FI-RADS: an imaging-based scoring system for skeletal fluorosis-adjunctive clinical and biochemical analysis

Skeletal fluorosis is a chronic bone disease caused by excessive fluoride exposure, leading to osteosclerosis, osteoporosis, and ectopic ossifications. Its radiological features and their links to age, joint function, and metabolism remain unclear, necessitating further study. This study aimed to develop a Fluorosis Imaging Reporting and Data System (FI-RADS) based on X-ray radiographs of multiple skeletal sites for the diagnosis of skeletal fluorosis in endemic regions.

This cross-sectional study aimed to establish the FI-RADS grading system by analyzing radiographic data from 942 individuals with suspected skeletal fluorosis in the endemic area. Univariate and multivariate logistic regression analyses were conducted to evaluate associations between imaging features and demographic/metabolic factors [age, sex, total bile acids (TBA), bilirubin, etc.]. Employing spline-based non-linear regression analysis, this study examined the association of age and TBA levels with radiological alterations indicative of fluorosis.

Radiographic analysis (n=942) revealed characteristic bilateral symmetric changes, including radial crest abnormalities (48.15%), interosseous membrane ossification (6.47%), and pronator teres alterations (36.16%). Multivariate analysis identified age [odds ratio (OR) =1.060, 95% confidence interval (CI): 1.033-1.088] and TBA (OR =1.102, 95% CI: 1.004-1.210) as independent risk factors for advanced FI-RADS (>3). Nonlinear analysis revealed that multiple radiological ossification phenomena exhibit significant age- and TBA-dependent thresholds, with the loosening of the pronator teres insertion (R2=0.154, age) and ossification of the interosseous membrane of the ulna and radius (R2=0.072, bile acids) demonstrating the strongest effects.

We developed an X‑ray-based FI‑RADS imaging scoring system that standardizes and quantifies skeletal fluorosis. Age showed a pronounced nonlinear influence on periosteal and ligamentous ossification around the knee and elbow (FI‑RADS >3), whereas TBA exhibited only a weak threshold‑dependent effect within a narrow physiological range. Integrating FI‑RADS with age and bile acids may facilitate precise detection and sharper risk stratification; performance validation is pending.