Abstract 3479: Prediction of Outcome in Cerebellar Infarction by Diffusion MRI

Abstract

Background and Purpose - Early identification of patients at risk for neurological deterioration after stroke remains a challenge. MRI based volumetric assessments of infarct have been shown to predict poor outcome amongst patients with large middle cerebral artery infarcts. However, no such quantitative imaging measure exists to predict outcome following cerebellar stroke. In this study, we tested the hypothesis that the volume of cerebellar infarction and its associated apparent diffusion coefficient (ADC), as estimated from diffusion-weighted images (DWI), can provide reliable information for the prediction of neurological deterioration. Methods - We retrospectively identified twenty-eight consecutive patients (age 58.2±13) with cerebellar stroke who underwent MRI (median 40.8 hours) with DWI. Patients were divided into poor (N=7) and good (N=21) outcomes. Patients with poor outcome were defined as those who required posterior fossa decompression or died as a result of their cerebellar infarct. Lesion and cerebellar volume were defined on ADC maps. The ratio of the lesion volume to the whole cerebellum volume was calculated (rVolume) as well as the ratio of the lesion mean ADC to normal cerebellum mean ADC (rADC). The National Institutes of Health Stroke Scale (NIHSS) score, as well as age, sex and race were collected. Binary logistic regression was used to identify predictors of poor outcome. The sensitivity and specificity of rVolume and lesion volume were determined in predicting poor outcome using ROC analysis. Results - Logistic regression analysis revealed that rVolume (p=0.018), lesion volume (p=0.034), NIHSS score (p=0.020) and whole cerebellar ADC value (p=0.043) were significant predictors of poor outcome. The rADC (p=0.073) approached significance. The rVolume (p=0.047) remained a significant predictor of outcome even when controlling for age, NIHSS score, and hours to scan, while a trend towards significance for lesion volume (p=0.072), cerebellar ADC (p=0.069) and rADC (p=0.061) were also observed. ROC analysis revealed that both rVolume as a percentage and lesion volume had a specificity of 95% and a sensitivity of 57% in predicting poor outcome. The respective cut-off values were 30% and 41 cc. Conclusions - Quantitative measurement of rVolume, lesion volume, and cerebellar ADC values are significantly correlated with patient outcomes following cerebellar stroke. Although cerebellar ADC provided additional sensitivity, the rVolume appears to be a strong predictor of outcome when controlling for age, NIHSS score and time to scan. Quantitative analysis of diffusion MRI may assist in identification of patients with cerebellar stroke at highest risk of deterioration. Prospective validation is warranted.