A Simple Blood Test Score Could Catch Pre-Cancer in Stubborn H. Pylori Infections

When H. pylori survives two consecutive rounds of standard eradication therapy, the infection earns the label "refractory" and the stomach pays a steep price. Persistent bacterial activity keeps the gastric lining in a state of chronic inflammatory stress, and that stress is one of the clearest known pathways toward gastric cancer. A new study published in Cancer Medicine by Jiang TT and colleagues now offers the first disease-specific tool for identifying which of these hard-to-treat patients are most likely to develop chronic atrophic gastritis (CAG), the precancerous condition that sits squarely on the road from inflammation to malignancy.

Chronic Atrophic Gastritis Rate Reaches 64% in Refractory H. Pylori Patients

The numbers from this retrospective cohort are sobering. Among 367 patients with refractory H. pylori infection (RHPI) enrolled between January 2015 and December 2024, 236 (64.3%) already had chronic atrophic gastritis at the time of evaluation. That figure sits well above the 10% to 60% range estimated for the general population, suggesting that failed eradication meaningfully accelerates gastric mucosal damage. CAG is not simply a nuisance diagnosis; it is the most important precancerous background for gastric dysplasia and eventual malignancy. Gastric cancer accounts for roughly 1.089 million new cases and 769,000 deaths per year worldwide, and projections place new annual cases at 1.82 million by 2050 as populations age. Within that global burden, H. pylori-driven chronic gastric inflammation remains the dominant modifiable risk factor, with infection rates approaching 50% across East Asia alone.

Four Accessible Variables Predict Who Develops Pre-Cancer

Using univariate and multivariate logistic regression alongside LASSO variable selection, the researchers narrowed a broad field of clinical indicators down to four independent risk factors for CAG in RHPI patients. Older age carried an odds ratio of 1.05, meaning each additional year of life modestly but consistently raised CAG risk, likely because aging immune systems shift toward chronic low-grade inflammation that accelerates intestinal metaplasia. Smoking came in with an odds ratio of 2.37, the strongest single factor, reflecting its capacity to elevate TNF-alpha and IL-10, drive gastric mucosal cell proliferation, and blunt immune surveillance. Elevated CA199, a tumor-associated marker that rises when inflammatory immune responses upregulate its expression on abnormal cells, carried an odds ratio of 1.12. The fourth factor was the lymphocyte ratio (LYMPH%), with an odds ratio of 1.08, a variable the authors describe as rarely reported in this context before. The underlying mechanism they propose is that refractory H. pylori activates NOD receptors, recruits lymphocytes and myeloid cells, and pushes macrophages toward an M2 polarization state associated with immune suppression and tissue remodeling rather than bacterial clearance.

These four variables were combined into a nomogram, a visual scoring tool that clinicians can complete at the bedside using routine blood work and patient history, with no additional invasive procedures required. The model achieved an area under the ROC curve (AUC) of 0.833, with sensitivity of 94.50% and specificity of 78.40%. Calibration curves and decision curve analysis confirmed both good statistical fit and real clinical utility across a range of probability thresholds.

How the Team Validated the Inflammatory Mechanism in the Lab

The clinical model alone would not explain why RHPI is particularly dangerous, so the team ran parallel in vitro experiments using both wild-type and refractory H. pylori strains in contact with GES-1 gastric mucosal epithelial cells. Refractory strains inhibited cell growth more aggressively than wild strains, and the effect was dose-dependent, meaning higher bacterial concentrations produced proportionally greater damage. The strains also drove measurable increases in TNF-alpha, IL-10, IL-33, TGF-beta, and VEGF, while Notch1 expression declined. The elevated IL-33 points to activation of what the authors call the innate immune alarm system, while the TGF-beta and VEGF rises suggest simultaneous promotion of fibrosis and tumor-associated angiogenesis through the Notch signaling pathway. Taken together, the lab data support the clinical observation that refractory infections create an immunosuppressive, pro-fibrotic microenvironment that ordinary eradication-responsive infections do not. This kind of inflammatory biomarker profiling is an active frontier; inflammatory ratios and cytokine signatures are increasingly studied as both prognostic and diagnostic tools in gastric cancer research, though standardization remains a challenge.

Limitations That Come With a Single-Center Retrospective Design

The study's authors are candid about where the model needs more work. The cohort is drawn from a single center, which limits generalizability and introduces the possibility of selection bias. Important potential confounders were not captured, including specific comorbidities such as diabetes or autoimmune conditions and medications like NSAIDs or statins that could independently influence gastric inflammation or CAG risk. The CAG diagnoses were also not stratified by severity, so the model predicts presence or absence of the condition rather than distinguishing mild from severe disease, which would matter for prioritizing endoscopic surveillance intensity. These gaps are not trivial. Current diagnostic tools for gastric pre-cancer and early cancer, including endoscopy and traditional serum markers, already struggle with sensitivity and are often applied too late, so a single-center nomogram entering that space needs multicenter prospective validation before it can be broadly trusted.

Next Steps Toward a Dynamic, Multicenter Prediction System

The investigators call for multicenter prospective cohorts drawn from diverse geographic regions to externally validate the model, a necessary step given both the single-center limitation and the substantial variation in H. pylori strains and dietary patterns across East Asia and beyond. They also flag the need for subgroup analyses stratified by CAG severity, which would let the tool do more than binary triage. Longer term, they envision dynamic prediction systems that incorporate longitudinal data, updating a patient's risk score over time as their inflammatory markers, smoking status, or treatment history changes rather than capturing a single clinical snapshot. For now, the nomogram offers something genuinely useful within its validated scope: a non-invasive, clinically actionable score that could help gastroenterologists decide which refractory H. pylori patients need closer endoscopic follow-up sooner rather than later.

Evidence rating

EMERGING — 80/100

Caveats:

• Single-center retrospective study — selection bias cannot be ruled out and results need external validation in independent populations.
• The model was not externally validated; performance in other clinical settings is unknown.
• Important potential confounders (diabetes, autoimmune disease, NSAIDs, statins) were not accounted for.
• CAG severity was not stratified, limiting granularity of risk predictions.
• Funding source was not disclosed by the authors.

Retrospective single-center cohort (n=367) using logistic regression and LASSO to build a nomogram predicting chronic atrophic gastritis in refractory H. pylori-infected patients; AUC 0.833 in the derivation cohort. The design establishes association, not causation, and the model has not been validated in an independent external cohort. In vitro mechanistic experiments support the inflammatory pathway findings but do not substitute for prospective replication.

References

1. Tian‐tong Jiang, Liu‐feng Yi, Ya‐xuan Wang, et al. (2026). Exploring the High‐Risk Factors of Gastric “Inflammatory Cancer Transformation” Based on Inflammatory Immune Response: Construction of RHPI Complicated With CAG Risk Prediction Model. Cancer Medicine. https://doi.org/10.1002/cam4.72006
2. Wen-Juan Yang, He-Ping Zhao, Yan Yu, et al. (2023). Updates on global epidemiology, risk and prognostic factors of gastric cancer. World Journal of Gastroenterology. https://doi.org/10.3748/wjg.v29.i16.2452
3. Inflammation and Gastric Cancer – PMC. NIH. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9218659/
4. Catalin Vladut Ionut Feier, Calin Muntean, Alaviana Monique Faur, et al. (2024). An Exploratory Assessment of Pre-Treatment Inflammatory Profiles in Gastric Cancer Patients. Diseases. https://doi.org/10.3390/diseases12040078
5. Duygu Kirkik, Yesim Uygun Kızmaz, Sevgi Kalkanli Tas (2025). Advancing early detection of gastric cancer through serum cytokine profiling. World Journal of Clinical Oncology. https://doi.org/10.5306/wjco.v16.i10.109736