This innovative study introduces the Mucosa Touch Rate (MTR) as a novel, objective, real-time quality control indicator for colonoscopy. The authors developed a ConvNeXt-B2 deep learning model to quantify the proportion of withdrawal frames that exhibit significant visual field loss due to mucosal contact. The work is comprehensive, encompassing model development, retrospective validation, and a large-scale prospective clinical trial, thereby providing strong evidence for the potential clinical utility of MTR in improving colonoscopy quality.