LAUSR

OBJECTIVE The latest World Health Organization data showed that stroke was the highest mortality in China, accounting for 23.7% of the total mortality from 2000 to 2012. Intracerebral hemorrhage (ICH) was the most deadly and incurable type of stroke. In the Qinghai-Tibetan Plateau, the incidence of stroke was relatively higher. Several studies showed that the shape and heterogeneity of hematoma and image markers on brain computed tomography scan had predictive effects on hematoma expansion (HE). The study aimed to find relative factors and established a nomogram model to predict the HE of ICH. METHODS All patients with ICH in Qinghai Provincial People's Hospital from January 1, 2012, to May 22, 2018, were continuously collected. A total of 402 patients were included in the study. This was a single-center retrospective study. Univariate and binary logistic regression analysis were performed to screen out the independent predictors that were significantly associated with HE. RESULTS The total incidence of HE in ICH was 30.9%, whereas the incidence of HE in the basal ganglia and nonbasal ganglia was 36.4% and 17.2%, respectively. Diabetes, basal ganglia hemorrhage, time of onset to baseline computed tomography, island sign, blend sign, black hole sign, and swirl sign were independent predictors of HE. Based on these predictors, a nomogram model was established and the accuracy was 81.6%, the sensitivity was 91.1%, and the specificity was 70.5%. CONCLUSIONS This model had a high accuracy of predicting HE in the Qinghai-Tibetan Plateau. Because this model is noninvasive, rapid, and low cost, it is easy to promote and has wide application prospects in clinical practice.