AUTOMATIC SUBPIXEL MEASUREMENT OF RADIOGRAPHIC FINGER JOINT SPACE NARROWING IN RHEUMATOID ARTHRITIS PATIENTS UNDER TOCILIZUMAB TREATMENT

BackgroundThe conventional scoring methods of radiographic joint space narrowing (JSN) in rheumatoid arthritis (RA) such as the Genant-modified Sharp score (GSS) are widely accepted but include subjective and time-consuming nature1. Therefore, we have developed an in-house software equipped with partial image phase-only correlation (PIPOC)2 which can automatically quantify joint space width (JSW) change.ObjectivesThe purpose of this study was to investigate whether the in-house software can predict the inhibitory effect of tocilizumab on joint destruction in a short period of time.MethodsThe study included 39 RA patients (35 female) who were treated with tocilizumab (Table 1). Radiological progression of the metacarpophalangeal and the proximal interphalangeal joints was evaluated according to the GSS at 0, 6, and 12 months. Automatic measurement was performed by the in-house software (Figure 1). We then validated in-house software in terms of accuracy in detecting the JSN.Figure 1.The algorithm flow of in-house software equipped with PIPOCTable 1.Clinical characteristics of RA patientsvariablebaseline6 months12 monthsTotal number of patients39Sex, female/male35/4Rheumatoid factor status, positive/negative29/10Age, mean (SD) years61.5 (14.6)Duration of disease, mean (SD) months111.4 (85.0)Swollen joint count, mean (SD)6.0 (4.7)3.5 (3.5)2.9 (4.0)Tender joint count, mean (SD)6.4 (3.5)2.9 (2.2)1.9 (2.0)DAS28-ESR, mean (SD)4.9 (1.2)3.0 (1.1)2.6 (1.0)DAS28-CRP, mean (SD)4.4 (1.1)3.0 (0.9)2.6 (0.8)RA, rheumatoid arthritis; SD, standard deviation; DAS28, disease activity score with 28 joints; ESR, erythrocyte sedimentation rate; CRP, C-reactive proteinResultsTo ensure homogeneity of the subjects, we targeted the joints with GSS = 0 at baseline in the software analysis. The success rate of the in-house software for JSW measurement was 96.8% (449/464). Here, the GSS (+)/PIPOC (+) were defined as joints with JSN progression according to the GSS and the software analysis, respectively. Otherwise, joints were defined as the GSS (-)/PIPOC (-) namely non-progressive JSN. The 0–12-month GSS with the 0–6-month GSS (+) group was significantly more JSN progression than the 0–6-month GSS (-) group (Mann-Whitney U test, p < 0.001). Similarly, the 0–12-month PIPOC with the 0–6-month PIPOC (+) group was significantly more JSN progression than the 0–6-month PIPOC (-) group (p < 0.001). The 0–12-month JSW change of finger joints with the 0–12-month GSS (+) detected by the in-house software was significantly greater than the 0–12-month GSS (-) (p = 0.02).ConclusionOur in-house software equipped with PIPOC might be able to predict the subsequent joint destruction with 6 months observations.References[1]Genant HK, et al: Assessment of rheumatoid arthritis using a modified scoring method on digitized and original radiographs. Arthritis & Rheumatism: Official Journal of the American College of Rheumatology 1998, 41(9):1583-1590.[2]Ou Y, et al: Automatic Radiographic Quantification of Joint Space Narrowing Progression in Rheumatoid Arthritis Using POC. In: 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019). 2019: 1183-1187.Disclosure of InterestsTaichi Okino: None declared, Yafei Ou: None declared, Masayuki Ikebe: None declared, Akira Furusaki: None declared, Akira Sagawa: None declared, Masaru Kato: None declared, Tatsuya Atsumi: None declared, Tamotsu Kamishima Grant/research support from: Tamotsu Kamishima reports grants from Chugai Pharmaceutical.