Automatic quantification of osteoarthritis in hand radiographs, by joint space width measurements

Graduation project in collaboration with Dr. Emile A. Hendriks
Information and Communication Theory Group, Technical University Delft.

Ronald van 't Klooster, Berend Stoel


Figure3Osteoarthritis (OA) is a painful disease, in which the joints are chronically inflamed. It is the most common chronic musculoskeletal disease, characterized by slowly progressive destruction of joints. Above the age of 65 years, it occurs in approximately 20% of the population, predominantly in women. The prevalence of OA is increasing therefore due to aging of the population. In clinical trials on possible treatments of OA, progress is assessed currently by visual comparison of radiographic images of joints, taken over varying periods of time. These methods are semi-quantitative, and do not necessarily reflect disease progression. Therefore, a more reproducible and quantitative method for measuring OA progression is required, in order to accelerate and improve evaluation of new treatments.


The develop automatic quantification of the joint space width. As the loss of hyaline cartilage can be observed by a decrease in the width of the joint seen on radiographs (see Figure 1). This method is to be validated against visual scoring of the same joints.


The joints are detected and assessed in the following steps:

  1. Identification of 12 joint locations  (see Figure 2)
    The individual fingers are detected, and through image filtering, the location of the joints are found.
  2. Detection of the joint margins  (see Figure 3)
    These contours are detected automatically, and can be adjusted manually, if necessary.
  3. Determination of the measurement interval, based on the thickness of the proximal phalanx as a reference, to compensate for the formation of osteophytes  (Figure 4).
  4. Quantification of the Joint Space
    The mean distance between the proximal and distal margin is determined.

In the validation study, it was found that the Joint Space Width measurement was significantly correlated with visual scoring (OARSI score).


Software program and its validation has been finalized. The system is now being applied in clinical trials.


van 't Klooster R, Hendriks EA, Watt I, Kloppenburg M, Reiber JH, Stoel BC, Automatic quantification of osteoarthritis in hand radiographs: validation of a new method to measure joint space width, Osteoarthritis Cartilage. 2008 16(1):18-25.


Figure 1 Example of a healthy and affected joint

Figure 2. Detection of joint locations

Figure 3. Detection of joint margins

Figure 4. Automatic determination of measurement intervals


For further information, please contact:
Berend C. Stoel, PhD.
Division of Image Processing
Department of Radiology, 1-C2S
Leiden University Medical Center
P.O. Box 9600
2300 RC Leiden
The Netherlands
Tel. +31 (0)71 526 1911
Fax. +31 (0)71 526 6801