Study of artificial intelligence to diagnose rheumatoid arthritis

Study of convolutional neural network for classification of two-dimensional array images generated by clinical information in rheumatoid arthritis - Study of convolutional neural network to diagnose rheumatoid arthritis

Status

Active, not recruiting

Phases

Unknown

Study type

Observational

Source

JPRN

Registry ID

JPRN-UMIN000039371

Enrollment

500

Registered

2020-02-10

Start date

2019-10-01

Completion date

Unknown

Last updated

2026-06-29

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Rheumatoid arthritis

Interventions

None listed

Eligibility

Sex/Gender

All

Inclusion criteria

Inclusion criteria: In this study, undiagnosed patients with arthralgia in their fingers or wrists who first visited hospital during 1 January to 31 May 2019 were enrolled.

Exclusion criteria

Exclusion criteria: The study outline was published on the hospital's homepag and an opt-out management strategy was used to collect the patient information. Patients who did not hope to participate in were excluded. Patients judged by doctors to be inappropriate for the study were excluded.

Design outcomes

Primary

Measure	Time frame
We semi-quantitatively converted several clinical information obtained from first visit in hospital to four color square images and arranged them as one image. Some modifications were added to clinical information in each patient to increase tuning data. These were used to fine-tune one of the pretrained CNNs, AlexNet. The fine-tuned AlexNet classified testing images that were independent of the tuning data.	—

Measure

Time frame

We semi-quantitatively converted several clinical information obtained from first visit in hospital to four color square images and arranged them as one image. Some modifications were added to clinical information in each patient to increase tuning data. These were used to fine-tune one of the pretrained CNNs, AlexNet. The fine-tuned AlexNet classified testing images that were independent of the tuning data.

—

Secondary

Measure	Time frame
We compared the classification ability of the CNN and diagnosis of rheumatologists.	—

Countries

Japan

Contacts

Public ContactJun Fukae

Hokkaido Medical Center for Rheumatic Diseases Rheumatology

jun.fukae@ryumachi-jp.com81-11-611-1373

Outcome results