• Mona Aarenstrup Karlsen, Noreen Sandhu, Claus Høgdall, Ib Jarle Christensen, Lotte Nedergaard, Lene Lundvall, Svend A Engelholm, Anette T Pedersen, Dorthe Hartwell, Magnus Lydolph, Inga Alice Laursen, and Estrid V S Høgdall.
• Department of Pathology, Molecular Unit, Herlev University Hospital, University of Copenhagen, Denmark.
• Gynecol. Oncol. 2012 Nov 1; 127 (2): 379-83.

ObjectiveDiagnostic factors are needed to improve the currently used serum CA125 and risk of malignancy index (RMI) in differentiating ovarian cancer (OC) from other pelvic masses, thereby achieving precise and fast referral to a tertiary center and correct selection for further diagnostics. The aim was to evaluate serum Human Epididymis protein 4 (HE4) and the risk of ovarian malignancy algorithm (ROMA) for these purposes.MethodsSerum from 1218 patients in the prospective ongoing pelvic mass study was collected prior to diagnosis. The HE4 and CA125 data were registered and evaluated separately and combined in ROMA and compared to RMI.Results809 benign tumors, 79 borderline ovarian tumors, 252 OC (64 early and 188 late stage), 9 non-epithelial ovarian tumors and 69 non-ovarian cancers were evaluated. Differentiating between OC and benign disease the specificity was 62.2 (CA125), 63.2 (HE4), 76.5 (ROMA) and 81.5 (RMI) at a set sensitivity of 94.4 which corresponds to RMI=200. The areas under the curve (AUC) were 0.854 (CA125), 0.864 (HE4), 0,897 (ROMA) and 0.905 (RMI) for benign vs. early stage OC. For premenopausal benign vs. OC AUC were 0.925 (CA125), 0.905 (HE4), 0.909 (ROMA) and 0.945 (RMI).ConclusionHE4 and ROMA helps differentiating OC from other pelvic masses, even in early stage OC. ROMA performs equally well as the ultrasound depending RMI and might be valuable as a first line biomarker for selecting high risk patients for referral to a tertiary center and further diagnostics. Further improvements of HE4 and ROMA in differentiating pelvic masses are still needed, especially regarding premenopausal women.Copyright © 2012. Published by Elsevier Inc.

36 keywords...

hide…