Background and Aims
Accurate risk stratification for gastric cancer is required for optimal endoscopic
surveillance in patients with chronic gastritis. We aimed to develop a machine learning
(ML) model that incorporates endoscopic and histologic findings for an individualized
prediction of gastric cancer incidence.
Methods
We retrospectively evaluated 1099 patients with chronic gastritis who underwent EGD
and biopsy sampling of the gastric mucosa. Patients were randomly divided into training
and test sets (4:1). We constructed a conventional Cox proportional hazard model and
3 ML models. Baseline characteristics, endoscopic atrophy, and Operative Link on Gastritis-Intestinal
Metaplasia Assessment (OLGIM)/Operative Link on Gastritis Assessment (OLGA) stage
at initial EGD were comprehensively assessed. Model performance was evaluated using
Harrel’s c-index.
Results
During a mean follow-up of 5.63 years, 94 patients (8.55%) developed gastric cancer.
The gradient-boosting decision tree (GBDT) model achieved the best performance (c-index
from the test set, .84) and showed high discriminative ability in stratifying the
test set into 3 risk categories (P < .001). Age, OLGIM/OLGA stage, endoscopic atrophy, and history of malignant tumors
other than gastric cancer were important predictors of gastric cancer incidence in
the GBDT model. Furthermore, the proposed GBDT model enabled the generation of a personalized
cumulative incidence prediction curve for each patient.
Conclusions
We developed a novel ML model that incorporates endoscopic and histologic findings
at initial EGD for personalized risk prediction of gastric cancer. This model may
lead to the development of effective and personalized follow-up strategies after initial
EGD.
Graphical abstract
Graphical Abstract
Abbreviations:
CPH (Cox proportional hazard), GBDT (gradient-boosting decision tree), OLGA (Operative Link on Gastritis Assessment), OLGIM (Operative Link on Gastritis-Intestinal Metaplasia Assessment), ML (machine learning)To read this article in full you will need to make a payment
Purchase one-time access:
Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online accessOne-time access price info
- For academic or personal research use, select 'Academic and Personal'
- For corporate R&D use, select 'Corporate R&D Professionals'
Subscribe:
Subscribe to Gastrointestinal EndoscopyAlready a print subscriber? Claim online access
Already an online subscriber? Sign in
Register: Create an account
Institutional Access: Sign in to ScienceDirect
References
- Global, regional, and national cancer incidence, mortality, years of life lost, years lived with disability, and disability-adjusted life-years for 29 cancer groups, 1990 to 2017: a systematic analysis for the Global Burden of Disease Study.JAMA Oncol. 2019; 5: 1749-1768
- Gastric cancer.Lancet. 2020; 396: 635-648
- Guidelines for endoscopic diagnosis of early gastric cancer.Dig Endosc. 2020; 32: 663-698
- The role of endoscopy in the management of premalignant and malignant conditions of the stomach.Gastrointest Endosc. 2015; 82: 1-8
- Prediction model for gastric cancer incidence in Korean population.PloS One. 2015; 10e0132613
- Development and validation of a prediction rule for estimating gastric cancer risk in the Chinese high-risk population: a nationwide multicentre study.Gut. 2019; 68: 1576-1587
- Applications of machine learning models in the prediction of gastric cancer risk in patients after Helicobacter pylori eradication.Aliment Pharmacol Ther. 2021; 53: 864-872
- Machine-learning approach for the development of a novel predictive model for the diagnosis of hepatocellular carcinoma.Sci Rep. 2019; 9: 7704
- Validation of a machine learning model that outperforms clinical risk scoring systems for upper gastrointestinal bleeding.Gastroenterology. 2020; 158: 160-167
- Helicobacter pylori infection and the development of gastric cancer.N Engl J Med. 2001; 345: 784-789
- Histologic intestinal metaplasia and endoscopic atrophy are predictors of gastric cancer development after Helicobacter pylori eradication.Gastrointest Endosc. 2016; 84: 618-624
- Predictive model for gastric cancer after eradication of Helicobacter pylori—a survival analysis using a deep learning algorithm [letter].Aliment Pharmacol Ther. 2021; 54: 528-529
- Association between gastric cancer and the Kyoto classification of gastritis.J Gastroenterol Hepatol. 2017; 32: 1581-1586
- A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.J Chronic Dis. 1987; 40: 373-383
- Nonsteroidal anti-inflammatory drugs prevent gastric cancer associated with the use of proton pump inhibitors after Helicobacter pylori eradication.JGH Open. 2021; 5: 770-777
- Distinct chemopreventive effects of aspirin in diffuse and intestinal-type gastric cancer.Cancer Prev Res. 2018; 11: 279-286
- Long-term proton pump inhibitor use is a risk factor of gastric cancer after treatment for Helicobacter pylori: a retrospective cohort analysis.Gut. 2018; 67: 1908-1910
- An endoscopic recognition of the atrophic border and its significance in chronic gastritis.Endoscopy. 1969; 1: 87-97
- The staging of gastritis with the OLGA system by using intestinal metaplasia as an accurate alternative for atrophic gastritis.Gastrointest Endosc. 2010; 71: 1150-1158
- Gastritis staging in clinical practice: the OLGA staging system.Gut. 2007; 56: 631-636
Arai J, Niikura R, Hayakawa Y, et al. OLGIM staging and proton pump inhibitor use predict the risk of gastric cancer. Gut. Epub 2021 Aug 3.
- DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network.BMC Med Res Methodol. 2018; 18: 24
- Gradient boosting machines, a tutorial.Front Neurorobot. 2013; 7: 21
- Extreme gradient boosting as a method for quantitative structure-activity relationships.J Chem Inf Model. 2016; 56: 2353-2360
- The nature of statistical learning theory.New York: Springer Science and Business Media, 1995
- Support vector machines.IEEE Intell Syst Their Appl. 1998; 13: 18-28
- PySurvival: open source package for survival analysis modeling, 2019.(Available at:)https://square.github.io/pysurvival/Date accessed: February 16, 2022
- scikit-survival: A Library for Time-to-Event Analysis Built on Top of scikit-learn.Journal of Machine Learning Research. 2020; 21: 1-6
- Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors.Stat Med. 1996; 15: 361-387
- All models are wrong, but many are useful: learning a variable’s importance by studying an entire class of prediction models simultaneously.J Mach Learn Res. 2019; 20: 1-81
- Linear model: overview.Wiley STATSRef Stat Ref Online, 2014 (Hoboken: John Wiley & Sons. Available at:)https://onlinelibrary.wiley.com/doi/10.1002/9781118445112.stat06948Date accessed: February 16, 2022
- Data mining with big data.IEEE Trans Knowl Data Eng. 2014; 26: 97-107
- Data-based identifiability analysis of non-linear dynamical models.Bioinformatics. 2007; 23: 2612-2618
- Machine learning in medicine.N Engl J Med. 2019; 380: 1347-1358
- Predictive accuracy of novel risk factors and markers: a simulation study of the sensitivity of different performance measures for the Cox proportional hazards regression model.Stat Methods Med Res. 2017; 26: 1053-1077
- Comparing three machine learning approaches to design a risk assessment tool for future fractures: predicting a subsequent major osteoporotic fracture in fracture patients with osteopenia and osteoporosis.Osteoporos Int. 2021; 32: 437-449
- Risk factors: new risk-assessment guidelines—more or less personalized?.Nat Rev Cardiol. 2014; 11: 136-137
- Epidemiology of gastric cancer in Japan.Postgrad Med J. 2005; 81: 419-424
Article info
Publication history
Published online: January 05, 2022
Accepted:
December 29,
2021
Received:
September 9,
2021
Footnotes
DISCLOSURE: The following author received research support for this study from a KAKENHI Grant-in-Aid for Scientific Research (grant nos. 20H03656 and 21K19474), P-CREATE from AMED, Japan, and the Advanced Research and Development Programs for Medical Innovation (PRIME), Japan: Y. Hayakawa. In addition, the following author disclosed financial relationships: M. Fujishiro: Speaker for Olympus and Fujifilm; research grant from Olympus, HOYA Pentax, and Fujifilm. Y. Tsuji: Lecture fees and research grant from Olympus; research grants from HOYA Pentax, Gunze, Nipro, and AI Medical. All other authors disclosed no financial relationships.
DIVERSITY, EQUITY, AND INCLUSION: We worked to ensure gender balance in the recruitment of human subjects. We worked to ensure ethnic or other types of diversity in the recruitment of human subjects. We worked to ensure that the language of the study questionnaires reflected inclusion. We worked to ensure sex balance in the selection of nonhuman subjects
If you would like to chat with an author of this article, you may contact Dr Hayakawa at [email protected] or Dr Aoki at [email protected]
Identification
Copyright
© 2022 by the American Society for Gastrointestinal Endoscopy
ScienceDirect
Access this article on ScienceDirectLinked Article
- What is the clinical value of prediction models in the management of gastric cancer?Gastrointestinal EndoscopyVol. 96Issue 1
