A feature that could be integrated into AssisDent’s "Paro View" in the future.
Roosa Rintamäki has been developing an innovative AI-powered tool that uses large language models to automatically summarize dental patient histories. 
➡️ What’s the Goal?
The aim is to create clear, concise summaries that illustrate how a patient's periodontal health has evolved over time. This includes tracking key indicators like Bleeding on Probing (BOP), and periodontal pocket depth - crucial metrics for understanding gum health and treatment progress.
➡️ Why It Matters
Dentists often face the challenge of navigating extensive and scattered patient data. This AI-powered summary could streamline clinical workflows, save valuable time, and ultimately enhance the quality of care by presenting essential information in one easily accessible view.
➡️ Thesis to Technology
The project began in January as part of Roosa’s master’s thesis in Information Technology (majoring in Data Science) at Tampere University, Faculty of Information Technology and Communication Sciences with support from Teknologiateollisuus - Technology Industries of Finland. The concept originated in house, where the team envisioned how this solution could bring added value to AssisDent. Recent studies have further supported the use of AI in automating medical data summarization.
➡️ How It Works
Roosa developed a dedicated microservice using Azure OpenAI Service, which provides secure and scalable access to OpenAI’s large language models. The summarization process is guided by carefully crafted prompts to ensure accuracy and reliability. Initial testing with synthetic data has yielded impressive results.
➡️ Current Status
The system is currently a prototype, demonstrating the technical feasibility and benefits of AI-based summarization. It’s built to be further developed and potentially integrated into AssisDent in the future.
➡️ What’s Next?
There’s significant potential to expand AI applications within AssisDent from summarizing tooth-specific or filling/fracture history to supporting shift planning. In fact, a new project in this direction is already underway.