Enhancing diagnostic accuracy in liver transplant rejection: The role of imaging and clinical data

Abstract

Background: Rejection is a major complication after liver transplant, but its diagnosis remains challenging.

Aims: To assess the diagnostic value of imaging features in identifying liver transplant rejection, with and without the addition of clinical data.

Methods: In this retrospective study, 70 liver transplant recipients (35 with biopsy-proven rejection, 35 without) underwent CT and MRI evaluation. Three radiologists independently assessed imaging findings while blinded to clinical information. Diagnostic metrics were compared before and after integrating laboratory results and time since transplant. Multivariate analysis identified predictors of graft rejection, and receiver operating characteristic analysis evaluated diagnostic performance.

Results: Imaging alone yielded 64 % diagnostic accuracy (95 % CI: 57-70 %), increasing to 72 % (95 % CI: 65-78 %) with clinical data. Specificity remained high (82-89 %), but sensitivity was limited (46-55 %). Abnormal laboratory results were the strongest independent predictor (OR: 203; 95 % CI: 21-6986; p < 0.001; AUC: 0.95), and simplified models using laboratory values alone maintained strong performance (AUC: 0.86; p < 0.001).

Conclusions: Imaging reliably identifies liver transplant rejection when present, but its sensitivity is limited. Diagnostic performance improves markedly with clinical context, particularly abnormal laboratory values, which are the strongest predictors. These findings highlight the importance of a multidisciplinary approach over imaging alone.

Keywords: Diagnostic accuracy; Graft rejection; Liver transplantation; Transplant monitoring.