POS0849 DEVELOPMENT AND VALIDATION OF A MACHINE LEARNING FOR MORTALITY IN THAI SYSTEMIC SCLEROSIS

Autor: C. Foocharoen, B. Thinkhamrop, W. Thinkhamrop, N. Chaichaya, A. Mahakkanukrauh, S. Suwannaroj
Rok vydání: 2022
Předmět:
Zdroj: Annals of the Rheumatic Diseases. 81:717.1-717
ISSN: 1468-2060
0003-4967
DOI: 10.1136/annrheumdis-2022-eular.485
Popis: BackgroundClinical predictors of mortality in systemic sclerosis (SSc) are diversely reported due to different healthcare conditions and populations. A universal and simplified predictive model for SSc mortality is needed so that practitioners can be used for managing their patients appropriately.ObjectivesWe aimed to develop and validate a simple predictive model for predicting mortality among patients with SSc.MethodsPrognostic research with a historical cohort study design was conducted between January 1, 2013, and December 31, 2019, in adult SSc patients and attending the Scleroderma Clinic at a university hospital in Thailand. The data were extracted from the Scleroderma Registry Database. A deep learning algorithm with Adam optimizer and different machine learning algorithms (including Decision tree, AdaBoost, Random Forest, Gradient Boosting, and XGBoost) was used to classify SSc mortality. In addition, the model’s performance was evaluated using the area under the receiver operating characteristic curve (auROC) and its 95% confidence interval (CI) and values in the confusion matrix.ResultsThe analysis and predictive model development included 658 SSc patients, 416 (63.2%) females, 452 (69.1%) had dcSSc, and 218 died. The final model included the modified Rodnan skin score (mRSS) and the WHO functional class (WHO-FC) ≥II (model 1). The final model provided the highest predictive performance, followed by model 2 (mRSS and WHO-FC ≥III). After internal validation, the accuracy and auROC were good, and the specificity was high in models 1 and 2 (81.1%, 0.84, and 95.5% in model 1 vs. 82.7%, 0.82, and 87.1% in model 2).Table 1.Generalizability of selected model(s) presented as accuracy, area under ROC, positive predictive value, positive likelihood ratio, specificity, and sensitivitySelected ModelAccuracyAUC (95%)PPV (95%)+LR (95%CI)Specificity (95%)Sensitivity (95%)Model 1 mRSS and WHO FC ≥ II81.183.6 (77.5 – 89.6)84.6 (69.5 - 94.1)11.3 (5.0-25.7)95.5 (90.4 - 98.3)51.6 (38.7 - 64.2)Model 2 mRSS and WHO FC ≥ III82.782.4 (75.8 - 88.9)73.4 (60.9 - 83.7)5.7 (3.6-9.187.1 (87.2 - 92.3)73.4 (60.9 - 83.7)95%CI 95% confidence interval, AUC Area Under the receiver operating characteristics (ROC), mRSS modified Rodnan skin scoreConclusionThis simplified machine learning model for predicting mortality among patients with SSc could guide early referrals to specialists and help rheumatologists with close monitoring and management planning. External validation across multi-SSc clinics should be considered for further study.References[1]Ferri C, Valentini G, Cozzi F, Sebastiani M, Michelassi C, La Montagna G, et al. Systemic sclerosis: demographic, clinical, and serologic features and survival in 1,012 Italian patients. Medicine (Baltimore). 2002;81(2):139–53.[2]Rubio-Rivas M, Royo C, Simeón CP, Corbella X, Fonollosa V. Mortality and survival in systemic sclerosis: systematic review and meta-analysis. Semin Arthritis Rheum. 2014;44(2):208–19.[3]Foocharoen C, Peansukwech U, Mahakkanukrauh A, Suwannaroj S, Pongkulkiat P, Khamphiw P, et al. Clinical characteristics and outcomes of 566 Thais with systemic sclerosis: A cohort study. Int J Rheum Dis. 2020;23(7):945–57.[4]Tyndall AJ, Bannert B, Vonk M, Airò P, Cozzi F, Carreira PE, et al. Causes and risk factors for death in systemic sclerosis: a study from the EULAR Scleroderma Trials and Research (EUSTAR) database. Ann Rheum Dis. 2010;69(10):1809–15.[5]Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373–9.[6]Elhai M, Meune C, Boubaya M, Avouac J, Hachulla E, Balbir-Gurman A, et al. Mapping and predicting mortality from systemic sclerosis. Ann Rheum Dis. 2017;76(11):1897–905.[7]Wangkaew S, Prasertwitayakij N, Phrommintikul A, Puntana S, Euathrongchit J. Causes of death, survival and risk factors of mortality in Thai patients with early systemic sclerosis: inception cohort study. Rheumatol Int. 2017;37(12):2087–94.AcknowledgementsThe authors thank (a) Thailand’s National Science, Research, and Innovation Fund for funding support, (b) the Scleroderma Research Group for research assistance, and (c) Mr. Bryan Roderick Hamman—under the aegis of the Publication Clinic Khon Kaen University, Thailand—for assistance with the English-language presentation.Disclosure of InterestsChingching Foocharoen Speakers bureau: By Boeringer Ingelheim, Bandit Thinkhamrop: None declared, Wilaiphorn Thinkhamrop: None declared, Nathaphop Chaichaya: None declared, Ajanee Mahakkanukrauh Speakers bureau: By Boeringer Ingelheim, Norvatis, Johnson & Johnson, Siraphop Suwannaroj Speakers bureau: By Boeringer Ingelheim, Johnson & Johnson, Norvatis
Databáze: OpenAIRE
Externí odkaz: