Jeffrey Duryea

Trainable rule-based algorithm for the measurement of joint space width in digital radiographic images of the knee.

The progression of osteoarthritis (OA) can be monitored by measuring the minimum joint space width (mJSW) between the edges of the femoral condyle and the tibial plateau on radiographs of the knee. This is generally performed by a trained physician using a graduated magnifying lens and is prone to the subjectivity and variation associated with observer measurement. We have developed software that performs this measurement automatically on digitized radiographs. The test data consisted of 180 digitized radiographs of the knee (90 duplicate acquisitions) from 18 normal (nonarthritic) subjects and 38 images from 10 subjects with OA. These were digitized and manually cropped so that the images were free of nonanatomical structures and the knee was approximately centered. The software first determined the edge of the femoral condyle on 400 microm pixel subsampled images. Contours marking the location of the tibial plateau in the medial compartment were found on 100 microm images using the femoral edge as a reference. The algorithm was trained using an independent but similar data set and using a jackknife approach with the test data. The results were compared to contours drawn by a trained reader and the duplicate acquisitions were used to measure the reproducibility of the mJSW measurement. The reproducibility was 0.16 mm and 0.18 mm for normal and osteoarthritic knees, respectively, representing an improvement of approximately a factor of 2 over manual measurement. The algorithm also showed excellent agreement with the hand-drawn contours and with mJSW determined by the manual method.

Comparison of radiographic joint space width with magnetic resonance imaging cartilage morphometry: analysis of longitudinal data from the Osteoarthritis Initiative.

Magnetic resonance imaging (MRI) and radiography are established imaging modalities for the assessment of knee osteoarthritis (OA). The objective of our study was to compare the responsiveness of radiographic joint space width (JSW) with MRI‐derived measures of cartilage morphometry for OA progression in participants from the Osteoarthritis Initiative (OAI).

Digital tomosynthesis of hand joints for arthritis assessment

The two principal forms of hand arthritis, rheumatoid arthritis (RA) and osteoarthritis (OA) have large clinical and economic costs. Radiography has been shown to be a useful tool to assess the condition of the disease. A hand radiograph, however, is a two-dimensional projection of a three-dimensional object. In this report we present the results of a study that applied digital tomosynthesis to hand radiography in order to extract three-dimensional outcome measures that should be more sensitive to arthritis progression. The study was performed using simulated projection radiographs created using micro computed tomography (microCT) and a set of five dry-bone hand skeletons. These simulated projection images were then reconstructed into tomographic slices using the matrix inversion tomosynthesis (MITS) algorithm. The accuracy of the tomosynthesis reconstruction was evaluated by comparing the reconstructed images to a gold standard created using the microCT data. A parameter from image registration science, normalized mutual information, provided a quantifiable figure of merit. This study examined the effects of source displacement, number of reconstructed planes, number of acquisitions, noise added to the gray scale images, and errors in the location of a fiducial marker. We also optimized the reconstruction as a function of two variables k and alpha, that controlled the mixing of MITS with conventional shift-and-add tomosynthesis. A study using hand delineated joint margins demonstrated that MITS images provided a better measurement of average joint space width. We found good agreement between the MITS slices and the true planes. Both joint margins and trabecular structure were visible and the reconstructed slices showed additional structures not visible with the standard projection image. Using hand-delineated joint margins we compared the average joint space width of the gold standard slices to the MITS and projection images. A root-mean square deviation (RMSD), calculated for this comparison, gave RMSDproj = 0.18 mm and RMSDMITS = 0.14 mm for the projection and MITS images, respectively. We have demonstrated the potential of digital tomosynthesis for imaging of the hand to assess arthritic changes. We have also developed a methodology that can be used to optimize the technique and have studied the issues that will control the feasibility of clinical implementation.

Location specific radiographic joint space width for osteoarthritis progression

OBJECTIVE To establish the performance of location specific computer measures of radiographic joint space width (JSW) compared to measurements of minimum joint space width (mJSW) for the assessment of medial compartment knee osteoarthritis (OA). The study also investigated the most disease-responsive location for measuring medial compartment JSW. METHODS Serial bilateral Posterior Anterior (PA) conventional radiographs acquired with a fixed flexion protocol were obtained 36 months apart in 118 persons with knee OA participating in the Health, Aging and Body Composition (Health ABC) Study. Measurements of medial compartment mJSW and JSW at seven fixed locations were facilitated by the use of semi-automated software that delineated the femoral and tibial margins of the joint. A human reader operated custom software to verify and correct the software-drawn margins where necessary. Paired images were displayed with the reader blinded to the chronological order. The amount of joint space narrowing was measured and the standardized response mean (SRM) was used as a metric to quantify performance. RESULTS For all subjects, the mJSW SRM value was 0.42 while, for the most responsive location specific measure of JSW, it was SRM=0.46. For subjects with a Kellgren-Lawrence (KL) score less than or equal to 1, mJSW (SRM=0.40) was more responsive than the new measures (Maximum SRM=0.30). For KL=2or3, SRM=0.49 for mJSW, and SRM=0.74 for the most responsive location specific measure of JSW. Improved responsiveness was observed in the more central portion of the joint on the more diseased knees. CONCLUSIONS Location specific computer measures of JSW are feasible and potentially provide a superior method to assess radiographic OA for more diseased subjects. This new measure has the potential to improve the power of clinical studies that use a fixed flexion protocol.

Radiographic evaluation of osteoarthritis

Osteoarthritis is the most widespread form of arthritis in the United States. Classically, osteoarthritis has been grouped into primary and secondary types. Primary or idiopathic osteoarthritis is believed to be a sequela of altered biomechanical stresses across joints in susceptible individuals. Secondary osteoarthritis is a consequence of underlying cartilage damage, such as from preceding inflammatory arthritis, metabolic abnormality, or injury. The radiographic hallmark of osteoarthritis is asymmetric loss of cartilage space. Osteophytosis bony eburnation, subchondral cysts, and eventual subluxation follow. Osteoporosis and erosions are not usual features of this disease.

Neural network based algorithm to quantify joint space width in joints of the hand for arthritis assessment

Arthritis diseases are widespread with enormous societal costs. The two most common forms, rheumatoid arthritis and osteoarthritis, affect joints of the hand and cause narrowing of the joint spaces as the disease destroys the articular cartilage. Radiographic assessment is one of the most promising tools to detect subtle changes in joint space width (JSW), and therefore disease progression. Currently radiographic assessment of arthritis in joints of the hand is accomplished though semiquantitative subjective scoring systems which do not provide a quantitative measurement of the JSW. We describe here an automated method which calculates the average JSW of the metacarpophalangeal (MCP), proximal interphalangeal (PIP), and distal interphalangeal (DIP) joint spaces for fingers 2 to 5 (index, middle, ring, and little) on digitized hand radiographs. The method was tested with a set of 54 hand radiographs on joints with mild to moderate rheumatoid arthritis. Performance was evaluated by comparing algorithm measured JSW to a gold standard determined from expertly hand-drawn joint margins. The agreement was quantified by a measurement of root mean square deviation, 0.148 mm, 0.089 mm, and 0.114 mm for the MCP, PIP, and DIP joints, respectively. In addition, the algorithm measured JSW strongly correlated with the gold standard: R2=0.80 (MCP), R2= 0.82 (PIP), and R2= 0.84 (DIP). This is an accurate and robust algorithm and should provide a more quantitative measure of disease progression than current methods.

Semiautomated three-dimensional segmentation software to quantify carpal bone volume changes on wrist CT scans for arthritis assessment.

Rapid progression of joint destruction is an indication of poor prognosis in patients with rheumatoid arthritis. Computed tomography (CT) has the potential to serve as a gold standard for joint imaging since it provides high resolution three-dimensional (3D) images of bone structure. The authors have developed a method to quantify erosion volume changes on wrist CT scans. In this article they present a description and validation of the methodology using multiple scans of a hand phantom and five human subjects. An anthropomorphic hand phantom was imaged with a clinical CT scanner at three different orientations separated by a 30-deg angle. A reader used the semiautomated software tool to segment the individual carpal bones of each CT scan. Reproducibility was measured as the root-mean-square standard deviation (RMMSD) and coefficient of variation (CoV) between multiple measurements of the carpal volumes. Longitudinal erosion progression was studied by inserting simulated erosions in a paired second scan. The change in simulated erosion size was calculated by performing 3D image registration and measuring the volume difference between scans in a region adjacent to the simulated erosion. The RMSSD for the total carpal volumes was 21.0 mm3 (CoV = 1.3%) for the phantom, and 44.1 mm3 (CoV = 3.0%) for the in vivo subjects. Using 3D registration and local volume difference calculations, the RMMSD was 1.0-3.0 mm3 The reader time was approximately 5 min per carpal bone. There was excellent agreement between the measured and simulated erosion volumes. The effect of a poorly measured volume for a single erosion is mitigated by the large number of subjects that would comprise a clinical study and that there will be many erosions measured per patient. CT promises to be a quantifiable tool to measure erosion volumes and may serve as a gold standard that can be used in the validation of other modalities such as magnetic resonance imaging.

Automated measurement of radiographic hip joint‐space width

Radiographic joint-space narrowing (JSN) is the principle indicator of cartilage loss in osteoarthritis (OA). JSN is usually assessed qualitatively by visual inspection or in clinical research, is measured manually with a graduated handheld lens directly applied to the x-ray film, or from digitized radiographs by hand tracing the joint margins with a mouse. The minimum joint-space width (mJSW) and joint-space area (JSA) are recorded as the indices of OA progression in epidemiological studies and clinical drug trials. We present a computerized method that automatically finds the articular margins of the hip to improve determination of mJSW and JSA. The algorithm requires that three seed points are manually identified on the femoral head and uses three steps to process each digitized hip x-ray. First, a Hough transform finds the center and radius (R) of a circle that approximates the femoral head. Finding R indicates whether magnification differences must be corrected on repeat exams. Second, a gradient algorithm finds the edge of the femoral head and acetabulum. Third, the mid-line of the femoral neck is automatically found and used to define the joint portion (theta) that is assessed for narrowing. theta is fixed for follow-up exams of the same subject. The algorithm was evaluated in three ways to determine its performance characteristics. First, the inter-reader and intra-reader variability for mJSW and JSA associated with the selection of the seed points was found to be negligible (< 1%) compared to the variability associated with manual scoring with a lens or by tracing the joint margins with a mouse. Second, from duplicate hip x-rays of 19 subjects with OA, the Root Mean Square Standard Deviation and coefficient of variation for mJSW and JSA defined by the algorithm was determined to be better than manual techniques by at least a factor of 2. Third, the algorithm correctly identified the joint margin in more than 85% of the 105 cases tested. Automated measures of radiographic hip joint-space narrowing is less subjective than manual methods and may be applicable for monitoring OA progression in clinical research.

Patient Repositioning Reproducibility of Joint Space Width Measurements on Hand Radiographs

Computer‐based methods to measure radiographic joint space width (JSW) have the potential to improve the longitudinal assessment of rheumatoid arthritis (RA). The purpose of this report was to measure the long‐term patient repositioning reproducibility of software‐measured radiographic JSW.

Automated algorithm for the identification of joint space and phalanx margin locations on digitized hand radiographs

Rheumatoid arthritis (RA) of the hand can be characterized and assessed by the narrowing of the joint spaces which are ordinarily scored semiquantitatively by a radiologist using radiographs of the hand. Software which delineates and measures the joint spaces would be a useful tool for assessment. The first part of such an algorithm has been developed which determines the locations of the distal interphalangeal (DIP), proximal interphalangeal (PIP), and metacarpophalangeal (MCP) joint spaces for fingers 2–5 (index, middle, ring, and little) on digitized hand radiographs. In addition, points on the medial and lateral margins of each phalanx are identified which can be used as starting points for edge detection algorithms to provide segmentation of the phalanges. The algorithm is a C-language program running on a UNIX computer, uses a multiresolution approach operating, and requires approximately 10 CPU seconds per image. It was tested on a set of 54 radiographs taken from a multicenter rheumatoid arthritis study where a study protocol was followed. In addition, radiographs of individuals wearing rings and where nonanatomical structures contacted the anatomy proximal to the midpoint of the distal phalanx and distal to the MCP joint were eliminated from the data set. In order to make a quantitative assessment, regions of interest drawn by a trained radiologist were used as a gold standard. The algorithm had a success rate of 100% for the identification of each digit and over 99% for the identification of joint space locations and phalanx margins. Quantitative tests indicated excellent algorithm robustness. We have developed fully automated software which accurately identifies anatomical landmarks on digital images of the hand.