Diagnostic accuracy and cost-effectiveness of different strategies to triage women with adnexal masses: a prospective study

Objective Transvaginal sonography (TVS) and serum biomarkers are used widely in clinical practice to triage women with adnexal masses, but the effectiveness of current biomarkers is weak. The aim of this study was to determine the best method of diagnosing patients with adnexal masses, in terms of diagnostic accuracy and economic costs, among four triage strategies: (1) the International Ovarian Tumor Analysis group's simple rules (SR) for interpretation of TVS with subjective assessment (SA) by an experienced ultrasound operator when TVS results are inconclusive (referred to hereafter as SR +/- SA), (2) SR +/- SA and cancer antigen 125 (CA 125), (3) SR +/- SA and human epididymis protein 4 (HE4) and (4) SR +/- SA and the risk of malignancy algorithm (ROMA). Our main hypothesis was that the addition of the biomarkers to SR +/- SA could improve triaging of these patients in terms of diagnostic accuracy (i.e. malignant vs benign). As secondary analyses, we estimated the cost effectiveness of the four strategies and the diagnostic accuracy of SR +/- SA at the study hospitals. Methods Between February 2013 and January 2015, 447 consecutive patients who were scheduled for surgery for an adnexal mass at the S. Anna and Mauriziano Hospitals in Turin were enrolled in this multicenter prospective cohort study. Preoperative TVS was performed and preoperative CA125 and HE4 levels were measured. Pathology reports were used to assess the diagnostic accuracy of the four triage strategies and the cost of each strategy was calculated. Results A total of 391 patients were included in the analysis: 57% (n = 221) were premenopausal and 43% (n = 170) were postmenopausal. The overall prevalence of malignancy was 21%. SR were conclusive in 89% of patients and thus did not require SA; the overall performance of SR +/- SA showed a sensitivity of 82%, specificity of 92% and positive and negative predictive values and positive and negative likelihood ratios of 74%, 95%, 10.5 and 0.19, respectively. In premenopausal women, mean cost among the four triage strategies varied from (sic)36.41 for SR +/- SA to (sic)70.12 for SR +/- SA + ROMA. The addition of biomarkers to SR +/- SA showed no diagnostic advantage compared with SR +/- SA alone and was more costly. Among postmenopausal women, mean cost among the four triage strategies varied from (sic)39.52 for SR +/- SA to (sic)73.23 for SR +/- SA + ROMA. Among these women, SR +/- SA + CA125 and SR +/- SA + ROMA had a higher sensitivity (both 92% (95% CI, 85-99%)) than SR +/- SA (81% (95% CI, 71-91%)), but SR +/- SA had a higher specificity (84% (95% CI, 77-91%)). SR +/- SA + CA125 and SR +/- SA + ROMA improved diagnostic accuracy, each diagnosing a third more malignant adnexal masses. In postmenopausal women, compared with SR +/- SA alone, SR +/- SA + CA125 showed a net reclassification improvement (NRI) of 28.8% at an extra cost of (sic)13.00, while the extra cost for SR +/- SA + ROMA was (sic)33.71, with a comparable gain, in terms of NRI, as that of SR +/- SA + CA125. Conclusions In our study sample, SR +/- SA seems to be the best strategy to triage women with adnexal masses for surgical management. Among postmenopausal women, SR +/- SA + CA125 increased the NRI at a reasonable extra cost. Our data do not justify the use of HE4 and ROMA in the initial triage of women with adnexal masses. Copyright (C) 2016 ISUOG. Published by John Wiley & Sons Ltd.