Dataset Information

Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis.

ABSTRACT:

Objective

The purpose of this research was to develop a deep-learning model to assess radiographic finger joint destruction in RA.

Methods

The model comprises two steps: a joint-detection step and a joint-evaluation step. Among 216 radiographs of 108 patients with RA, 186 radiographs were assigned to the training/validation dataset and 30 to the test dataset. In the training/validation dataset, images of PIP joints, the IP joint of the thumb or MCP joints were manually clipped and scored for joint space narrowing (JSN) and bone erosion by clinicians, and then these images were augmented. As a result, 11 160 images were used to train and validate a deep convolutional neural network for joint evaluation. Three thousand seven hundred and twenty selected images were used to train machine learning for joint detection. These steps were combined as the assessment model for radiographic finger joint destruction. Performance of the model was examined using the test dataset, which was not included in the training/validation process, by comparing the scores assigned by the model and clinicians.

Results

The model detected PIP joints, the IP joint of the thumb and MCP joints with a sensitivity of 95.3% and assigned scores for JSN and erosion. Accuracy (percentage of exact agreement) reached 49.3-65.4% for JSN and 70.6-74.1% for erosion. The correlation coefficient between scores by the model and clinicians per image was 0.72-0.88 for JSN and 0.54-0.75 for erosion.

Conclusion

Image processing with the trained convolutional neural network model is promising to assess radiographs in RA.

SUBMITTER: Hirano T

PROVIDER: S-EPMC6921374 | biostudies-literature | 2019

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis.

Hirano Toru T Nishide Masayuki M Nonaka Naoki N Seita Jun J Ebina Kosuke K Sakurada Kazuhiro K Kumanogoh Atsushi A

Rheumatology advances in practice 20191122 2

<h4>Objective</h4>The purpose of this research was to develop a deep-learning model to assess radiographic finger joint destruction in RA.<h4>Methods</h4>The model comprises two steps: a joint-detection step and a joint-evaluation step. Among 216 radiographs of 108 patients with RA, 186 radiographs were assigned to the training/validation dataset and 30 to the test dataset. In the training/validation dataset, images of PIP joints, the IP joint of the thumb or MCP joints were manually clipped and ...[more]

PMID: 31872173

Similar Datasets

Project description:ObjectiveRheumatoid arthritis (RA) is a chronic disease leading to joint destruction. Although many studies have addressed factors potentially correlated with the speed of joint destruction, less attention has been paid to the distribution of joint destruction in patients with RA. In this study, destruction of the hand bones in patients with RA was classified into 2 anatomic subgroups, the fingers and the non-fingers, with the aim of analyzing which factors are associated with destruction of the finger joints.MethodsA total of 1,215 Japanese patients with RA were recruited from 2 different populations. The degree of joint destruction was assessed using the total modified Sharp/van der Heijde score (SHS) of radiographic joint damage. The SHS score of joint damage in the finger joints was used as the dependent variable, and the SHS score in the non-finger joints was used as a covariate. Age, sex, disease duration, smoking, C-reactive protein level, treatment for RA, and positivity for and levels of anti-citrullinated protein antibodies and rheumatoid factor (RF) were evaluated as candidate correlates. Overall effect sizes were assessed in a meta-analysis. In addition, associations observed in the Japanese patients were compared to those in a cohort of 157 Dutch RA patients in the BeSt study (a randomized, controlled trial involving 4 different strictly specified treatment strategies for early RA).ResultsNot surprisingly, disease duration in Japanese patients with RA was associated with the finger SHS score (P ≤ 0.00037). Both positivity for and levels of RF showed significant associations with the finger SHS score after adjustment for covariates (P = 0.0022 and P = 8.1 × 10(-7) , respectively). These associations were also true in relation to the time-averaged finger SHS score. An association between RF positivity and the finger SHS score was also observed in Dutch patients with RA in the BeSt study (P = 0.049).ConclusionPositivity for and levels of RF are associated with finger joint destruction independent of non-finger joint destruction and other covariates. Our findings suggest that there are different mechanisms of joint destruction operating in the finger joints of patients with RA.

Dataset Information

Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis.

Objective

Methods

Results

Conclusion

Publications

Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets