The value of a nomogram for predicting the outcome of intracerebral hemorrhage based on clinical characteristics and diffusion-weighted imaging of hyperintense lesions

Abstract

Objective: To investigate the value of a nomogram predicting the outcome of intracerebral hemorrhage (ICH) based on clinical characteristics and diffusion-weighted imaging (DWI) of hyperintense lesions. Methods: A case-control study. Consecutive patients, aged 30-88(59±13) years old, with ICH were recruited at the Stroke Center of Zhengzhou People's Hospital from January 2018 to August 2021. Patients were divided into a group with DWI lesions and a group without DWI lesions depending on whether there were DWI hyperintense lesions distant from the hematoma. Prognosis was evaluated at 90 days via the modified Rankin Scale (mRS). Univariate and multivariable logistic regression models were used to identify independent predictors of a poor ICH outcome (mRS score≥4), and a nomogram model was developed. The performance of the nomogram was validated via the area under the receiver operating characteristic curve (AUC) and a calibration chart. Results: Of the 303 patients included in the study, 24.8% presented with DWI lesions; 17.5% with asymptomatic DWI lesions and 7.3% with symptomatic DWI lesions. Poor outcomes were significantly more frequent in the group with DWI lesions than in the group without DWI lesions (χ2=21.32, P<0.001). In multivariable regression analysis, age [odds ratio (OR)=1.032, 95% confidence interval (CI) 1.002-1.063, P=0.035], hematoma volume (OR=1.050, 95%CI 1.011-1.090, P=0.012), hematoma location (OR=3.839, 95%CI 1.248-11.805, P=0.019), DWI lesions (OR=3.955, 95%CI 1.906-8.206, P<0.001), and baseline NIHSS scores (OR=1.102, 95%CI 1.038-1.170, P=0.001) were independent predictors of a poor outcome. In subgroup analysis patients with asymptomatic DWI lesions had a 3-fold greater risk of a poor outcome compared to those without DWI lesions (OR=3.135, 95%CI 1.382-7.112, P=0.006), and patients with symptomatic DWI lesions had a 7-fold greater risk of a poor outcome compared to those without DWI lesions (OR=7.126, 95%CI 2.279-22.277, P=0.001). A nomogram model was established based on the independent predictors for a poor outcome. The AUC of the nomogram was 0.846 (95%CI 0.795-0.898), and a calibration chart indicated good consistency between values predicted by the nomogram and actual observed values. Conclusions: DWI lesions are an independent risk factor for a poor outcome in patients with ICH-particularly symptomatic DWI lesions. A nomogram model based on clinical characteristics and DWI lesions exhibited good efficacy when predicting the outcome of ICH.