An objective diagnosis of gout and calcium pyrophosphate deposition disease with machine learning of Raman spectra acquired in a point-of-care setting.
Tom Niessink, Tim L Jansen, Frank A W Coumans, Tim J M Welting, Matthijs Janssen, Cees Otto
{"title":"An objective diagnosis of gout and calcium pyrophosphate deposition disease with machine learning of Raman spectra acquired in a point-of-care setting.","authors":"Tom Niessink, Tim L Jansen, Frank A W Coumans, Tim J M Welting, Matthijs Janssen, Cees Otto","doi":"10.1093/rheumatology/keae472","DOIUrl":null,"url":null,"abstract":"<p><strong>Objective: </strong>Raman spectroscopy is proposed as a next-generation method for the identification of monosodium urate (MSU) and calcium pyrophosphate (CPP) crystals in synovial fluid. As the interpretation of Raman spectra requires specific expertise, the method is not directly applicable for clinicians. We developed an approach to demonstrate that the identification process can be automated with the use of machine learning techniques. The developed system is tested in a point-of-care-setting at our outpatient rheumatology department.</p><p><strong>Methods: </strong>We collected synovial fluid samples from 446 patients with various rheumatic diseases from three centres. We analysed all samples with our Raman spectroscope and used 246 samples for training and 200 samples for validation. Trained observers classified every Raman spectrum as MSU, CPP or other. We designed two one-against-all classifiers, one for MSU and one for CPP. These classifiers consisted of a principal component analysis model followed by a support vector machine.</p><p><strong>Results: </strong>The accuracy for classification of CPP using the 2023 ACR/EULAR CPPD classification criteria was 96.0% (95% CI: 92.3, 98.3), while the accuracy for classification of MSU using the 2015 ACR/EULAR gout classification criteria was 92.5% (95% CI: 87.9, 95.7). Overall, the accuracy for classification of pathological crystals was 88.0% (95% CI: 82.7, 92.2). The model was able to discriminate between pathological crystals, artifacts and other particles such as microplastics.</p><p><strong>Conclusion: </strong>We here demonstrate that potentially complex Raman spectra from clinical patient samples can be successfully classified by a machine learning approach, resulting in an objective diagnosis independent of the opinion of the medical examiner.</p>","PeriodicalId":21255,"journal":{"name":"Rheumatology","volume":" ","pages":"1791-1798"},"PeriodicalIF":4.7000,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11962871/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Rheumatology","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1093/rheumatology/keae472","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"RHEUMATOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Objective: Raman spectroscopy is proposed as a next-generation method for the identification of monosodium urate (MSU) and calcium pyrophosphate (CPP) crystals in synovial fluid. As the interpretation of Raman spectra requires specific expertise, the method is not directly applicable for clinicians. We developed an approach to demonstrate that the identification process can be automated with the use of machine learning techniques. The developed system is tested in a point-of-care-setting at our outpatient rheumatology department.
Methods: We collected synovial fluid samples from 446 patients with various rheumatic diseases from three centres. We analysed all samples with our Raman spectroscope and used 246 samples for training and 200 samples for validation. Trained observers classified every Raman spectrum as MSU, CPP or other. We designed two one-against-all classifiers, one for MSU and one for CPP. These classifiers consisted of a principal component analysis model followed by a support vector machine.
Results: The accuracy for classification of CPP using the 2023 ACR/EULAR CPPD classification criteria was 96.0% (95% CI: 92.3, 98.3), while the accuracy for classification of MSU using the 2015 ACR/EULAR gout classification criteria was 92.5% (95% CI: 87.9, 95.7). Overall, the accuracy for classification of pathological crystals was 88.0% (95% CI: 82.7, 92.2). The model was able to discriminate between pathological crystals, artifacts and other particles such as microplastics.
Conclusion: We here demonstrate that potentially complex Raman spectra from clinical patient samples can be successfully classified by a machine learning approach, resulting in an objective diagnosis independent of the opinion of the medical examiner.
期刊介绍:
Rheumatology strives to support research and discovery by publishing the highest quality original scientific papers with a focus on basic, clinical and translational research. The journal’s subject areas cover a wide range of paediatric and adult rheumatological conditions from an international perspective. It is an official journal of the British Society for Rheumatology, published by Oxford University Press.
Rheumatology publishes original articles, reviews, editorials, guidelines, concise reports, meta-analyses, original case reports, clinical vignettes, letters and matters arising from published material. The journal takes pride in serving the global rheumatology community, with a focus on high societal impact in the form of podcasts, videos and extended social media presence, and utilizing metrics such as Altmetric. Keep up to date by following the journal on Twitter @RheumJnl.