Early Prediction of Cerebral Computed Tomography under Intelligent Segmentation Algorithm Combined with Serological Indexes for Hematoma Enlargement after Intracerebral Hemorrhage

Abstract

The aim of this study was to explore the application value of brain computed tomography (CT) images under intelligent segmentation algorithm and serological indexes in the early prediction of hematoma enlargement in patients with intracerebral hemorrhage (ICH). Fuzzy $C$ -means (FCM) intelligence segmentation algorithm was introduced, and 150 patients with early ICH were selected as the research objects. Patient cerebral CT images were intelligently segmented to assess the diagnostic value of this algorithm. According to different hematoma volumes during CT examination, patients were divided into observation group (hematoma enlargement occurred, $n=48$ ) and control group (no hematoma enlargement occurred, $n=102$ ). The predicative value of hematoma enlargement after ICH was investigated by assessing CT image quality and measuring intracerebral edema, hematoma volume, and serological indicators of the patients of the two groups. The results demonstrated that the sensitivity, specificity, and accuracy of CT images processed by intelligence segmentation algorithm amounted to 0.894, 0.898, and 0.930, respectively. Besides, early edema enlargement and hematoma of patients in the observation group were more significant than those of patients in the control group. Relative edema volume was 0.912, which was apparently lower than that in the control group (1.017) ( $P<0.05$ ). In terms of CT signs of ICH patients, the incidence of blend sign, low density sign, and stroke of the observation group was evidently higher than those of the control group ( $P<0.05$ ). Besides, absolute lymphocyte count (ALC) and hemoglobin (HGB) concentration of the patients in the observation group were $6.23\times 109$ /L and $6.29\times 109$ /L, respectively, both of which were higher than those of the control group ( $6.08\times 109$ /L and $4.25\times 109$ /L). Neutrophil to lymphocyte ratio (NLR) was $0.99\times 109$ /L, which was apparently lower than that in the control group ( $1.43\times 109$ /L) ( $P<0.05$ ). To sum up, cerebral CT images processed by FCM algorithm showed good diagnostic effect on ICH and high clinical values in the early prediction of hematoma among ICH patients.