To explore the prediction effect of hematoma enlargement in the early stage of spontaneous intracerebral hemorrhage (SICH) by combining the morphology and density classification method of CT plain scan with multiple risk factors proposed by Barras et al.

A total of 150 patients with cerebral hemorrhage treated conservatively from January 2015 to December 2019 in the cerebrovascular ward were retrospectively included, and all patients completed their first CT scan within 3 hours of the onset of disease, in the Sixth Affiliated Hospital of Guangzhou Medical University. According to whether the relative volume of hematoma increased by 33% or absolute volume increased by 6 ml on CT plain scan within 24 hours, the patients were divided into the increased hematoma group (65 cases) and the non-increased hematoma group (85 cases). Logistics regression was used to analyze the value of various indicators, including the general situation of patients, medication history, laboratory indicators, unenhanced CT plain scan, examination results, and other indicators, as well as the hematoma grading proposed by Barras et al, for the prediction of early hematoma enlargement in patients with spontaneous intracerebral hemorrhage. By drawing receiver operating characteristic (ROC) curve, the prediction efficiency of CT plain scan hematoma morphology and density classification combined with multiple risk factors was analyzed.

Totally, 149 cases were included (64 in the hematoma enlargement group and 85 in the non-hematoma enlargement group). Logistic regression analysis showed that previous use of anticoagulants (OR=4.855, 95%CI: 1.102-21.38, P=0.037), previous anti-platelet aggregation (OR=3.831,95%CI:1.089-13.472, P=0.036), Glasgow coma scale (OR=0.797, 95%CI: 0.671-0.947, P=0.01), high-density lipoprotein (OR 0.116, 95%CI:0.025-0.534, P=0.006), non-enhanced CT scan results of hematoma showed high-grade Barras hematoma morphology (OR=2.481, 95%CI: 1.429-4.308, P=0.001), and density grading results (OR=2.28, 95%CI:1.312-3.963, P=0.003) were independent risk factors for early hematoma enlargement. ROC curve analysis indicated that the regression equation (AUC=0.907, specificity 80.0%, sensitivity 89.1%) proposed by Barras combined with multiple risk factors showed better predictive efficiency. Barras morphology grading (AUC=0.746, specificity 55.3%, sensitivity 82.8%) or density grading (AUC=0.694, specificity 55.3%, sensitivity 76.6%) or a combination of the two (AUC=0.799, specificity 81.3%, sensitivity 62.4%) was used.

Long-term history of anticoagulant and anti-platelet medication, Glasgow coma score, serum high-density lipoprotein concentration, and hematoma CT scan grading score proposed by Barras were independent predictors of increased risk of early hematoma in SICH patients. The application of CT plain scan hematoma grading method proposed by Barras combined with multiple risk factors to predict hematoma enlargement in the early stage of spontaneous intracerebral hemorrhage is more effective than a single index.