Table 3:

Correlation between individual biomarkers and pathology-proved spondylodiscitis^a

	Positive Pathologyy
	r	P	PPV	NPV	Sensitivity	Specificity
Epidural enhancement
Observer 1	0.52	<.001c	61.3	87.8	79.2	75.0
Observer 2	0.36	.002c	42.1	100	100	31.3
Observer 3	0.33	.004c	40.7	100	100	27.1
Majority consensus	0.41	.001c	44.4	100	100	37.5
Paraspinal enhancement
Observer 1	0.27	.02c	40.7	88.9	91.7	33.3
Observer 2	0.23	.05c	36.9	100	100	14.6
Observer 3	0.24	.04c	43.6	78.8	70.8	54.2
Majority consensus	0.33	.01c	40.7	100	100	27.1
Hyperintense T2 disc signal
Observer 1	0.37	.02c	55.5	80.1	86.5	45.0
Observer 2	0.19	.03c	33.0	100	100	20.2
Observer 3	0.21	.03c	41.2	79.2	73.1	51.0
Majority consensus	0.29	.03c	51.0	100	100	33.2
Vertebral endplate erosion
Observer 1	0.333	.01c	46.0	83.4	88.8	30.2
Observer 2	0.22	.05c	35.6	95.5	93.2	25.7
Observer 3	0.30	.0c	39.4	80.6	75.6	56.7
Majority consensus	0.31	.02c	42.1	92.3	92.0	35.6
Blood culture growth	0.06	.63	50	66.7	4.2	97.9
Fever status	0.26	.03c	48.3	76.7	58.3	68.8
Leukocytosis	0.032	.80	40	64.3	16.7	85.7
Hemoglobin countb	–0.05	.67	NA	NA	NA	NA
Platelet countb	0.20	.11	NA	NA	NA	NA
ALP levelb	–0.03	.84	NA	NA	NA	NA
ESRb	–0.09	.49	NA	NA	NA	NA
CRPb	0.29	.02c	NA	NA	NA	NA

Note:—NA indicates not applicable; ALP, alkaline phosphatase.

^aThe 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.

^bContinuous variables in which the Pearson correlation was used.

^cStatistically significant P values (<.05).