
KFBio
AI module for analysis of gastric histology KFBio AI Stomach
AI analysis of gastric biopsies: consensus with experts over 96%
An AI module to automatically detect stomach cancer and precancerous changes in biopsy drugs, localizes suspicious areas, suggests the type and extent of the lesion, and speeds up the preparation of a standardized conclusion.
Request a quotationThe algorithm recognizes adenocarcinoma, mucinous and perineal cell cancer, low and high intraepithelial neoplasia, as well as inflammation, intestinal metaplasia, atrophy and hemorrhage, tumor segmentation, heat maps and structured reports with the type and extent of the lesion are available.
The system is trained on more than 50,000 slides from 10 medical centers. KFBio indicates consistency with experts from leading clinics over 96% and reduces the time for preparing a report to 80%. The local configuration supports a GPU server, integration with digital scanners and transferring results to the laboratory information system.
Analysis opportunities
- Detection of adenocarcinoma, mucinous and periform cell carcinoma
- Separation of low and high degree of intraepithelial neoplasia
- Recognition of inflammation, intestinal metaplasia, atrophy and hemorrhage
Accuracy and validation
- Training on more than 50,000 slides from 10 medical centers
- Consistency with experts of leading clinics over 96%
- Reduced report preparation time to 80%
Accountability and integration
- Tumor segmentation and heat maps
- Structured report with type and degree of damage
- Integration with digital scanners and LIS
Diagnostics
| Parameter | Stomach |
|---|---|
| Identified changes | adenocarcinoma, mucinous and perineal cell carcinoma, low/high degree of intraepithelial neoplasia, inflammation, intestinal metaplasia, atrophy, hemorrhages |
precision
| Parameter | Stomach |
|---|---|
| Consistency with experts | >96% |
| Reducing reporting time | 80% |
| Training sample | More than 50,000 slide from 10 centers |
Accountability
| Parameter | Stomach |
|---|---|
| Visualization | tumor segmentation, heat maps |
| Integration | Digital scanners, LIS |
