Skip to main content
. 2020 Jan 21;9(2):299. doi: 10.3390/jcm9020299

Table 4.

Subgroup analyses for validation of performance of RMI, CA125, GOT-1 (RMI + HE4), and GOT-2 (CA125 + HE4) comparing benign disease with EOC in dualistic model Type I and Type II tumors.

Group (n) Model Type I Type II
ROC SN% (75% SP) SP% (Target SN) ROC SN% (75% SP) SP% (Target SN)
AUC 95% CI AUC 95% CI
Benign (445) vs. Type I (36)/Type II (99) RMI (cut-off <200) 0.89 0.84–0.94 94 84 0.97 0.95–0.98 98 84
GOT-1 (RMI + HE4) 0.90 0.84–0.95 94 85 0.97 0.95–0.99 97 89
Benign (445) vs. Type I (36)/Type II (99) CA125 (cut-off >35 U/mL) 0.85 0.79–0.91 75 68 0.94 0.91–0.97 93 68
GOT-2 (CA125 + HE4) 0.87 0.82–0.93 83 80 0.96 0.94–0.99 96 71

Specificity was calculated using target SN. Target SN = target sensitivity of RMI at cut-off >200 was 83% (type I) and 95% (type II); target sensitivity of CA125 at cut-off >35 was 83% (type I) and 96% (type II). AUC = area under the curve; EOC = epithelial ovarian cancer; GOT-1 = Gothenburg index 1; GOT-2 = Gothenburg index 2; Pre-M = premenopausal; Post-M = postmenopausal; ROC = receiver operating characteristics; SN = sensitivity; SP = specificity; Type I = low-grade tumors; Type II = high-grade tumors.