Establishment of a Feasibility Model for NOSE Surgery Based on Machine Learning

Sun Yat-sen University

Status: Not yet enrolling

Conditions: Rectosigmoid Cancer; Surgery; Machine Learning; Natural Orifice Specimen Extraction Surgery

Treatments: Procedure: Natural Orifice Specimen Extraction Surgery

Study type: Observational

Funder types: Other

Identifiers

NCT05797064

1010PY(2022)-09

Details and patient eligibility

About

The goal of this observational study is to test in patients with resectable rectosigmoid cancers. The main question it aims to answer is establishment of a feasibility model for predicting natural orifice specimen extraction surgery (NOSES) based on machine learning.

Enrollment: 460 estimated patients

Sex: All

Ages: 18 to 80 years old

Volunteers: No Healthy Volunteers

Inclusion criteria

1. Patients diagnosed with colorectal cancer or large adenoma who are suitable for laparoscopic colorectal surgery;
2. Tumor staging ≤ T3 without invasion of surrounding organs;
3. No abdominal seeding or distant organ metastasis;
4. Clear and complete imaging data (CT, pelvic MRI) that can be processed by a computer;
5. Feasible evaluation and determination for obtaining specimens through the rectal channel during preoperative and intraoperative assessments.

Exclusion criteria

1. Contraindications for laparoscopic colorectal surgery;
2. Tumor staging is T4, or there are cancer nodules;
3. Presence of metastasis or distant organ metastasis;
4. Incomplete imaging data;
5. Preoperative intestinal obstruction;
6. Tumor or specimen diameter larger than the transverse diameter of the pelvic outlet;
7. Previous rectal radiotherapy;
8. Unsuitable evaluation and determination for obtaining specimens through the rectal channel during preoperative and intraoperative assessments.

Trial design

460 participants in 2 patient groups

Training set

Description:

The training set is a dataset used to train the model, which includes randomly enrolled patients with colon and rectal cancer. The inputs include data such as gender, age, height, weight, BMI, tumor stage, tumor pathology type, and the output information is whether NOSES surgery was successful or not. During training, the model learns from this dataset to make predictions on whether new patients with colon and rectal cancer can undergo NOSES surgery successfully.

Treatment:

Procedure: Natural Orifice Specimen Extraction Surgery

test set

Description:

The test set is a dataset used to evaluate the performance of a trained machine learning model. It includes another randomly enrolled group of patients with colon and rectal cancer, as well as their clinical and pathological data and surgical outcomes. The outputs are not used during training, but are used to test the trained model to evaluate its predictive ability on unknown data. The purpose is to evaluate the model's generalization ability, that is, its performance on new and unknown data.

Treatment:

Procedure: Natural Orifice Specimen Extraction Surgery