TraibunHealth_Platform Icon - Small - AIA-1


Artificial intelligence for health: the power of machine learning and AI powered automation

Calopix 5 is built to seamlessly run both Tribun Health proprietary and third party automation applications to accelerate and improve image quantification for diagnostics, prognosis and biomarker analysis. First applications include: Breast molecular signature (MoSI4Bca), colon polyp prescreening (AID), cervix lesion detection, non-small cell lung carcinoma outcome prediction (NSCLC), prostate Gleason grading, breast cancer relapse prediction,...and more to come.

tsribun-health-ai-apps (1)

The power of automated analysis

Accelerate and enhance workflow and digital confidence

Automated tissue detection, segmentation and classification.

Group 131-1

Results are worth 1,000 words.

Tribun Health has been competing in a range of algorithm performance contests and is proud to have ranked high in several of them. Specifically:

1. Data challenge SFP 2020 - First place award. Cervix biopsies grading.

2. Data challenge MICCAI - MIDOG 2021 - 3rd place. Mitosis detection (Breast Cancer).

3. Innov’UP PIA 2019 - Digestive biopsies sorting, labeled Medicine (received financing).

4. Grand Défi IA - Clinical validation of MoSI4Bca application.

5. iTech Awards 2020-2021 (Sanofi) - AI model for oncology clinical trials patient stratification.

6. Trophée de la Health Tech France Biotech 2021 - One of 4 nominated companies in E-health award.

Mask Gssroup 7-1


With both proprietary
and third party apps.


MoSI4Bca Breast - Molecular signature (KI67, ER/PR, HER-2, Mitosis count).


AID-1 Gastro Intestinal - Colonoscopy auto tissue segmentation and screening.


AID-2 Gastro Intestinal - Polyp, tissue diagnosis.


Uterine cervical lesions auto-detection.

Non-small cell lung carcinoma

(NSCLC) outcome prediction.

Prostate gleason grading

Gleason grading - A DeepBio solution.


Immunohistochemically (IHC) stained human breast cancer tissue (Ki-67, ER and PR) - A Mindpeak solution.




1. Breast Molecular Signature

Breast Molecular Signature (MoSI4Bca) is an AI application allowing pathologists to automatically and precisely score the known prognostic biomarkers in breast cancer. The goal is dual: (1) ensure accurate patient therapy by providing tools enabling the pathologist to choose the best therapy according to the patient molecular profile and (2) save the pathologist time and favor his comfort by avoiding him to manually count thousands of cells.

Mask Group 1-1
Mask Group 2-1

2&3. AID-1 and AID-2 Gastro Intestinal

AID-1 (binary segmentation) is a solution for automatic colonoscopy tissue segmentation and screening. It aims at highlighting malignant tissues via probability map giving the AI model a certainty level for tissue malignity. The slides are then tagged with a label (benign vs. malignant) and sorted allowing for the pathologist an optimized and structured workflow.

AID-2 (multi-class) is a solution to assist pathologists in diagnosing polyp tissues. Based on the latest AI technologies, our solution displays a class-map indicating the different tissue types composing the image, and helps the pathologist to quickly and easily find the risk areas and their grades.

4. Cervix lesion detection

Cervix lesion detection is an application for the automatic detection of epithelial lesions in uterine cervical slides. It visualizes a class map as an overlay on the slide image indicating 4 types of tissues : 1) benign tissue, 2) low-grade squamous intraepithelial lesion, 3) high-grade squamous intraepithelial lesion, 4) invasive cancer.

Mask Group 3-1
Mask Group 4-1

5. Non-small cell lung carcinoma outcome prediction

Lung cancer is the leading cause of carcinoma-related death worldwide, with non-small cell lung cancer (NSCLC) accounting for approximately 85% of all lung cancer patients. Therefore, prognostic stratification is crucial to categorize patients to help physicians make decisions about treatment strategies.

6. Prostate Gleason grading by DeepBio

We are integrating the third-party DeepDx® Prostate algorithm into our Image Management System (IMS). It is a clinically validated artificial intelligence algorithm for the analysis of prostate needle biopsy tissue images for cancer detection and classification.

Mask Group 5-1
Breast cancer algorithms

7. Breast cancer algorithms by Mindpeak

We integrate Mindpeak's third-party algorithm into our IMS. This algortihms analyze whole-slide-images (WSI) of immunohistochemically (IHC) stained human breast cancer tissue (biomarkers: Ki-67, ER and PR) to detect, classify, count cancer cells and to provide the respective quantitative score. Both our IMS and the Mindpeak algorithm are CE-IVD marked, allowing their use as medical devices for both clinical diagnosis and research.

Mindpeak logo