Research progress of machine learning applications in gastric cancer diagnosis and therapy

Clinical Snapshot

Key Findings

• P Value: Not reported at review level

• Effect Size: Not reported — no pooled effect size calculated; individual study metrics (e.g., AUC, sensitivity, specificity for diagnostic ML models) may be cited narratively but are not synthesised

• Primary Outcome: Narrative overview of machine learning applications in gastric cancer diagnosis (including endoscopic image analysis, pathological diagnosis, radiomics) and therapy (treatment selection, prognosis prediction, drug response modelling) — no primary quantitative outcome is defined or reported

• Nnt Or Sensitivity: Not calculable from this review — individual ML model performance metrics from primary studies (e.g., AUC values for endoscopic AI systems) are referenced narratively but no summary diagnostic accuracy statistics are provided

• Confidence Interval: Not reported

Clinical Application

Clinical implementation of ML tools in GC diagnosis remains variable. Endoscopic AI systems (e.g., for polyp/lesion detection) are commercially available in some jurisdictions. Pathology AI platforms are in early clinical adoption. Therapeutic decision-support ML tools remain largely investigational. Infrastructure requirements (digital pathology, AI-enabled endoscopy platforms, validated software), regulatory approval, and clinician training represent significant implementation barriers in most healthcare systems. Gastric cancer is not among Australia's most prevalent cancers (approximately 2,300 new cases annually per AIHW data), but incidence is disproportionately higher in Aboriginal and Torres Strait Islander peoples and in communities with high H. pylori prevalence. No ML-based GC diagnostic tools are currently listed on the Australian Register of Therapeutic Goods (ARTG) as Class III medical devices for this specific indication, though the TGA's Digital Health regulatory framework is evolving. The RACGP does not currently recommend population-based GC screening in Australia. PBS-listed treatments for GC (including trastuzumab for HER2-positive disease, ramucirumab, and nivolumab) are not directly addressed by this review. Clinicians should note that ML tools validated predominantly in East Asian populations may not perform equivalently in Australian cohorts with different GC epidemiology, endoscopic practice patterns, and pathological subtypes. Adult patients undergoing investigation or treatment for gastric cancer, particularly in settings with access to advanced endoscopic, pathological, or radiological AI-assisted tools. Most directly applicable to high-incidence populations (East Asia), though emerging relevance exists in Western settings with increasing GC burden in high-risk subgroups (H. pylori-infected, familial GC, immigrant populations from high-incidence regions).

Abstract

Gastric cancer (GC), a malignant neoplasm originating from the gastric mucosal epithelium, represents one of the most prevalent cancers worldwide. Early detection is critical for improving treatment outcomes and patient prognosis. Recent advances in artificial intelligence (AI), particularly in machine learning, have introduced powerful computational and analytical capabilities that are increasingly being applied in GC research. Machine learning algorithms have shown considerable promise in enhancing the accuracy of GC diagnosis and optimizing therapeutic strategies. This review provides a concise overview of progress in machine learning applications within oncology, examines their current role and clinical utility in GC diagnosis and treatment, and highlights the transformative potential of machine learning in advancing GC management and patient care.

References

1. 1.Cui, W.-Z., Wen, C.-Q., Li, C.-Q., Zhang, Q.-J., Yu, Q.-Q., & Sun, W.-W. (2026). Research progress of machine learning applications in gastric cancer diagnosis and therapy. Clinical & Translational Oncology: Official Publication of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico. Advance online publication. https://doi.org/10.1109/TMI.2018.2823083 [Note: DOI as supplied — bibliographic inconsistency identified; independent verification recommended]