. 2024 Jan 17;15(2):e02867-23. doi: 10.1128/mbio.02867-23

TABLE 2.

Logistic regression and CART prediction models

Variable	N	Cut-point^c	Sensitivity^e	Specificity^e	Accuracy^e	AUC^b	P value^a
Siderophore	49	>100^f	.82 (13/16)	0.33 (11/33)	0.49 (24/49)	0.502	0.860
Mucoviscosity LB	49	>0.11	.63 (10/16)	0.76 (25/33)	0.71 (35/49)	0.723	0.034
Mucoviscosity c-M9-CA-te	49	>0.088	1.0 (16/16)	0.70 (23/33)	0.80 (39/49)	0.869	0.001
Marker count	49	5	0.94 (15/16)	0.94 (31/33)	0.94 (46/49)	0.962	0.004
Marker count	49	≥4^d	1 (16/16)	0.76 (25/33)	0.84 (41/49)
Kleborate score	49	>4.05	0.44 (7/16)	0.94 (31/33)	0.78 (38/49)	0.768	0.003
Mash	29	<0.008	0.77 (7/9)	0.9 (18/20)	0.86 (25/29)	0.900	<0.001
Jaccard	29	<0.318	0.77 (7/9)	0.95 (19/20)	0.90 (26/29)	0.919	<0.001

^{^a}

Exact test about the odds ratio from logistic regression model.

^{^b}

Area under the ROC curve.

^{^c}

Derived from classification and regression tree (CART) prediction model.

^{^d}

Assessment at a potentially clinically relevant cut-point.

^{^e}

Computed using cut-point values.

^{^f}

Determined from ROC curve.