The Predictive Value of Non-contrast CT for Intracerebral Hemorrhage Expansion

The primary purpose of the present meta-analysis was to systematically explore the non-contrast computed tomography (NCCT) signs for predicting hematoma expansion (HE) in patients with spontaneous intracerebral hemorrhage (sICH). Through a search of Pubmed, Web of Science and Embase, the diagnostic value of 10 NCCT signs, including heterogeneous density, irregular shape, blend sign, swirl sign, hypodensity, black hole sign, island sign, satellite sign, fluid levels, ultra-early hematoma growth, were combined and analyzed separately. After processing the data with the Stata 14.0 software, the study presented the sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and posterior probability of various signs. The results showed none of the signs has high sensitivity and specificity at the same time, which may be the defect of using NCCT to predict HE. Nevertheless, the island sign may have a better predictive value in practice due to the higher positive likelihood ratios of 6.6 (95% CI, 3.4-12.9).本项荟萃分析的主要目的是系统探讨自发性脑内出血（spontaneous intracerebral hemorrhage, sICH）患者中，用于预测血肿扩大（hematoma expansion, HE）的非增强计算机断层扫描（non-contrast computed tomography, NCCT）征象。通过检索PubMed、Web of Science及Embase数据库，本研究对10项NCCT征象的诊断价值分别进行合并与分析，所纳入的征象包括密度不均、形态不规则、混杂征、漩涡征、低密度征、黑洞征、岛征、卫星征、液平征及超早期血肿生长。随后采用Stata 14.0统计软件处理数据，本研究给出了各项征象的灵敏度、特异度、阳性似然比、阴性似然比及后验概率。结果显示，尚无任何一项征象可同时具备较高的灵敏度与特异度，这或为利用NCCT预测HE所存在的局限性。尽管如此，岛征在实际应用中可能具备更优的预测价值，因其阳性似然比可达6.6（95%置信区间：3.4~12.9）。