Frailty Estimation in Elderly Oncology Patients Using Multimodal Wearable Data and Multi-Instance Learning

2026-04-08

AI Summary

This study addresses the inadequate monitoring of frailty in older oncology patients during clinical follow-up intervals due to the lack of continuous assessment. To overcome this, the authors propose an attention-based multiple instance learning framework that integrates multimodal wearable data—collected via smartwatches and ECG chest straps—to enable continuous prediction of frailty-related functional status under real-world conditions characterized by missing data and weak supervision. The method employs modality-specific MLP encoders (128-dimensional embeddings), HRV analysis, and activity/sleep feature extraction, with an attention mechanism to effectively fuse irregular and partially missing longitudinal data. Under leave-one-subject-out cross-validation, the model achieved balanced accuracies of 0.68/0.70 for grip strength and 0.59/0.64 for FACIT-F scores at the M3/M6 timepoints, respectively, demonstrating the efficacy of multimodal fusion for dynamic frailty assessment.

Abstract

Frailty and functional decline strongly influence treatment tolerance and outcomes in older patients with cancer, yet assessment is typically limited to infrequent clinic visits. We propose a multimodal wearable framework to estimate frailty-related functional change between visits in elderly breast cancer patients enrolled in the multicenter CARDIOCARE study. Free-living smartwatch physical activity and sleep features are combined with ECG-derived heart rate variability (HRV) features from a chest strap and organized into patient-horizon bags aligned to month 3 (M3) and month 6 (M6) follow-ups. Our innovation is an attention-based multiple instance learning (MIL) formulation that fuses irregular, multimodal wearable instances under real-world missingness and weak supervision. An attention-based MIL model with modality-specific multilayer perceptron (MLP) encoders with embedding dimension 128 aggregates variable-length and partially missing longitudinal instances to predict discretized change-from-baseline classes (worsened, stable, improved) for FACIT-F and handgrip strength. Under subject-independent leave-one-subject-out (LOSO) evaluation, the full multimodal model achieved balanced accuracy/F1 of 0.68 +/- 0.08/0.67 +/- 0.09 at M3 and 0.70 +/- 0.10/0.69 +/- 0.08 at M6 for handgrip, and 0.59 +/- 0.04/0.58 +/- 0.06 at M3 and 0.64 +/- 0.05/0.63 +/- 0.07 at M6 for FACIT-F. Ablation results indicated that smartwatch activity and sleep provide the strongest predictive information for frailty-related functional changes, while HRV contributes complementary information when fused with smartwatch streams.

Problem

• frailty
• elderly oncology
• functional decline
• wearable data
• multimodal sensing

Innovation

• multimodal wearable data
• multiple instance learning
• attention mechanism
• frailty estimation
• real-world missingness