ABSTRACT
INTRODUCTION: Diffusion-weighted imaging (DWI) in magnetic resonance imaging (MRI) has been proposed as the first line of neuroimaging for acute ischaemic stroke. The reliability of DWI in detecting intracranial haemorrhage, however, is still unproven, compared with susceptibility-weighted imaging (SWI) and CT scan which being considered the gold standard. This study seeks to establish the reliability of DWI as a first-line imaging modality to detect the intracranial haemorrhage in the patients present within the thrombolysis window. MATERIALS AND METHODS: A retrospective cross-sectional analysis was performed on patients who presented to our institution from April 2020 until July 2021 for acute stroke and had MRI brain as first-line neuroimaging. A total of 31 subjects were included in this study. Two radiologists assessed the signal patterns in DWI sequence and compared them with SWI and CT Brain, whenever available, as the gold standard for observing the presence of intracranial haemorrhage. RESULTS: The majority of patients with hyperacute bleed proven to be revealed on SWI or CT, thus showed characteristics of central hyperintensity and peripheral hypointense rim, on DWI. Slightly more than half (51.6%) presented with mild to moderate NIHSS scores (1-15). The sensitivity, specificity, positive predictive value and negative predictive value of DWI in detecting intracranial intra-axial haemorrhages were exceptionally high. There is strong interobserver level of agreement in identifying central haemorrhagic signal intensity [kappa = 0.94 (0.06), p < 0.05]. CONCLUSION: This study supported the DWI sequence as a reliable sequence in MRI, to detect intracranial haemorrhage in hyperacute stroke.
