Background: The diagnosis of small deep infarcts (SDIs) based on clinical and noncontrast CT findings is often employed to triage stroke patients for emergent endovascular interventions. Few studies have evaluated the accuracy of <6 hours presentation with lacunar clinical syndromes in predicting the presence of a SDI and the absence of large vessel occlusion.
Methods: Consecutive patients were identified with symptoms suggestive of acute stroke, diffusion-weighted magnetic resonance imaging (DW-MRI) and CT performed within 6 hours of onset, and absence of hemorrhage on imaging. Patients were given a diagnosis of SDI if they met the defined clinical and CT criteria. Diagnoses of large vessel occlusion absent (LVOA) and final diagnoses of SDI versus non-SDI incorporated hyperacute multimodal MRI data.
Results: Of 64 patients, 15 were diagnosed as SDI at <6 hours based on clinical/CT criteria. Fourteen of 15 patients (93%) had a final diagnosis of LVOA. Ten of 15 (67%) had a final MR-DWI-confirmed diagnosis of SDI. Clinical/CT diagnosis of SDI showed sensitivity of 44%, specificity 97%, positive predictive value (PPV) 93%, and negative predictive value (NPV) of 61% in predicting LVOA and sensitivity 83%, specificity 90%, PPV 67%, and NPV 96% compared to the final diagnosis of SDI. Most patients incorrectly diagnosed with SDIs harbored small cortical branch infarcts.
Conclusions: Diagnoses of SDI <6 hours from symptom onset based on a classic lacunar syndrome and CT criteria show only modest predictive value in predicting the presence of a small-vessel territory infarct but perform fairly well in identifying patients without large-vessel occlusions amenable to endovascular therapy.