Hao Peng,1,2,* Jixue Li,1,* Xiaoru Zhou,1 Zhewen Nong,1 Ruiying Zhang,1 Pei Lu,1 Shasha Ye,1 Liping Lei,3 Chuang Qin,2 Jiangfa Li1,4,5 

1Department of Hepatobiliary and Pancreatic Surgery, Affiliated Hospital of Guilin Medical University, Guilin, 541001, People’s Republic of China; 2Department of Hepatobiliary and Pancreatic Surgery, Liuzhou People’s Hospital, Liuzhou, 545001, People’s Republic of China; 3Department of Geriatric Medicine, The Affiliated Hospital of Guilin Medical University, Guilin, 541001, People’s Republic of China; 4Key Laboratory of Early Prevention and Treatment for Regional High Frequency Tumor (Guangxi Medical University), Ministry of Education, Nanning, 530021, People’s Republic of China; 5Guangxi Key Laboratory of Early Prevention and Treatment for Regional High Frequency Tumor, Nanning, 530021, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Chuang Qin, Department of Hepatobiliary and Pancreatic Surgery, Liuzhou People’s Hospital, No. 8, Wenchang Road, Central District, Liuzhou, 545001, People’s Republic of China, Email 382944957@qq.com Jiangfa Li, Department of Hepatobiliary and Pancreatic Surgery, Affiliated Hospital of Guilin Medical University, No. 15, Lequn Road, Xiufen District, Guilin, Guangxi, 541001, People’s Republic of China, Email 247546160@qq.com

Objective: To explore a more effective and practical comprehensive index for differentiating benign from malignant obstructive jaundice by analyzing the clinical data of patients with benign obstructive jaundice (BJ) group and malignant obstructive jaundice (MJ) group.
Methods: A retrospective analysis was conducted on the clinical data of 339 patients with obstructive jaundice. The cases were divided into two data sets: training cohort and validation cohort. The cases were divided into two groups: malignant and benign obstructive jaundice group. Logistic regression analysis was used to build a prediction model for judging the nature of obstructive jaundice, and the prediction model was verified using the validation cohort.
Results: Multivariate analysis revealed that CEA, TBIL, and NLR were independent factors in malignant obstructive jaundice. A comprehensive index for differentiating benign from malignant obstructive jaundice was established based on these indicators. The sensitivity, specificity, and receiver operating characteristic curve of this model for differentiating benign from malignant obstructive jaundice were 79.57%, 93.26%, and 0.920, respectively.
Conclusion: The prediction model based on the comprehensive index of CEA, TBIL, and NLR has a higher accuracy in differentiating malignant obstructive jaundice.

Keywords: obstructive jaundice, comprehensive index, differentiation diagnosis