Correlation Between Quantitative Multi-parameters Of Energy Spectrum CT And Short-term Prognosis Of Acute-on-chronic Liver Failure

Objective:To analyze the correlation between clinical features and quantitative multiparametric tests of imaging（conventional CT and energy-spectral CT）and the prognosis of acute-on-chronic liver failure（ACLF）and screened for parameters that were independently associated with short-term patient progression.A comprehensive prediction model for the progression of ACLF based on quantitative multi-parameters of energy spectrum CT was further constructed,and the value of the model was investigated in the clinical context and the nomogram was drawn to assist in clinical decision making.Materials and Method:Retrospective collection of 107 cases of ACLF seen at the Second Hospital of Lanzhou University between June 2019 and June 2022,ACLF was diagnosed in accordance with the Asia-Pacific Association for the Study of the Liver guidelines,and all patients underwent liver energy-spectrum CT scans within one week before and after the diagnosis of ACLF.Patients with ACLF were grouped according to their short-term progressive status（stable and progressive groups）,with the primary outcome being the progression of the patient’s disease（deterioration, death,or receipt of liver transplantation）at 90 days from the diagnosis of ACLF.Independent risk factors associated with disease progression in clinical characteristics and quantitative imaging（conventional CT and energy-spectral CT）parameters were identified by univariate and multifactorial logistic regression analysis.And predictive models of disease progression were constructed by combining clinical and imaging parameters and compared with existing clinical scores（MELD and MELD-Na）.The effectiveness of the model was evaluated by applying the area under the subject operating characteristic（ROC）curve（AUC）,C-index,calibration curve,and decision curve（DCA）.Results:A total of 107 patients were included,with 44 cases in the stable group（age 51.70±11.92 years;24 men）and 63 cases in the progressive group（49.29±10.76 years;38 men）.The results of univariate analysis showed that among clinical characteristics,there were statistically significant differences between groups for the occurrence of hepatic encephalopathy,BUN,TB and WBC in patients（all P<0.05）;among imaging parameters,there were statistically significant differences between groups for VC_sufra,VCR,single energy CT values of 50-100 ke V for portal phase parameters,K₁₄₀-VP,Z-VP,IC-VP and NIC-VP and There were statistically significant differences between groups for the delayed phase parameters 40-80 ke V of single energy CT values,K₁₄₀-DP,Z-DP,IC-DP and NIC-DP（all P<0.05）.Multifactorial logistic regression analysis showed that the clinical characteristics of BUN（OR=1.774,95%CI:1.617-1.971;P=0.027）and the quantitative imaging parameters VCR（OR=0.003,95%CI:0.001-0.087;P=0.010）,IC-VP（OR=0.820,95%CI:0.728-0.924;P=0.001）and NIC-VP（OR=0.002,95%CI:0.001-0.026;P=0.004）were independent risk factors for 90-day disease progression in patients with ACLF,and a comprehensive prediction model-nomogram was developed based on the results of regression analysis.The ROC curve showed the highest predictive accuracy of the nomogram with an AUC value of 0.893,sensitivity of 84.13%,and specificity of 84.09%,both higher than the clinical model（MELD and MELD-Na scores）.The C-index and corrected C-index of the Nomogram were 0.893（95%CI:0.833-0.953）and 0.876,respectively,showing good model discrimination;the calibration curve indicated good agreement between the predicted probability and the actual outcome of the nomogram;the DCA indicated the high clinical utility value of the nomogram.Conclusions:1.Among clinical characteristics,hepatic encephalopathy,BUN,TB and WBC were statistically significantly different in the progressive and stable groups;among conventional CT parameters,there were statistically significant differences in VCsufra and VCR between patients in the progressive and stable groups;among quantitative parameters of energy spectrum CT,there were significant differences in 70 ke V,IC-VP,NIC-VP in the portal phase,and 50 ke V and NIC-DP in the delayed phase in the progressive status of ACLF patients at 90 days,all of which contributed to the early assessment of the prognosis of ACLF.2.BUN in clinical features and VCR,IC-VP and NIC-VP in imaging parameters were independent risk factors for the progression of the patient’s disease.3.The integrated prediction model-Nomogram constructed based on independent risk factors（BUN,VCR,IC-VP and NIC-VP）helps to predict the progressive status of ACLF patients at 90 days early for risk stratification of patients and clinical decision making.