Let Curves Speak: A Continuous Glucose Monitor based Large Sensor Foundation Model for Diabetes Management

2024-12-12

arXiv.org

AI Summary

Existing CGM-AI models suffer from poor generalizability, limiting cross-patient, cross-disease (T1D/T2D), and cross-temporal transfer. To address this, we propose CGM-LSM—the first large-scale sensor foundation model for continuous glucose monitoring—treating each patient as a temporal glycemic trajectory and leveraging a Transformer architecture for self-supervised pretraining on 15.96 million real-world CGM records from 592 patients. This work pioneers the application of large language model paradigms to glycemic time-series modeling. CGM-LSM uncovers implicit physiological dynamics embedded in sensor data, enabling robust, generalizable predictions across diverse populations, disease types, and diurnal cycles. On the OhioT1DM dataset, it achieves an rMSE of 15.64 mg/dL for 1-hour-ahead prediction—improving upon the state-of-the-art by 51%. For 2-hour-ahead prediction, rMSE reaches 29.81 mg/dL (T1D) and 23.49 mg/dL (T2D), markedly enhancing clinical utility.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Other Foundations of Cognitive Modeling & Systems
• Natural Language Processing: (Large) Language Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

While previous studies of AI in diabetes management focus on long-term risk, research on near-future glucose prediction remains limited but important as it enables timely diabetes self-management. Integrating AI with continuous glucose monitoring (CGM) holds promise for near-future glucose prediction. However, existing models have limitations in capturing patterns of blood glucose fluctuations and demonstrate poor generalizability. A robust approach is needed to leverage massive CGM data for near-future glucose prediction. We propose large sensor models (LSMs) to capture knowledge in CGM data by modeling patients as sequences of glucose. CGM-LSM is pretrained on 15.96 million glucose records from 592 diabetes patients for near-future glucose prediction. We evaluated CGM-LSM against state-of-the-art methods using the OhioT1DM dataset across various metrics, prediction horizons, and unseen patients. Additionally, we assessed its generalizability across factors like diabetes type, age, gender, and hour of day. CGM-LSM achieved exceptional performance, with an rMSE of 29.81 mg/dL for type 1 diabetes patients and 23.49 mg/dL for type 2 diabetes patients in a two-hour prediction horizon. For the OhioT1DM dataset, CGM-LSM achieved a one-hour rMSE of 15.64 mg/dL, halving the previous best of 31.97 mg/dL. Robustness analyses revealed consistent performance not only for unseen patients and future periods, but also across diabetes type, age, and gender. The model demonstrated adaptability to different hours of day, maintaining accuracy across periods of various activity intensity levels. CGM-LSM represents a transformative step in diabetes management by leveraging pretraining to uncover latent glucose generation patterns in sensor data. Our findings also underscore the broader potential of LSMs to drive innovation across domains involving complex sensor data.

Problem

• Improving glucose forecasting accuracy for diabetes management
• Enhancing generalization across diverse patient populations
• Developing a pretrained large sensor model for CGM data

Innovation

• Transformer decoder-based Large Sensor Model
• Pretrained on 1.6 million CGM records
• Improves glucose prediction accuracy significantly