AUTHOR=Lin Yen-Sheng , Kim Hyungtaek , Seitz Amee L. , Tsai Tsung-Yuan , Jain Nitin
TITLE=Computer-aided quantitative ultrasound algorithm of acromiohumeral distance among individuals with spinal cord injury
JOURNAL=Frontiers in Physics
VOLUME=11
YEAR=2023
URL=https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2023.1075753
DOI=10.3389/fphy.2023.1075753
ISSN=2296-424X
ABSTRACT=Background: Shoulder ultrasound is a well-established point-of-care diagnostic modality in orthopaedic and sports medicine. Despite offering measurements of high-quality morphology, this methodology has faced several challenges, including variability in ultrasound systems, operator dependency, and lack of reliable and objective quantitative measures to track disease progression and responses to therapeutic interventions. Computer-aided quantitative ultrasound algorithm (CAQUSA) is an emerging novelty that automates the detection of normal and abnormal structures. Although CAQUSA has been shown to improve detections and diagnoses of soft tissue lesions, the proof-of-concept of utilizing CAQUSA to measure subacromial space width and its encroachment to the rotator cuff tendon have not been tested to assist in clinical decision-making for subacromial pain syndrome.
Objective: This study aimed to develop a CAQUSA that measured the acromiohumeral distance (AHD) and test the algorithm’s reliability and agreement with manual measurements in wheelchair users with spinal cord injury (SCI).
Methods: 116 ultrasound video clips recorded from 10 manual wheelchair users with SCI were evaluated manually by an experienced examiner with expertise in AHD examination and by the CAQUSA, which was developed for bone segmentation with probability mapping. The reliability and agreement of the diagnostic performance between the examiner and the CAQUSA were calculated and compared in both groups of AHD measurements.
Results: The CAQUSA achieved a satisfactory agreement between computer-aided (11.95 ± 2.29 mm) and manual (11.33 ± 2.48 mm) measurements. The intraclass correlation coefficient between the two measures was excellent (ICC = 0.95). The sensitivity was 0.2 mm, with 95% true positives and 5% false positives at the fixed threshold of CAQUSA.
Conclusion: AHD measurements made by the clinical observer were similar to that of the CAQUSA for assessing the clinical metrics related to subacromial pain syndrome. However, the CAQUSA showed greater reliability in its unclear or misleading image analysis. Based on the findings of this proof-of-concept study, the CAQUSA has promise for clinical utilization and interchangeability to minimize examiner-dependent errors and potentially reduce the cost of care.