🫀 Announcing the CardI-HACK Data Challenge: Advancing Hypertrophic Cardiomyopathy (HCM) progression Through AI

About the Challenge

Trustii.io is excited to announce the upcoming CardI-HACK Data Challenge, set to launch on our platform on November 4th, 2025. This groundbreaking competition invites data scientists from our community and around the world to contribute to a transformative project in the field of cardiogenetics.

A total of €14,000 in prize money is awarded to the top teams developing the most accurate AI models for cardiac risk prediction.

Led by the IHU ICAN in partnership with the French Health Data Hub, this challenge is based on the analysis of a fraction of the genome of patients suffering from hypertrophic cardiomyopathy.

The objective of the project is to develop a tool to help cardiogeneticists to identify patients at high risk of MACE (serious complications and progress) towards personalized patient care, based on this new knowledge.

This project is supported by the plan France 2030 in partnership with the Health Data Hub.

The Challenge: Automating Blood Smear Analysis

The objective is to identify which combination(s) of SNPs (associated or not with clinical data obtained during diagnosis of the disease) can predict which patients will experience the most serious complications. Competitors will have access to clinical data, SNPs, and Patient’s disease severity at baseline from more than 1,000 patients, and they will be asked to predict adverse outcome : serious complications during follow-up

The Vision Behind CardI-HACK

Hypertrophic cardiomyopathies (HCM) are hereditary heart diseases that are major causes of sudden death and/or heart failure in young people (<40 years old).

HCM is defined by an increase in the thickness of the left ventricle wall of the heart that cannot be explained by load conditions. The disease, which is relatively rare (1 in 500 people in the general population), is caused in about half of cases by a mutation in the sarcomere genes, one of the basic units of the contractile muscle cell (MYL3, TPMI, TNNI3, TNNT2, MYH7, and MYBPC3).

The drop in sequencing costs has led to an explosion in its use for all patients who need it. However, the sheer volume of results generated by sequencers has revealed an unexpected variety of genes involved, with each gene itself capable of being mutated in different ways, some of which are encountered very often (these are alleles with a so-called “frequent” variant, which can be combined in a polygenic risk score in the context of multifactorial background), and others much less often (alleles with a rare variant, in the classic sense of mendelian inheritance related to a pathogenic causal mutation). Given the number of genes that may be responsible for the disease and “ordinary” genetic variability : Single Nucleotide Polymorphisms (SNPs), many combinations of genetic variants are unique and therefore of limited clinical utility.

Adding to the confusion, clinicians are faced with patients' relatives who, despite having a similar rare pathogenic variant, may have completely different clinical presentations or disease progression, that may be related to each and every heterogeneous complex genetic background.

Eventually, current prognostic risk scores are only based on cardiac features, not on genetic features, and the relevance of the current model remains modest. Moreover, the distinction between high-risk and low-risk patients is highly insufficient and also controversial.

Risk calculation from ESC guidelines 2023[1] :

[1] Eur Heart J. 2023 Oct 1;44(37):3503-3626. doi: 10.1093/eurheartj/ehad194 /// PMID: 37622657

Building a Comprehensive Database

A critical component of this challenge is the creation of a diverse, high-quality database of patients in a relatively rare disease (1 patient for 500 births). The CardI-HACK team has explored the clinical records of hundreds of patients (who allowed it) at Pitié-Salpêtrière Faculty hospital in Paris. In order to protect patients from any reidentification risks, synthetic data were created from the original clinical data base.

The organizers of the challenge therefore offers the following dataset Features for each synthetical patient :

• 11 clinical data points describing the patient at the time of diagnosis

• Up to 288 common genetic variants (SNPs), split in one batch of priority SNPs

• One outcome defining the severity of the disease in the patient at the time of diagnosis => to be predicted

• One outcome column summarizing the most serious complications (composite endpoint of major advserse cardiovascular events  - MACE) suffered by patients during follow-up => to be predicted

This extended dataset provides an ideal foundation for training and evaluating machine learning models, offering participants a great opportunity to work with synthetic data generated from real-world medical data.

The Data Challenge: Your Role in Advancing Healthcare

We invite data scientists, machine learning enthusiasts, and researchers to participate in this ambitious project. Your mission is to develop and optimize algorithms capable of accurately predict patients’ adverse events, based on the provided dataset, and using mainly SNPs data from patients.

Objectives:

Enhance Diagnostic Tools: Create models that can assist clinicians in personalizing more efficiently care for HCM patients.

• Promote Equitable Healthcare: Contribute to solutions that reduce disparities in diagnostic capabilities across different regions.

• Advance Scientific Knowledge: Support the development of open-source AI tools and datasets for broader research applications.

Benefits of Participation:

• Access to Exclusive Data: Work with one of the most comprehensive datasets available from a low prevalence disease.

• Collaborative Platform: Utilize Trustii.io's resources to develop, test, and refine your models.

• Community Engagement: Connect with other professionals and experts in the field.

• Recognition and Impact: Your contributions could significantly influence patient care and set new standards in medical diagnostics.

Timeline and Important Dates

• Challenge Launch: November 4^th, 2025

• Submission Deadline: December 30^th, 2025

• Winners Announced: Early 2026

To recognize and reward excellence, a total of €14,000 will be awarded to the top three teams or individual participants who demonstrate outstanding performance in the challenge:

• 1st Prize: € 8,000

• 2nd Prize: € 4,000

• 3rd Prize: € 2,000

These prizes are intended to encourage innovation and acknowledge the significant effort required to develop effective the most unfavorable combination of SNPs. Winners will also have the unique opportunity to collaborate with our partners, further refining their solutions and exploring potential applications in clinical settings.

Detailed technical guidelines and evaluation criteria will be provided upon the launch of the challenge.

Looking Ahead: Broader Impacts and Future Prospects

The CardI-HACK project aspires not only to enhance and personalize care but also to foster continuous learning and collaboration within the medical and data science communities. By making the dataset and developed AI models openly accessible under the MIT license, we aim to:

• Encourage Further Research: Facilitate additional studies on the impact of SNPs and the integration with rare pathogenic genes as well as usual cardiologic features

• Promote International Collaboration: Lay the groundwork for similar initiatives on a global scale.

• Enhance Medical Training: Provide resources that can aid in the education and upskilling of healthcare professionals.

 

Join Us in Making a Difference

The CardI-HACK Data Challenge represents a significant step toward integrating artificial intelligence into essential healthcare services. Your expertise and innovative solutions are crucial for the success of this initiative.

We can't wait to see how your contributions will help revolutionize care for patients suffering from deadly cardiac diseases, and improve patient outcomes worldwide.

For any questions or further information, please contact us at challenges@trustii.io.

Stay tuned for updates, and prepare to embark on this exciting journey with us at Trustii.io!