Early detection system for Alzheimer's wins SAS Hackathon

Katalyze Data's DementAI, powered by SAS Viya, helps clinical teams find subtle patterns of decline earlier, improving care

DALLAS, April 29, 2026 /PRNewswire/ -- SAS INNOVATE -- The overall winner of the fifth global SAS Hackathon has developed an AI‑driven system designed to identify Alzheimer's disease up to two years sooner than existing approaches.

The DementAI team from UK‑based consultancy Katalyze Data, a long-time SAS Partner, was named the 2025 SAS Hackathon Grand Champion at SAS Innovate, the company's global data and AI conference.

The team's system, also called DementAI, analyzes existing clinical records and detects early risk signals for Alzheimer's. By flagging patients who show subtle patterns of decline before they have been referred to specialists, the solution gives doctors and clinical teams additional time to intervene with care and treatments when they are more effective.

"We didn't build DementAI just to make predictions; we built it to buy patients time," said Tamás Bosznay, Principal Consultant at Katalyze Data. "By surfacing the signals already hiding in plain sight within clinical records, the system is designed to help ensure that when care teams are ready to act, the right patients are identified earlier and more consistently."

Earlier diagnosis to improve Alzheimer's care
Dementia is a broad term for a decline in mental ability, often coming with age. Alzheimer's Disease is the most common cause of dementia. According to Alzheimer's Disease International, someone in the world develops dementia every 3 seconds. In 2020, there are 55 million people worldwide living with dementia. That number is expected to reach 78 million by 2030 and 139 million by 2050.

People living with Alzheimer's too often remain undiagnosed until the disease has significantly progressed. DementAI is designed to address this gap. Healthcare providers can use the system to uncover risk signals present in existing clinical records rather than needing to perform additional screening.

The team estimates that the tool could help identify Alzheimer's up to two years earlier than existing approaches, based on retrospective analysis of patient records, while also reducing the time and resources needed for diagnosis.

Built and refined during the SAS Hackathon as an end-to-end working prototype, the DementAI system works across multiple stages of care that are typically handled separately, from analyzing clinical data to deploying models within decision pathways.

It combines structured medical records and brain scans with unstructured clinical information such as physician notes, using synthetic data where appropriate. The AI-powered tool blends these signals to surface subtle patterns of decline that can be difficult to identify during brief consultations.

"What was notable about the DementAI project was not only the accuracy of the models but how the team approached development using governed workflows and privacy‑preserving techniques such as synthetic data," said Iain Brown, SAS' Global Head of AI & Data Science and the SAS Hackathon mentor for the DementAI team. "In regulated healthcare environments, that foundation is critical to moving from prototype to real‑world pilot deployment."

Built with SAS Viya
To develop the DementAI solution, the team used SAS^® Viya^® to manage data, develop models and operationalize analytics within its decision workflows. SAS Viya is SAS' cloud-native and cloud-agnostic data and AI platform.

DementAI was designed with AI governance and responsible AI in mind. Its data and models include explainability, audit trails, ongoing monitoring and bias reporting, which are particularly important in sensitive healthcare settings.

As they selected the 2025 SAS Hackathon Grand Champion, a panel of judges comprised of data and AI experts noted that DementAI stood out for its focus on practical deployment and clinical relevance. The judges lauded the Katalyze team for applying AI and advanced analytics to improve patient outcomes while helping health systems manage growing demand with limited resources.

What is the SAS Hackathon?
Katalyze developed and refined DementAI during the SAS Hackathon, a monthlong annual global event that brings together teams from around the world to collaborate, compete and innovate using data and AI.

During the SAS Hackathon, teams of data scientists, developers, business analysts and students work together – often virtually across countries, regions and continents – to experiment with technology, learn from one another and apply advanced analytics to real-world business and social challenges.

The most recent 2025 SAS Hackathon was supported by SAS in partnership with Microsoft and Intel. It attracted a record 2,058 registrants from 708 organizations, universities and SAS Partners across 66 countries. From those entries, 125 teams were approved to compete, gaining hands-on experience while supported by 134 SAS mentors, many of whom worked with multiple teams.

In December, SAS announced category winners for the 2025 SAS Hackathon across 24 regional, industry, technology and specialty categories. DementAI was named champion in three categories: Health Care & Life Sciences (industry), Models (technology) and Data for Good (specialty).

About Katalyze Data
For over 35 years, Katalyze Data (formerly Amadeus Software) has been unlocking the power of data-driven decisions. During that time, it has developed an unrivalled reputation for exceeding the expectations of its customers in the pharmaceutical, financial and public sectors. A long-time SAS Partner, Katalyze was named SAS Global Solution Provider Partner of the Year for 2025.

About SAS
SAS is a global leader in data and AI, helping organizations make confident decisions with AI they can trust. For decades, SAS has set the standard for delivering software that drives meaningful impact, incorporating deep industry expertise, transparency and governance. SAS gives you THE POWER TO KNOW^®.

SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. Copyright © 2026 SAS Institute Inc. All rights reserved.

Editorial Contact:
Mike Nemecek
[email protected]
919-531-5140
sas.com/news

SOURCE SAS