Abstract. Computerized methods promise quick, objective, and sensitive tools to quantify progression of radiological damage in rheumatoid arthritis (RA). Measurement of joint space width (JSW) in finger and wrist joints with these systems performed comparable to the Sharp-van der Heijde score (SHS). A next step toward clinical use, validation of precision and accuracy in hand joints with minimal damage, is described with a close scrutiny of sources of error. A recently developed system to measure metacarpophalangeal (MCP) and proximal interphalangeal (PIP) joints was validated in consecutive hand images of RA patients. To assess the impact of image acquisition, measurements on radiographs from a multicenter trial and from a recent prospective cohort in a single hospital were compared. Precision of the system was tested by comparing the joint space in mm in pairs of subsequent images with a short interval without progression of SHS. In case of incorrect measurements, the source of error was analyzed with a review by human experts. Accuracy was assessed by comparison with reported measurements with other systems. In the two series of radiographs, the system could automatically locate and measure 1003/1088 (92.2%) and 1143/1200 (95.3%) individual joints, respectively. In joints with a normal SHS, the average (SD) size of MCP joints was 1.7 AE 0.2 and 1.6 AE 0.3 mm in the two series of radiographs, and of PIP joints 1.0 AE 0.2 and 0.9 AE 0.2 mm. The difference in JSW between two serial radiographs with an interval of 6 to 12 months and unchanged SHS was 0.0 AE 0.1 mm, indicating very good precision. Errors occurred more often in radiographs from the multicenter cohort than in a more recent series from a single hospital. Detailed analysis of the 55/1125 (4.9%) measurements that had a discrepant paired measurement revealed that variation in the process of image acquisition (exposure in 15% and repositioning in 57%) was a more frequent source of error than incorrect delineation by the software (25%). Various steps in the validation of an automated measurement system for JSW of MCP and PIP joints are described. The use of serial radiographs from different sources, with a short interval and limited damage, is helpful to detect sources of error. Image acquisition, in particular repositioning, is a dominant source of error. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.
Objective. To further simplify the simple erosion narrowing score (SENS) by removing scored areas that contribute the least to its measurement precision according to analysis based on item response theory (IRT) and to compare the measurement performance of the simplified version to the original. Methods. Baseline and 18-month data of the Combinatietherapie Bij Reumatoide Artritis (COBRA) trial were modeled using longitudinal IRT methodology. Measurement precision was evaluated across different levels of structural damage. SENS was further simplified by omitting the least reliably scored areas. Discriminant validity of SENS and its simplification were studied by comparing their ability to differentiate between the COBRA and sulfasalazine arms. Responsiveness was studied by comparing standardized change scores between versions. Results. SENS data showed good fit to the IRT model. Carpal and feet joints contributed the least statistical information to both erosion and joint space narrowing scores. Omitting the joints of the foot reduced measurement precision for the erosion score in cases with below-average levels of structural damage (relative efficiency compared with the original version ranged 35-59%). Omitting the carpal joints had minimal effect on precision (relative efficiency range 77-88%). Responsiveness of a simplified SENS without carpal joints closely approximated the original version (i.e., all D standardized change scores were 0.06). Discriminant validity was also similar between versions for both the erosion score (relative efficiency 5 97%) and the SENS total score (relative efficiency 5 84%). Conclusion. Our results show that the carpal joints may be omitted from the SENS without notable repercussion for its measurement performance.
Joint damage in rheumatoid arthritis is frequently assessed using radiographs of hands and feet. Evaluation includes measurements of the joint space width (JSW) and detection of erosions. Current visual scoring methods are timeconsuming and subject to inter-and intra-observer variability. Automated measurement methods avoid these limitations and have been fairly successful in hand radiographs. This contribution aims at foot radiographs. Starting from an earlier proposed automated segmentation method we have developed a novel model based image analysis algorithm for JSW measurements. This method uses active appearance and active shape models to identify individual bones. The model compiles ten submodels, each representing a specific bone of the foot (metatarsals 1-5, proximal phalanges 1-5). We have performed segmentation experiments using 24 foot radiographs, randomly selected from a large database from the rheumatology department of a local hospital: 10 for training and 14 for testing. Segmentation was considered successful if the joint locations are correctly determined. Segmentation was successful in only 14%. To improve results a step-by-step analysis will be performed. We performed JSW measurements on 14 randomly selected radiographs. JSW was successfully measured in 75%, mean and standard deviation are 2.30±0.36mm. This is a first step towards automated determination of progression of RA and therapy response in feet using radiographs.
Background Current clinical scoring methods for hand radiographs in RA are time consuming and subject to intra and inter-reader variance. Several methods for partially automatic radiographic assessment of hand radiographs were proposed [1]. These methods depend on detection (segmentation) of bones and joints. Their success rate is unpublished. We developed a semi-automated pattern recognition method [2]. To model bone shape variation independent from hand positions, this method used connected submodels and an iterative search to find the bones of the hand. These models were combined into a single model of the entire hand. The wide variation in the position of hands on subsequent radiographs was an obstacle in joint recognition and measurement of joint space width. We have therefore developed a positioning frame with 7 pins to reduce positioning error. Objectives Compare the success rate of joint recognition on hand radiographs made without and with a positioning aid. Methods The positioning frame was introduced in 2011 for hand radiographs of RA patients. A random selection of 91 images made before, and 87 made after introduction of the frame was analyzed using the previous described model [2]. Radiographs made with the frame were analyzed with the same model, with the addition to check for correct detection of hands relative to the pins. Processed images are intended for measurements of joint damage and have bone outlines surrounded by boxes. For 14 joints per hand (MCP, PIP, DIP) segmentation was judged by a rheumatologist and considered correct when at least part of the joint space was included in the overlapping boxes around the adjoining bones. Results Conclusions The use of a hand positioning frame improves the performance of previously developed segmentation software for analysis of hand radiographs. This brings us closer to a fully automated system that measures joint damage. Training of the model using a wider range of hands may further improve segmentation. References J.T.Sharp et al, Computer based methods for measurement of joints space width: update of an ongoing OMERACT project, Journal of Rheumatology, 34(4):874-883,2007 J.A.Kauffman et al, Segmentation of hand radiographs by using multi-level connected active appearance models, Proceedings of Medical Imaging2005: Image Processing, 5747:1571-1581,2005 Disclosure of Interest None Declared
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2025 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.