Rheumatoid Arthritis Disease Activity Assessment Model Data from a Two-Center Study

A prospective diagnostic study involving 203 patients with rheumatoid arthritis, conducted from December 2023 to September 2025. The dataset was used to develop and validate three machine learning models for assessing disease activity, integrating musculoskeletal ultrasound and radiomics features. The study was authored by Jinjin Peng and shared under a CC-BY-4.0 license.

Use Cases

• Training diagnostic models for rheumatoid arthritis based on ultrasound and radiomics features described in the study.
• Comparing the performance of combined imaging models against single-modality approaches for disease activity assessment.
• Validating machine learning models for clinical use using an independent external test cohort from a different hospital.

Strengths

• Includes data from 203 patients prospectively enrolled in a clinical study.
• Model performance is validated with an independent external test cohort of 61 patients from a separate institution.
• The combined model achieved a high area under the ROC curve (AUC) of 0.998 in the training cohort and 0.929 in the test cohort.

Limitations

• The dataset is very small at 15.8 KB, suggesting limited raw data or a summary document.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect geographic bias inherent to the two specific hospitals in China that provided the patient cohorts.

Provenance

Source

figshare, authored by Jinjin Peng.

Collection Method

Prospective diagnostic study with patient data collected from two hospitals.

Time Range

December 2023 to September 2025

Freshness

Last updated 2026-04-27 05:16:31; freshness should be verified.

Geography

Patients were enrolled from hospitals in Henan and Xinjiang, China.