Background
Septic arthritis (SA), although rare, is a critical joint-threatening emergency. The shoulder, being the third most common site after knee and hip joints, is predominantly affected by hematogenous spread of infection. 2-deoxy-2-[18F]fluoro-D-glucose (F-18-FDG) positron emission tomography (PET)/computed tomography (CT) has shown promise in identifying infectious foci and modifying treatment plans. This study aims to differentiate the metabolic patterns of septic shoulder arthritis (SSA) from various stages of shoulder osteoarthritis (OA) using F-18-FDG PET/CT and assess its diagnostic performance.
Methods
We retrospectively included subsequent patients diagnosed with SSA between November 2017 and October 2023, who had undergone whole-body F-18-FDG PET/CT scans within 2 weeks before and after diagnosis. The control group included noninfected contralateral shoulder joints and patients who underwent F-18-FDG PET/CT for malignant melanoma staging without evidence of acute infections. Visual and quantitative analyses of F-18-FDG uptake, measured as maximum standardized uptake value (SUVmax), were conducted. OA severity was categorized using the Kellgren and Lawrence system. To establish a cut-off for SSA, a ratio of joint uptake to liver uptake of 18F-FDG was calculated using the SUVmax of the shoulder joint and liver. Intraobserver and interobserver reliability were assessed through repeated measurements.
Results
Of 150 shoulders, 149 were included in the study: 13 into the experimental group with confirmed SA and 136 into the control group. One shoulder was excluded from the control group due to metastasis. Thirteen shoulders had confirmed SA. In the control group, F-18-FDG uptake measured by SUVmax increased significantly with OA severity (P = .001). SSA exhibited markedly higher F-18-FDG uptake compared to controls (P < .001). Visual intensity assessments corroborated these findings (P < .001). For a cut-off value of 1 for the joint-liver-ratio, sensitivity, specificity, positive predictive value, and negative predictive value for distinguishing between SSA and OA were 92.3%, 80.9%, 31.6%, and 99.1%, respectively. Interobserver and intraobserver reliability were moderate to high. The intraclass correlation coefficients for standardized uptake value measurements in the total shoulder were 0.994 and 0.996, respectively, while Cohen’s kappa coefficients for visual analysis were 0.570 and 0.891.
Conclusion
F-18-FDG PET/CT effectively differentiates SSA from varying stages of OA. The SUVmax in liver can be used as a cut-off value with high sensitivity and specificity. A negative F-18-FDG PET/CT excludes SSA.