A groundbreaking study has harnessed the power of electronic health records (EHRs) to predict the risk of noncardia gastric cancer (NCGC), offering a promising tool for earlier detection and improved patient outcomes. Gastric cancer ranks among the top causes of cancer-related deaths globally, often slipping under the radar until it reaches advanced, harder-to-treat stages. This new model, developed by researchers, aims to identify individuals at high risk using data already collected during routine medical care, potentially paving the way for targeted screening that could save lives.
The research team, based at the Cleveland Clinic, conducted a retrospective case-control study by diving into a massive EHR database containing records from over 12 million patients across multiple states. They focused on 614 patients aged 40 to 80 diagnosed with NCGC between 2010 and 2021, comparing them to 6,331 individuals without the disease. Using a statistical approach called logistic regression, enhanced by a technique known as multiple imputation to address gaps in the data, they crafted a model to estimate NCGC risk based on factors like age, sex, race, smoking history, and specific health conditions. The model’s accuracy was assessed with a metric called the 0.632 estimator, ensuring its reliability.
The findings are compelling. The model scored a 0.731 on the 0.632 estimator, a strong indicator of its ability to distinguish between those at risk and those not. It pinpointed key risk factors: for every 10-year increase in age, the odds of NCGC rose by 16% (odds ratio [OR] 1.16, 95% confidence interval [CI] 1.04-1.30). Men faced nearly double the risk (OR 1.97, 95% CI 1.64-2.36), while Black and Asian individuals had over three and four times the odds, respectively (OR 3.07, 95% CI 2.46-3.83; OR 4.39, 95% CI 2.60-7.42), compared to White individuals. Smokers (OR 1.61, 95% CI 1.34-1.94), those with anemia (OR 1.35, 95% CI 1.09-1.68), and those with pernicious anemia (OR 6.12, 95% CI 3.42-10.95) also showed heightened risk.
This model could transform how doctors approach gastric cancer. By identifying high-risk patients early, it opens the door to targeted screening—like endoscopy—that could catch the disease when it’s more treatable. In the U.S., where widespread screening isn’t feasible due to lower incidence rates, this approach could be a cost-effective way to focus efforts on those who need it most, especially minority groups disproportionately affected by NCGC. It’s a step toward reducing cancer disparities and boosting survival rates.
The study’s authors are optimistic about its potential but cautious about next steps. “We demonstrate the feasibility and good performance of an electronic health record-based logistic regression model for estimating the probability of NCGC,” they noted, highlighting its practical foundation in everyday clinical data. However, they stress that refinement and validation through future studies are essential to perfect the tool and define a high-risk group eligible for screening.
Gastric cancer remains a formidable challenge worldwide, claiming numerous lives due to late diagnoses. In the U.S., where screening isn’t routine, this study stands out by leveraging existing health data to tackle a pressing need. It builds on a growing trend of using EHRs to predict disease risk, offering a fresh, accessible strategy to combat a cancer that hits certain populations—like Black and Asian communities—harder. Published in Gastro Hep Advances 2024;3:910-916, this research lights a path toward smarter, more equitable healthcare solutions.