The quantifying of radiographic axial psoriatic arthritis through currently available scoring systems has moderate sensitivity but high specificity for identifying true change, according to results.
Researchers obtained radiographs of 105 patients with axial PsA from the University of Toronto PsA clinic. Patients had serial radiographs, separated by at least 2 years, and obtained since 2003. Researchers extracted two sets of axial radiographs. Three radiologists blinded to name and order of examination scored the radiographs using the scoring model of axial radiograph Toronto (SMART) application. Using the application, they automatically calculated Bath ankylosing spondylitis radiology index for the spine (BASRI-s), the modified stroke ankylosing spondylitis spine score (mSASSS), the radiographic ankylosing spondylitis spine score (RASSS) and PsA spondylitis radiology index (PASRI) scores. The existence of true radiographic progression was determined by an independent expert who viewed the radiographs with knowledge of the chronology. Researchers identified true change by determining the sensitivity, specific and odds ratios for every 1-unit increase in scores.
They found a significant change in mean BASRI-s, mSASSS and PASRI scores between the two radiographic evaluations. For the RASSS scores, there was a trend at 5% significance. The external expert identified radiographic progression in 24% of patients.
In terms of patients with worsening scores across time points, there were changes in PASRI score in the largest percentage of patients; while changes in RASSS score were seen in the lowest percentage of patients.
All scoring instruments showed moderate sensitivity, but high specificity in detecting change related to increases of at least one. The following sensitivity values were noted for the ability of the instrument to detect true change based on increases in score: 0.48 for BASRI-s; 0.52 for mSASSS; 0.44 for RASSS and 0.52 for PASRI. Specificity values for these instruments were as follows: 0.78 for BASRI-s; 0.84 for RASSS and 0.74 for PASRI. In logistic regression analyses, the researchers found that a 1-point increase in the scores correlated with the following odds ratios for detecting true progression: 3.0 for BASRI-s; 5.27 for mSASSS; 3.70 for RASSS and 3.06 for PASRI.
For increases of at least two in scores, a lower sensitivity and a higher specificity was seen. The most notable change was seen in BASRI-s, which showed a significant reduction in sensitivity and an increase in specificity.
“This study demonstrates that available scoring systems for quantifying radiographic axial PsA have moderate sensitivity and high specificity in detecting true change,” the researchers wrote. “All measures performed equally well in detecting change.”-by Jennifer Byrne
Disclosures: The authors report no relevant disclosures.