CT Perfusion in Acute Lacunar Stroke: Detection Capabilities Based on Infarct Location

BACKGROUND AND PURPOSE: Recent studies demonstrated superiority of CTP to NCCT/CTA at detecting lacunar infarcts. This study aimed to assess CTP's capability to identify lacunae in different intracranial regions. MATERIALS AND METHODS: Over 5.5 years,1085 CTP examinations were retrospectively reviewed in patients with acute stroke symptoms with CTP within 12 hours and MRI within 7 days of symptom onset. Patients had infarcts <= 2 cm or no acute infarct on DWI; patients with concomitant infarcts >2 cm on DWI were excluded. CTP postprocessing was automated by a delay-corrected algorithm. Three blinded reviewers were given patient NIHSS scores and symptoms; infarcts were recorded based on NCCT/CTA, CTP (CBF, CBV, MTT, and UP), and DWI. RESULTS: One hundred thirteen patients met inclusion criteria (53.1% female). On DWI, lacunar infarcts were present in 37 of 113 (32.7%), and absent in 76 of 113 (67.3%). On CTP, lacunar infarcts typically appeared as abnormalities larger than infarct size on DWI. Interobserver kappa for CTP ranged from 0.38 (CBF) (P<.0001) to 0.66 (UP) (P<.0001); interobserver kappa for DWI was 0.88 (P < 0.0001). In all intracranial regions, sensitivity of CTP ranged from 18.9% (CBV) to 48.7% (UP); specificity ranged from 97.4% (CBF and TTP) to 98.7% (CBV and MU). CTP's sensitivity was highest in the subcortical white matter with or without cortical involvement (21.7%-65.2%) followed by periventricular white matter (12.5%-37.5%); sensitivity in the thalami or basal ganglia was 0%. CONCLUSIONS: CTP has low sensitivity and high specificity in identifying lacunar infarcts. Sensitivity is highest in the subcortical white matter with or without cortical involvement, but limited in the basal ganglia and thalami.