Value of a logistic regression model based on the clinical features of liver cancer in judging the traditional Chinese medicine syndrome types of primary liver cancer
Dongyi LIN,Bo PENG,Jinghui ZHENG,Zongyu WANG,Hanzhong LI,Jianbo HE,Kezhi LI,WuGuobin,Chuang CHEN
DOI: https://doi.org/10.3969/j.issn.1001-5256.2020.06.021
2020-01-01
Abstract:ObjectiveTo investigate the association of the clinical features of liver cancer patients, including albumin-bilirubin (ALBI) and neutrophil-lymphocyte ratio (NLR), with the traditional Chinese medicine (TCM) syndrome types of primary liver cancer, and to establish a clinical judgment model for TCM syndrome differentiation of primary liver cancer. MethodsA total of 289 previously untreated patients who were admitted to The Affiliated Tumor Hospital of Guangxi Medical University from November 1, 2016 to October 31, 2018 and were diagnosed with primary liver cancer based on pathology or clinical examination were enrolled, and TCM syndrome differentiation was performed for all patients. A one-way analysis of variance was used for comparison of normally distributed continuous data between groups, and the Kruskal-Wallis H test was used for comparison of non-normally distributed continuous data between groups; the chi-square test was used for comparison of categorical data between groups. The data with significant difference in the univariate analysis were included in the logistic regression analysis, and the receiver operating characteristic (ROC) curve was used to evaluate the efficiency of these clinical features in the TCM syndrome differentiation of liver cancer. ResultsThere were significant differences between the patients with different TCM syndrome types of liver cancer in ALBI (F=5.487, P＜0.001), NLR (χ2=30.146, P＜0.001), BCLC stage (χ2=71.973, P＜0.001), albumin (Alb) (χ2 = 18.887, P＜0.001), total bilirubin (TBil) (χ2=12.138, P=0.007), alanine aminotransferase (ALT) (χ2=18.001, P＜0.001), aspartate aminotransferase (χ2=12.067, P=0.007), absolute neutrophil count (F=6.262, P＜0.001), absolute lymphocyte count (F=2.934, P=0.034), diameter of intrahepatic primary tumor (F=4.905, P=0.002), ascites (χ2=9034, P=0.021), portal vein tumor thrombus (χ2=13.434, P=0.004), and number of extrahepatic metastatic lesions (χ2=2.529, P=0.002). The logistic regression analysis showed that ALT (odds ratio ［OR］=1.002, 95% confidence interval ［CI］: 1.003-1.021, P＜0.05) and BCLC stage (OR=0.591, 95% CI: 0.413-0.845, P＜0.05) were independent factors for judging damp-heat accumulation; ALT (OR=0.985, 95% CI: 0.974-0.997, P＜0.05) and BCLC stage (OR=3.191, 95% CI: 2.042-4.986, P＜0.05) were also independent factors for judging liver depression and spleen deficiency; TBil (OR=0.966, 95% CI: 0.937-0.995, P＜0.05), Alb (OR=1.259, 95% CI: 1.064-1.490, P＜0.05), and ALBI (OR=0.088, 95% CI: 0.013-0.607, P＜0.05) were independent factors for judging Qi stagnation and blood stasis. The ROC curve analysis showed that ALT and BCLC stage had an area under the ROC curve (AUC) of 0.662 (95% CI: 0.605-0.717), a sensitivity of 69.4%, and a specificity of 58% in judging damp-heat accumulation, at the cut-off values of 36 U/L for ALT and stage C for BCLC stage; ALT and BCLC stage had an AUC of 0.753 (95% CI: 0.699-0.801), a sensitivity of 72.7%, and a specificity of 68.2% in judging liver depression and spleen deficiency, at the cut-off values of 64 U/L for ALT and stage B for BCLC stage; TBil, Alb, and ALBI had an AUC of 0.634 (95% CI: 0.576-0.690), a sensitivity of 567%, and a specificity of 65.3% in judging Qi stagnation and blood stasis, at the cut-off values of 28.4 μmol/L for TBil, 37.8 g/L for Alb, and 1.95 for ALBI. ConclusionThe clinical judgment model based on ALT, BCLC stage, TBil, Alb, and ALBI can differentiate between the TCM syndrome types of damp-heat accumulation, liver depression and spleen deficiency, and Qi stagnation and blood stasis for liver cancer, and this model is simple, convenient, and objective and thus holds promise for clinical application.