Septic Arthritis Modeling Using Sonographic Fusion with Attention and Selective Transformation: a Preliminary Study

Overview

Journal J Imaging Inform Med

Publisher Springer Nature

Specialty Radiology

Date 2024 Sep 16

PMID 39284980

Authors

Chung-Ming Lo

Kuo-Lung Lai

Affiliations

Soon will be listed here.

Abstract

Conventionally diagnosing septic arthritis relies on detecting the causal pathogens in samples of synovial fluid, synovium, or blood. However, isolating these pathogens through cultures takes several days, thus delaying both diagnosis and treatment. Establishing a quantitative classification model from ultrasound images for rapid septic arthritis diagnosis is mandatory. For the study, a database composed of 342 images of non-septic arthritis and 168 images of septic arthritis produced by grayscale (GS) and power Doppler (PD) ultrasound was constructed. In the proposed architecture of fusion with attention and selective transformation (FAST), both groups of images were combined in a vision transformer (ViT) with the convolutional block attention module, which incorporates spatial, modality, and channel features. Fivefold cross-validation was applied to evaluate the generalized ability. The FAST architecture achieved the accuracy, sensitivity, specificity, and area under the curve (AUC) of 86.33%, 80.66%, 90.25%, and 0.92, respectively. These performances were higher than using conventional ViT (82.14%) and significantly better than using one modality alone (GS 73.88%, PD 72.02%), with the p-value being less than 0.01. Through the integration of multi-modality and the extraction of multiple channel features, the established model provided promising accuracy and AUC in septic arthritis classification. The end-to-end learning of ultrasound features can provide both rapid and objective assessment suggestions for future clinic use.

References

Weston V, Jones A, Bradbury N, Fawthrop F, Doherty M . Clinical features and outcome of septic arthritis in a single UK Health District 1982-1991. Ann Rheum Dis. 1999; 58(4):214-9. PMC: 1752863. DOI: 10.1136/ard.58.4.214. View

Silvagni E, Callegher S, Mauric E, Chiricolo S, Schreiber N, Tullio A . Musculoskeletal ultrasound for treating rheumatoid arthritis to target-a systematic literature review. Rheumatology (Oxford). 2022; 61(12):4590-4602. PMC: 9707059. DOI: 10.1093/rheumatology/keac261. View

Cardinal E, Bureau N, AUBIN B, Chhem R . Role of ultrasound in musculoskeletal infections. Radiol Clin North Am. 2001; 39(2):191-201. DOI: 10.1016/s0033-8389(05)70272-4. View

Chau C, Griffith J . Musculoskeletal infections: ultrasound appearances. Clin Radiol. 2005; 60(2):149-59. DOI: 10.1016/j.crad.2004.02.005. View

Legome E, Pancu D . Future applications for emergency ultrasound. Emerg Med Clin North Am. 2004; 22(3):817-27. DOI: 10.1016/j.emc.2004.04.011. View

Tayal V, Hasan N, Norton H, Tomaszewski C . The effect of soft-tissue ultrasound on the management of cellulitis in the emergency department. Acad Emerg Med. 2006; 13(4):384-8. DOI: 10.1197/j.aem.2005.11.074. View

Loffler C, Sattler H, Peters L, Tuleweit A, Loffler U, Wadsack D . In arthritis the Doppler based degree of hypervascularisation shows a positive correlation with synovial leukocyte count and distinguishes joints with leukocytes greater and less than 5/nL. Joint Bone Spine. 2016; 83(5):517-23. DOI: 10.1016/j.jbspin.2015.09.007. View

Lo C, Lai K . Deep learning-based assessment of knee septic arthritis using transformer features in sonographic modalities. Comput Methods Programs Biomed. 2023; 237:107575. DOI: 10.1016/j.cmpb.2023.107575. View

Lo C, Hung P . Computer-aided diagnosis of ischemic stroke using multi-dimensional image features in carotid color Doppler. Comput Biol Med. 2022; 147:105779. DOI: 10.1016/j.compbiomed.2022.105779. View

10.

Holzmeister A, Frazzetta J, Yuan F, Cherones A, Summers H, Cohen J . Evaluation for septic arthritis of the native adult knee is aided by multivariable assessment. Am J Emerg Med. 2020; 46:614-618. DOI: 10.1016/j.ajem.2020.11.048. View

11.

Lo C, Hung P . Predictive stroke risk model with vision transformer-based Doppler features. Med Phys. 2023; 51(1):126-138. DOI: 10.1002/mp.16861. View

12.

Backhaus M, Burmester G, Gerber T, Grassi W, Machold K, Swen W . Guidelines for musculoskeletal ultrasound in rheumatology. Ann Rheum Dis. 2001; 60(7):641-9. PMC: 1753749. DOI: 10.1136/ard.60.7.641. View

13.

Lo C, Yang Y, Lin J, Lin T, Chen W, Yang S . Modeling the survival of colorectal cancer patients based on colonoscopic features in a feature ensemble vision transformer. Comput Med Imaging Graph. 2023; 107:102242. DOI: 10.1016/j.compmedimag.2023.102242. View

14.

Zhang J, He X, Liu Y, Cai Q, Chen H, Qing L . Multi-modal cross-attention network for Alzheimer's disease diagnosis with multi-modality data. Comput Biol Med. 2023; 162:107050. DOI: 10.1016/j.compbiomed.2023.107050. View

15.

Reddy A, Rao K, Soora N, Shailaja K, Kumar N, Sridharan A . Multi-modal fusion of deep transfer learning based COVID-19 diagnosis and classification using chest x-ray images. Multimed Tools Appl. 2022; 82(8):12653-12677. PMC: 9483263. DOI: 10.1007/s11042-022-13739-6. View