Convolutional neural network for classification of two-dimensional array images generated from clinical information may support diagnosis of rheumatoid arthritis.

This research aimed to study the application of deep learning to the diagnosis of rheumatoid arthritis (RA). Definite criteria or direct markers for diagnosing RA are lacking. Rheumatologists diagnose RA according to an integrated assessment based on scientific evidence and clinical experience. Our novel idea was to convert various clinical information from patients into simple two-dimensional images and then use them to fine-tune a convolutional neural network (CNN) to classify RA or nonRA. We semi-quantitatively converted each type of clinical information to four coloured square images and arranged them as one image for each patient. One rheumatologist modified each patient's clinical information to increase learning data. In total, 1037 images (252 RA, 785 nonRA) were used to fine-tune a pretrained CNN with transfer learning. For clinical data (10 RA, 40 nonRA), which were independent of the learning data and were used as testing data, we compared the classification ability of the fine-tuned CNN with that of three expert rheumatologists. Our simple system could potentially support RA diagnosis and therefore might be useful for screening RA in both specialised hospitals and general clinics. This study paves the way to enabling deep learning in the diagnosis of RA.

Active synovitis in the presence of osteitis predicts residual synovitis in patients with rheumatoid arthritis with a clinical response to treatment.

AIM: To clarify the relationship between active synovitis/osteitis and subsequent residual synovitis (R-synovitis) in patients with rheumatoid arthritis (RA). METHODS: Three hundred and twenty finger joints of 16 patients with active RA at baseline (Disease Activity Score with 28 joints - erythrocyte sedimentation rate > 3.2) who subsequently achieved clinical low disease activity or remission afterwards were analyzed. Synovial vascularity (SV) was assessed according to a semi-quantitative ultrasound score (grades 0-3). Active synovitis was defined by SV positivity at baseline. R-synovitis was defined by the presence of grade > 2 SV at the 24th week. Osteitis was detected by magnetic resonance imaging (MRI) at baseline as trabecular bone lesions with water content and indistinct margins. RESULTS: Ultrasonography detected active synovitis in 116 joints at baseline. Forty-seven joints had R-synovitis at the 24th week. MRI detected osteitis in 12 joints at baseline. The presence of active synovitis with osteitis at baseline was significantly correlated with R-synovitis at the 24th week. CONCLUSIONS: Active synovitis in the presence of osteitis predicted R-synovitis regardless of whether there was a clinical improvement in RA.

Structural deterioration of finger joints with ultrasonographic synovitis in rheumatoid arthritis patients with clinical low disease activity.

OBJECTIVE: In this study we investigated the relationship between synovial vascularity (SV) and structural alteration of finger joints in patients with RA and long-term sustained clinical low disease activity (CLDA). METHODS: RA patients with CLDA of >2 years (minimum 1 year of CLDA for study entry plus 1 year of observation) were analysed. Quantitative SV values were sequentially measured in each finger joint using power Doppler ultrasonography (0, 8, 20 and 52 weeks). Radiological progression of local finger joints was evaluated according to the Genant-modified Sharp score (0-52 weeks). RESULTS: Of the 25 patients enrolled, 15 patients were finally analysed after excluding 10 patients who failed to maintain CLDA during the observational period. Changes in radiological progression of MCP and PIP joints with positive SV were significantly greater than those in joints with negative SV. Joint space narrowing (JSN) was strongly related to structural alteration of finger joints. In joints with positive SV, changes in structural alteration did not relate to total SV values, which reflect total exposure to inflammation in an observational period. CONCLUSION: Even in patients with a long period of CLDA, finger joints with positive SV showed structural alteration, especially in the progression of JSN. TRIAL REGISTRATION: University Hospital Medical Information Network Clinical Trials Registry, http://www.umin.ac.jp/ctr/, UMIN000007305.

Positive synovial vascularity in patients with low disease activity indicates smouldering inflammation leading to joint damage in rheumatoid arthritis: time-integrated joint inflammation estimated by synovial vascularity in each finger joint.

OBJECTIVE: To investigate the relationship between synovial vascularity and joint damage progression in each finger joint of patients with RA under low disease activity during treatment with biologic agents. METHODS: We studied 310 MCP and 310 PIP joints of 31 patients with active RA who were administered adalimumab (ADA) or tocilizumab (TCZ). Patients were examined with clinical and laboratory assessments. Power Doppler sonography was performed at baseline and at weeks 8, 20 and 40. Synovial vascularity was evaluated according to quantitative measurement. Hand and foot radiography was performed at baseline and at week 50. RESULTS: Composite scores of the DAS with 28 joints and the Simplified Disease Activity Index (SDAI) were significantly decreased from baseline to week 8, being sustained at a low level by biologic agents during the observational period. MCP and PIP joints with positive synovial vascularity after week 8 showed more subsequent joint damage progression than joints without synovial vascularity throughout the follow-up. The changes in radiographic progression in these joints were independent of the sum of synovial vascularity from baseline to week 40 or the occasional occurrence of positive synovial vascularity. CONCLUSION: Smouldering inflammation reflected by positive synovial vascularity under low disease activity was linked to joint damage. The damage progressed irrespective of the severity of positive synovial vascularity. Even with a favourable overall therapeutic response, monitoring of synovial vascularity has the potential to provide useful joint information to tailor treatment strategies. Trial registration. University Hospital Medical Information Network Clinical Trials Registry; http://www.umin.ac.jp/ctr/; UMIN000004476.

Radiographic prognosis of finger joint damage predicted by early alteration in synovial vascularity in patients with rheumatoid arthritis: Potential utility of power doppler sonography in clinical practice.

OBJECTIVE: To investigate the relationship between synovial vascularity and progression of structural bone damage in each finger joint in patients with rheumatoid arthritis (RA) and to demonstrate synovial vascularity as a potential therapeutic marker. METHODS: We studied 250 metacarpophalangeal (MCP) and 250 proximal interphalangeal (PIP) joints of 25 patients with active RA who were administered adalimumab or tocilizumab. Patients were examined with clinical and laboratory assessments. Power Doppler sonography was performed at baseline and at the fourth and eighth weeks. Synovial vascularity was evaluated according to quantitative measurement. Hand and foot radiography was performed at baseline and the twentieth week. RESULTS: Clinical indices such as the 28-joint Disease Activity Score, the Clinical Disease Activity Index, and the Simplified Disease Activity Index were significantly decreased by biologic agents. The MCP and PIP joints with no response in synovial vascularity between baseline and the eighth week (vascularity improvement of ≤70% at the eighth week) showed a higher risk of radiographic progression compared with responsive joints (vascularity improvement of >70% at the eighth week; relative risk 2.33-9). Radiographic progression at the twentieth week was significantly lower in responsive joints than in nonresponsive joints. CONCLUSION: The improvement of synovial vascularity following treatment with biologic agents led to suppression of radiographic progression of RA in each finger joint. The alteration in synovial vascularity numerically reflected therapeutic efficacy. Using vascularity as a marker to determine the most suitable therapeutic approach would be beneficial for patients with active RA.