Introduction/Background The aim of this study is to evaluate the predictive performance of OncoOVARIAN Dx algorithm, which takes into account tumor markers (beta HCG, CA 19.9, CEA, AFP, CA 125, HE4), general biochemistry and clinical data (age, menopause, comorbidities) in patients scheduled for surgical removal of a suspicious adnexal tumor in comparison with the ROMA model.

Methodology Consecutive women diagnosed with an adnexal tumor mass and scheduled for surgical intervention at a single tertiary cancer center between October 2018–April 2019 were enrolled. Preoperative values of tumor markers and general biochemistry (ASAT, ALAT, GGT, total bilirubin, creatinine) were determined. Following surgery, histological type and stage were determined based on FIGO classification.

Results A total of 44 patients were selected, including 19 benign (BEN), 5 borderline (BOR) and 20 malignant (MAL) epithelial ovarian cancer (EOC) cases on final pathology. BOR tumors comprised 3 serous (S) and 2 mucinous (M) FIGO stage I cases. MAL tumors included 13 high grade S, 4 endometrioid and 3 M histotypes, FIGO stage IA–IIIC. The two models demonstrated very good correlation (Phi 0.78, p<0.001), however among the 30 cases deemed ROMA low risk there were 5 (1 BEN, 2 BOR/M, 2 MAL/high grade S) cases that had OncoOVARIAN Dx moderate/high risk. Among the 7 BOR/MAL cases labeled OncoOVARIAN Dx low risk 71% were FIGO IA and the remaining two cases were FIGO IB and IIA. The sensibility (Se), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV) of OncoOVARIAN Dx versus ROMA model were 72% vs. 56%, 94.73% vs. 100%, 94.73% vs. 100%, 72% vs. 63.33%, respectively.

Conclusion OncoOVARIAN Dx model demonstrated higher Se and NPV compared to ROMA and could be a useful marker in the preoperative management of adnexal masses however larger studies are warranted to validate and further refine this algorithm.

Disclosure This work was supported by a National Grant PN-III-P2-2.1-CI-2018-1509.