ABSTRACT
Objective:To obtain the parameters associated with hematoma morpholoy by finite element analysis(FEA) and investigated their performance on predicting and diagnosis hematoma expansion(HE) in patients with spontaneous intracrebral hemorrhage(SICH).Methods:Patients with SICH who met research criteria were retrospective enrolled between June 2015 and December 2017. Clinical parameters on admission were collected, Perform 2 independent methodology on same patient to analysis the hematoma shape base on computed tomography(CT): Clinical routine method that performed by clinical investigator to identified margin irregularity of hematoma by CT ,and calculated the volume of hematoma by simplify Tada formula(ABC/2);The FEA method performed by FEA investigator and gain the hematoma 3 dimensional morphology and variables, include Volume, Surface area, and The quantity of triangles per square milimet surface(TQOT/mm 2). The HE was defined as volume enlargement of >33% compared with that on addmission. All patients were divided into HE and none HE group ,respectively, ABC/2 and FEA generated thire own HE and none HE group as different volume calcuation. The HE risk factors of ABC/2 and FEA were assessed in univariate and multivariable Logistic regression models. and the risk fators diagnosis value for HE were determined by the receiver operating characteristic(ROC) curves. Results:Total of 127 patients were enrolled, The mean time of symptom onset to hospital admitted was 3.08±1.34 h. There were 34(26.77%) cases HE identifed by ABC/2 and 31(24.41%)by FEA. Althought there are significant different (pearson χ2=53.66, P<0.01) of HE identification between ABC/2 and FEA, the 2 methods has moderate consistency (Kappa=0.65). All patients’ hematoma 3D reconstruction were performed by FEA and general observation show that TQOT/mm 2 most likely correlate to irregularity of hematoma 3D shape. Multivariable Logistic regression models indicated that ICH score( OR=1.79, 95% CI:1.19~2.68)was independent HE risk factor for ABC/2, respectively, TQOT/mm 2≥1.95/mm 2 ( OR=16.99,95% CI:5.98~48.33)and Ultraearly Hematoma Growth,(uHG) ( OR=1.05, 95% CI:1.01~1.09)were independent HE risk factor for FEA. With ROC analysis, both the ICH score of ABC/2 and uHG of FEA have low HE predictive and diagnosis value ,the area under the curve (AUC) were 0.64 and 0.67 respectively. However, TQOT/mm 2 was found to have excellent diagnosis value (AUC:0.9), sensitivity and specificity were 77% and 83% when the cut-off value was 1.95. Panel parameter model (TQOT/mm 2+uHG) was not be found to have a significant higher AUC than single parameter on FEA and the clinical routine parameters panel model (ICH +SB P>180 mmHg on addmission) have a unacceptable AUC(<0.7) as well as single parameters. Conclusions:Hematoma shape could be reconstructed and analysis by FEA and TQOT/mm 2 was likely relevance to hematoma morphology. TQOT/mm 2≥1.95 was indicate to have a better HE predicting and diagnosis value than any other risk factors and clinical parameters panel models in our reaserch.