Infarct volume predicts outcome after decompressive hemicraniectomy for malignant hemispheric stroke

Abstract

The decision to perform decompressive hemicraniectomy (DHC) by default in malignant hemispheric stroke (MHS) remains controversial. Even under ideal conditions, DHC usually results in moderate to severe disability. The present study for the first time uses neuroimaging to identify independent outcome predictors in a prospective cohort of 96 MHS patients undergoing DHC. The primary outcome was functional status according to the modified Rankin Scale (mRS) at 12 months and categorized as favorable (mRS 0-3) or unfavorable (mRS 4-6). At 12 months, 19 patients (20%) reached favorable and 77 patients (80%) unfavorable outcome. The overall mean infarct volume was 328 ± 114 ml. Multivariable logistic regression identified age per year (OR 1.14, 95% CI 1.04-1.24; p = 0.005), infarct volume per cm³ (OR 1.012, 95% CI 1.003-1.022; p = 0.013), thalamic involvement (OR 8.65, 95% CI 1.04-72.15; p = 0.046) and postoperative pneumonia (OR 5.52, 95% CI 1.03-29.57; p = 0.046) as independent outcome predictors, which was confirmed by multivariable ordinal regression for age ( p = 0.004) and infarct volume ( p = 0.015). The infarct volume threshold for reasonable prediction of unfavorable outcome in our patients was 270 cm³, which in the future may help prognostication and development of clinical trials on DHC and outcome in MHS.

Keywords: Acute stroke; MRI; cerebrovascular disease; neurosurgery; stroke.

Figure 1.

ROC analysis of the infarct volume and unfavorable outcome defined as mRS 4–6 yielded a good classification result with an area under the curve (AUC) of 0.80. The Youden-Index at the vertex of the curve identified an infarct volume threshold of 270 cm³ for outcome prediction with the highest sensitivity and specificity in our study population (sensitivity 81.8%, specificity 73.7%; Supplemental Table 2; http://jcbfm.sagepub.com/content/by/supplemental-data). The dashed line represents the identity line.

Figure 2.

Illustration of the infarct volume distribution in each patient according to his or her 12-month mRS score. The dashed line represents the threshold infarct volume of 270 cm³, determined at the Youden-Index of the ROC analysis presented in Figure 1.