Hepatobiliary & Pancreatic Diseases International最新文献

Background: The rate of distant metastasis in patients with pancreatic neuroendocrine tumors (PNETs) is 20%-50% at the time of initial diagnosis. However, whether tumor size can predict distant metastasis for PNETs remains unknown up to date.

Methods: We used Surveillance, Epidemiology, and End Results (SEER) population-based data to collect 6089 patients with PNETs from 2010 to 2019. The optimal cut-off point of tumor size to predict distant metastasis was calculated by Youden's index. Multivariate logistic regression analysis was used to figure out the association between tumor size and distant metastasis patterns.

Results: The most common metastatic site was liver (27.2%), followed by bone (3.0%), lung (2.3%) and brain (0.4%). Based on an optimal cut-off value of tumor size (25.5 mm) for predicting distant metastasis determined by Youden's index, patients were categorized into groups of tumor size < 25.5 mm and ≥ 25.5 mm. Multivariate logistic regression analyses showed that, compared with < 25.5 mm, tumor size ≥ 25.5 mm was an independent risk predictor of overall distant metastasis [odds ratio (OR) = 4.491, 95% confidence interval (CI): 3.724-5.416, P < 0.001] and liver metastasis (OR = 4.686, 95% CI: 3.886-5.651, P < 0.001).

Conclusions: Tumor size ≥ 25.5 mm was significantly associated with more overall distant and liver metastases. Timely identification of distant metastasis for tumor size ≥ 25.5 mm may provide survival benefit for timely and precise treatment.

Background: Liver transplantations (LTs) with extended criteria have produced surgical results comparable to those obtained with traditional standards. However, it is not sufficient to predict hepatocellular carcinoma (HCC) recurrence after LT according to morphological criteria alone. The present study aimed to construct a nomogram for predicting HCC recurrence after LT using extended selection criteria.

Methods: Retrospective data on patients with HCC, including pathology, serological markers and follow-up data, were collected from January 2015 to April 2020 at Huashan Hospital, Fudan University, Shanghai, China. Logistic least absolute shrinkage and selection operator (LASSO) regression and multivariate Cox regression analyses were performed to identify and construct the prognostic nomogram. Receiver operating characteristic (ROC) curves, Kaplan-Meier curves, decision curve analyses (DCAs), calibration diagrams, net reclassification indices (NRIs) and integrated discrimination improvement (IDI) values were used to assess the prognostic capacity of the nomogram.

Results: A total of 301 patients with HCC who underwent LT were enrolled in the study. The nomogram was constructed, and the ROC curve showed good performance in predicting survival in both the development set (2/3) and the validation set (1/3) (the area under the curve reached 0.748 and 0.716, respectively). According to the median value of the risk score, the patients were categorized into the high- and low-risk groups, which had significantly different recurrence-free survival (RFS) rates (P < 0.01). Compared with the Milan criteria and University of California San Francisco (UCSF) criteria, DCA revealed that the new nomogram model had the best net benefit in predicting 1-, 3- and 5-year RFS. The nomogram performed well for calibration, NRI and IDI improvement.

Conclusions: The nomogram, based on the Milan criteria and serological markers, showed good accuracy in predicting the recurrence of HCC after LT using extended selection criteria.