Marzieh Ebrahimi - Department of Medical Physics, School of Medicine, Iran University of Medical Sciences, Tehran, Iran
Zeinab Paymani - Department of Nuclear Medicine, Children’s Medical Center, Tehran University of Medical Sciences, Tehran, Iran
Mostafa Nazari - Department of Nuclear Medicine, Children’s Medical Center, Tehran University of Medical Sciences, Tehran, Iran
Hossien Kian Ara - Department of Mathematics and Computer Science, Shahed University, Tehran, Iran
Nafiseh Alemohammad - Department of Mathematics and Computer Science, Shahed University, Tehran, Iran
Fatemeh Tahghigi Sharabian - Department of Pediatric Rheumatology, Children’s Medical Center, Tehran University of Medical Sciences, Tehran, Iran
Molood Gooniband Shooshtari - Department of Medical Physics, School of Medicine, Iran University of Medical Sciences, Tehran, Iran

Introduction: Diagnosing Juvenile Idiopathic Arthritis (JIA) presents challenges due to symptom variations, clinical-radiologic delays, and the absence of definitive diagnostic tools. This study aimed to evaluate the diagnostic capability of radiomic features derived from blood pool phase images obtained through bone scintigraphy in JIA.Methods: A cohort of ۱۹۰ patients was included, utilizing the area between knee growth plates as the region of interest (ROI) for extracting image features. After preprocessing, quantitative features were extracted from original and filtered images. A recursive feature elimination (RFE) algorithm identified significant features, subsequently employed in training a random forest classifier.Results: In the validation phase, our radiomic model, comprising ۱۴ features (۴ original and ۱۰ filtered image features), achieved an area under the receiver operating characteristic curve (AUC) of ۰.۸۹ (۹۵% CI: ۰.۸۸–۰.۹۲). This robust performance confirmed the efficacy of radiomics in identifying active knee arthritis using technetium–۹۹m-methyl diphosphonate blood pool images in JIA patients.Conclusion: This study highlights the diagnostic accuracy of radiomics in discerning arthritic joints, suggesting its potential as an alternative to conventional quantification techniques. The robustness of radiomics in diagnosing arthritic joints signifies a promising avenue for future research in JIA diagnosis and treatment.

Juvenile Idiopathic Arthritis, Nuclear medicine, Machine learning, Bone scintigraphy