We aimed to identify quality indicator for esophagogastroduodenoscopy for detecting upper gastrointestinal (UGI) cancer. Data from 43,526 consecutive health checkups from August 2012 to January 2022 were retrospectively collected. The study ultimately analyzed 42,387 examinations by 12 endoscopists who performed more than 1000 examinations, including all cancers detected. These endoscopists were classified either into fast/slow group based on their mean examination time for a normal finding of esophagogastroduodenoscopy during their first year of the examination, or small/large group based on number of endoscopic images, respectively. The association between UGI cancer detection rate and examination time or the number of images was analyzed, using 5 minutes or 50 images as cutoff values. The detection rate of overall (8 pharyngeal, 39 esophageal, 69 gastric) cancers in the fast, slow, small, and large groups were 0.17%, 0.32%, 0.21%, and 0.31%, respectively. On multivariable analysis, endoscopists in the fast group or the small group were less likely to detect overall UGI cancer (OR: 0.596, 95% CI: 0.373-0.952, P = .030; OR: 0.652, 95% CI: 0.434-0.979, P = .039). Additionally, repeated endoscopy within 2 years had a higher overall cancer detection rate, compared with repeated screening after 2 years. In a sub-analysis, a significant negative relationship was found between the detection rate of gastric cancer and the number of gastric images < 35 (OR: 0.305, 95% CI: 0.189-0.492, P = .000). There was also a negative correlation trend between the detection rate of pharyngeal and esophageal cancers and the number of esophageal images < 11 (OR: 0.395, 95% CI: 0.156-1.001, P = .050). The optimal number of images and screening 2-year interval are considered useful quality indicators for detecting UGI cancer. This study also suggests that a total of 50 images, or 35 images of the stomach are suitable for detecting UGI cancer, or gastric cancer, during screening endoscopy.