GlucoFM 发布:双流架构解析连续血糖监测数据

原标题:GlucoFM: Foundation model for continuous glucose monitoring
68AI 研判分

研究人员推出 GlucoFM,一种针对连续血糖监测(CGM)的基础模型。该模型创新性地采用双流架构,将缓慢的基线趋势与瞬态事件动态分离处理,克服了现有单流模型在捕捉复杂生理信号时的局限性,旨在提升稀疏标签下的血糖数据解读能力。

报道正文

本文目录10 个章节

消费级可穿戴设备利用运动和生理传感器来估算活动量和睡眠情况,但这些信号仅能提供对血糖调节的间接视角。连续血糖监测仪(CGM)通过植入皮下的微型传感器每隔几分钟追踪一次组织间液葡萄糖,从而补充这些测量数据,捕捉空腹、夜间及餐后模式。然而,解读这些轨迹仍然充满挑战,尤其是在有助于解释这些数据的高质量临床标签稀缺且获取成本高昂的情况下。

许多现有的 CGM 基础模型——包括 CGMformer、GluFormer 和 CGM-JEPA——通过单一表示流处理葡萄糖数据,而非明确分离缓慢的基线动态与瞬态事件动态。但 CGM 并非无差别的数据流:它包含相对缓慢的基线模式,其间穿插着可能反映进食、活动或传感器伪影的短期偏差。如果我们能利用每日 CGM 数据,在有限标注数据下估算糖尿病风险、胰岛素抵抗和β细胞功能障碍等情况,那会怎样?

这就是我们构建 GlucoFM 的原因。它是一个采用双流设计的自监督基础模型,将较慢的血糖趋势与短期偏差分离开来,同时保留时间上下文和缺失值信息。随后,通过潜在预测目标学习其日常上下文和时间演化规律。我们在四个多样化的队列中,针对七项临床预测任务对 GlucoFM 进行了评估——包括糖尿病风险、胰岛素抵抗、β细胞功能障碍、高脂血症、低血糖症、肥胖以及血糖表型——共计 14 项“队列–任务”评估。在这些评估中,GlucoFM 的平均 PR-AUC 比表现最佳的 GluFormer 变体高出 5.8 个百分点,且两者均在相同语料库上进行预训练。在 PR-AUC 指标上,GlucoFM 在所有糖尿病风险和 β 细胞功能障碍评估中均领先,并在四项胰岛素抵抗评估中的三项中取得最优成绩。我们还评估了 GlucoFM 在餐后血糖反应(PPGR)预测方面的表现。在输入条件和评估协议一致的情况下,GlucoFM 实现了最低的平均绝对误差(MAE),该结果基于两种连续血糖监测设备(Dexcom 和 Libre)的数据平均值计算得出。此外,GlucoFM 展现了最佳的跨数据集迁移性能,并表现出强大的少样本适应能力,即使在新队列或标注受试者数据极其有限的情况下也是如此。

训练 GlucoFM 以理解代谢

我们在来自 Wear-CGM[8e73d8] 及四个已发布数据集的共计 109,066 小时无标注 CGM(连续血糖监测)数据上对 GlucoFM 进行了预训练,这些数据涵盖 477 份参与者/会话记录。

CGM 记录可能包含数据缺口、不同的采样间隔以及传感器伪影。GlucoFM 将每条记录对齐到 24 小时、5 分钟的时间网格,并保留观测掩码(observation mask),以区分实际测量位置与未观测位置。其双流编码器将代表较缓血糖趋势的低频状态分量,与捕捉由生理变化、行为因素或传感伪影引起的短期偏差的残差事件分量分离开来。

鉴于原始血糖读数易受测量噪声和传感器伪影的影响,GlucoFM 并未直接重建精确的原始数值,而是采用潜在预测预训练策略,包含两个互补任务:

  • 上下文预测: 我们对每日血糖序列的部分数据进行掩码处理(即隐藏),并要求模型根据周围上下文预测这些被隐藏部分的潜在表示。通过在潜在空间中进行预测,模型能够捕捉更广泛的每日血糖模式,而无需重建每一个传感器读数。
  • 时间动态: 我们还训练模型预测个体的稳定基线和短期偏差如何随时间从一小时过渡到下一小时。这促使模型捕捉血糖变化的连续性特征,而非将读数视为孤立的时间快照。

最后,CGM 感知增强技术引入了基线漂移、类似压缩的信号骤降、稀疏采样以及短暂断连等情况,使模型能够适应真实 CGM 记录中常见的变异性和数据缺失现象。

GlucoFM 的功能

我们在四个队列(CGMacros、Stanford、Hall 和 ShanghaiT2DM)以及七项临床预测任务上对 GlucoFM 进行了评估,并单独评估了两小时餐后血糖反应预测。具体而言,我们探讨了以下问题:其冻结表征对于未见受试者的单个 24 小时窗口是否具有信息量;它们是否为预测餐后葡萄糖轨迹提供了有用的历史背景;结合多日数据是否能改善受试者级别的预测;这些表征向新队列的迁移效果如何;以及在标注数据有限时,它们的适应效果如何。

代谢任务的准确性

首先,我们使用了受试者不相交的窗口级 线性探测。我们冻结每个模型的编码器,在单个 24 小时表征上训练线性分类器,并确保没有受试者同时出现在训练折和测试折中。这旨在检验单日表征对于未见受试者是否具有表型信息量,同时保留日间变异性。

GlucoFM 在所有评估方法中取得了最强的任务平均 PR-AUC。在 14 次队列–任务评估中,它将平均 PR-AUC 从在同一数据上重新训练的最强 CGM 特定基线的 54.7 提升至 58.8——绝对提升 4.1 个百分点,相对于该基线约为 7.5% 的相对提升。GlucoFM 在所有糖尿病风险和 β 细胞功能障碍评估中均取得了最高的 PR-AUC,并在四项胰岛素抵抗评估中的三项中表现最佳。

预测餐后血糖反应

为了在动态预测任务上测试 GlucoFM,我们利用每次记录进餐前可用的信息,预测相对于进餐起始值的完整两小时葡萄糖变化轨迹。我们使用受试者不相交的 交叉验证 评估了来自 34 名受试者的 874 个配对进餐事件,其中 Dexcom 和 Libre(两种 CGM 设备)在相同的划分下分别建模。

[注:本文篇幅超长,以上为核心前篇精译,后续内容保留原文呈现]

We progressively combined each frozen model representation with one hour of pre-meal CGM, meal nutrition — including energy, carbohydrate, fat, protein, and dietary fiber — and participant-level information such as fasting glucose, BMI, and diabetes status. With the full context, GlucoFM achieved the lowest mean MAE among the evaluated models: 21.88 mg/dL, compared with 22.90 mg/dL for the best baseline and 27.69 mg/dL for the train-fold mean baseline. These results suggest that GlucoFM provides complementary historical context for predicting postprandial glucose changes.

Looking beyond a single day

A single 24-hour trace may not fully capture a person’s glucose patterns, so we tested whether combining multiple days improves subject-level prediction. GlucoFM encoded each day separately and averaged representations across up to seven days, with each participant contributing equally.

As the chart below shows, additional days improved PR-AUC in most settings across most datasets, including gains of 9.6 points for Stanford beta-cell dysfunction and 14.0 points for Hall diabetes prediction. CGMacros also showed mostly positive gains across Dexcom, Libre, and fused sensor data. ShanghaiT2DM insulin resistance was the main exception under simple averaging, indicating that the best aggregation strategy can vary by task. Overall, GlucoFM’s frozen daily representations can be combined to strengthen subject-level prediction without retraining the encoder.

Crossing the cohort divide

Next, we wanted to know if the physiological patterns our model learns can generalize. If we train a downstream classifier to spot diabetes risk using data from one clinical cohort, will it still work on patients from an entirely different study? The cross-dataset transfer bar chart highlights how GlucoFM handles this challenge, specifically plotting its direct improvement over the second-best model for diabetes risk and insulin resistance.

In the chart below, positive bars indicate that GlucoFM outperformed the strongest competing method: it led in 11 of 12 evaluations by 0.5 – 8.6 PR-AUC points and trailed once by 0.6 points. Its absolute PR-AUC ranged from 61.6% for both Stanford-to-Hall tasks to 90.0% for Hall-to-CGMacros insulin resistance, showing that focus on underlying physiology helps the frozen representations look past cohort-specific noise to find universal metabolic patterns.

Learning with less data

Labeled clinical data are expensive to obtain, so we also tested GlucoFM under two few-shot settings: the left plot varies the number of labeled participants per class, while the right varies the fraction of observations available from each participant. Moving right adds labeled data, and higher points indicate better task-averaged PR-AUC.

The orange GlucoFM markers are highest at every evaluated data budget, including the most limited settings of one per class and 1% of observations. The advantage is especially clear when labeled subjects are scarce, showing the model is highly efficient at picking up the right signals even with just a handful of examples.

Modeling glucose dynamics at two timescales

We also looked closely at whether splitting the signal into two streams actually made a difference. We compared the full dual-stream design with simpler alternatives: one that processes raw glucose directly, one designed to emphasize slower trends, and one designed to emphasize faster, short-term deviations.

As our encoder design analysis showed, the "event-only" version was the weakest, proving that transient fluctuations alone are not enough for a stable metabolic picture. While the raw-input and "state-only" versions were competitive, the full dual-stream model consistently came out on top. These results support organizing slower and faster glucose dynamics as complementary streams before combining them, rather than relying on either stream alone.

Conclusion

Our results suggest that CGM models can benefit from explicitly accounting for the multiscale structure of glucose dynamics, including slower trends, short-term deviations, daily timing, and sensor missingness. By learning reusable patterns from unlabeled CGM, GlucoFM produced representations that performed strongly across the evaluated prediction, transfer, and few-shot settings, offering a way to make better use of limited labeled clinical data.

Metabolic responses vary across people, cohorts, and sensor devices, while our current pre-training population remains modest. Our next steps are to train on larger and more diverse populations and extend GlucoFM beyond independently processed 24-hour windows toward native multi-day modeling to capture trends that unfold over weeks or months, and explore how these representations handle real-time changes. There is still so much to learn about metabolic health, and we are excited to see where these tools take us next.

Acknowledgements

The following researchers contributed to this work: Zechen Li, Keerthana Natarajan, Weizhi Zhang, Simon A. Lee, Yuwei Zhang, Maxwell A Xu, Menglian Zhou, Zeinab Esmaeilpour, Flora D. Salim (from the University of New South Wales), Mark Malhotra, Lindsey Sunden, Shwetak Patel, Yuzhe Yang, and Ahmed A. Metwally.

We gratefully acknowledge Bobak J. Mortazavi and Ricardo Gutierrez-Osuna (Texas A&M University) for providing theCGMacrosdataset used in this study.