Case Study
Non-alcoholic steatohepatitis (NASH)
as a risk factor
in hepatocellular carcinoma (HCC)
development
The Question
HCC is a leading cause of mortality and morbidity worldwide that often occurs in conjunction with other liver diseases. It is suspected that, due to underdiagnosis, poor surveillance, and a lack of treatment options, NASH patients present with more severe liver disease and are at a higher risk of developing HCC sooner, but this relationship is not well understood.
The Approach
The Medicus team used SEER-Medicare data to estimate the differential time-to- and time-varying probability of HCC development between NASH and non-NASH patients. An extension of the traditional Kaplan-Meier (KM) estimator using inverse probability of treatment weighting (IPTW) to achieve covariate balance was implemented.
Identifying assumptions were tested and alternative approaches to confounder adjustment, including direct standardization through g-computation and doubly robust estimation via augmented IPTW KM, were explored in sensitivity analyses. Differences in average time-to-HCC development were derived from survival curves based on adjusted restricted mean survival time. Standard errors and 95% confidence intervals (CI) were generated via bootstrap resampling.
The Results
A total of 12,610 HCC patients were included in the final sample, 229 with NASH and 12,381 non-NASH (other liver or metabolic disease) patients. Of the 229 NASH patients, 27% (n=62) had confirmed cirrhosis at HCC diagnosis compared to 35% (n=4,339) of non-NASH patients. Median time-to-HCC was 1.13 years (95% CI: 0.60 to 1.71) in NASH patients and occurred an average of 1.88 years (95% CI: 1.41 to 2.35) faster than in non-NASH patients.
This difference was more pronounced in the five years following index liver disease diagnosis, with the probability of developing HCC in NASH patients remaining at least 20-percentage points higher than non-NASH patients during this time.
The Long And
Short Of it
By applying advanced survival methods to SEER-Medicare data, we provided novel evidence on the comparative risk of HCC by liver disease etiology. Results suggesting that NASH patients develop HCC sooner than those with other underlying liver diseases highlight the potential role of both expanded surveillance and new therapies that can delay progression in reducing HCC disease burden in this population.