Esophageal cancer is the sixth leading cancer in the world cause of mortality, nearly half of the new cases found in China with most of the type: esophageal squamous cell carcinoma (ESCC). Detection of early ESCC is essential to improving a patient's survival. 1. We apply magnifying endoscopy narrow-band imaging (ME-NBI) and Lugol's chromoendoscopy (LCE) in ESCC detection, where ME-NBI is more accurate than conventional white light, and LCE has higher sensitivity than HD-WLE. 2. We mark the ME-NBI lesion object in two different types of classes: Real and Distinct boundaries. The label region of distinct boundary covers all lesion areas, including blurred borders, while the label region of real boundary only covers clear lesions. The main benefit of incorporating various types of classes is to help us to identify the border of the lesion area more clearly since the scenario that blurred borders caused by the imaging reconstruction error can be excluded. 3. By combining different recently published advanced modules, we construct multiple instance segmentation frameworks and obtain the most optimal one for ESCC detection. The modules mainly include (1) InstaBoost, a widely used image random augmentation technique; (2) Cascade RCNN, a multistage target detection architecture; (3) DCN that enhances the transformation modeling capabilities. 4. The machine learning performance demonstrates the efficacy of our application and shows that it has great potential to apply to real-time clinical Computer-Aided Diagnosis (CAD).

The authors have declared no competing interest.

This study was funded by Jinan healthcare science and technology plan(201907046), Shandong Province key research and development program (2019GSF108028) and Youth Natural Science Foundation of Shandong Province (ZR2020QH040).

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Ethics committee/IRB of The first affiliated hospital of Shandong first Medical University (Shandong Provincial Qianfoshan hospital) gave ethical approval for this work.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.

Yes

Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.