Defining and predicting early recurrence after liver resection of hepatocellular carcinoma: a multi-institutional study.

ABSTRACT

BACKGROUND: A clear definition of "early recurrence" after hepatocellular carcinoma (HCC) resection is still lacking. This study aimed to determine the optimal cutoff between early and late HCC recurrence, and develop nomograms for pre- and postoperative prediction of early recurrence. METHODS: Patients undergoing HCC resection were identified from a multi-institutional Chinese database. Minimum P-value approach was adopted to calculate optimal cut-off to define early recurrence. Pre- and postoperative risk factors for early recurrence were identified and further used for nomogram construction. The results were externally validated by a Western cohort. RESULTS: Among 1501 patients identified, 539 (35.9%) were recurrence-free. The optimal length to distinguish between early (n = 340, 35.3%) and late recurrence (n = 622, 64.7%) was 8 months. Multivariable logistic regression analyses identified 5 preoperative and 8 postoperative factors for early recurrence, which were further incorporated into preoperative and postoperative nomograms (C-index 0.785 and 0.834). The calibration plots for the probability of early recurrence fitted well. The nomogram performance was maintained using the validation dataset (C-index 0.777 for preoperative prediction and 0.842 for postoperative prediction). CONCLUSIONS: An interval of 8 months was the optimal threshold for defining early HCC recurrence. The two web-based nomograms have been published to allow accurate pre- and postoperative prediction of early recurrence. These may offer useful guidance for individual treatment or follow up for patients with resectable HCC.