Retrospective Assessment of Radiological Patterns in Stroke Patients Using CT and MRI

Abstract

Background: Stroke is a leading cause of morbidity and mortality worldwide, with ischemic and hemorrhagic subtypes requiring fundamentally different management strategies. Rapid and accurate differentiation between these subtypes is essential for initiating appropriate therapy, yet diagnostic performance may vary depending on imaging modality and healthcare setting, particularly in low- and middle-income countries where resource limitations impact access to advanced imaging technologies. Objective: To compare the diagnostic performance of computed tomography (CT) and magnetic resonance imaging (MRI) for ischemic and hemorrhagic stroke subtyping in a tertiary care hospital in Pakistan, and to evaluate diagnostic agreement, timing intervals, and associations with clinical severity. Methods: A retrospective cross-sectional study was conducted among 200 adult stroke patients who underwent CT, MRI, or both at Sir Ganga Ram Hospital, Lahore, between January and June 2024. Clinical data, imaging findings, and timing intervals were extracted from medical records and analyzed using descriptive statistics, group comparisons, and Cohen’s kappa for inter-modality agreement. Results: CT detected hemorrhagic stroke with high specificity (hyperdensity in 50.0% of hemorrhagic patients) but was normal in 70.3% of ischemic cases. MRI demonstrated superior performance for ischemic strokes, with diffusion restriction in 61.5% and chronic infarct in 69.2% of cases, while hemorrhage was confirmed in 100% of hemorrhagic strokes. Agreement between CT and MRI was substantial (kappa = 0.664). Diagnostic accuracy was robust across time intervals but higher for MRI in all subgroups. Conclusion: CT is a reliable frontline tool for excluding hemorrhage, while MRI provides superior sensitivity for ischemic stroke detection and confirms subtype classification even in delayed presentations. A multimodal imaging strategy is recommended to optimize acute stroke diagnosis in resource-constrained settings.