Evaluation of AL Prediction for Rectal Cancer

Evaluation of a Machine Learning Based Anastomotic Leakage Prediction Model After Anterior Resection for Rectal cancer-a Multicenter, Prospective, Randomized Controlled Study

Anastomotic leakage is one of the most serious postoperative complications of low rectal cancer, with an incidence of 3%-21%. The occurrence of anastomotic leakage is related to many factors, and the occurrence of anastomotic leakage can be predicted by building a prediction model. Most of the anastomotic leakage prediction models constructed in the past are nomograms, which have limitations in the fitting of model creation. In the previous study, the center took the lead in building a random forest anastomotic leakage prediction model based on machine learning. This study intends to prospectively enroll patients with rectal cancer undergoing anterior abdominal resection and use their clinical data to prospectively verify the efficacy of the anastomotic leakage prediction model, and further improve and promote the prediction model.

Study Overview

• Status: Not yet recruiting
• Conditions: Anastomotic Leak Rectum
• Intervention / Treatment: Diagnostic test: Prediction model evaluation

• Study Type: Interventional
• Enrollment (Anticipated): 418
• Phase: Not Applicable

Contacts and Locations

Study Locations

• China
  • Shanghai
    • Shanghai, Shanghai, China, 200433
      • Department of Colorectal Surgery in Changhai Hospital

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years to 75 years (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

Description

Inclusion Criteria:

1. Patients aged 18-75 years
2. Adenocarcinoma confirmed by pathology
3. Colonoscopy or imaging examination confirmed that the distance between the lower edge of the tumor and the anal edge was less than or equal to 12cm
4. Preoperative imaging diagnosis was cTxNxM0
5. No local complications (no obstruction, incomplete obstruction, no massive active bleeding, no perforation, abscess formation, and no invasion of adjacent organs)
6. The hematopoietic functions of heart, lung, liver, kidney and bone marrow meet the requirements of surgery and anesthesia
7. Voluntarily sign the informed consent form

Exclusion Criteria:

1. Previous history of malignant tumor
2. Simultaneous multiple primary colorectal cancer
3. Previous multiple abdominal and pelvic surgeries or extensive abdominal adhesions
4. Patients with intestinal obstruction, intestinal perforation, intestinal bleeding, etc., requiring emergency surgery
5. Patients with familial adenomatous polyposis and active inflammatory bowel disease
6. A history of severe mental illness
7. pregnant or lactating women
8. Patients with uncontrolled infection before operation
9. The investigator did not consider the patient to be eligible for the trial

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Prevention
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Single

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
No Intervention: Surgeon evaluation
Experimental: Surgeon combining with model evaluation	Diagnostic test: Prediction model evaluation; a machine learning based anastomotic leakage prediction model

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Accuracy of stoma implementation	Accuracy of stoma implementation: the number of anastomotic leakage patients with stoma and none anastomotic leakage patients without stoma to the number of total patients.	1 months after surgery

Secondary Outcome Measures

Outcome Measure	Time Frame
Sensitivity and specificity in the prediction of anastomotic leakage	1 months after surgery
Grade C leakage rate	1 months after surgery
Preventive stoma rate	1 months after surgery
Rate of stoma reverse	3-6 months after surgery

Collaborators and Investigators

• Sponsor: Changhai Hospital

Study record dates

Study Major Dates

• Study Start (Anticipated): December 10, 2022
• Primary Completion (Anticipated): October 10, 2024
• Study Completion (Anticipated): October 10, 2025

Study Registration Dates

• First Submitted: October 30, 2022
• First Submitted That Met QC Criteria: November 3, 2022
• First Posted (Actual): November 9, 2022

Study Record Updates

• Last Update Posted (Actual): November 9, 2022
• Last Update Submitted That Met QC Criteria: November 3, 2022
• Last Verified: November 1, 2022

More Information

Terms related to this study

• Keywords: machine learning; rectal cancer; Anastomotic Leak
• Additional Relevant MeSH Terms: Digestive System Diseases; Pathologic Processes; Neoplasms; Neoplasms by Site; Postoperative Complications; Gastrointestinal Neoplasms; Digestive System Neoplasms; Gastrointestinal Diseases; Intestinal Diseases; Intestinal Neoplasms; Rectal Diseases; Colorectal Neoplasms; Rectal Neoplasms; Anastomotic Leak
• Other Study ID Numbers: CHALP001

Drug and device information, study documents

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