Related Experiment Video
Updated: May 9, 2026
Kinematic Analysis Using 3D Motion Capture of Drinking Task in People With and Without Upper-extremity Impairments
Published on: March 28, 2018
Augmenting Performance Status: A Preliminary Study of Objective Kinematic Assessment Using Motion Capture
Christopher Wang1, Ajay Prasad2, Elizabeth Qi3
1Division of Medical Oncology, Department of Medicine, USC Norris Comprehensive Cancer Center, Los Angeles, CA, USA. Christopher.wang2@med.usc.edu.
Introduction:
Understanding performance status (PS) is an essential aspect of oncologic care. PS is traditionally evaluated using the Karnofsky Performance Status (KPS) or Eastern Cooperative Oncology Group (ECOG) performance scales. Despite their ubiquity, these tools rely on subjective clinician interpretation and broad clinical categories, which contribute to poor inter-rater reliability and imprecise patient characterization. This study explores digital approaches to PS by leveraging three-dimensional video recordings and kinematic movement data to provide more objective, granular evaluations.
Methods:
In this single-center observational study, 34 oncology patients performed a standardized chair-to-table (CTT) movement. Of these participants, 52.9% were male and 47.1% were female. Patient movement was recorded using the Microsoft Kinect v2 sensor, which captures both color video (RGB) and depth-based skeletal tracking (Skeleton). Four oncologists independently reviewed the recordings and assigned KPS scores. Inter-rater reliability was assessed using the intraclass correlation coefficient (ICC). Pelvic acceleration metrics extracted from the Kinect data were correlated with summed KPS scores using Spearman's rank correlation.
Results:
The ICC for KPS scoring was 0.844 for RGB and 0.790 for Skeleton recordings, indicating good inter-rater reliability despite notable inter-observer variability. Vertical pelvic acceleration (mean and median) strongly correlated with summed KPS scores in both RGB (ρ ≈ 0.73-0.75, p < 0.0001) and Skeleton recordings (ρ ≈ 0.69-0.72, p < 0.0001). Horizontal and maximum acceleration showed weaker or nonsignificant associations.
Conclusion:
Kinect-derived motion capture offers an objective and reproducible approach to characterizing a patient's PS. Strong correlations between vertical pelvic acceleration and clinician-assigned KPS scores support its potential to augment both the accuracy and granularity of PS evaluation.
Trial Registry:
ClinicalTrials.gov ID: NCT07082257.
