Objective <p>White light endoscopy (WLE) is an important method for upper gastrointestinal cancer screening. However, studies have demonstrated that magnifying endoscopy with image-enhanced endoscopy (ME-IEE) achieves a significantly higher detection rate of early cancer compared with WLE. Current research has primarily focused on evaluating the diagnostic performance of ME-IEE, while evidence regarding its feasibility as a routine opportunistic screening mode for early upper gastrointestinal cancer in daily clinical practice remains limited.</p> Methods <p>A prospective, single-center, randomized controlled trial was conducted between October 15, 2023, and October 7, 2024. Patients aged ≥ 18 years who underwent painless anesthesia esophagogastroduodenoscopy(EGD) at our endoscopy center were enrolled. Recruited patients were randomly assigned to either the routine magnifying endoscopy group (r-ME group) or the white light endoscopy group (WLE group). All suspicious lesions were biopsied. The primary endpoint was the detection rate of early neoplasms; secondary endpoints included biopsy implementation rate, biopsy positivity rate, mean screening time, and mean number of biopsies.</p> Results <p>A total of 3,000 patients were randomized (1:1). The detection rate of early neoplasms and biopsy positivity rate were significantly higher in the r-ME group than the WLE group (3.87% vs. 2.27% and 6.34% vs. 3.58%, respectively; both <i>P</i> &lt; 0.05). However, the mean screening time was slightly longer in the r-ME group, by less than a minute (643s vs. 600s, <i>P</i> &lt; 0.05). Among expert endoscopists, the detection rate of early neoplasms was significantly higher than that of non-expert endoscopists (4.54% vs. 2.50%, <i>P</i> &lt; 0.05), while the biopsy submission rate was significantly lower (56.5% vs. 64.4%, <i>P</i> &lt; 0.05). For non-expert endoscopists, both the detection rate of early neoplasms and biopsy positivity rate were significantly higher in the r-ME group than in the WLE group (3.32% vs. 1.67% and 5.29% vs. 2.53%, respectively; both <i>P</i> &lt; 0.05). In contrast, no significant differences were observed of the above endpoints between the two groups for expert endoscopists (5.28% vs. 3.81% and 9.36% vs. 6.72%, both <i>P</i> &gt; 0.05). The detection rate of early cancer using WLE mode was significantly lower for non-expert endoscopists compared with expert endoscopists (1.67% vs. 3.81%, <i>P</i> &lt; 0.05). Notably, when r-ME mode was used, the detection rate achieved by non-expert endoscopists did not differ significantly from that of expert endoscopists (3.32% vs. 5.28%, <i>P</i> &gt; 0.05), nor from that of expert endoscopists using WLE mode (3.32% vs. 3.81%, <i>P</i> &gt; 0.05).</p> Conclusion <p>The r-ME model enhances diagnostic accuracy and consistency in opportunistic screening for early upper gastrointestinal cancer, offering particular value for non-expert endoscopists.</p> Trial registration <p>This study was registered in the Chinese Clinical Trial Registry on 03/10/2023 (clinical trial registration number: ChiCTR2300076327).</p>