Table 3:

Correlation between individual biomarkers and pathology-proved spondylodiscitisa

Positive Pathologyy
rPPPVNPVSensitivitySpecificity
Epidural enhancement
 Observer 10.52<.001c61.387.879.275.0
 Observer 20.36.002c42.110010031.3
 Observer 30.33.004c40.710010027.1
 Majority consensus0.41.001c44.410010037.5
Paraspinal enhancement
 Observer 10.27.02c40.788.991.733.3
 Observer 20.23.05c36.910010014.6
 Observer 30.24.04c43.678.870.854.2
 Majority consensus0.33.01c40.710010027.1
Hyperintense T2 disc signal
 Observer 10.37.02c55.580.186.545.0
 Observer 20.19.03c33.010010020.2
 Observer 30.21.03c41.279.273.151.0
 Majority consensus0.29.03c51.010010033.2
Vertebral endplate erosion
 Observer 10.333.01c46.083.488.830.2
 Observer 20.22.05c35.695.593.225.7
 Observer 30.30.0c39.480.675.656.7
 Majority consensus0.31.02c42.192.392.035.6
Blood culture growth0.06.635066.74.297.9
Fever status0.26.03c48.376.758.368.8
Leukocytosis0.032.804064.316.785.7
Hemoglobin countb–0.05.67NANANANA
Platelet countb0.20.11NANANANA
ALP levelb–0.03.84NANANANA
ESRb–0.09.49NANANANA
CRPb0.29.02cNANANANA
  • Note:—NA indicates not applicable; ALP, alkaline phosphatase.

  • a The Spearman correlation was used for rank/categoric variables. PPV, NPV, sensitivity, and specificity numbers are represented in percentages. Continuous variables will not have PPV, NPV, sensitivity, or specificity values without established thresholds.

  • b Continuous variables in which the Pearson correlation was used.

  • c Statistically significant P values (<.05).