AB1343 ON THE DEVELOPMENT OF NEW DISEASE ACTIVITY SCORES FOR REMOTE ASSESSMENT OF PATIENT WITH RHEUMATOID ARTHRITIS USING THERMOGRAPHY AND MACHINE LEARNING

Autor: I. Morales-Ivorra, J. Narváez, C. Gómez Vaquero, J. M. Nolla, C. Moragues Pastor, D. Grados Canovas, J. A. Narvaez, M. A. Marin-López
Rok vydání: 2022
Předmět:
Zdroj: Annals of the Rheumatic Diseases. 81:1778.1-1778
ISSN: 1468-2060
0003-4967
DOI: 10.1136/annrheumdis-2022-eular.1567
Popis: BackgroundDisease activity scores are used in the follow-up of patients with rheumatoid arthritis (RA). These scores include variables obtained through physical examination, such as the tender and swollen joint count. In telematic consultations it is not possible to determine these variables. Thermography is a safe and fast technique that measures heat through infrared imaging. Inflammation of the joints causes an increase in temperature and could therefore be detected by thermography. Machine learning methods are highly accurate in analyzing medical images, and could be used to analyze thermal images automatically. Thermography of hands, patient global health (PGH) and acute phase reactants could be combined to develop new activity scores that facilitate remote assessment of RA patients.ObjectivesTo develop new disease activity scores based on the machine learning analysis of thermal images of the hands, PGH and acute phase reactants.MethodsMulticenter observational study conducted in the rheumatology and radiology service of two hospitals. Patients with RA, psoriatic arthritis, undifferentiated arthritis and arthritis of hands secondary to other diseases that attended the follow-up visits were recruited. Companions of patients and healthcare professionals were also recruited as healthy subjects. In all cases, a thermographic image of the hands was taken using a Flir One Pro or a Thermal Expert TE-Q1 camera connected to a smartphone. Ultrasound (US) of both hands was performed. The degree of synovial hypertrophy (SH) and power doppler (PD) was assessed for each joint (score from 0 to 3). Machine learning was used to quantify joint inflammation from the thermal images using US (SH+PD) as ground truth. This score has been named ThermoJIS. RA patients whose thermal image was taken with the Thermal Expert TE-Q1 camera were used to evaluate the performance (test dataset). The other participants were used as training dataset. The PGH, C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) were also assessed in the test dataset. ThermoDAS (ThermoJIS + PGH), ThermoDAS-ESR (ThermoJIS + PGH + ESR) and ThermoDAS-CRP (ThermoJIS + PGH + CRP) activity scores were developed using a linear regression. The Spearman’s correlation coefficient of ThermoJIS, ThermoDAS, ThermoDAS-ESR and ThermoDAS-CRP vs. SH, PD, PGH, ESR and CRP were used to characterize the new developed disease activity scores. The study was approved by the Clinical Ethics and Research Committee of both centers.ResultsThe total number of recruited subjects were 616 (475 for the training and 141 for the testing dataset). The correlations obtained between the different activity scores (ThermoJIS, ThermoDAS, ThermoDAS-ESR and ThermoDAS-CRP) vs. SH, PD, PGH, ESR and CRP are shown in Table 1. All correlations are statistically significant.Table 1.Spearman’s correlations of the developed scores vs synovial hypertrophy (SH); vs power doppler (PD); vs patient global health (PGH); vs erythrocyte sedimentation rate (ESR) and vs C-reactive protein (CRP).SHPDPGHESRCRPThermoJIS0.420.430.180.160.12ThermoDAS0.500.530.870.160.19ThermoDAS-ESR0.540.530.790.490.33ThermoDAS-CRP0.600.600.770.490.54ConclusionThermoJIS shows moderate correlation with US but weak correlation with PGH and acute phase reactants, suggesting that ThermoJIS is non-redundant with symptoms and laboratory assessment. Adding PGH and acute phase reactants to ThermoJIS improves all correlations, including correlation with US. These thermographic scores do not require a physical examination, opening an opportunity to facilitate remote consultations in RA patients.References[1]Lynch CJ et al. New machine-learning technologies for computer-aided diagnosis. Nat Med. 2018 Sep;24(9):1304-1305.[2]Tan YK et al. Thermography in rheumatoid arthritis: a comparison with ultrasonography and clinical joint assessment. Clin Radiol. 2020 Dec;75(12): 963.Disclosure of InterestsNone declared
Databáze: OpenAIRE
Externí odkaz: