUND

The Oxfordshire Community Stroke Project (OCSP) classification has been widely used to assess ischemic stroke patients based on clinical characteristics alone. However, the correlation between the clinical presentation evaluated using OCSP and imaging findings is yet to be determined. Our study aimed to describe the baseline characteristics of the OCSP subtypes of patients admitted with ischemic stroke and evaluate the predictors of the relationship between clinical and neuroimaging findings.

METHODS

Patients with a confirmed diagnosis of ischemic stroke admitted to a comprehensive stroke center in Brazil between February 2015 and October 2017 were eligible for the study. All participants underwent computed tomography (CT) at admission and follow-up neuroimaging within seven days, per the institutional protocol. Trained staff classified patients according to the OCSP at hospital admission. The radiographic OCSP classification was retrospectively assessed based on the last follow-up neuroimaging by investigators unaware of the clinical classification.

RESULTS

The overall agreement rates ranged from 65.5% to 88.7%. Lower NIHSS scores, absent hyperdense MCA sign, higher ASPECTS, and absent brainstem symptoms were related to a higher risk of misclassification. Treatment with intravenous tPA was associated with reclassification in patients with total anterior circulation syndrome. For predicting radiographic posterior circulation involvement, vertigo (OR 2.9, 95% CI 1.7-5.1, p < 0.001) and brainstem symptoms (OR 35, 95% CI 20.5-60.2, p < 0.001) were directly associated with correct classification, but motor and higher cerebral function were not correlated.

CONCLUSION

The clinical OCSP classification presented good congruency rates with the neuroimaging findings. However, patients with lacunar syndromes are often misclassified when radiological criteria are considered.