Skip to main content
. 2021 Apr 14;13(8):1875. doi: 10.3390/cancers13081875

Table 2.

Performance of developed models identifying platinum-resistant cases.

No. of Variables List Machine
Learning
AUC Sensitivity Specificity Balanced
Accuracy
Threshold
1 FIGO stage LR 0.556 1.000 0.111 0.556 0.025
RF 0.500 0 1.000 0.500 0
SVM 0.556 1.000 0.167 0.583 0.241
DNN 0.558 1.000 0.122 0.561 0.214
1 Residual tumor size after PDS/IDS LR 0.586 0.605 0.564 0.584 0.172
RF 0.500 0 1.000 0.500 0
SVM 0.586 0.605 0.564 0.584 0.295
DNN 0.587 0.605 0.564 0.584 0.355
2 FIGO stage + Residual tumor size after PDS/IDS LR 0.611 0.605 0.570 0.588 0.203
RF 0.500 0 1.000 0.500 0
SVM 0.611 0.605 0.570 0.588 0.309
DNN 0.611 0.605 0.570 0.588 0.252
6 Age + Serum CA125 levels * + NAC + Pelvic LN status + Involvement of pelvic tissue other than uterus and tube + Involvement of small bowel and mesentery LR 0.741 0.778 0.622 0.700 0.175
RF 0.738 0.538 0.887 0.713 0.185
SVM 0.733 0.731 0.745 0.738 0.232
DNN 0.721 0.857 0.556 0.706 0.357
7 Age + Serum CA125 levels * + NAC + Pelvic LN status + Involvement of pelvic tissue other than uterus and tube + Involvement of small bowel and mesentery + FIGO stage LR 0.748 0.920 0.476 0.698 0.141
RF 0.704 0.800 0.524 0.662 0.034
SVM 0.745 0.920 0.457 0.689 0.133
DNN 0.646 0.655 0.625 0.640 0.218
7 Age + Serum CA125 levels * + NAC + Pelvic LN status + Involvement of pelvic tissue other than uterus and tube + Involvement of small bowel and mesentery + Residual tumor size after PDS/IDS LR 0.741 0.793 0.563 0.678 0.144
RF 0.719 0.960 0.385 0.672 0.021
SVM 0.740 0.517 0.883 0.700 0.259
DNN 0.735 0.680 0.654 0.667 0.461
8 Age + Serum CA125 levels * + NAC + Pelvic LN status + Involvement of pelvic tissue other than uterus and tube + Involvement of small bowel and mesentery + FIGO stage + Residual tumor size after PDS/IDS LR 0.738 0.769 0.648 0.708 0.211
RF 0.738 0.897 0.519 0.708 0.065
SVM 0.729 0.519 0.883 0.701 0.293
DNN 0.740 0.852 0.561 0.707 0.088

Abbreviations: AUC, area under the receiver operating characteristic curve; CA-125, cancer antigen 125; DNN, deep neural network; FIGO, International Federation of Gynecology and Obstetrics; IDS, interval debulking surgery; LN, lymph node; LR, logistic regression; NAC, neoadjuvant chemotherapy; PDS, primary debulking surgery; RF, random forest; SVM, support vector machine. * Ln-transformed.