Open Access Research article

Fully automated system for the quantification of human osteoarthritic knee joint effusion volume using magnetic resonance imaging

Wei Li1, François Abram1, Jean-Pierre Pelletier2, Jean-Pierre Raynauld2, Marc Dorais3, Marc-André d'Anjou4 and Johanne Martel-Pelletier2*

Author Affiliations

1 ArthroVision Inc., 1871 Sherbrooke Street East, Montreal, Quebec H2K 1B6, Canada

2 Osteoarthritis Research Unit, University of Montreal Hospital Research Centre (CRCHUM), Notre-Dame Hospital, 1560 Sherbrooke St. East, Montreal, Quebec H2L 4M1, Canada

3 StatSciences Inc., 60 rue Sylvio Leduc, Notre-Dame de l'Île Perrot, Quebec J7V 7P2, Canada

4 Department of Clinical Sciences, Faculty of Veterinary Medicine, University of Montreal, 3200 Sicotte, Saint-Hyacinthe, Quebec J2 S 2M2, Canada

For all author emails, please log on.

Arthritis Research & Therapy 2010, 12:R173  doi:10.1186/ar3133

Published: 16 September 2010

Abstract

Introduction

Joint effusion is frequently associated with osteoarthritis (OA) flare-up and is an important marker of therapeutic response. This study aimed at developing and validating a fully automated system based on magnetic resonance imaging (MRI) for the quantification of joint effusion volume in knee OA patients.

Methods

MRI examinations consisted of two axial sequences: a T2-weighted true fast imaging with steady-state precession and a T1-weighted gradient echo. An automated joint effusion volume quantification system using MRI was developed and validated (a) with calibrated phantoms (cylinder and sphere) and effusion from knee OA patients; (b) with assessment by manual quantification; and (c) by direct aspiration. Twenty-five knee OA patients with joint effusion were included in the study.

Results

The automated joint effusion volume quantification was developed as a four stage sequencing process: bone segmentation, filtering of unrelated structures, segmentation of joint effusion, and subvoxel volume calculation. Validation experiments revealed excellent coefficients of variation with the calibrated cylinder (1.4%) and sphere (0.8%) phantoms. Comparison of the OA knee joint effusion volume assessed by the developed automated system and by manual quantification was also excellent (r = 0.98; P < 0.0001), as was the comparison with direct aspiration (r = 0.88; P = 0.0008).

Conclusions

The newly developed fully automated MRI-based system provided precise quantification of OA knee joint effusion volume with excellent correlation with data from phantoms, a manual system, and joint aspiration. Such an automated system will be instrumental in improving the reproducibility/reliability of the evaluation of this marker in clinical application.