Skip to main content
. 2021 Nov 22;11:22636. doi: 10.1038/s41598-021-02168-4

Table 3.

Comparison of patient-level accuracy between classification tasks. The performance in detecting EBV + MSI/dMMR subgroup is shown for “EBV + MSI vs. others” and “EBV vs. MSI vs. others” tasks, whereas the values of “EBV vs. MSI + others” and “MSI vs. EBV + others” tasks are performances for detecting EBV and MSI/dMMR, respectively. “Combination of 3) and 4)” shows the performance for detecting EBV + MSI/dMMR based on the results of 3) and 4) tasks. EBV, Epstein-Barr virus; MSI, Microsatellite instability; dMMR, Mismatch repair deficiency; AUC, Area under the curve; CI, Confidence interval.

Classification task Sensitivity Specificity AUC (95%CI)
(1) EBV + MSI vs. others

0.879

(29/33)

0.878

(43/49)

0.947

(0.901–0.992)

(2) EBV vs. MSI vs. others

0.909

(30/33)

0.837

(41/49)

0.931

(0.876–0.986)

(3) EBV vs. MSI + others

0.857

(18/21)

0.951

(58/61)

0.980

(0.956–1.000)

(4) MSI vs. EBV + others

0.833

(10/12)

0.814

(57/70)

0.880

(0.759–1.000)

Combination of (3) and (4)

0.879

(29/33)

0.735

(36/49)

 −