Subtle radiographic progression in axial SpA cannot be reliably distinguished from error

Sacroiliitis observed in patients with axial spondyloarthritis more often regressed rather than progressed on radiography over nearly 5 years of follow-up of the Assessment of SpondyloArthritis international Society (ASAS) cohort, which lead author Dr. Alexandre Sepriano and his colleagues called “strange” and “sobering.”

The findings call into question the reliability of plain pelvic radiographs for detecting subtle change in sacroiliitis and should prompt the evaluation of alternative imaging modalities such as MRI and low-dose CT, according to Dr. Sepriano of Leiden (the Netherlands) University Medical Center and his associates.

©Rocky89/Thinkstock.com

Determining the presence of radiographic sacroiliitis is prognostically relevant and can pave the way for treatment with biologics, but ambiguity in making this decision and in tracking progression has been revealed in the large inter-and intrareader variability found in previous studies. Furthermore, previous studies tracking progression of nonradiographic axial spondyloarthritis (axSpA) to radiographic axSpA have addressed only disease progression and ignored regression. While regression is likely to be rare, it cannot be ignored from a methodologic standpoint, the investigators wrote.

The researchers therefore set out in the current study to assess positive and negative changes in sacroiliitis on plain pelvic radiographs over time in 975 patients from the ASAS cohort who had chronic back pain of unknown origin or undiagnosed peripheral symptoms (Ann Rheum Dis. 2016 Feb 22. doi: 10.1136/annrheumdis-2015-208964).

Of the 357 of the patients who had paired plain pelvic radiographs available at baseline and follow-up, 17.4% (62/357) fulfilled the criteria for radiographic axSpA at baseline, as defined by modified New York criteria (mNY). At a mean follow-up of 4.4 years, this figure had risen to 22.4% (80/357), suggesting a net progression of 5%.

However, when the authors cross-tabulated their figures, more than half (36/62) of the patients considered mNY positive at baseline were assessed as mNY negative at follow-up. This would mean that radiographic sacroiliitis would have regressed in 58% of the cases; conversely, only 54 of 295 patients (18.3%) became mNY positive at follow-up.

“If only positive change (progression) is valued and negative change is ignored, one would disregard measurement error and spuriously attribute part of the observed positive change to real progression,” the research team explained. “The most likely explanation of our strange and extreme observation is that subtle radiographic progression (the signal) – if truly present – cannot be reliably distinguished from measurement error (the noise). These sobering data clearly illustrate that more research is needed in visualising progression in axSpA.”

ASAS funded the study. The authors had no competing interests to declare.

Sacroiliitis observed in patients with axial spondyloarthritis more often regressed rather than progressed on radiography over nearly 5 years of follow-up of the Assessment of SpondyloArthritis international Society (ASAS) cohort, which lead author Dr. Alexandre Sepriano and his colleagues called “strange” and “sobering.”

The findings call into question the reliability of plain pelvic radiographs for detecting subtle change in sacroiliitis and should prompt the evaluation of alternative imaging modalities such as MRI and low-dose CT, according to Dr. Sepriano of Leiden (the Netherlands) University Medical Center and his associates.

©Rocky89/Thinkstock.com

Determining the presence of radiographic sacroiliitis is prognostically relevant and can pave the way for treatment with biologics, but ambiguity in making this decision and in tracking progression has been revealed in the large inter-and intrareader variability found in previous studies. Furthermore, previous studies tracking progression of nonradiographic axial spondyloarthritis (axSpA) to radiographic axSpA have addressed only disease progression and ignored regression. While regression is likely to be rare, it cannot be ignored from a methodologic standpoint, the investigators wrote.

The researchers therefore set out in the current study to assess positive and negative changes in sacroiliitis on plain pelvic radiographs over time in 975 patients from the ASAS cohort who had chronic back pain of unknown origin or undiagnosed peripheral symptoms (Ann Rheum Dis. 2016 Feb 22. doi: 10.1136/annrheumdis-2015-208964).

Of the 357 of the patients who had paired plain pelvic radiographs available at baseline and follow-up, 17.4% (62/357) fulfilled the criteria for radiographic axSpA at baseline, as defined by modified New York criteria (mNY). At a mean follow-up of 4.4 years, this figure had risen to 22.4% (80/357), suggesting a net progression of 5%.

However, when the authors cross-tabulated their figures, more than half (36/62) of the patients considered mNY positive at baseline were assessed as mNY negative at follow-up. This would mean that radiographic sacroiliitis would have regressed in 58% of the cases; conversely, only 54 of 295 patients (18.3%) became mNY positive at follow-up.

“If only positive change (progression) is valued and negative change is ignored, one would disregard measurement error and spuriously attribute part of the observed positive change to real progression,” the research team explained. “The most likely explanation of our strange and extreme observation is that subtle radiographic progression (the signal) – if truly present – cannot be reliably distinguished from measurement error (the noise). These sobering data clearly illustrate that more research is needed in visualising progression in axSpA.”

ASAS funded the study. The authors had no competing interests to declare.

Sacroiliitis observed in patients with axial spondyloarthritis more often regressed rather than progressed on radiography over nearly 5 years of follow-up of the Assessment of SpondyloArthritis international Society (ASAS) cohort, which lead author Dr. Alexandre Sepriano and his colleagues called “strange” and “sobering.”

The findings call into question the reliability of plain pelvic radiographs for detecting subtle change in sacroiliitis and should prompt the evaluation of alternative imaging modalities such as MRI and low-dose CT, according to Dr. Sepriano of Leiden (the Netherlands) University Medical Center and his associates.

©Rocky89/Thinkstock.com

Determining the presence of radiographic sacroiliitis is prognostically relevant and can pave the way for treatment with biologics, but ambiguity in making this decision and in tracking progression has been revealed in the large inter-and intrareader variability found in previous studies. Furthermore, previous studies tracking progression of nonradiographic axial spondyloarthritis (axSpA) to radiographic axSpA have addressed only disease progression and ignored regression. While regression is likely to be rare, it cannot be ignored from a methodologic standpoint, the investigators wrote.

The researchers therefore set out in the current study to assess positive and negative changes in sacroiliitis on plain pelvic radiographs over time in 975 patients from the ASAS cohort who had chronic back pain of unknown origin or undiagnosed peripheral symptoms (Ann Rheum Dis. 2016 Feb 22. doi: 10.1136/annrheumdis-2015-208964).

Of the 357 of the patients who had paired plain pelvic radiographs available at baseline and follow-up, 17.4% (62/357) fulfilled the criteria for radiographic axSpA at baseline, as defined by modified New York criteria (mNY). At a mean follow-up of 4.4 years, this figure had risen to 22.4% (80/357), suggesting a net progression of 5%.

However, when the authors cross-tabulated their figures, more than half (36/62) of the patients considered mNY positive at baseline were assessed as mNY negative at follow-up. This would mean that radiographic sacroiliitis would have regressed in 58% of the cases; conversely, only 54 of 295 patients (18.3%) became mNY positive at follow-up.

“If only positive change (progression) is valued and negative change is ignored, one would disregard measurement error and spuriously attribute part of the observed positive change to real progression,” the research team explained. “The most likely explanation of our strange and extreme observation is that subtle radiographic progression (the signal) – if truly present – cannot be reliably distinguished from measurement error (the noise). These sobering data clearly illustrate that more research is needed in visualising progression in axSpA.”

ASAS funded the study. The authors had no competing interests to declare.