Performance of the IOTA ADNEX model in preoperative discrimination of adnexal masses in a gynecological oncology center

Transcription

1 Ultrasound Obstet Gynecol 2017; 49: Published online 12 April 2017 in Wiley Online Library (wileyonlinelibrary.com). DOI: /uog Performance of the IOTA ADNEX model in preoperative discrimination of adnexal masses in a gynecological oncology center K. G. ARAUJO 1,2, R. M. JALES 1,2,P.N.PEREIRA 1,A.YOSHIDA 1, L. DE ANGELO ANDRADE 3, L. O. SARIAN 1 and S. DERCHAIN 1 1 Department of Obstetrics and Gynecology, Faculty of Medical Sciences, University of Campinas, Unicamp, Campinas, São Paulo, Brazil; 2 Section of Ultrasonography, Prof. José Aristodemo Pinotti Women s Hospital, CAISM, University of Campinas, Unicamp, Campinas, São Paulo, Brazil; 3 Department of Pathologic Anatomy, Faculty of Medical Sciences, University of Campinas, Unicamp, Campinas, São Paulo, Brazil KEYWORDS: adnexal masses; ovarian cancer; ovarian neoplasm; preoperative; ultrasonography ABSTRACT Objective To evaluate the performance of the International Ovarian Tumor Analysis (IOTA) ADNEX model in the preoperative discrimination between benign ovarian (including tubal and para-ovarian) tumors, borderline ovarian tumors (BOT), Stage I ovarian cancer (OC), Stage II IV OC and ovarian metastasis in a gynecological oncology center in Brazil. Methods This was a diagnostic accuracy study including 131 women with an adnexal mass invited to participate between February 2014 and November Before surgery, pelvic ultrasound examination was performed and serum levels of tumor marker CA 125 were measured in all women. Adnexal masses were classified according to the IOTA ADNEX model. Histopathological diagnosis was the gold standard. Receiver operating characteristics (ROC) curve analysis was used to determine the diagnostic accuracy of the model to classify tumors into different histological types. Results Of 131 women, 63 (48.1%) had a benign ovarian tumor, 16 (12.2%) had a BOT, 17 (13.0%) had Stage I OC, 24 (18.3%) had Stage II IV OC and 11 (8.4%) had ovarian metastasis. The area under the ROC curve (AUC) was 0.92 (95% CI, ) for the basic discrimination between benign vs malignant tumors using the IOTA ADNEX model. Performance was high for the discrimination between benign vs Stage II IV OC, BOT vs Stage II IV OC and Stage I OC vs Stage II IV OC, with AUCs of 0.99, 0.97 and 0.94, respectively. Performance was poor for the differentiation between BOT vs Stage I OC and between Stage I OC vs ovarian metastasis with AUCs of Conclusion The majority of adnexal masses in our study were classified correctly using the IOTA ADNEX model. On the basis of our findings, we would expect the model to aid in the management of women with an adnexal mass presenting to a gynecological oncology center. Copyright 2016 ISUOG. Published by John Wiley & Sons Ltd. INTRODUCTION Although infrequent, ovarian cancer is the fifth most common cause of death from cancer among women in the USA, excluding non-melanoma skin cancers 1. Whilst survival rate in women with ovarian cancer has increased since the 1970s, further advances are necessary to improve patient outcome 2. As far as diagnostic imaging of adnexal masses is concerned, the International Ovarian Tumor Analysis (IOTA) group conducted research of unprecedented accuracy 3. Since 2005, this group has proposed logistic regression models and ultrasound-based predictive rules (LR1, LR2 and simple rules) for the preoperative characterization of adnexal masses 4,5,which have been validated prospectively and externally 6,7. IOTA studies have demonstrated that these models have better performance than preexisting predictive risk models, such as the risk of malignancy index (RMI) 7,8. More recently, the IOTA group developed a polytomous risk prediction model for more detailed characterization of adnexal masses, the Assessment of Different NEoplasias in the adnexa (ADNEX). This model considers three clinical and six ultrasound features to predict the risk of benign ovarian tumor, borderline ovarian tumor (BOT), Stage I ovarian cancer (OC), Stage II IV OC and ovarian metastasis 9,10. As well as differentiating Correspondence to: Dr S. Derchain, Department of Obstetrics and Gynecology, Faculty of Medical Sciences, PO Box 6111, University of Campinas, Unicamp, , Campinas, São Paulo, Brazil ( derchain@fcm.unicamp.br) Accepted: 11 May 2016 Copyright 2016 ISUOG. Published by John Wiley & Sons Ltd. ORIGINAL PAPER

2 Performance of the IOTA ADNEX model 779 benign from malignant tumors, the IOTA ADNEX model attempts to identify preoperatively various histological types and extents of tumor spread. This comprehensive classification is clinically important. For example, BOTs confined to the ovary are associated with excellent survival, almost 100% within 10 years. They often affect young women 11 and, indeed, one-third of BOTs are diagnosed under the age of 40 years, when fertility-sparing treatments should be considered 12. As far as invasive disease is concerned, a meta-analysis showed that women with apparent early-stage OC who underwent laparoscopy had lower rates of complications and shorter hospital stays compared with women who underwent laparotomy, with no significant difference in recurrence rate, although studies with long-term follow-up are still needed to assess recurrence rates and survival outcomes 13. Conversely, women with ovarian metastasis require highly specific management and usually need additional diagnostic assessment before treatment. Pathological specimens are often obtained from percutaneous biopsy, and primary cytoreductive surgery is indicated in many cases A diagnostic model that stratifies adnexal masses into four categories according to malignancy might improve triage and be useful for therapeutic decision-making 9. The aim of this study was to evaluate the performance of the IOTA ADNEX model in the preoperative discrimination between benign ovarian (including tubal and para-ovarian) tumors, BOT, Stage I OC, Stage II IV OC and ovarian metastasis in a Brazilian gynecological oncology center. METHODS Patients This was a diagnostic accuracy study conducted in the gynecological oncology center of Prof. José Aristodemo Pinotti Women s Hospital, CAISM, Campinas, Brazil. Three hundred and seventy-three women who had been referred to the ovarian pelvic oncology outpatient clinic due to the presence of an adnexal mass were invited consecutively between February 2014 and November 2015 to participate in the study. After pelvic examination, an ultrasound evaluation was scheduled and peripheral blood was collected for serum measurement of tumor marker CA 125. In 196 women, surgery was not indicated due to the benign/functional characteristics of the adnexal mass and they were excluded from the study. These patients were scheduled for sonographic follow-up at 3, 6 and 12 months. If the patient remained asymptomatic and the imaging studies of the mass remained stable or regressed, patients were scheduled for annual follow-up. Forty-six women were excluded from the study for the following reasons: ultrasound scan was performed by examiners not trained in IOTA terms (n = 15), adnexal mass was actually a uterine lesion (n = 11) or an extragynecological tumor (n = 8), doubtful histology of biopsy sample (n = 4), ectopic pregnancy (n = 2), incomplete ultrasound data (n = 2), no measurement of CA 125 (n = 1), > 120 days between the ultrasound scan and surgery (n = 1), previous diagnosis of OC (n = 1) and death before surgery (n = 1). Figure 1 illustrates the enrolment of women into the study. Among the 131 women included, 114 were scheduled for diagnostic and/or treatment surgical procedures and histopathological assessment was carried out by surgical excision of the adnexal mass through laparoscopy or laparotomy. In 17 cases of non-resectable tumors, pathological specimens were obtained from percutaneous biopsy of pelvic mass or abdominal implant. Non-resectable tumors were defined as Stage IIIc or Stage IV OC according to the International Federation of Gynecology and Obstetrics (FIGO) 17 and were deemed not amenable for surgical removal after clinical evaluation by a trained gynecological oncology surgeon. Therapeutic decisions were based on a set of factors: epidemiological and clinical features, subjective assessment in the ultrasound report, biomarkers and, in some cases, magnetic resonance imaging. Pelvic examination, ultrasound assessment, surgery and histological analysis were all performed at this single institution. For histological classification, the guideline of the World Health Organization International Classification of Ovarian Tumors was followed 18. More than one adnexal mass was detected in 20 women. In cases of multiple masses, the mass with the most complex morphology was included in the statistical analysis. If they presented similar morphology, the largest mass was chosen for inclusion in the study. Written informed consent was obtained from all included women. The study was approved by the University s Research Ethics Committee (protocol no. 008/2010). The STARD statement 19 was followed for reporting the study. Ultrasound examination Ultrasound examinations were performed at the Ultrasound Technical Section of CAISM, using a Voluson Expert 730 (GE Healthcare Ultrasound, Milwaukee, WI, USA) or a Toshiba Xsario SSA 660A (Toshiba Medical Systems Corp., Otawara, Japan). Ultrasound examinations were performed by two experienced gynecologists (R.M.J. and K.G.A.) with further training in gynecological ultrasonography. Each had more than 10 years experience in ultrasound and had performed > 5000 female pelvic ultrasound assessments. R.M.J. obtained his IOTA certificate (London, 2014) before the study commenced. Although K.G.A. obtained her certificate in Leuven, April 2015, she was supervised by R.M.J. during the entire data collection period. In the majority of cases, ultrasound evaluation was performed initially by a transabdominal approach, with the patient s bladder full, then a transvaginal scan, with the patient s bladder empty, was performed. Adnexal masses were described according to IOTA terms and definitions 20. Color Doppler was applied (pulse repetition frequency = khz) and a color score (1 = no blood flow; 2 = minimal blood flow; 3 = moderate blood flow; 4 = marked blood flow) was determined. Doppler parameters were not used for the current analysis. For clinical purposes, the nature of the adnexal mass

3 780 Araujo et al. Women referred to ovarian pelvic oncology outpatient clinic due to adnexal mass (n = 373) Women included (n = 131) Benign tumor (n = 63) Malignant tumor (n = 68) Borderline ovarian tumor (n = 16) Stage I ovarian cancer (n = 17) Stage II IV ovarian cancer (n = 24) Ovarian metastasis (n = 11) Women excluded (n = 242) No surgery or biopsy indication (n = 196) US performed by examiners not trained in IOTA terms (n = 15) Uterine lesion (n = 11) Extragynecological tumor (n = 8)* Doubtful histology of biopsy sample (n = 4) Ectopic pregnancy (n = 2) Incomplete US data (n = 2) No CA 125 measurement (n = 1) > 120 days between US and surgery (n = 1) Previous diagnosis of ovarian cancer (n = 1) Death before surgery (n = 1) Figure 1 Flowchart showing enrolment of women with adnexal mass and reasons for exclusion. *Majority were gastrointestinal tract tumors. IOTA, International Ovarian Tumor Analysis; US, ultrasound. was assessed systematically by subjective assessment on ultrasound as: benign, probably benign, uncertain but more likely benign, uncertain but more likely malignant, probably malignant, malignant. The researchers applied theiotaadnexmodela posteriori, just before statistical analysis, without knowledge of the histological results. CA 125 Serum levels of CA 125 (U/mL) were measured in each patient by the CA 125 II test, which applies the chemoluminescence technique, using an automatic analyzer (Cobas e411; Roche Diagnostics GmbH, Mannheim, Germany), according to the manufacturer s instructions. Statistical analysis We analyzed data using R software (R Core Team, Vienna, Austria) 21. Statistical calculations were performed using 95% CIs, with P < 0.05 considered to be significant. Sample power for the proposed analyses was 91.7%. The study that introduced the IOTA ADNEX model 9 was published 9 months after commencement of our other study on the accuracy of different IOTA models and RMI. In the present analysis, we used data derived from that study. The IOTA ADNEX model was calculated based on clinical and sonographic data obtained before surgery. However, these calculations were performed a posteriori and IOTA ADNEX results were therefore not available when clinical decisions were made. For the purpose of evaluating the performance of the IOTA ADNEX model, each case was reported based on the following parameters: age (years), type of center (gynecological oncology center for all women in our study), maximum diameter of the lesion (mm), maximum diameter of the largest solid part of the lesion (mm), > 10 locules (yes/no), number of papillary projections (0, 1, 2, 3, > 3), acoustic shadowing (yes/no), ascites (yes/no) and CA 125 level (U/mL) 9,10,20.AllIOTA ADNEX parameters were logged objectively when the ultrasound examination was performed. A data collection sheet in Excel was used. The parameters were entered into an IOTA ADNEX mobile application for each case and the risk stratification was acquired. We compared mean values of the components of the IOTA ADNEX model using Student s t-test or the Kruskal Wallis test, as appropriate. Sensitivity, specificity and diagnostic odds ratios for the IOTA ADNEX model were calculated at progressive cut-off points for risk of malignancy and at the cut-off point determined by the receiver operating characteristics (ROC) curve analyses of our data. Areas under the ROC curves (AUC) were determined for different discriminating scenarios. The gold standard was histopathological diagnosis and masses were classified according to the IOTA ADNEX model as benign ovarian tumor, BOT, Stage I OC, Stage II IV OC or ovarian metastasis. RESULTS Histological results of the study group of 131 women with an adnexal mass are provided in Tables 1 and 2. Sixty-three (48.1%) women had a benign ovarian tumor, 16 (12.2%) had a BOT, 17 (13.0%) had Stage I OC, 24 (18.3%) had Stage II IV OC and 11 (8.4%) had ovarian metastasis. Among Stage I IV OC, 80.5% were epithelial, 14.6% originated in the stroma and 4.9% were of mixed histological type. Teratomas were the most frequent benign tumor (31.7%) followed by serous cystadenomas (20.6%). Clinical and sonographic features of adnexal masses in our study cohort are shown in Table 3. Women with Stage I IV OC were significantly older than those with benign, BOT and metastatic tumors. Besides age, the features able to distinguish the category of tumor were serum levels of CA 125, presence of solid tissue within the mass, maximum diameter of the solid tissue, maximum diameter of the solid tissue : maximum diameter of the lesion ratio, > 10 locules and presence of ascites (P < 0.05). The diagnostic performance of the IOTA ADNEX model is shown by the ROC curve (Figure 2) and the performances at progressive cut-off points for probability of malignancy are presented in Table 4. For the basic discrimination between benign and malignant tumors, the AUC was 0.92 (95% CI, ), sensitivity was 88.2% and specificity was 85.7% at an optimal cut-off of 23.6% probability of malignancy. High sensitivity (>90%) was found at cut-offs of 10% and 15%; however,

4 Performance of the IOTA ADNEX model 781 Table 1 Histological type of tumor in 131 women with adnexal mass Histological type Benign (n = 63) Borderline (n = 16) Stage I OC (n = 17) Malignant Stage II IV OC (n = 24) Metastasis (n = 11) Epithelial 23 (36.5) 16 (100) 10 (58.8) 23 (95.8) 10 (90.9) Stromal 2 (3.2) 0 (0) 6 (35.3) 0 (0) 0 (0) Germ cell tumor 21 (33.3) 0 (0) 0 (0) 0 (0) 0 (0) Mixed 0 (0) 0 (0) 1 (5.9) 1 (4.2) 0 (0) Lymphoma 0 (0) 0 (0) 0 (0) 0 (0) 1 (9.1) Non-neoplastic ovarian tumor 15 (23.8) 0 (0) 0 (0) 0 (0) 0 (0) Extraovarian tumor 2 (3.2) 0 (0) 0 (0) 0 (0) 0 (0) Data are given as n (%). OC, ovarian cancer. Table 2 Histological subtype of tumor in 131 women with adnexal mass Histological subtype n (%) Benign disease (n = 63) Teratoma 20 (31.7) Serous cystadenoma 13 (20.6) Endometrioma 7 (11.1) Mucinous cystadenoma 4 (6.3) Cystadenofibroma 3 (4.8) Fibroma 2 (3.2) Brenner tumor 2 (3.2) Ovarian torsion 2 (3.2) Functional cyst 2 (3.2) Other ovarian benign lesion 6 (9.5) Extraovarian tumor 2 (3.2) Malignant disease (n = 68) Serous adenocarcinoma 21 (30.9) Clear cell carcinoma 5 (7.4) Granulosa cell tumor 5 (7.4) Mucinous adenocarcinoma 4 (5.9) Carcinosarcoma 2 (2.9) Serous/mucinous adenocarcinoma 1 (1.5) Endometrioid adenocarcinoma 1 (1.5) Endometrioid/clear cell carcinoma 1 (1.5) Sertoli leydig 1 (1.5) Serous borderline tumor 7 (10.3) Mucinous borderline tumor 6 (8.8) Other borderline tumor 3 (4.4) Ovarian metastasis 11 (16.2) there was an important gain in specificity at a cut-off point of 15% compared with at 10%. Among the most common histological results, considering a cut-off of 15% probability of malignancy, teratomas were classified correctly as benign tumors in 17/20 (85.0%) of the cases, serous cystadenomas in 9/13 (69.2%) and endometriomas in 7/7 (100%). Serous adenocarcinomas were classified correctly as malignant tumors in 20/21 (95.2%) of the cases. Incorrect negative prediction was given for a Stage I serous adenocarcinoma that presented on ultrasound as a 94-mm unilocular cyst in a 75-year-old woman with CA 125 level of 28.4 U/mL. All (5/5) clear cell carcinomas and 4/5 (80.0%) granulosa cell tumors were classified correctly as malignant. The ADNEX model classified correctly 7/7 (100%) cases of serous BOT and 3/6 (50.0%) cases of mucinous BOT as malignant. Table 5 shows the performance of the IOTA ADNEX model in differentiating between the different types of tumors. Performance was high for the discrimination between benign vs Stage II IV OC, benign vs metastasis, BOT vs Stage II IV OC and Stage I OC vs Stage II IV OC, with AUCs of 0.99 (95% CI, ), 0.97 (95% CI, ), 0.97 (95% CI, ) and 0.94 (95% CI, ), respectively. Performance was poor for the differentiation between BOT vs Stage I OC (AUC, 0.64 (95% CI, )) and between Stage I OC vs ovarian metastasis (AUC, 0.64 (95% CI, )). DISCUSSION The IOTA ADNEX model performed well in discriminating between benign and malignant adnexal masses using data obtained from a gynecological oncology center. The model showed good to excellent results in distinguishing Stage II IV OC from the other four tumor classifications and in distinguishing benign tumors from ovarian metastasis, Stage I OC and BOT. Previous studies have demonstrated that, using subjective ultrasound evaluation, more than 50% of BOT were classified incorrectly as benign or unclassifiable and more than 60% were confused with other histopathological diagnoses 22,23. In addition, IOTA simple rules and non-iota models had poorer performance for identifying BOT and Stage I OC 5,8,24 than did the IOTA ADNEX model in our study. We found that the IOTA ADNEX model did not perform well in discriminating BOT and ovarian metastasis from Stage I OC. Ovarian metastasis is usually confined to the ovaries, and affected women have lower CA 125 levels and are younger in comparison with cases of Stage II IV OC. Clinical history and subjective assessment of the morphological and Doppler features of the adnexal mass, in addition to non-gynecological findings, are essential when there is suspicion of secondary disease 14,15. The preoperative diagnosis of BOT, Stage I OC and ovarian metastasis still remains a challenge in the context of ultrasound-based predictive models. In our gynecological oncology center, the malignancy rate was similar to that found in IOTA oncology centers, and was approximately 52%. In our study sample, the cut-off for malignancy probability of 15% had a sensitivity of 91.2%, comparable to the 94.5% found by Van Calster et al. in their pooled data analysis 9.Our specificity of 71.4% was slightly lower than theirs, which

5 782 Araujo et al. Table 3 Sonographic features of tumor in 131 women with adnexal mass Malignant Benign Borderline Stage I OC Stage II IV OC Metastasis Characteristic (n = 63) (n = 16) (n = 17) (n = 24) (n = 11) P Age (years) 44.1 ± ± ± ± ± 12.0 < 0.01 CA 125 (U/mL) 56.5 ± ± ± ± ± < 0.01 Max diameter of lesion (mm) ± ± ± ± ± Presence of solid tissue 16 (25.4) 9 (56.3) 14 (82.4) 24 (100) 10 (90.9) < 0.01 Max diameter of solid tissue (mm) 44.8 ± ± ± ± ± 59.0 < 0.01 Max diameter of solid tissue : max 44.0 ± ± ± ± ± diameter of lesion ratio (%) > 10 locules 10 (15.9) 4 (25.0) 9 (52.9) 4 (16.7) 3 (27.3) 0.03 Number of papillary projections NA 0 61 (96.8) 11 (68.8) 16 (94.1) 22 (91.7) 11 (100) 1 0 (0) 2 (12.5) 0 (0) 1 (4.2) 0 (0) 2 1 (1.6) 0 (0) 0 (0) 0 (0) 0 (0) 3 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) > 3 1 (1.6) 3 (18.8) 1 (5.9) 1 (4.2) 0 (0) Acoustic shadowing 19 (30.2) 0 (0) 0 (0) 0 (0) 0 (0) NA Ascites 2 (3.2) 1 (6.3) 2 (11.8) 14 (58.3) 5 (45.5) < 0.01 Data are given as mean ± SD or n (%). Groups compared using Student s t-test or Kruskal Wallis test, as appropriate. Max, maximum; NA, not applicable; OC, ovarian cancer. Sensitivity % 15.0% 10.0% 5.0% 3.0% Table 4 Performance of International Ovarian Tumor Analysis ADNEX model in discriminating between benign and malignant tumors at progressive cut-offs for probability of malignancy (Prob) Prob (%) AUC (95% CI) Sens (%) Spec (%) dor * 0.92 ( ) Specificity Figure 2 Receiver operating characteristics (ROC) curve analysis of performance of International Ovarian Tumor Analysis ADNEX model in discriminating between benign and malignant adnexal masses. Optimal cut-off was 23.6% for probability of malignancy, at which sensitivity was 88.2%, specificity was 85.7%, positive predictive value was 12.9%, negative predictive value was 13.0% and area under ROC curve was Cut-offs of 3.0%, 5.0%, 10.0% and 15.0% are also indicated. was 78.7%. Considering these results, the cut-off value of 15% might be suitable in an oncology center, contributing to the best indication of the surgical method required. In comparison with the pooled data in the original IOTA study 9, the IOTA ADNEX model in our study showed similar or better performance for distinguishing between benign ovarian tumors vs Stage II IV OC, benign ovarian tumor vs ovarian metastasis, BOT vs Stage II IV OC, Stage I OC vs Stage II IV OC and Stage II IV OC vs ovarian metastasis. We noted a significant difference in the maximum diameter of the adnexal masses in our study in comparison with the IOTA ADNEX study, for all types of tumors; mean maximum diameter of benign tumors was 113 mm in our *Optimum cut-off. AUC, area under receiver operating characteristics curve; dor, diagnostic odds ratio; Sens, sensitivity; Spec, specificity. study vs 63 mm in the IOTA study and of Stages II IV OC was 139 mm in our study vs 85 mm in the IOTA study. This difference could indicate delayed access to tertiary care in the healthcare system of a developing country. The IOTA group found poorer performance of subjective assessment, LR1, LR2, IOTA simple rules and RMI in very small and very large adnexal masses, with the poorest performance found for large tumors with diameter 10 cm 25. It is possible that larger tumors are more difficult to image for ultrasound morphological assessment, particularly in the finding of papillary projections, which might have contributed to the marginally lower performance found in our study in comparison with the original one 9. In our study, acoustic shadowing was 100% associated with benignity and was observed in 30.2% of benign lesions. This sonographic feature is a predictive criterion of benignity in all aforementioned IOTA models 4,5,9. Nevertheless, in a recent study, Timmerman et al. observed that acoustic shadowing was found in 17.4% of benign tumors and in a small number (4.2%) of malignant tumors 26. A limitation of our study is the relatively small sample, especially for the number of BOT, Stage I OC and

6 Performance of the IOTA ADNEX model 783 Table 5 Performance of International Ovarian Tumor Analysis ADNEX model in polytomous discriminations between different types of tumor Discrimination AUC (95% CI) Benign vs malignant 0.92 ( ) Benign vs BOT 0.83 ( ) Benign vs Stage I OC 0.88 ( ) Benign vs Stage II IV OC 0.99 ( ) Benign vs metastasis 0.97 ( ) BOT vs Stage I OC 0.64 ( ) BOT vs Stage II IV OC 0.97 ( ) BOT vs metastasis 0.77 ( ) Stage I OC vs Stage II IV OC 0.94 ( ) Stage I OC vs metastasis 0.64 ( ) Stage II IV OC vs metastasis 0.89 ( ) AUC, area under receiver operating characteristics curve; BOT, borderline ovarian tumor; OC, ovarian cancer. ovarian metastasis, which might preclude a reliable conclusion being made concerning the performance of the IOTA ADNEX model in these specific histological types. Moreover, different from the original study that proposed the IOTA ADNEX model, our data were collected in a single gynecological oncology center with a high rate of malignancy, which could introduce bias. Even though ultrasound evaluation was performed by well-trained medical doctors specialized in ultrasonography, it is worth highlighting that these professionals were non-iota investigators and that the study was conducted at a center located in a middle-resource country. In conclusion, the IOTA ADNEX model had good to excellent performance in discriminating different types of ovarian tumors. We consider this model to be a useful diagnostic tool in the management of patients with adnexal masses presenting to a gynecological oncology center. The classification system stratified according to malignancy might allow more accurate triaging of cases. It should also assist in the prioritization of patients for surgery and therapeutic decision-making, adding diagnostic information to cases with an indication for minimally invasive surgery. Further analyses are needed to determine the use of the IOTA ADNEX model by less experienced ultrasound examiners. ACKNOWLEDGMENTS This study was supported by the Research Support Foundation of the State of São Paulo Fapesp (number 2012/ ) and by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). REFERENCES 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, CA Cancer J Clin 2016; 66: Wright JD, Chen L, Tergas AI, Patankar S, Burke WM, Hou JY, Neugut AI, Ananth CV, Hershman DL. Trends in relative survival for ovarian cancer from 1975 to Obstet Gynecol 2015; 125: Kaijser J. Towards an evidence-based approach for diagnosis and management of adnexal masses: findings of the International Ovarian Tumour Analysis (IOTA) studies. FactsViewsVisObgyn2015; 7: Timmerman D, Testa AC, Bourne T, Ferrazzi E, Ameye L, Konstantinovic ML, Van Calster B, Collins WP, Vergote I, Van Huffel S, Valentin L; International Ovarian Tumor Analysis Group. Logistic regression model to distinguish between the benign and malignant adnexal mass before surgery: a multicenter study by the International Ovarian Tumor Analysis Group. J Clin Oncol 2005; 23: Timmerman D, Testa AC, Bourne T, Ameye L, Jurkovic D, Van Holsbeke C, Paladini D, Van Calster B, Vergote I, Van Huffel S, Valentin L. Simple ultrasound-based rules for the diagnosis of ovarian cancer. Ultrasound Obstet Gynecol 2008; 31: Kaijser J, Bourne T, Valentin L, Sayasneh A, Van Holsbeke C, Vergote I, Testa AC, Franchi D, Van Calster B, Timmerman D. Improving strategies for diagnosing ovarian cancer: a summary of the International Ovarian Tumor Analysis (IOTA) studies. Ultrasound Obstet Gynecol 2013; 4: Kaijser J, Sayasneh A,Van Hoorde K, Ghaem-Maghami S, Bourne T, Timmerman D, Van Calster B. Presurgical diagnosis of adnexal tumours using mathematical models and scoring systems: a systematic review and meta-analysis. Hum Reprod Update 2014; 20: Van Holsbeke C, Van Calster, Bourne T, Ajossa S, Testa AC, Guerriero S, Fruscio R, Lissoni AA, Czekierdowski A, Savelli L, Van Huffel S, Valentin L, Timmerman D. External validation of diagnostic models to estimate the risk of malignancy in adnexal masses. Clinical Cancer Res 2012; 18; Van Calster B, Van Hoorde K, Valentin L, Testa AC, Fischerova D, Van Holsbeke C, Savelli L, Franchi D, Epstein E, Kaijser J, Van Belle V, Czekierdowski A, Guerriero S, Fruscio R, Lanzani C, Scala F, Bourne T, Timmerman D; International Ovarian Tumour Analysis IOTA Group. Evaluating the risk of ovarian cancer before surgery using the ADNEX model to differentiate between benign, borderline, early and advanced stage invasive, and secondary metastatic tumours: prospective multicenter diagnostic study. BMJ 2014; 349: g Van Calster B, Van Hoorde K, Froyman W, Kaijser J, Wynants L, Landolfo C, Anthoulakis C, Vergote I, Bourne T, Timmerman D. Practical guidance for applying the ADNEX model from the IOTA group to discriminate between different subtypes of adnexal tumors. FactsViewsVisObgyn2015; 7: Sherman ME, Mink PJ, Curtis R, Cote TR, Brooks S, Hartge P, Devesa S. Survival among women with borderline ovarian tumors and ovarian carcinoma: a population-based analysis. Cancer 2004; 100: Tropé CG, Kaern J, Davidson B. Borderline ovarian tumours. Best Pract Res Clin Obstet Gynaecol 2012; 26: Zhang Y, Fan S, Xiang Y, Duan H, Sun L. Comparison of the prognosis and recurrence of apparent early-stage ovarian tumors treated with laparoscopy and laparotomy: a meta-analysis of clinical studies. BMC Cancer 2015; 15: Testa AC, Ferrandina G, Timmerman D, Savelli L, Ludovisi M, Van Holsbeke C, Malaggese M, Scambia G, Valentin L. Imaging in gynecological disease (1): ultrasound features of metastases in the ovaries differ depending on the origin of the primary tumor. Ultrasound Obstet Gynecol 2007; 29: Lee SJ, Bae JH, Lee AW, Tong SY, Park YG, Park JS. Clinical characteristics of metastatic tumors to the ovaries. J Korean Med Sci 2009; 24: Epstein E, Van Calster B, Timmerman D, Nikman S. Subjective ultrasound assessment, the ADNEX model and ultrasound-guided tru-cut biopsy to differentiate disseminated primary ovarian cancer from metastatic non-ovarian cancer. Ultrasound Obstet Gynecol 2016; 47: Prat J; FIGO Committee on Gynecologic Oncology. Staging classification for cancer of the ovary, fallopian tube, and peritoneum. Int J Gynaecol Obstet 2014; 124: Kurman RJ, Carcangiu ML, Herrington CS, Young RH. World Health Organization Classification of Tumours of Female Reproductive Organs (4 th edn). International Agency for Research on Cancer: Lyon, Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig L, Lijmer JG, Moher D, Rennie D, de Vet HC, Kressel HY, Rifai N, Golub RM, Altman DG, Hooft L, Korevaar DA, Cohen JF; STARD Group. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ 2015; 28: Timmerman D, Valentin L, Bourne TH, Collins WP, Verrelst H, Vergote I. Terms, definitions and measurements to describe the sonographic features of adnexal tumors: a consensus opinion from the International Ovarian Tumor Analysis (IOTA) Group. Ultrasound Obstet Gynecol 2000; 16: R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing: Vienna, Austria, [Accessed 28 January 2016]. 22. Valentin L, Ameye L, Jurkovic D, Metzger U, Lécuru F, Van Huffel S, Timmerman D. Which extrauterine pelvic masses are difficult to correctly classify as benign or malignant on the basis of ultrasound findings and is there a way of making a correct diagnosis? Ultrasound Obstet Gynecol 2006; 27: Sokalska A, Timmerman D, Testa AC, Van Holsbeke C, Lissoni AA, Leone FP, Jurkovic D, Valentin L. Diagnostic accuracy of transvaginal ultrasound examination for assigning a specific diagnosis to adnexal masses. Ultrasound Obstet Gynecol 2009; 34: Campos C, Sarian LO, Jales RM, Hartman C, Araujo KG, Pitta D, Yoshida A, Andrade L, Derchain S. Performance of the risk of malignancy index for discriminating malignant tumors in women with adnexal masses. J Ultrasound Med 2016; 35: Di Legge A, Testa AC, Ameye L, Van Calster B, Lissoni AA, Leone FP, Savelli L, Franchi D, Czekierdowski A, Trio D, Van Holsbeke C, Ferrazzi E, Scambia G, Timmerman D, Valentin L. Lesion size affects diagnostic performance of IOTA logistic regression models, IOTA simple rules and risk of malignancy index in discriminating between benign and malignant adnexal masses. Ultrasound Obstet Gynecol 2012; 40: Timmerman D, Van Calster B, Testa A, Savelli L, Fischerova D, Froyman W, Wynants L, Van Holsbeke C, Epstein E, Franchi D, Kaijser J, Czekierdowksi A, Guerriero S, Fruscio R, Leone FP, Rossi A, Landolfo C, Vergote I, Bourne T, Valentin L. Predicting the risk of malignancy in adnexal masses based on the Simple Rules from the International Ovarian Tumor Analysis (IOTA) group. Am J Obstet Gynecol 2016; 214: