Dataset Information

Development and validation of a simple risk model to predict major cancers for patients with nonalcoholic fatty liver disease.

ABSTRACT:

Objective

To recognize risk factors and build up and validate a simple risk model predicting 8-year cancer events after nonalcoholic fatty liver disease (NAFLD).

Methods

This was a retrospective cohort study. Patients with NAFLD (n = 5561) were randomly divided into groups: training (n = 1254), test (n = 627), evaluation (n = 627), and validation (n = 3053). Risk factors were recognized by statistical method named as a Cox model with Markov chain Monte Carlo (MCMC) simulation. This prediction score was established based on the training group and was further validated based on the testing and evaluation group from January 1, 2007 to December 31, 2009 and another 3053 independent cases from January 1, 2010 to February 13, 2014.

Results

The main outcomes were NAFLD-related cancer events, including those of the liver, breast, esophagus, stomach, pancreas, prostate and colon, within 8 years after hospitalization for NAFLD diagnosis. Seven risk factors (age (every 5 years),LDL, smoking, BMI, diabetes, OSAS, and aspartate aminotransferase (every 5 units)) were identified as independent indicators of cancer events. This risk model contained a predictive range of 0.4%-37.7%, 0.3%-39.6%, and 0.4%-39.3% in the training, test, evaluation group, respectively, with a range 0.4%-30.4% for validation groups. In the training group, 12.6%, 76.9%, and 10.5% of patients, which corresponded to the low -, moderate -, and high-risk groups, had probabilities of, <0.01, <0.1, and 0.23 for 8-year events.

Conclusions

Seven risk factors were recognized and a simple risk model were developed and validated to predict the risk of cancer events after NAFLD based on 8 years. This simple risk score system may recognize high-risk patients and reduce cancer incidence.

SUBMITTER: Wei Z

PROVIDER: S-EPMC6997093 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Some patients with nonalcoholic fatty liver disease (NAFLD) develop liver-related complications and have higher mortality than other patients with NAFLD. We determined the accuracy of simple, noninvasive scoring systems in identification of patients at increased risk for liver-related complications or death.We performed a retrospective, international, multicenter cohort study of 320 patients diagnosed with NAFLD, based on liver biopsy analysis through 2002 and followed through 2011. Patients were assigned to mild-, intermediate-, or high-risk groups based on cutoff values for 2 of the following: NAFLD fibrosis score, aspartate aminotransferase/platelet ratio index, FIB-4 score, and BARD score. Outcomes included liver-related complications and death or liver transplantation. We used multivariate Cox proportional hazard regression analysis to adjust for relevant variables and calculate adjusted hazard ratios (aHRs).During a median follow-up period of 104.8 months (range, 3-317 months), 14% of patients developed liver-related events and 13% died or underwent liver transplantation. The aHRs for liver-related events in the intermediate-risk and high-risk groups, compared with the low-risk group, were 7.7 (95% confidence interval [CI]: 1.4-42.7) and 34.2 (95% CI: 6.5-180.1), respectively, based on NAFLD fibrosis score; 8.8 (95% CI: 1.1-67.3) and 20.9 (95% CI: 2.6-165.3) based on the aspartate aminotransferase/platelet ratio index; and 6.2 (95% CI: 1.4-27.2) and 6.6 (95% CI: 1.4-31.1) based on the BARD score. The aHRs for death or liver transplantation in the intermediate-risk and high-risk groups compared with the low-risk group were 4.2 (95% CI: 1.3-13.8) and 9.8 (95% CI: 2.7-35.3), respectively, based on the NAFLD fibrosis scores. Based on aspartate aminotransferase/platelet ratio index and FIB-4 score, only the high-risk group had a greater risk of death or liver transplantation (aHR = 3.1; 95% CI: 1.1-8.4 and aHR = 6.6; 95% CI: 2.3-20.4, respectively).Simple noninvasive scoring systems help identify patients with NAFLD who are at increased risk for liver-related complications or death. NAFLD fibrosis score appears to be the best indicator of patients at risk, based on HRs. The results of this study require external validation.

Project description:ImportanceMachine-learning algorithms offer better predictive accuracy than traditional prognostic models but are too complex and opaque for clinical use.ObjectiveTo compare different machine learning methods in predicting overall mortality in cirrhosis and to use machine learning to select easily scored clinical variables for a novel cirrhosis prognostic model.Design, setting, and participantsThis prognostic study used a retrospective cohort of adult patients with cirrhosis or its complications seen in 130 hospitals and affiliated ambulatory clinics in the integrated, national Veterans Affairs health care system from October 1, 2011, to September 30, 2015. Patients were followed up through December 31, 2018. Data were analyzed from October 1, 2017, to May 31, 2020.ExposuresPotential predictors included demographic characteristics; liver disease etiology, severity, and complications; use of health care resources; comorbid conditions; and comprehensive laboratory and medication data. Patients were randomly selected for model development (66.7%) and validation (33.3%). Three different statistical and machine learning methods were evaluated: gradient descent boosting, logistic regression with least absolute shrinkage and selection operator (LASSO) regularization, and logistic regression with LASSO constrained to select no more than 10 predictors (partial pathway model). Predictor inclusion and model performance were evaluated in a 5-fold cross-validation. Last, the predictors identified in the most parsimonious (the partial path) model were refit using maximum-likelihood estimation (Cirrhosis Mortality Model [CiMM]), and its predictive performance was compared with that of the widely used Model for End Stage Liver Disease with sodium (MELD-Na) score.Main outcomes and measuresAll-cause mortality.ResultsOf the 107 939 patients with cirrhosis (mean [SD] age, 62.7 [9.6] years; 96.6% male; 66.3% white, 18.4% African American), the annual mortality rate ranged from 8.8% to 15.3%. In total, 32.7% of patients died within 3 years, and 46.2% died within 5 years after the index date. Models predicting 1-year mortality had good discrimination for the gradient descent boosting (area under the receiver operating characteristics curve [AUC], 0.81; 95% CI, 0.80-0.82), logistic regression with LASSO regularization (AUC, 0.78; 95% CI, 0.77-0.79), and the partial path logistic model (AUC, 0.78; 95% CI, 0.76-0.78). All models showed good calibration. The final CiMM model with machine learning-derived clinical variables offered significantly better discrimination than the MELD-Na score, with AUCs of 0.78 (95% CI, 0.77-0.79) vs 0.67 (95% CI, 0.66-0.68) for 1-year mortality, respectively (DeLong z = 17.00; P < .001).Conclusions and relevanceIn this study, simple machine learning techniques performed as well as the more advanced ensemble gradient boosting. Using the clinical variables identified from simple machine learning in a cirrhosis mortality model produced a new score more transparent than machine learning and more predictive than the MELD-Na score.