VisioMel Challenge: Predicting Melanoma Relapse

Use digitized microscopic slides to predict the likelihood of melanoma relapse within the next five years. #health

€25,000 in prizes
may 2023
542 joined


About the project

The search for new predictive markers of relapse for non-metastatic primary melanomas using artificial intelligence would complement the pathologist’s analysis and therefore, assist the dermatologist in adapting patient care. Moreover, this identification for thin melanomas, whose relapse is particularly complex to predict, would constitute a major advance in the management of melanomas.

— VisioMel

The data for this challenge come from RIC-Mel, the French national database on melanoma. This data collection effort began in 2012 and now spans over 40,000 patients from over 49 sites in France.

For this challenge, all patients were initially diagnosed between 2012 and 2016 and have 5 years of subsequent observations. In general, the greatest risk of disease recurrence and mortality occurs in the first 5 years following diagnosis. All data has been anonymized and labeled by the pathological anatomy and cytology laboratories that initially diagnosed the patients' tumors.

About the organizers

VisioMel began in 2021 when the French Society of Pathology, the French Society of Dermatology, the French Cutaneous Cancers Group, and the National Professional Council of Pathologists decided to build a challenge using clinical variables from a national database on melanoma (RIC-Mel).

The French Society of Pathology promotes research and teaching in Pathology in France and organizes the annual Carrefour Pathologie meeting in Paris.

The French Society of Dermatology promotes research and teaching in France and organizes the annual Journées dermatologiques de Paris.

The French Cutaneous Cancers Group coordinates French cutaneous oncology research and helps organize national collaborative studies.

The National Professional Council of Pathologists organizes the training of French pathologists and promotes the quality of practices.

The Health Data Hub (HDH), a key component of French President Macron’s AI strategy, is a public one-stop health data access point and technological platform for public interest research conducted by academic and industrial research teams. Officially created in December 2019, the HDH seeks to provide simple, unified, transparent and secure access to health data in compliance with regulations and citizens' rights. The HDH provides infrastructure, a database catalogue and related services. These services include providing access to health data, supporting data collection, and facilitating data consolidation and use.

The primary sponsor of the challenge is the French Ministerial Delegation for Digital Health (Délégation Ministérielle au Numérique en Santé) in the context of the "Digital Health Acceleration Strategy" of the French government.

Additional references on AI and pathology

  • Video presentation of the VisioMel project by Dr. Frédéric Staroz, VisioMel project manager, during an event organized by the Health Data Hub in May 2022 (in French with English slides).
  • Podcast in which Dr. Frédéric Staroz presents the benefits of data science competitions for the health ecosystem and entrepreneurs (in French).