Validated nomogram for the prediction of disease-free survival after hepatectomy for hepatocellular carcinoma within the Milan criteria: individualizing a surveillance strategy

The prognostic role of an optimal machine learning model based on clinical available indicators in HCC patients

Although methods in diagnosis and therapy of hepatocellular carcinoma (HCC) have made significant progress in the past decades, the overall survival (OS) of liver cancer is still disappointing. Machine learning models have several advantages over traditional cox models in prognostic prediction. This study aimed at designing an optimal panel and constructing an optimal machine learning model in predicting prognosis for HCC. A total of 941 HCC patients with completed survival data and preoperative clinical chemistry and immunology indicators from two medical centers were included. The OCC panel was designed by univariate and multivariate cox regression analysis. Subsequently, cox model and machine-learning models were established and assessed for predicting OS and PFS in discovery cohort and internal validation cohort. The best OCC model was validated in the external validation cohort and analyzed in different subgroups. In discovery, internal and external validation cohort, C-indexes of our optimal OCC model were 0.871 (95% CI, 0.863-0.878), 0.692 (95% CI, 0.667-0.717) and 0.648 (95% CI, 0.630-0.667), respectively; the 2-year AUCs of OCC model were 0.939 (95% CI, 0.920-0.959), 0.738 (95% CI, 0.667-0.809) and 0.725 (95% CI, 0.643-0.808), respectively. For subgroup analysis of HCC patients with HBV, aged less than 65, cirrhosis or resection as first therapy, C-indexes of our optimal OCC model were 0.772 (95% CI, 0.752-0.792), 0.769 (95% CI, 0.750-0.789), 0.855 (95% CI, 0.846-0.864) and 0.760 (95% CI, 0.741-0.778), respectively. In general, the optimal OCC model based on RSF algorithm shows prognostic guidance value in HCC patients undergoing individualized treatment.

A new method for predicting the microvascular invasion status of hepatocellular carcinoma through neural network analysis

AIMS

To determine the relationship between microvascular invasion (MVI) and the clinical features of hepatocellular carcinoma (HCC) and provide a method to evaluate MVI status by neutral network analysis.

METHODS

The patients were divided into two groups (MVI-positive group and MVI-negative group). Univariate analysis and multivariate logistic regression analysis were carried out to identify the independent risk factors for MVI positivity. Neural network analysis was used to analyze the different importance of the risk factors in MVI prediction.

RESULTS

We enrolled 1697 patients in this study. We found that the independent prognostic factors were age, NEU, multiple tumors, AFP level and tumor diameter. By neural network analysis, we proposed that the level of AFP was the most important risk factor for HCC in predicting MVI status (the AUC was 0.704). However, age was the most important risk factor for early-stage HCC with a single tumor (the AUC was 0.605).

CONCLUSION

Through the neutral network analysis, we could conclude that the level of AFP is the most important risk factor for MVI-positive patients and the age is the most important risk factor for early-stage HCC with a single tumor.

The appropriate method of hepatectomy for hepatocellular carcinoma within University of California San Francisco (UCSF) criteria through neural network analysis

BACKGROUND

This study aimed to find effective treatments for the patient within UCSF criteria.

METHODS

This study enrolled 1006 patients meeting UCSF criteria, undergoing hepatic resection (HR), divided into two groups: single tumor group and multiple tumors group. We compared and analyzed the risk factors between these two groups' long-term outcomes, through log-rank test, cox proportional hazards model and using neural network analysis to identify the independent risk factors.

RESULTS

The 1-, 3-, and 5-year OS rates in single tumor were significantly higher than multiple tumors (95.0%, 73.2% and 52.3% versus 93.9%, 69.7% and 38.0%, respectively, p < 0.001). The 1-, 3- and 5-year RFS rates were 90.3%, 60.7%, and 40.1% in single tumor and 83.4%, 50.7% and 23.8% in multiple tumors, respectively (p < 0.001). And tumor type, anatomic resection and MVI were the independent risk factors for the patient within UCSF criteria. MVI was the most important risk factor affecting OS and RFS rates in neural network analysis. The method of hepatic resection and the number of tumors were also affected OS and RFS rates.

CONCLUSION

The anatomic resection should be applied to the patient within UCSF criteria, especially for the patient was in single tumor with MVI-negative.

A Novel Nomogram to Predict Prolonged Survival After Hepatectomy in Repeat Recurrent Hepatocellular Carcinoma

Background: Repeat hepatectomy is an important treatment for patients with repeat recurrent hepatocellular carcinoma (HCC).

Methods: This study was a multicenter retrospective analysis of 1,135 patients who underwent primary curative liver resection for HCC. One hundred recurrent patients with second hepatectomy were included to develop a nomogram to predict the risk of post-recurrence survival (PRS). Thirty-eight patients in another institution were used to externally validate the nomogram. Univariate and multivariate Cox regression analyses were used to identify independent risk factors of PRS. Discrimination, calibration, and the Kaplan–Meier curves were used to evaluate the model performance.

Results: The nomogram was based on variables associated with PRS after HCC recurrence, including the tumor, node, and metastasis (TNM) stage; albumin and aspartate aminotransferase levels at recurrence; tumor size, site, differentiation of recurrences; and time to recurrence (TTR). The discriminative ability of the nomogram, as indicated by the C statistics (0.758 and 0.811 for training cohort and external validation cohorts, respectively), was shown, which was better than that of the TNM staging system (0.609 and 0.609, respectively). The calibration curves showed ideal agreement between the prediction and the real observations. The area under the curves (AUCs) of the training cohort and external validation cohorts were 0.843 and 0.890, respectively. The Kaplan–Meier curve of the established nomogram also performed better than those of both the TNM and the BCLC staging systems.

Conclusions: We constructed a nomogram to predict PRS in patients with repeat hepatectomy (RH) after repeat recurrence of HCC.

15-lncRNA-Based Classifier-Clinicopathologic Nomogram Improves the Prediction of Recurrence in Patients with Hepatocellular Carcinoma

Background

Our study aims to develop a lncRNA-based classifier and a nomogram incorporating the genomic signature and clinicopathologic factors to help to improve the accuracy of recurrence prediction for hepatocellular carcinoma (HCC) patients.

Methods

The lncRNA profiling data of 374 HCC patients and 50 normal healthy controls were downloaded from The Cancer Genome Atlas (TCGA). Using univariable Cox regression and least absolute shrinkage and selection operator (LASSO) analysis, we developed a 15-lncRNA-based classifier and compared our classifier to the existing six-lncRNA signature. Besides, a nomogram incorporating the genomic classifier and clinicopathologic factors was also developed. The predictive accuracy and discriminative ability of the genomic-clinicopathologic nomogram were determined by a concordance index (C-index) and calibration curve and were compared with the TNM staging system by the C-index and receiver operating characteristic (ROC) analysis. Decision curve analysis (DCA) was performed to estimate the clinical value of our nomogram.

Results

Fifteen relapse-free survival (RFS-) related lncRNAs were identified, and the classifier, consisting of the identified 15 lncRNAs, could effectively classify patients into the high-risk and low-risk subgroups. The prediction accuracy of the 15-lncRNA-based classifier for predicting 2-year and 5-year RFS was 0.791 and 0.834 in the training set and 0.684 and 0.747 in the validation set, respectively, which was better than the existing six-lncRNA signature. Moreover, the AUC of genomic-clinicopathologic nomogram in predicting RFS were 0.837 in the training set and 0.753 in the validation set, and the C-index of the genomic-clinicopathologic nomogram was 0.78 (0.72-0.83) in the training set and 0.71 (0.65-0.76) in the validation set, which was better than the traditional TNM stage and 15-lncRNA-based classifier. The decision curve analysis further demonstrated that our nomogram had a larger net benefit than the TNM stage and 15-lncRNA-based classifier. The results were confirmed externally.

Conclusion

Compared to the TNM stage, the 15-lncRNAs-based classifier-clinicopathologic nomogram is a more effective and valuable tool to identify HCC recurrence and may aid in clinical decision-making.

Novel Models Predict Postsurgical Recurrence and Overall Survival for Patients with Hepatitis B Virus-Related Solitary Hepatocellular Carcinoma ≤10 cm and Without Portal Venous Tumor Thrombus

BACKGROUND

The predictive model of postsurgical recurrence for solitary early hepatocellular carcinoma (SE-HCC) is not well established. The aim of this study was to develop a novel model for prediction of postsurgical recurrence and survival for patients with hepatitis B virus (HBV)-related SE-HCC ≤10 cm.

PATIENTS AND METHODS

Data from 1,081 patients with HBV-related SE-HCC ≤10 cm who underwent curative liver resection from 2003 to 2016 in our center were collected retrospectively and randomly divided into the derivation cohort (n = 811) and the internal validation cohort (n = 270). Eight hundred twenty-three patients selected from another four tertiary hospitals served as the external validation cohort. Postsurgical recurrence-free survival (RFS) and overall survival (OS) predictive nomograms were generated. The discriminatory accuracies of the nomograms were compared with six conventional hepatocellular carcinoma (HCC) staging systems.

RESULTS

Tumor size, differentiation, microscopic vascular invasion, preoperative α-fetoprotein, neutrophil-to-lymphocyte ratio, albumin-to-bilirubin ratio, and blood transfusion were identified as the risk factors associated with RFS and OS. RFS and OS predictive nomograms based on these seven variables were generated. The C-index was 0.83 (95% confidence interval [CI], 0.79-0.87) for the RFS-nomogram and 0.87 (95% CI, 0.83-0.91) for the OS-nomogram. Calibration curves showed good agreement between actual observation and nomogram prediction. Both C-indices of the two nomograms were substantially higher than those of the six conventional HCC staging systems (0.54-0.74 for RFS; 0.58-0.76 for OS) and those of HCC nomograms reported in literature.

CONCLUSION

The novel nomograms were shown to be accurate at predicting postoperative recurrence and OS for patients with HBV-related SE-HCC ≤10 cm after curative liver resection.

IMPLICATIONS FOR PRACTICE

This multicenter study proposed recurrence or mortality predictive nomograms for patients with hepatitis B virus-related solitary early hepatocellular carcinoma ≤10 cm after curative liver resection. A close postsurgical surveillance protocol and adjuvant therapy should be considered for patients at high risk of recurrence.

Establishment of a prognostic model for predicting short-term disease-free survival in cases of hepatitis B-related hepatocellular carcinoma with the TP53 249Ser mutation in southern China

Background

Hepatitis B virus (HBV) infection and dietary aflatoxin exposure are two major and synergistic carcinogenic factors of hepatocellular carcinoma (HCC) in southern China. Mutation of the TP53 gene at codon 249 (TP53 249Ser) is recognized as a fingerprint of aflatoxin B1 (AFB1) exposure.

Methods

A total of 485 HCC patients positive for serum hepatitis B surface antigen were enrolled. The clinicopathological information and survival time were collected. TP53 249Ser mutations in HCC were detected by Sanger DNA sequencing after PCR amplification. Immunohistochemical staining was used to evaluate TP53 expression. Propensity score matching (PSM) and Cox proportional hazards regression (CPHR) were conducted to identify independent risk factors for prognosis that were incorporated into the nomogram. Univariate logistic regression analysis was used to compare differences in clinical factors between the TP53 249Ser mutation group and the non-mutation group. A Kaplan-Meier plot, univariate and multivariate Cox proportional hazards models were used to assess the association between clinicopathological characteristics and survival outcomes.

Results

After PSM, a total of 322 cases were included in the analysis of clinical prognosis. Results of CPHR showed that the mutation group had a relatively higher risk of tumor recurrence within 2 years after undergoing hepatectomy (P=0.039, HR =1.47, 95% CI: 1.02–2.18). The prognostic model performed better in terms of 2-year DFS prediction than BCLC stage. Patients who had a nomogram score of more than 160 were considered to have a higher risk of recurrence within 2 years.

Conclusions

Our study found that the TP53 249Ser mutation may be a high risk factor of HBV-related HCC recurrence in the short term. And we initially established a nomogram scoring system for predicting 2-year recurrence in HBV-related HCC patients in southern China.