Novel Technologies for Gastrointestinal Cancer Detection: A Systematic Review and Meta-Analysis of Diagnostic Accuracy.

INTRODUCTION

Gastrointestinal (GI) cancers account for a quarter of all cancers and one-third of cancer-related deaths worldwide. Novel diagnostic approaches are crucial to enhance early detection and improve patient outcomes. This review systematically identified and evaluated novel GI cancer detection approaches.

METHODS

A systematic review was conducted using PubMed, EMBASE, Scopus, and the Cochrane Library on December 11, 2024, and updated on June 2, 2025. Cohort studies and randomized controlled studies, published in English, that evaluated a novel diagnostic approach for incident GI cancer in a clinical setting, included ≥300 participants, and reported diagnostic accuracy metrics were eligible for inclusion. Two reviewers independently screened studies and assessed risk of bias using the Quality Assessment of Diagnostic Accuracy Studies 2 tool.

RESULTS

Of 9,604 records screened, 72 studies were included: by site, 51 on colorectal, 1 on upper GI, 1 on esophageal, 7 on gastric, 9 on pancreatic, and 5 on hepatocellular carcinoma. Stool DNA testing in patients with a positive fecal immunochemical test showed the highest performance for colorectal cancer (area under the receiver operating characteristic curve [AUC] 0.94-0.98). The best performing technologies across other cancer types were Q-Cancer for pancreatic (AUC 0.89-0.92), the GALADUS model for hepatocellular carcinoma in cirrhotic patients (AUC 0.94), and an artificial intelligence model for esophageal cancer (AUC 0.87).

DISCUSSION: This review highlights several promising novel diagnostic approaches for GI cancer detection. External validation, population-level studies, and health economic evaluation are crucial to establish clinical utility across different settings and populations. Clinicians and policy makers must weigh the added value of novel technologies against current practice, regulatory requirements, clinical understanding, interpretation of results, and public trust.