Table 4.

Classifier performance in the validation set (n = 130)

Biomarker performance metric	Clinical model	Clinicogenomic model	P
Area under the curve (95% CI)*	0.74 (0.66 to 0.83)	0.81 (0.74 to 0.89)	.01
Sensitivity (95% CI)†	0.79 (0.67 to 0.88)	0.91 (0.81 to 0.97)	.03
Specificity (95% CI)†	0.58 (0.45 to 0.70)	0.52 (0.39 to 0.64)	.42
Negative predictive value (95% CI)‡	0.73 (0.58 to 0.84)	0.85 (0.69 to 0.94)	.03
Positive predictive value (95% CI)‡	0.66 (0.54 to 0.76)	0.66 (0.55 to 0.76)	.97
Accuracy (95% CI)§	0.68 (0.60 to 0.76)	0.72 (0.63 to 0.79)	.68

^*P value comparing models calculated using Delong’s two-sided test. CI = confidence interval.

^†P value comparing models calculated using McNemar’s two-sided chi-square test.

^‡P value comparing models calculated using two-sided generalized score statistic.

^§P value comparing models calculated using two-sided Fisher Exact test.