Predictive value of clinical lacunar syndromes for lacunar infarcts on magnetic resonance brain imaging

OBJECTIVE: We prospectively investigated the predictive value of clinical and CT-supported lacunar syndromes for lacunar infarcts on magnetic resonance (MR) brain imaging. PATIENTS AND METHODS: The 54 prospective, consecutive patients had clinical lacunar syndromes of acute onset and early computed tomography (CT; on admission day, i.e. < or =48 h after onset of symptoms) showing either a small deep infarct or no corresponding lesion. Taking MR (at day 2 to 4 after admission) as the gold standard, the positive predictive value of the CT-supported clinical syndrome for corresponding lacunar lesions was calculated. RESULTS: In 27 (50%) patients, early CT showed a lacunar infarct corresponding to the clinical syndrome, a further 27 (50%) patients had no fresh ischemic lesion. In 51 patients (94%), MR showed a recent lacunar infarct (hyperintense lacune in T2-weighted scans, no demarcation on T1-weighted scans and/or positive gadolinium-enhancement) corresponding to the clinical syndrome (positive predictive value 0.94, 95%, CI: 0.88 to 0.98). In 3 (6%) patients MR was normal. Aside from old unrelated ischemic (macro- and/or microangiopathic) lesions, MR revealed no acute non-lacunar infarct. CT and MR sites of lacunar lesions were matching. Compared to gold standard MR, the sensitivity of early CT for suspected lacunar lesions was 0.53 (95% CI: 0.38 to 0.67). CONCLUSION: Lacunar syndromes were highly predictive for small deep infarcts on MR. Magnetic resonance brain imaging may be redundant in the setting of a lacunar syndrome supported by a CT that excludes non-ischemic causes of stroke; it may therefore be abandoned in order to reduce costs in the health care system.