top of page


Dernière mise à jour : 18 avr. 2023

Artificial intelligence helping diagnosis of rare diseases with oral expressions

This Data Challenge is led by Pr Agnès Bloch-Zupan (PU-PH professor, clinician, rare disease expert, coordinator of the Strasbourg University Hospital, Reference center for rare oral and dental diseases and associated network, researcher at IGBMC, Université de Strasbourg, CNRS- UMR7104, INSERM U1258) and supported by the D-IA-GNO-DENT consortium. This consortium (members below), in partnership with the Health Data Hub, proposes a Data Challenge on the theme of rare dental diseases.

Introduction of the DIAGNODENT data challenge by Pr Agnès Bloch-Zupan (video from Health Data Hub)

Moving from wandering to diagnosis

Join the D-IA-GNO-DENT Challenge and develope an AI solution to aid in the diagnosis of rare dental diseases based on the analysis of photographs and X-rays of the mouth and teeth. These genetic diseases, which affect multiple organs and systems, are recognised through abnormalities in teeth fixed over time by mineralisation.

The goal : promote access to early detection of rare oral diseases for patients and health professionals

The expected benefit : to reduce the diagnostic wandering of affected individuals.

High level overview on the challenge goal

The context

Oral pathologies or anomalies are often the little known expression of rare diseases of genetic origin. These anomalies are receiving increasing attention because of their diagnostic or even predictive character. They are classified into dental anomalies of number, shape, size, structure, root formation and eruption and correspond to specific developmental and genetic problems. They are fixed in time by mineralisation. They are isolated or associated with other symptoms in syndromes.

900 rare diseases have a dento/oro/facial component. Clinical diagnoses and identification of the genes involved are difficult to make and to implement, leading patients and their families into diagnostic wandering. However, there are diagnostic signatures recognised by experts in the field and supported by phenotype/genotype correlations (D4/phenodent database/NGS GenoDENT panel).




Designed for rare disease patients

Used by doctors and medical research

Input and output data

Decrease diagnostic wandering Accelarate care in the « right » care pathway

Access to enhanced knowledge of rare diseases Acces to rapid expertise

Inputs : images, reports, diagnostic signatures... Output : a (pre-)diagnosis of suspected rare disease with oral manifestations


We have to make a limited set of data speak for itself; we are in the field of rare tooth enamel diseases, but also of precision medicine.

Data / Cohorts

  • Amelogenesis imperfecta --> 166

  • Dentine anomalies --> 50

  • Control --> 100

  • Intraoral colour photographs (4 to 10/individuals)

  • Panoramic radiographs

  • Colour

  • Surface

  • texture

  • Shape/size

Challenge & Rewards

Timeline : The D-IA-GNO-DENT Challenge will take place in April 2023 and will be hosted on our platform,

To participate in the D-IA-GNO-DENT, sign-up at (as data scientist) and complete this form to gain access to the data challenge once it's available.

3 prizes will be awarded to the teams that provide the most effective* algorithms, including a first prize of €20 000

Apply Now

To participate in the DIAGNODENT Challenge, apply by completing this form and sign-up at (as data scientist).

Fully managed development Environment

DIAGNODENT challenge will be hosted at data challenge platform, Data scientists participating in the DIAGNODENT Challenge will have access to a hosted Jupyter notebook with free GPU resources to develop their models.

3 213 vues0 commentaire

Posts récents

Voir tout


bottom of page