Progress Report
Challenge for Eradication of Diabetes and Comorbidities through Understanding and Manipulating Homeostatic Systems[3] Development of technology for easy acquisition of biometric data in humans and analysis of human data
Progress until FY2024
1. Outline of the project
This R&D Item is responsible for the development and social implementation of a method to detect and predict the early stages of diabetes and its co-morbidities as simply and non-invasively as possible based on the analysis of biological information, genome, and hepatic glucose uptake capacity using contact and non-contact devices (See figure below).
To achieve this, we are working on creating a highly accurate early diabetes detection algorithm, improving the accuracy of the Diabetes Omnigenic Model, and collecting data from the 13CO2 breath test as challenging themes. We are working on the concept of early detection of diabetes and heart failure from non-invasive devices only, which is completely different from conventional methods, using high-speed spectral cameras, AI, cohort data analysis, and other methods.
2. Outcome so far
(1) Development of AI system for early detection of heart failure at home
We have successfully developed an AI model that can determine the severity of heart failure using I-lead electrocardiograms measured by portable electrocardiographs (ECGs and smartwatches that can be used at home). With this system, heart failure patients can easily monitor their condition at home, which will reduce the risk of re-hospitalization and enable early therapeutic intervention.
(2) Examination of glucose oxidation and hepatic glucose disposal capacity using the 13CO2 breath test
We have intellectualized the 13C-glucose breath test, which tests glycoxidation capacity at the individual level by measuring 13CO2 emitted in the exhaled breath after ingestion of 13C-glucose.
(3) Construction of an algorithm to detect diabetes
Using a non-contact device (high-speed spectral camera), we have developed an algorithm for detecting diabetes in addition to an algorithm for detecting hypertension at an early stage. We are working on IP (Intellectual Property).
3. Future plans
In the future, we will further collect data from humans to enable early detection of hypertension and diabetes by non-invasive devices, and will try to tune the algorithm. By doing so, we aim to build algorithms that can withstand social implementation for the general public.
For the 13CO2 breath test, we will also try to accumulate data on a 75 g 13C -glucose load in order to link it to the data related to life expectancy obtained in the Ohasama cohort. This will allow for further matching with the life span-related results obtained in the Ohasama cohort.