Table 2 Results of multivariate modelling.

From: Large-scale proteomics reveals precise biomarkers for detection of ovarian cancer in symptomatic women

  

AUCa

Sensb

Specb

B vs I–II

Discovery

0.94 (0.88–1.00)

1.00 (1.00–1.00)

0.70 (0.55–0.85)

Replication

0.90 (0.81–0.98)

0.80 (0.60–1.00)

0.79 (0.68–0.89)

p-valuec

0.45

0.08

0.43

B vs III–IV

Discovery

0.97 (0.94–1.00)

0.96 (0.91–1.00)

0.85 (0.73–0.97)

Replication

0.93 (0.88–0.97)

0.88 (0.81–0.94)

0.79 (0.66–0.89)

p-valuec

0.12

0.14

0.57

B vs I–IV

Discovery

0.97 (0.94–0.99)

0.96 (0.91–1.00)

0.73 (0.58–0.88)

Replication

0.92 (0.88–0.97)

0.93 (0.88–0.97)

0.77 (0.64–0.89)

p-valuec

0.12

0.53

0.79

  1. aNumbers for discovery and replication are given as a point estimate and 95% confidence intervals. bThe point estimate and 95% confidence interval is given at a cut-off defined in the discovery cohort at the point on the ROC (receiver operating characteristics) curve closest to perfect classification. cp-values for difference in AUCs were calculated using the DeLong’s method. For differences of sensitivity and specificity, a Fisher’s exact test was used. B—Benign. Roman numerals specify ovarian cancer stage.
Back to article page