Joint research with Science Tokyo: Wearable device data improves prediction accuracy
Selected for poster tour at European rheumatology conference

TechDoctor Inc. (CEO: Kazumichi Minato, Headquarters: Chuo-ku, Tokyo; hereinafter "TechDoctor") presented research results on predicting disease activity in systemic lupus erythematosus (SLE) at the European Alliance of Associations for Rheumatology (EULAR 2025) held in Barcelona, Spain, from June 11-14, 2025.

This presentation reports results from research adopted by the Japan Agency for Medical Research and Development (AMED) FY2024 "Medical Equipment Research Outcome Development Project (Challenge Type)." Science Tokyo (formerly Tokyo Medical and Dental University) is the lead institution (Research Representative:Dr. Tadashi Hosoya, Department of Rheumatology), and TechDoctor participates as a research development partner institution.

We reported on the development of an algorithm to predict achievement of the SLE disease activity indicator "Lupus Low Disease Activity State (LLDAS)" with higher accuracy using biometric data obtained from wrist-worn wearable devices (Google Fitbit) in addition to clinical indicators, medications, and patient-reported outcomes (PROs). This presentation was selected for a poster tour and received high attention.

--------------------------------------------------------------------------------

BACKGROUND

--------------------------------------------------------------------------------

Systemic lupus erythematosus (SLE) is an autoimmune disease that causes chronic inflammation in various organs throughout the body. While treatment outcomes in the acute phase have improved with advances in drug therapy in recent years,standardization of maintenance therapy considering relapse risk and judgment of safe drug reduction/discontinuation remain clinical challenges.In particular, there are limitations to the information physicians can grasp during examinations, and it has been pointed out that there is a possibility of not being able to fully understand the overall picture of disease activity. In this context, attention is being paid to the possibility that PROs (patient's own subjective symptoms) and continuous biometric data such as heart rate, sleep, and steps obtained through wearable devices can contribute to more comprehensive disease understanding.In this research, we aimed to construct a model to predict LLDAS achievement in SLE patients using such new data.

--------------------------------------------------------------------------------

RESEARCH OVERVIEW

--------------------------------------------------------------------------------

AMED Medical Equipment Research Outcome Development Project Challenge Type Research Development Project Name: Development of software to realize safe and appropriate treatment management by visualizing disease activity in SLE

Title: PREDICTION OF DISEASE ACTIVITY STATE IN PATIENTS WITH SYSTEMIC LUPUS

ERYTHEMATOSUS BY UTILIZING WRIST-WORN WEARABLE DEVICES
Citation: Niwano T. et al. Annals of the Rheumatic Diseases, Volume 84, 455-456
URL: https://ard.eular.org/article/S0003-4967(25)01590-0/abstract
Research Period: July 2023 - September 2024
Research Implementation Institution: Science Tokyo (formerly Tokyo Medical and Dental University)

Research Development Partner Institution: TechDoctor Inc.
Research Subjects: 274 SLE patients aged 15 and above (of which 112 consented to wearing Fitbit Inspire3)

Evaluation Indicators:
• Medical records (treatment history, test data)
• Patient-reported subjective symptoms (Visual Analog Scale (VAS), LupusPRO questionnaire)

• Heart rate, sleep, step count data obtained from Fitbit, etc.

Prediction Model Construction Method: Four machine learning methods were employed: Random Forest, CatBoost, XGBoost, and LightGBM
Biometric data from 7 days before and after the examination date were averaged daily and integrated with clinical information and PROs. A prediction model was constructed using LLDAS determination by physicians at examination as the correct label.

--------------------------------------------------------------------------------

RESEARCH RESULTS

--------------------------------------------------------------------------------

Among 112 individuals who wore Fitbit and 162 non-wearers, no significant differences were observed in gender, age, LLDAS achievement rate, etc., and both groups were considered equivalent populations.

For machine learning prediction model construction, we evaluated two types of models: [1] models using only "clinical indicators + treatment details + PROs" and [2] models adding Fitbit data to these. In all four machine learning models, the latter models with added Fitbit data showed improved ROC-AUC, confirming the additional value of health-related data in predicting disease activity.

These results suggest that combining continuous biometric data reflecting patients' own physical condition changes in addition to medical information obtained in clinical settings can contribute to more accurate understanding of SLE disease status.

--------------------------------------------------------------------------------

SIGNIFICANCE AND FUTURE PROSPECTS

--------------------------------------------------------------------------------

These results demonstrate new possibilities where utilization of patients' daily data can contribute to treatment optimization and coexistence with disease in the rare and intractable disease of SLE, where the full picture of the disease and treatment methods are not yet established.

Science Tokyo and TechDoctor will continue to advance development of digital biomarkers and social implementation of remote monitoring technology in diverse disease areas including autoimmune diseases, aiming to realize individually optimized medicine and patient-centered care based on these research results.

--------------------------------------------------------------------------------

REFERENCE INFORMATION

--------------------------------------------------------------------------------

Conference Name: EULAR 2025 (European Alliance of Associations for Rheumatology)Conference Dates: June 11-14, 2025 (Wednesday-Saturday)Location: Barcelona, Spain

Official Website: https://congress.eular.org/
EULAR Conference Abstract Collection: https://congress.eular.org/abstractarchive
*EULAR is one of the world's largest conferences in the field of rheumatology,
with approximately 14,000 participants from around the world.
Press Release at AMED Adoption (July 9, 2024):
https://www.technology-doctor.com/news/20240709

--------------------------------------------------------------------------------

ABOUT SYSTEMIC LUPUS ERYTHEMATOSUS (SLE)

--------------------------------------------------------------------------------

Systemic lupus erythematosus (SLE) is the third most common intractable disease in Japan. It is an autoimmune disease where the immune system attacks the body's own tissues, and is a type of collagen disease. It causes chronic inflammation in various organs throughout the body, with diverse symptoms. While treatment methods have improved in recent years and acute phase management has advanced, management of relapse risk and judgment of medication reduction/discontinuation remain challenges.

--------------------------------------------------------------------------------

ABOUT TECHDOCTOR INC.

--------------------------------------------------------------------------------

TechDoctor operates under the vision "Towards an era where data improves well-being," developing and promoting social implementation of "digital biomarkers" that derive health-related insights from daily sensing data collected by wearable devices and other means. We aim to realize AI-driven medicine based on data in collaboration with medical, pharmaceutical, food-related companies, and research institutions.

Company Name: TechDoctor Inc.
Website: https://www.technology-doctor.com/
Founded: June 21, 2019
Location: Kyobashi Ed-Grand 4F, 2-2-1 Kyobashi, Chuo-ku, Tokyo
CEO: Kazumichi Minato
Chief Medical Officer: Keisuke Izumi
Business: Development and operation of digital biomarker development platform
"SelfBase," provision of digital medical solutions

* About Digital Biomarkers

Digital biomarkers are indicators for continuously and objectively evaluating the presence or absence of disease, changes in disease status, and treatment effects based on daily biometric data obtained from smartphones, wearable devices, and other sources.

While traditional biomarkers were primarily "point data" measured temporarily at medical institutions, digital biomarkers are characterized by the ability to continuously acquire "line data" from daily life. Based on indicators such as exercise, sleep, and heart rate, they are expected to be used for early disease detection, treatment monitoring, and as new endpoints in drug development.

Development has been progressing overseas since around 2019, and attention is also increasing in Japan.