Clinical Application of an AI-based Dissection Trajectory Prediction System (ADTPS) in Endoscopic Submucosal Dissection

Clinical Application of an AI-based Dissection Trajectory Prediction System (ADTPS) in Endoscopic Submucosal Dissection: A Prospective Paired Diagnostic Study and a Randomized Controlled Clinical Trial

In this prospective paired diagnostic study and single-center, randomized controlled trial, patients with early esophageal squamous neoplasia or high-grade intraepithelial neoplasia meeting the inclusion and exclusion criteria will be enrolled in a paired diagnostic cohort (60 patients) and subsequently randomly assigned (1:1) to receive endoscopic submucosal dissection (ESD) with AI-based Dissection Trajectory Prediction System (ADTPS) guidance or conventional ESD (without AI). Clinical data and operator workload scores (NASA-TLX) are collected during the procedure, and postoperative follow-up assessments are performed at days 1, 3, 7, and 14. The study aims to analyze the impact of ADTPS on the mean single-dissection time and operator workload in patients undergoing ESD by comparing the efficacy differences between the experimental and control groups. Additionally, the study investigates the effects of ADTPS on other postoperative complications including R0 resection rate, muscularis propria injury, intraoperative bleeding, perforation (acute and delayed), and total procedure time; conducts a comparative analysis of the safety and efficiency of AI-assisted versus conventional ESD; and develops effective clinical strategies for optimizing dissection trajectory and reducing complications in endoscopic submucosal dissection.

Study Overview

• Status: Active, not recruiting
• Conditions: Esophageal Squamous Cell Carcinoma (ESCC); High-Grade Intraepithelial Neoplasia; AI-Based Dissection Trajectory Prediction System
• Intervention / Treatment: Device: Conventional ESD Procedure; Device: AI-based Dissection Trajectory Prediction System (ADTPS)

• Study Type: Interventional
• Enrollment (Estimated): 160
• Phase: Not Applicable

Contacts and Locations

Study Locations

• China
  • Shandong
    • Jinan, Shandong, China, 250012
      • 山东大学齐鲁医院

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Chinese patients aged 18-80 years.

Lesions meeting ESD indications for esophageal, gastric, or colorectal early cancer or high-grade intraepithelial neoplasia, as defined by:

Non-invasive tumors regardless of size; or

Differentiated-type intramucosal carcinoma without ulceration, regardless of size; or

Differentiated-type intramucosal carcinoma with ulceration and a diameter ≤3 cm; or

Undifferentiated-type intramucosal carcinoma without ulceration and a diameter ≤2 cm.

Planned to undergo ESD treatment.

No prior treatment for the lesion (including ESD, surgery, radiotherapy, chemotherapy, etc.).

Platelet count >100 × 10⁹/L and PT-INR <1.5, with antiplatelet agents (aspirin, clopidogrel, etc.) discontinued for at least 5 days.

American Society of Anesthesiologists (ASA) physical status grade I or II.

Voluntarily signed informed consent.

Exclusion Criteria:

• Patients currently undergoing dialysis.

Patients with severe cardiopulmonary disease or other severe comorbidities that may increase the risk of the ESD procedure.

Pregnant or breastfeeding women.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Treatment
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Double

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Active Comparator: Conventional ESD without AI	Device: Conventional ESD Procedure; The same endoscopic hardware platform is used, but the AI-based Dissection Trajectory Prediction System (ADTPS) is turned off. The endoscopist performs the ESD procedure based solely on clinical judgment and personal experience, following standard conventional ESD steps including marking, injection, circumferential incision, and submucosal dissection.
Experimental: AI-assisted ESD with ADTPS	Device: AI-based Dissection Trajectory Prediction System (ADTPS); The ADTPS is an artificial intelligence software system that analyzes endoscopic images in real time during ESD. It automatically identifies lesion boundaries and generates a recommended dissection trajectory overlaid on the endoscopic view. The endoscopist follows the AI-generated trajectory to perform submucosal dissection. The system provides visual guidance only and does not alter the standard surgical workflow.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Time Frame
NASA-TLX Workload Score	Immediately after the ESD procedure
Mean Single Dissection Time	During the ESD procedure

Secondary Outcome Measures

Outcome Measure	Time Frame
R0 Resection Rate	At the time of pathological examination
Muscularis Propria Injury Rate	During the ESD procedure
Intraoperative Bleeding Episodes	During the ESD procedure
Intraoperative Perforation Rate	During the ESD procedure
Delayed Bleeding Rate	Within 14 days after procedure
Delayed Perforation Rate	Within 14 days after procedure
Total Procedure Time	During the ESD procedure

Collaborators and Investigators

• Sponsor: Qilu Hospital of Shandong University

Study record dates

Study Major Dates

• Study Start (Actual): July 29, 2026
• Primary Completion (Estimated): July 28, 2027
• Study Completion (Estimated): October 1, 2027

Study Registration Dates

• First Submitted: August 6, 2026
• First Submitted That Met QC Criteria: August 6, 2026
• First Posted (Actual): August 11, 2026

Study Record Updates

• Last Update Posted (Actual): September 18, 2026
• Last Update Submitted That Met QC Criteria: September 16, 2026
• Last Verified: July 1, 2026

More Information

Terms related to this study

• Keywords: Endoscopic submucosal dissection; Randomized controlled trial; Artificial intelligence; R0 resection; NASA-TLX; Procedural efficiency; Dissection trajectory prediction
• Additional Relevant MeSH Terms: Neoplasms by Site; Neoplasms; Neoplasms by Histologic Type; Gastrointestinal Neoplasms; Digestive System Neoplasms; Digestive System Diseases; Gastrointestinal Diseases; Head and Neck Neoplasms; Neoplasms, Glandular and Epithelial; Esophageal Diseases; Carcinoma; Neoplasms, Squamous Cell; Carcinoma, Squamous Cell; Esophageal Neoplasms; Esophageal Squamous Cell Carcinoma
• Other Study ID Numbers: 2026SDU-QILU-4

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No