Application of Hyperspectral Imaging Analysis Technology in the Diagnosis of Colorectal Cancer Based on Colonoscopic Biopsy
Study Details
Study Description
Brief Summary
The purpose of this study is to develop and validate a deep learning algorithm for the diagnosis of colorectal cancer other colorectal disease by marking and analyzing the characteristics of hyperspectral images based on the pathological results of colonoscopic biopsy, so as to improve the objectiveness and intelligence of early colorectal cancer diagnosis.
Condition or Disease | Intervention/Treatment | Phase |
---|---|---|
|
Detailed Description
Prospectively collect the hyperspectral image information of ordinary colonoscopic biopsy tissue. The colonoscopic biopsy tissue is from the Endoscopy Center of Qilu Hospital of Shandong University. The hyperspectral images are marked based on the biopsy pathological results, and the deep convolutional neural network (DCNN) model is used. With training and verification, develop the Hyperspectral Imaging Artificial Intelligence Diagnostic System (HSIAIDS) .A portion of colonoscopic biopsy tissue will be collected as a prospective test set to prospectively test the diagnostic performance of the HSIAIDS algorithm.
Study Design
Arms and Interventions
Arm | Intervention/Treatment |
---|---|
Deep learning algorithm group After the patient has passed the screening, a routine colonoscopy will be performed, and the target tissue with suspected inflammation or neoplasia will be biopsied. The clinical investigators use the hyperspectral microscope to collect image information of the biopsy tissue in the endoscopy room. After collecting information, biopsy specimens will be routinely processed and sent for pathological diagnosis. |
Outcome Measures
Primary Outcome Measures
- Accuracy of HSI artificial intelligence model to identify colorectal adenoma and cancer [1 year]
Accuracy of hyperspectral imaging (HSI) artificial intelligence model to identify colorectal hyperplastic polyp, adenoma, SSL and colorectal cancer. Accuracy of artificial intelligence models Accuracy = (true positives + true negatives) / total number of subjects * 100%
- Sensitivity [1 year]
Sensitivity of HSI artificial intelligence model Sensitivity = number of true positives / (number of true positives + number of false negatives) * 100%.
- Specificity [1 year]
Specificity of HSI Artificial Intelligence Model Specificity = number of true negatives / (number of true negatives + number of false positives))*100%
- Negative predictive values(NPV) [1 year]
Negative predictive values for HSI artificial intelligence model = number of true negatives / (number of true negatives + number of false negatives)*100%
- AUC (95% CI) [1 year]
area under the receiver operating characteristic curve (AUC)
Secondary Outcome Measures
- To record and evaluate any unknown risks and adverse events of hyperspectral imaging in specimen image acquisition [1 year]
To record and evaluate any unknown risks and adverse events of hyperspectral imaging in specimen image acquisition
Eligibility Criteria
Criteria
Inclusion Criteria:
- patients aged 18-75 years who undergo the colonoscopy examination and biopsy
Exclusion Criteria:
-
patients with severe cardiac, cerebral, pulmonary or renal dysfunction or psychiatric disorders who cannot participate in colonoscopy
-
patients with previous surgical procedures on the gastrointestinal tract.
-
patients with contraindications to biopsy
-
patients who refuse to sign the informed consent form
Contacts and Locations
Locations
Site | City | State | Country | Postal Code | |
---|---|---|---|---|---|
1 | Qilu hosipital | Jinan | Shandong | China | 250012 |
Sponsors and Collaborators
- Shandong University
Investigators
- Study Chair: Xiuli Zuo, MD,PhD, Study Principal investigator
Study Documents (Full-Text)
None provided.More Information
Publications
None provided.- 2022-SDU-QILU-G003