Literature DB >> 28141515

Automatic Quantification of Radiographic Wrist Joint Space Width of Patients With Rheumatoid Arthritis.

Yinghe Huo1, Koen L Vincken2, Desiree van der Heijde3, Maria J H de Hair4, Floris P Lafeber4, Max A Viergever2.   

Abstract

Objective: Wrist joint space narrowing is a main radiographic outcome of rheumatoid arthritis (RA). Yet, automatic radiographic wrist joint space width (JSW) quantification for RA patients has not been widely investigated. The aim of this paper is to present an automatic method to quantify the JSW of three wrist joints that are least affected by bone overlapping and are frequently involved in RA. These joints are located around the scaphoid bone, viz. the multangular-navicular, capitate-navicular-lunate, and radiocarpal joints.
Methods: The joint space around the scaphoid bone is detected by using consecutive searches of separate path segments, where each segment location aids in constraining the subsequent one. For joint margin delineation, first the boundary not affected by X-ray projection is extracted, followed by a backtrace process to obtain the actual joint margin. The accuracy of the quantified JSW is evaluated by comparison with the manually obtained ground truth.
Results: Two of the 50 radiographs used for evaluation of the method did not yield a correct path through all three wrist joints. The delineated joint margins of the remaining 48 radiographs were used for JSW quantification. It was found that 90% of the joints had a JSW deviating less than 20% from the mean JSW of manual indications, with the mean JSW error less than 10%.
Conclusion: The proposed method is able to automatically quantify the JSW of radiographic wrist joints reliably. SIGNIFICANCE: The proposed method may aid clinical researchers to study the progression of wrist joint damage in RA studies.Objective: Wrist joint space narrowing is a main radiographic outcome of rheumatoid arthritis (RA). Yet, automatic radiographic wrist joint space width (JSW) quantification for RA patients has not been widely investigated. The aim of this paper is to present an automatic method to quantify the JSW of three wrist joints that are least affected by bone overlapping and are frequently involved in RA. These joints are located around the scaphoid bone, viz. the multangular-navicular, capitate-navicular-lunate, and radiocarpal joints.
Methods: The joint space around the scaphoid bone is detected by using consecutive searches of separate path segments, where each segment location aids in constraining the subsequent one. For joint margin delineation, first the boundary not affected by X-ray projection is extracted, followed by a backtrace process to obtain the actual joint margin. The accuracy of the quantified JSW is evaluated by comparison with the manually obtained ground truth.
Results: Two of the 50 radiographs used for evaluation of the method did not yield a correct path through all three wrist joints. The delineated joint margins of the remaining 48 radiographs were used for JSW quantification. It was found that 90% of the joints had a JSW deviating less than 20% from the mean JSW of manual indications, with the mean JSW error less than 10%.
Conclusion: The proposed method is able to automatically quantify the JSW of radiographic wrist joints reliably. SIGNIFICANCE: The proposed method may aid clinical researchers to study the progression of wrist joint damage in RA studies.

Entities:  

Keywords:  Bones; Diagnostic radiography; Feature extraction; Indexes; Joints; Wrist

Mesh:

Year:  2017        PMID: 28141515     DOI: 10.1109/TBME.2017.2659223

Source DB:  PubMed          Journal:  IEEE Trans Biomed Eng        ISSN: 0018-9294            Impact factor:   4.538


  1 in total

Review 1.  Use of artificial intelligence in imaging in rheumatology - current status and future perspectives.

Authors:  Berend Stoel
Journal:  RMD Open       Date:  2020-01
  1 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.