ABSTRACT
Introduction: The aim of this study was to develop a noninvasive prediction model for histological stages in PBC that is simple, easy to implement, and highly accurate. Methods: A total of 114 patients with PBC were included in this study. Demographic, laboratory data and histological assessments were collected. The independent predictors of histological stages were selected to establish a noninvasive serological model. The scores of 22 noninvasive models were calculated and compared with the established model. Results: This study included 99 females (86.8%) and 15 males (13.2%). The number of patients in Scheuer's stage 1, 2, 3 and 4 was 33 (29.0%), 34 (29.8%), 16 (14.0%), and 31 (27.2%), respectively. TBA and RDW are independent predictors of PBC histological stages. The above indexes were used to establish a noninvasive model-TR score. When predicting early histological change (S1) or liver fibrosis and cirrhosis (S3-S4), the AUROC of TR score were 0.887 (95% CI, 0.809-0.965) and 0.893 (95% CI, 0.816-0.969), higher than all of the other 22 models included in this study. When predicting cirrhosis (S4), its AUROC is still as high as 0.921 (95% CI, 0.837-1.000). Conclusion: TR score is an easy, cheap and stable noninvasive model, without complex calculation formulas and tools, and shows good accuracy in diagnosing the histological stages of PBC.
