The comparative study of island sign and the spot sign in predicting short-term prognosis of patients with intracerebral hemorrhage.

OBJECTIVES

It is well known that early hematoma expansion is associated with short-term prognosis of patients with intracranial hemorrhage (ICH). And spot sign is recognized as a reliable computed tomography angiography (CTA) predictor for early hematoma expansion. Recently, island sign is also reported as a novel computed tomography (CT) predictor for early hematoma growth. Here, we compared the predictive abilities of these two signs for short-term outcomes of ICH patients.

PATIENTS AND METHODS

All the ICH patients were retrospectively identified. Clinical characteristics and radiological parameters were obtained from electronic medical records. Hematoma expansion, spot sign and island sign were assessed by two senior neurologists according to the initial and follow-up CT scans. 3-months prognoses were estimated according to Glasgow outcome scale (GOS). Multivariate logistic regression analyses were employed to explore the associations of short-term prognosis on island sign, spot sign and other clinical variables.

RESULTS

There were 283 ICH patients included. 113 of them presented with early hematoma expansions. 66 of them exhibited island sign, while spot sign occurred in 85 patients. Univariate analyses demonstrated that GCS score at admission (OR: 0.464, 95%CI: 0.395-0.547), hematoma volume (OR:1.062, 95%CI: 1.041-1.083), interventricular extension(OR:9.528, 95%CI: 3.915-23.187), island sign (OR: 4.595, 95%CI: 2.404-8.784) and spot sign (OR: 4.052, 95%CI: 2.297-7.147) were correlated with 3-months morbidity. Moreover, multivariate logistic regression analyses further revealed that both spot sign (OR: 3.413, 95%CI: 1.570-7.422) and island sign (OR: 7.564, 95%CI: 2.969-19.273) were strongly associated with 3-months poor outcome and have comparable predictive values (AUC: 0.636 vs. 0.622, P = .58). However, spot sign exhibited a superior predictive ability for 3-months mortality compared to island sign (OR: 2.713, 95%CI: 1.570-4.217 vs. OR: 2.362, 95%CI: 1.238-3.899, AUC: 0.700 vs. 0.603, P < .01).

CONCLUSIONS

Island sign is not just a convenient and reliable predictor for short-term prognosis of ICH patients, but also could be used as an indicator for accurate diagnosis and aggressive treatment.