In the Journals

Lumbar spine MRI limited in diagnosing ankylosing spondylitis

Inflammatory and fatty lesions seen on lumbar spine MRI had limited diagnostic value in distinguishing ankylosing spondylitis from other diseases associated with chronic back pain, according to research results.

Researchers in China reviewed lumbar spine MRI results of 400 patients reporting chronic back pain from 2007 to 2011, including 192 patients diagnosed with ankylosing spondylitis (AS; mean age, 28.4 years; 150 males; mean disease duration, 5.6 years) and 208 patients without AS. Lumbar spine MRI was used to determine lesions including vertebral corner inflammatory lesions (CIL), inflammation in posterior elements (PE) and fatty deposition lesions (FDL), which were blindly scored.

In the AS cohort, CIL and FDL frequencies were higher (mean, 11.3 and 0.5, respectively) than the non-AS cohort (mean 4.2, and 0.1; both P<.01); the positive rate of PE in the spine between groups, however, was not significantly different (31.8% vs. 27.9%; P=.103). All imaged lesions had higher scores in the AS cohort compared with patients without AS (P<.01, all).

As the cut-off score for distinguishing AS from other disease increased, the positive likelihood ratio also increased (range, 1.14-18.42).

“The biggest value of area under the receiver operating characteristic curve of all types of lesions was only 62.58%,” the researchers said.

“We found that the diagnostic value of [inflammatory and fatty deposition] lesions in lumbar spine MRI to AS was limited,” they concluded. “However, a high specificity of high scores of lesions found in lumbar spine MRI may be of diagnostic value in selected cases. The diagnosis of AS would be more convincing if patients had high scores of these three types of lesions (CIL≥16, and/or inflammation in PE of the spine≥5, and/or FDL≥2).”

Disclosure: The researchers report no relevant financial disclosures.

Inflammatory and fatty lesions seen on lumbar spine MRI had limited diagnostic value in distinguishing ankylosing spondylitis from other diseases associated with chronic back pain, according to research results.

Researchers in China reviewed lumbar spine MRI results of 400 patients reporting chronic back pain from 2007 to 2011, including 192 patients diagnosed with ankylosing spondylitis (AS; mean age, 28.4 years; 150 males; mean disease duration, 5.6 years) and 208 patients without AS. Lumbar spine MRI was used to determine lesions including vertebral corner inflammatory lesions (CIL), inflammation in posterior elements (PE) and fatty deposition lesions (FDL), which were blindly scored.

In the AS cohort, CIL and FDL frequencies were higher (mean, 11.3 and 0.5, respectively) than the non-AS cohort (mean 4.2, and 0.1; both P<.01); the positive rate of PE in the spine between groups, however, was not significantly different (31.8% vs. 27.9%; P=.103). All imaged lesions had higher scores in the AS cohort compared with patients without AS (P<.01, all).

As the cut-off score for distinguishing AS from other disease increased, the positive likelihood ratio also increased (range, 1.14-18.42).

“The biggest value of area under the receiver operating characteristic curve of all types of lesions was only 62.58%,” the researchers said.

“We found that the diagnostic value of [inflammatory and fatty deposition] lesions in lumbar spine MRI to AS was limited,” they concluded. “However, a high specificity of high scores of lesions found in lumbar spine MRI may be of diagnostic value in selected cases. The diagnosis of AS would be more convincing if patients had high scores of these three types of lesions (CIL≥16, and/or inflammation in PE of the spine≥5, and/or FDL≥2).”

Disclosure: The researchers report no relevant financial disclosures.