The primary objective is to investigate the clinical utility of a PET/CT-derived GMIS in quantifying the biological inflammatory burden of axSpA. Specifically, we aim to evaluate the correlation between this whole-body metabolic mapping and traditional serological monitoring indicators, validating the hypothesis that deep-tissue microenvironments maintain an active, subclinical immune activation state despite normal routine biochemical profiles. A secondary objective focuses on the ACW as a pathological node, elucidating its natural prevalence, independent metabolic characteristics, and spatial coupling with axial joint inflammation to determine its role as an early indicator of systemic entheseal vulnerability.

This retrospective cohort study continuously enrolled 38 adult axSpA patients who underwent standardized whole-body 18F-FDG PET/CT evaluation between January 2011 and December 2024. Patients with concomitant aortic arteritis were strictly excluded to eliminate radioactive spillover interference, ensuring the accuracy of the aortic arch blood pool as the systemic background parameter. The maximum standardized uptake value (SUVmax) of the aortic arch was used to calculate the TBR for independent regions across five core anatomical modules. The GMIS was mathematically constructed as the cumulative sum of TBRs for all active lesions (TBR ≥ 1.0). "Subclinical" lesions were defined as those exhibiting abnormal hypermetabolism (TBR ≥ 1.0) without spontaneous pain or palpable tenderness. Statistical analyses incorporated unsupervised hierarchical clustering (Ward's method) to map spatial associations without prior clinical grouping, and multidimensional bubble charts to map the phenotypic correlation between serological markers and localized metabolic intensity.

Based on systemic TBR data, the cohort was stratified into a PET-positive group (n = 18; having at least one lesion with TBR ≥ 1.0) and a PET-negative group (n = 20). Baseline demographic and disease duration characteristics were comparable. A "biochemical-imaging dissociation" was observed: ESR distribution differed significantly between the PET-positive and PET-negative groups (P = 0.002), whereas the core systemic marker, CRP, exhibited no statistical difference (P = 0.327). Linear regression demonstrated no significant correlation between GMIS and peripheral CRP. Spatial distribution analysis revealed the ACW had the highest regional prevalence, affecting 55.6% (10/18) of the positive cohort. Within these 10 patients, 35 independent hypermetabolic ACW lesions were identified, of which 34 (97.1%) were subclinical. Subgroup analysis demonstrated that patients with ACW involvement exhibited a higher total number of PET-detected active lesions (median 9.0 vs. 2.0, P = 0.023) and an elevated overall metabolic burden (median GMIS 17.2 vs. 3.2, P = 0.006). Unsupervised hierarchical clustering revealed early spatial coupling between the ACW and the central axial spine.

This study substantiates the "biochemical-imaging dissociation" in axSpA, mechanistically related to the anatomical "compartmentalization" of the entheseal microenvironment. Because avascular entheses impede the systemic release of macromolecular cytokines, traditional markers like CRP inherently underestimate deep osteitis. The PET/CT-derived GMIS circumvents this physiological barrier, providing an objective biological metric to quantify the hidden inflammatory burden independent of hepatic acute-phase responses. Furthermore, driven by continuous respiratory biomechanical stress, the ACW acts as a prevalent and subclinical pathological node. Incorporating a systemic molecular imaging framework that targets anatomical blind spots like the ACW is useful for the precise stratification of subclinical hypermetabolic states, providing a translational foundation for optimizing early targeted interventions before structural damage occurs.