A NEW study has demonstrated that computer vision technology applied to smartphone photographs can accurately identify early inflammatory arthritis, offering a promising screening tool for clinical practice. Conducted on an Indian patient cohort, the research utilised convolutional neural networks (CNNs) to differentiate between healthy individuals and patients with joint inflammation.
The study involved 200 patients with early inflammatory arthritis and 200 healthy controls. Standardised hand photographs were taken and analysed using fine-tuned CNNs with an Inception-ResNet-v2 backbone. Both uncropped and joint-specific images, focusing on the wrist and proximal interphalangeal joints, were evaluated for accuracy, sensitivity, and specificity.
The CNN model achieved remarkable results when distinguishing patients from controls, with 99% accuracy, 98% sensitivity, and 99% specificity. In joint-specific analysis, wrist images proved most indicative of inflammation, yielding an accuracy of 75% and an area under the curve (AUC) of 0.86. The middle finger (MFPIP) and index finger joints (IFPIP) achieved similar results, with AUC values of 0.87 and 0.88, respectively.
“These findings demonstrate the potential of smartphone-based AI tools for screening inflammatory arthritis, particularly in resource-limited settings,” the authors noted. The study highlights the technology’s efficiency, with high levels of accuracy comparable to traditional rheumatologic assessment.
Further research will focus on validating the approach across diverse populations, other joint regions, and integrating the tool into clinical workflows. If widely adopted, this non-invasive and cost-effective screening method could significantly enhance early diagnosis and management of inflammatory arthritis.
Aleksandra Zurowska, EMJ
Reference
Phatak S et al. Incorporating computer vision on smart phone photographs into screening for inflammatory arthritis: results from an Indian patient cohort. Rheumatology. 2024;DOI: 10.1093/rheumatology/keae678.
