PTH-119 Differences between radiology and histopathology: are we judging wrong? should we lower alpha-fetoprotein cut-off in the selection criteria?

Abstract

Introduction Current UK selection criteria for liver transplantation (LT) in patients with hepatocellular carcinoma (HCC) are based on pre-LT imaging and aFP values Aim Aim of the study was to identify potential discrepancies between radiology and histopathology in patients undergoing LT for HCC and identify predictors of outcome with HCC recurrence as endpoint. Method All consecutive patients transplanted for HCC between 1997–2014 were retrospectively evaluated: demographic and clinical data, aetiology, pre-LT aFP levels and HCC treatment, radiological staging of HCC, histological tumour characteristics. Cox regression was used to assess factors associated with HCC recurrence and AUROC and Youden index to access cut-offs. Results 185 patients: male 150(81%); mean age 55(27–68) years; aetiology HCV (54%), HBV (19%), ALD (15%). 126 (68%) underwent pre-LT treatment for HCC. Discrepancy between histopathological and radiological findings regarding the number of HCC nodules was found in 104/185(56%) patients: 78 were underestimated and 26 overestimated by radiology. 10/40 (25%) patients, who fell within radiological Milan criteria, resulted outside Milan criteria according to the explant due to discrepancy in the number of nodules. Considering the diameter of the biggest nodule on explant, discrepancy was observed in 158/182(85%) patients: 93 were underestimated, 65 overestimated. 20/40 (50%) patients fulfilling Milan criteria based on radiology, exceeded these criteria due to discrepancy in the size of the largest tumour. Recurrence was observed in 29/185(16%) patients post-LT in a median of 37 months (4–157 months). Cox regression showed that pre-LT factors associated with HCC recurrence post-LT were: aFP >100 IU/l (p = 0.007, OR=0.28, 95% CI=0.113–0.707) and radiological size of the largest nodule >3 cm (p = 0.001, OR=0.8, 95% CI = 0.78–0.97). Combined together the two parameters showed the best sensitivity (77%) and specificity (70%): AUROC=0.72, p = 0.0001, 95% CI=0.62–0.82. Conclusion In patients transplanted for HCC there is an important discrepancy between radiological findings and pathological description of the explant. The best predictors of HCC recurrence was the combination of pre-LT aFP >100 IU/l and diameter of the biggest nodule >3 cm. Disclosure of interest None Declared.