Quantifying Rheumatoid Arthritis Disease Activity using a Multimodal Sensing Knee Brace

Kristine L. Richardson, Caitlin N. Teague, Samer Mabrouk, Brandi Nicole Nevius, Goktug C. Ozmen, Rachel S Graham, Daniel Zachs, Adam Tuma, Erik J Peterson, Hubert Lim, Omer Inan

Research output: Contribution to journalArticlepeer-review

Abstract

Objective: Rheumatoid arthritis (RA) is a chronic inflammatory syndrome that features painful and destructive joint disease. Aggressive disease-modifying treatment can result in reduced symptoms and protection from irreversible joint damage; however, assessment of treatment efficacy is currently based largely on subjective measures of patient and physician impressions. In this work, we address this compelling need to provide an accurate and quantitative capability for monitoring joint health in patients with RA. Methods: Joint acoustic emissions (JAEs), electrical bioimpedance (EBI), and kinematics were measured noninvasively from 11 patients with RA over the course of three weeks using a custom multimodal sensing brace, resulting in 49 visits with JAE recordings and 43 with EBI recordings. Features derived from all sensing modalities were fed into a linear discriminant analysis (LDA) model to predict disease activity according to the validated disease activity index (the DAS28-ESR). Erythrocyte sedimentation rate (ESR) was predicted using ridge regression and classified into a high or low class using LDA. Results: DAS28-ESR level was predicted with an area under the receiver operating characteristic curve (AUC) of 0.82. With JAEs alone, we were able to track intrasubject differences in the disease activity score as well as classify ESR level with an AUC of 0.93. The majority of patients reported both an interest and ability to use the brace at home for longitudinal monitoring. Conclusion: This work demonstrates the ability to detect RA disease activity using noninvasive sensing. Significance: This system has the potential to improve RA disease activity monitoring by giving treating clinicians objective data that can be acquired independent of a face-to-face clinic visit.

Original languageEnglish (US)
Pages (from-to)1
Number of pages1
JournalIEEE Transactions on Biomedical Engineering
VolumePP
DOIs
StateAccepted/In press - 2022

Bibliographical note

Publisher Copyright:
IEEE

Keywords

  • Arthritis
  • Biomedical measurement
  • Electrical Bioimpedance
  • Joint Acoustic Emissions
  • Kinematics
  • Manganese
  • Monitoring
  • Multimodal sensors
  • Noninvasive Sensing
  • Recording
  • Rheumatoid Arthritis
  • Wearable

PubMed: MeSH publication types

  • Journal Article

Fingerprint

Dive into the research topics of 'Quantifying Rheumatoid Arthritis Disease Activity using a Multimodal Sensing Knee Brace'. Together they form a unique fingerprint.

Cite this