Use of double-echo ultrashort echo time magnetic resonance imaging to assess proximal femoral cortical bone changes in axial spondyloarthritis

Abstract

Objective

We quantitatively evaluated proximal femoral cortical bone changes associated with generalized bone loss in axial spondyloarthritis (axSpA) patients using double-echo ultrashort echo time (UTE) MRI. To achieve non-radiation, clinically available visualization of cortical microstructural deterioration in the proximal femur of axSpA and to determine the factors influencing it.

Materials and methods

Patients with axSpA (n = 83) and age- and sex-matched healthy controls (n = 61) were recruited and underwent double-echo UTE MR scan of the nondominant proximal femur. Porosity index (PI) and cortical bone thickness (CbTh) were measured by two radiologists and their average measurements were used for subsequent analyses. Additionally, demographic characteristics of all subjects and disease-specific characteristics of axSpA patients were recorded. Proximal femoral cortical bone PI and CbTh values were compared between axSpA patients and healthy controls using independent samples t-test. Correlation analyses (Pearson or Spearman or Point-biserial correlation coefficients) were conducted to investigate factors potentially associated with UTE measurements in axSpA patients, and Bonferroni correction was applied at the α = 0.002 level for more stringent correction. Multiple linear regression analyses were performed to further identify influencing factors of UTE measurements with multiple correlated variables.

Results

A total of 72 axSpA patients and 52 healthy control subjects were finally included. Patients and control subjects were comparable in sex, age, and body mass index (BMI). Proximal femoral cortical PI was higher (p < 0.001) and CbTh was lower (p < 0.001) in axSpA patients than in healthy controls. Sex (p < 0.001), BMI (p = 0.003), disease duration (p = 0.044), onset age (p = 0.01), alkaline phosphatase (ALP) (p < 0.001), radiographic axSpA (r-axSpA) (p = 0.02), Sacroiliac Joint Structural Score (SSS)-backfill (p = 0.03), SSS-ankylosis (p = 0.01), and nonsteroidal anti-inflammatory drugs (NSAIDs) use (p = 0.03) were potentially correlated to proximal femoral cortical PI, and among them, sex (p < 0.001) and BMI (p = 0.01) were independently associated demographic characteristics, and ALP (p = 0.02), SSS-backfill (p = 0.044) and SSS-ankylosis (p = 0.01) were independently associated disease-specific characteristics, and when all types of variables were considered, sex (p < 0.001), BMI (p = 0.016), and SSS-ankylosis (p = 0.042) were independently associated with PI. BMI (p = 0.043) and NSAIDs use (p = 0.041) were potentially negatively associated with CbTh.

Conclusion

Double-echo UTE measurements revealed deteriorated proximal femoral cortical bone microstructure in axSpA patients than controls, potentially associated with sex, BMI, disease duration, onset age, ALP, r-SpA, SSS-backfill, SSS-ankylosis, and NSAIDs use, where sex, BMI, ALP, SSS-backfill and SSS-ankylosis were the independently relevant factors. It offers a novel tool for axSpA's cortical bone assessment.