Accuracy of the Blend Sign on Computed Tomography as a Predictor of Hematoma Growth after Spontaneous Intracerebral Hemorrhage: A Systematic Review

Abstract

Hematoma growth is a strong independent predictor of poor outcome after intracerebral hemorrhage. However, there is no gold standard to accurately predict hematoma growth. Several noncontrast computed tomographic markers associated with hematoma growth have been reported recently. Blend sign, which is a new marker, has been reported in several studies and seems a particularly promising marker but lacks a standardized evaluation so far. A systematic review of published literature on blend sign and hematoma growth and clinical outcomes was conducted. Systematic review of best practices was followed, and study quality was assessed. The 6 studies involved 1573 participants in this review. The prevalence of blend sign ranged from 8.70% to 38.46%. The sensitivity of blend sign to predict hematoma growth varied from 13.0% to 42.86%; the specificity varied from 88.51% to 95.5%. Blend sign showed lower sensitivity but superior specificity for prediction of hematoma growth. Four studies indicated that the presence of blend sign was an independent predictor of hematoma growth. Four studies showed that the prevalence of blend sign was significantly higher in patients with hematoma growth compared with those without hematoma growth (odds ratio, 9.33; 95% confidence interval, 5.20-16.74). There was an association between blend sign and hematoma growth, but this finding is tentative in light of the fact that the number of included studies was relatively small.