M. J. KUDLA* and J. L. ALCÁZAR ABSTRACT

Ultrasound Obstet Gynecol 2010; 35: 602 608 Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/uog.7601 Does sphere volume affect the performance of three-dimensional power Doppler virtual vascular sampling for predicting malignancy in vascularized solid or cystic-solid adnexal masses? M. J. KUDLA* and J. L. ALCÁZAR *Department of Obstetrics and Gynecology, Medical University of Silesia, Katowice, Poland and Department of Obstetrics and Gynecology, University of Navarra, Pamplona, Spain KEYWORDS: adnexal mass; ovarian cancer; three-dimensional power Doppler ABSTRACT Objective To assess whether, when using spherical sampling with Virtual Organ Computer-Aided Analysis (VOCAL TM ) for calculating three-dimensional (3D) power Doppler angiography (PDA) indices, the sphere volume affects performance in the prediction of malignancy in vascularized cystic-solid or solid adnexal masses. Methods One hundred and thirty-eight women (mean ± SD age, 51.8 ± 14.1 years) diagnosed as having vascularized cystic-solid or solid adnexal masses on B-mode and two-dimensional (2D) power Doppler ultrasound were evaluated by 3D-PDA prior to surgery. Five women had bilateral masses, giving a total number of 143 masses analyzed. Vascularization was assessed using VOCAL software. 3D-PDA vascular indices (vascularization index (VI), flow index (FI) and vascularization flow index (VFI)) from the most vascularized area within papillary projections and solid areas were calculated automatically using spherical sampling. Five different volumes of sphere were used (1 cm 3,2cm 3,3cm 3,4cm 3 and 5 cm 3 )ineachcase.a definitive histological diagnosis was obtained in each case after surgical tumor removal. Results One hundred and seventeen (82%) masses were malignant and 26 (18%) were benign. Morphological evaluation revealed 34 (24%) unilocular solid masses, 49 (34%) multilocular solid masses and 60 (42%) mostly solid masses. The 1-cm 3 sphere could be used in 100% of the cases, the 2-cm 3 sphere could be used in 98.2% of the cases and the 3 5-cm 3 spheres could be used in 97.2% of the cases. The median VI, FI and VFI for all sphere volumes were significantly higher in malignant compared with non-malignant tumors. Receiver operating characteristics curve analysis showed that VI and VFI, independently of sphere volume, were better predictors of malignancy than was FI. The best cut-off values for the 3D-PDA indices differed depending on sphere volume. VI was significantly more specific than were VFI and FI. Conclusions Sphere volume does not affect the performance of 3D-PDA. We recommend the use of different cut-off values for 3D-PDA indices for discriminating between benign and malignant adnexal masses, depending on the sphere volume used. Use of VI is preferable due to its higher specificity. Copyright 2010 ISUOG. Published by John Wiley & Sons, Ltd. INTRODUCTION Three-dimensional (3D) power Doppler angiography (PDA) allows the objective assessment of tumor vascularization by analysis of power Doppler signals 1.Using Virtual Organ Computer-Aided AnaLysis (VOCAL TM ) software, three vascular indices (vascularization index (VI), flow index (FI) and vascularization flow index (VFI)) can be estimated from a given tissue volume 1. To date, two different methods for performing vascular sampling in ovarian tumors have been proposed. Alcázar et al. 2 suggested manual sampling by outlining vascularized solid areas within the tumor and calculating the vascular indices from the entire solid area. Jokubkiene et al. 3 proposed 5-cm 3 spherical sampling from the most Correspondence to: Dr J. L. Alcázar, Department of Obstetrics and Gynecology, Clinica Universitaria de Navarra, Avenida Pio XII, 36, 31008 Pamplona, Spain (e-mail: jlalcazar@unav.es) Accepted: 6 October 2009 Copyright 2010 ISUOG. Published by John Wiley & Sons, Ltd. ORIGINAL PAPER

Sphere volume in 3D-PDA 603 vascularized area of the tumor, selected subjectively by the examiner. This method is performed automatically by the VOCAL software and has been shown to save time compared with manual sampling when the examiner is less experienced 4. However, the sphere volume chosen by Jokubkiene et al. 3 was rather arbitrary. In fact, Kudla et al. 5 recently proposed the use of a smaller sphere volume (1 cm 3 ). To our knowledge there are no published data regarding which sphere volume should be used and it is not known whether this can affect the diagnostic performance of this method. The aim of our study, therefore, was to assess whether, when using spherical sampling with VOCAL for calculating 3D-PDA indices, the sphere volume affects performance in the prediction of malignancy in vascularized cystic-solid or solid adnexal masses. METHODS Stored 3D-PDA datasets collected prospectively from 138 women diagnosed on B-mode and two-dimensional (2D) power Doppler ultrasound as having centrally vascularized solid or cystic-solid adnexal masses were analyzed retrospectively. Thus, no tumor with a clear benign appearance on B-mode ultrasound, such as endometrioma, dermoid cyst, simple cyst, hemorrhagic cyst, paraovarian cyst or hydrosalpinx, or any tumor with solid areas or thick papillary projections but without vessels detected within these areas, was included in this study. Five women had bilateral masses, giving a total number of 143 masses analyzed. Datasets were obtained from two centers: the Department of Obstetrics and Gynecology, University of Navarra, Pamplona, Spain (n = 122) and the Department of Obstetrics and Gynecology, Medical University of Silesia, Katowice, Poland (n = 16). All women from the University of Navarra had been included in a previous study 6. The mean patient age was 51.7 (SD, 14.1; range, 18 82) years. Sixty-two (45%) women were premenopausal and seventy-six (55%) were postmenopausal. Data obtained after 3D-PDA examination were not used for the patients management decisions. Institutional review board approval for this study was obtained and all women gave verbal informed consent to participate. All ultrasound examinations were performed using a Voluson 730 Expert (GE Medical Systems, Milwaukee, WI, USA) machine equipped with a 5 7.5-MHz endovaginal probe and with color, power and pulsed Doppler as well as 3D ultrasound capabilities. The scanning technique has been described in detail previously 6.Briefly, on B-mode ultrasound a solid component of the mass was defined in the presence of thick papillary projections (> 3 mm), solid areas or mostly solid echogenicity. Lesions were classified as unilocular solid, multilocular solid and mostly solid (> 80% of tumor was solid). Following B-mode evaluation, the 2D power Doppler gate was activated to assess tumor vascularization. Power Doppler settings were adjusted to achieve maximum sensitivity to detect low velocity flow without noise (frequency, 5 MHz; power Doppler gain, 0.8 (range, 0.6 1.0); dynamic range, 20 40 db; edge, 1; persistence, 2; color map, 5; gate, 2; filter, L1; pulse repetition frequency, 0.6 khz). Central vascularization was defined in the presence of color spots within the suspicious areas of the tumor (thick papillary projections, solid areas or central part of solid tumors). 3D volumes were then obtained from these suspicious areas (from the whole tumor in the case of mostly solid tumors) and stored on the ultrasound machine s hard disk to be analyzed subsequently on a personal computer. All examinations were performed by one of the authors, each in his own center and using the same ultrasound machine model and settings. In all cases, on the day of ultrasound examination blood samples were collected to measure the CA-125 plasma concentration. Measurements were performed using an enzyme immunoassay with a monoclonal antibody (Cobas-Core CA-125 II, Laboratories Roche, Basel, Switzerland) and a sensitivity of < 5IU/mL. The intra- and interassay coefficients of variation were < 5.3% and < 7.5%, respectively. All 143 stored 3D volumes were analyzed by one of the authors (M.K.) using VOCAL software version 7.0.3 (GE Medical Systems). The most vascularized region within the solid area was selected subjectively and spherical sampling was performed (Figure 1), using the volume size software tool to select and calculate automatically the sphere volume. Five different concentric sphere volumes were used: 1 cm 3,2cm 3,3cm 3,4cm 3 and 5 cm 3, with an acceptable volume deviation of ± 0.05 cm 3. We have previously shown that reproducibility of this method is not affected by sphere volume 7. Following the ultrasound examination, all patients underwent surgery and definitive histological diagnosis was obtained in every case. Tumors were classified according to the World Health Organization (WHO) criteria 8 and ovarian cancers were staged according to International Federation of Gynecology and Obstetrics (FIGO) criteria 9. Statistical analysis The Kolmogorov Smirnov test was used to check whether the data were normally distributed, and showed that 3D- PDA indices had non-normal distributions. 3D power Doppler vascular indices according to sphere volume and histology were compared using the Mann Whitney U-test. Wilcoxon s rank test was used to compare 3D- PDA indices for each pair of sphere volumes. P < 0.05 was considered statistically significant for all tests. Receiver operating characteristics (ROC) curves were plotted for each vascular index and each sphere volume. ROC curves were compared using the Hanley method 10. Cut-offs for VI, VFI and FI for 1-cm 3 and 5-cm 3 sphere volumes, the volumes proposed by Kudla et al. 5 and Jokubkiene et al. 3, were selected according to the best sensitivity specificity relationship. and specificity as well as positive and negative likelihood ratios

604 Kudla and Alcázar were calculated for each index. and specificity were compared using McNemar s test. RESULTS Histology showed 117 (82%) of the adnexal masses to be malignant and twenty-six (18%) to be benign (Table 1). The diagnoses on B-mode ultrasound were: unilocular solid cyst, n = 34 (24%); multilocular solid cyst, n = 49 (34%); and mostly solid cyst, n = 60 (42%). All lesions had blood flow detected within thick papillary projections or solid areas, or in the central area in the case of solid tumors. The relationship between morphology and final diagnosis is summarized in Table 2. The median CA- 125 value was significantly higher in malignant compared with benign lesions (261 (range, 4 20 791) IU/mL vs. 29.3 (range, 1.3 161 IU/mL) (P < 0.001). The 1-cm 3 sphere could be used in 100% of cases, the 2-cm 3 sphere could be used in 98.2% of the cases and the 3-cm 3,4-cm 3 and 5-cm 3 spheres could be used in 97.2% of the cases. In cases in which the larger sphere volumes could not be used, this was due to the volume being greater than the vascularized area under analysis. The larger the sphere volumes were, the lower were the median values of 3D-PDA indices (Table 3). Median VI, FI and VFI for all sphere volumes were significantly higher in malignant compared with benign lesions (Table 4). (a) 1.0 0.8 0.6 0.4 0.2 (b) 1.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1 Specificity 0.8 0.6 0.4 0.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1 Specificity (c) 1.0 0.8 Figure 1 Spherical sampling with Virtual Organ Computer-aided AnaLysis from a solid adnexal mass using a 1-cm 3 sphere volume. Table 1 Final histological diagnosis of adnexal masses in the study group (n = 143) 0.6 0.4 Histology n % 0.2 Cystadenofibroma 4 2.8 Endometrioma 5 3.5 Teratoma 6 4.2 Ovarian fibroma 4 2.8 Fibrothecoma 1 0.7 Benign granulosa cell tumor 1 0.7 Hemorrhagic cyst 2 1.4 Tubo-ovarian abscess 3 2.1 Borderline tumor 7 4.9 Primary ovarian carcinoma 84 58.7 Metastatic carcinoma 26 18.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1 Specificity Figure 2 Receiver operating characteristics curves for vascularization index (a), vascularization flow index (b) and flow index (c) in the prediction of malignancy of adnexal masses, calculated using spherical sampling with Virtual Organ Computer-aided AnaLysis, using sphere volumes of 1 cm 3 ( ), 2cm 3 ( ), 3 cm 3 ( ), 4 cm 3 ( )and5cm 3 ( ).

Sphere volume in 3D-PDA 605 Table 2 Distribution of adnexal masses according to appearance on B-mode ultrasound Ultrasound appearance Histological diagnosis Benign (n = 26) Malignant (n = 117) Unilocular solid (n = 34) 13 (50) 21 (18) Multilocular solid (n = 49) 7 (27) 42 (36) Solid (n = 60) 6 (23) 54 (46) Values are n (%). Figure 2 shows ROC curve analysis for VI, VFI and FI for all sphere volumes analyzed. Areas under the curve for VI and VFI were significantly higher than those for FI (Table 5). The best index cut-offs for 1-cm 3 sphere volumes for discriminating between malignant and benign tumors were: VI = 24.015, VFI = 7.395 and FI = 33.731. The best cut-offs for 5-cm 3 sphere volumes were: VI = 10.490, VFI = 2.785 and FI = 31.505. The sensitivity, specificity and positive and negative likelihood ratios for VI, VFI and FI, obtained using 1cm 3 and 5 cm 3 sphere volumes, to discriminate between benign and malignant tumors are shown in Table 6. There were no differences in terms of sensitivity between VI, VFI and FI according to different sphere volumes. However, VI was significantly more specific compared with VFI and FI for both sphere volumes. DISCUSSION In 2005, Alcázar et al. 2 proposed the use of 3D-PDA as a new method for assessing tumor vascularization in the prediction of malignancy of solid and cystic-solid adnexal masses. This proposal was based on two facts: first, malignant ovarian tumors are known to have a higher microvessel density than do benign ones in immunohistochemical studies 11. Second, although it is not known exactly what 3D-PDA vascular indices reflect physiologically and hemodynamically 12, it is thought that they reflect tissue vascularization 1. Both the original publication and all papers subsequently published using this technique Table 4 Three-dimensional power Doppler angiography (3D-PDA) indices for each spherical sampling volume on Virtual Organ Computer-aided AnaLysis according to histological diagnosis of adnexal masses Histological diagnosis 3D-PDA index/ sphere volume Malignant Benign P VI (%) 1cm 3 45.920 15.975 < 0.001 (31.661 72.055) (9.132 27.384) 2cm 3 34.830 11.205 < 0.001 (24.143 59.095) (5.795 19.635) 3cm 3 29.772 9.113 < 0.001 (19.800 53.135) (4.980 17.211) 4cm 3 25.926 8.235 < 0.001 (16.735 48.745) (4.402 15.352) 5cm 3 22.940 7.705 < 0.001 (14.665 44.925) (4.122 14.350) FI 1cm 3 44.780 37.025 < 0.001 (38.160 51.445) (32.742 45.747) 2cm 3 42.903 35.812 < 0.001 (38.135 48.855) (32.655 44.270) 3cm 3 42.213 35.954 < 0.001 (36.795 47.890) (32.725 43.267) 4cm 3 41.497 35.840 < 0.001 (36.065 46.970) (32.700 42.622) 5cm 3 40.990 35.515 < 0.001 (35.921 46.725) (32.052 41.892) VFI 1cm 3 19.764 5.340 < 0.001 (13.185 36.150) (2.970 12.125) 2cm 3 14.957 3.653 < 0.001 (9.585 28.010) (1.835 8.002) 3cm 3 12.290 3.125 < 0.001 (7.915 24.615) (1.645 6.482) 4cm 3 10.460 2.820 < 0.001 (6.435 21.640) (1.477 5.685) 5cm 3 10.001 2.535 < 0.001 (5.685 19.545) (1.552 5.291) Values are median (25 th 75 th quartile). FI, flow index; VFI, vascularization flow index; VI, vascularization index. have reported similar results: malignant tumors exhibit significantly higher 3D-PDA vascular indices than do benign ones 2,3,5,6,13 15. Table 3 Three-dimensional power Doppler angiography (3D-PDA) indices according to spherical sampling volume on Virtual Organ Computer-aided AnaLysis of adnexal masses Spherical sampling volume 3D-PDA index 1 cm 3 2cm 3 3cm 3 4cm 3 5cm 3 P* VI (%) 41.720 29.992 25.750 23.120 21.173 < 0.001 (25.640 67.992) (20.010 56.658) (15.661 46.900) (13.902 44.340) (11.940 40.460) FI 43.002 42.073 41.180 40.012 40.103 < 0.001 (37.050 50.890) (36.430 48.263) (34.945 46.750) (34.572 46.291) (34.740 45.580) VFI 17.920 12.790 10.916 9.345 8.481 < 0.001 (10.550 35.172) (7.780 26.486) (5.965 21.865) (5.383 18.406) (5.030 16.431) Values are median (25 th 75 th quartile). *For all comparisons between different sphere volume pairs for each index. FI, flow index; VFI, vascularization flow index; VI, vascularization index.

606 Kudla and Alcázar Table 5 Areas under the receiver operating characteristics curve (AUC) for three-dimensional power Doppler angiography (3D-PDA) indices and each spherical sampling volume on Virtual Organ Computer-aided AnaLysis of adnexal masses 95% CI 3D-PDA index/ sphere volume AUC Upper Lower VI 1cm 3 0.854 0.758 0.949 2cm 3 0.856 0.761 0.951 3cm 3 0.852 0.757 0.946 4cm 3 0.848 0.753 0.944 5cm 3 0.842 0.745 0.938 FI* 1cm 3 0.712 0.595 0.828 2cm 3 0.711 0.599 0.822 3cm 3 0.697 0.586 0.807 4cm 3 0.690 0.579 0.800 5cm 3 0.681 0.569 0.793 VFI 1cm 3 0.851 0.759 0.942 2cm 3 0.852 0.760 0.944 3cm 3 0.850 0.759 0.941 4cm 3 0.847 0.753 0.941 5cm 3 0.844 0.750 0.938 *P < 0.001 for all FI-AUCs vs. all VI- and VFI-AUCs, comparing sphere volume pairs. There were no differences among VI- and VFI-AUCs. FI, flow index; VFI, vascularization flow index; VI, vascularization index. However, certain methodological questions about this technique remain unsolved. Some of these questions relate to machine settings, the relevance of which was pointed out recently by Raine-Fenning et al. 16. Others relate to how the calculations should be performed; Alcazar et al. proposed using the software s manual mode 2, while Jokubkiene et al. 3 proposed using automatic 5-cm 3 spherical sampling and Kudla et al. 5 proposed using 1-cm 3 spherical sampling. Spherical sampling has the advantage of being time-saving, because the software performs the calculation automatically and avoids the rotational approach of manual sampling 4.Furthermore, while in older versions of the software, the examiner had to position the calipers on the screen to obtain the desired volume, the newer version of the software allows the examiner to select automatically the volume desired, making the process even easier and quicker, and the calculation more reliable. However, to date, there are no published data regarding which sphere volume should be used and it is not known whether this choice can affect the diagnostic performance of the method. In this study we found that diagnostic performance was not affected by sphere volume. However, the larger was the sphere volume, the lower were the VI, VFI and FI values. The first two findings were not surprising: VI represents the proportion of color voxels present within the analyzed sphere volume, being the quotient of the number of color voxels and the total number of voxels contained in this volume, and VFI represents a combination of the proportion and the mean signal intensity of the color voxels in the analyzed volume, being the quotient of the number of weighted color values of the voxels and the total number of voxels contained in this volume. However, the finding that FI, which represents the mean signal intensity of the voxels in the analyzed volume, was lower with increasing sphere volume is rather surprising, because, theoretically, it does not depend on the volume analyzed 1,17. We also found that VI was a better predictor for malignancy than were VFI and FI. Overall, VI yielded significantly fewer false-positive cases than did VFI and FI, and this was unaffected by sphere volume. Table 6, specificity, positive likelihood ratio (LR+) and negative likelihood ratio (LR ) for three-dimensional power Doppler angiography (3D-PDA) indices and spherical sampling volumes on Virtual Organ Computer-aided AnaLysis of 1 cm 3 and5cm 3 in the prediction of adnexal mass malignancy 3D-PDA index/ sphere volume (% (95% CI)) Specificity* (% (95% CI)) LR + (95% CI) LR (95% CI) VI 1cm 3 91 77 3.96 0.10 (84 95) (58 89) (1.94 7.94) (0.07 0.22) 5cm 3 91 77 3.96 0.10 (84 95) (58 89) (1.94 7.94) (0.07 0.22) VFI 1cm 3 93 61.5 2.42 0.11 (87 96) (42 77) (1.49 3.95) (0.05 0.23) 5cm 3 95 61.5 2.71 0.09 (88 97) (46 88) (1.60 4.61) (0.04 0.20) FI 1cm 3 91 34.6 1.39 0.27 (84 95) (19 53) (1.04 1.84) (0.13 0.59) 5cm 3 90 34.6 1.39 0.27 (84 95) (19 53) (1.04 1.84) (0.13 0.59) *McNemar test: VI-1 cm 3 vs. VFI-1 cm 3, P = 0.016; VI-1cm 3 vs. FI-1 cm 3, P = 0.043; VI-5cm 3 vs. VFI-5 cm 3, P = 0.004; VI-5cm 3 vs. FI-5 cm 3, P = 0.046. FI, flow index; VFI, vascularization flow index; VI, vascularization index.

Sphere volume in 3D-PDA 607 Earlier studies 3,13 15, based on analysis of the whole mass of the ovary/tumor, generally found VI to be 20%. The values obtained in our study were higher, especially for benign lesions. This could be explained by our use of spherical sampling. By definition, this method, which preselects vascularized areas rather than calculating indices across both vascularized and non-vascularized areas of tissue, could lead to increased VI values. Taking this logic one step further, using smaller rather than larger sphere volumes could have a similar effect and be a potential source of false-positive cases, particularly in certain circumstances such as in benign tumors with scanty vascularization, for example fibromas or cystadenofibromas. Thus, it might seem logical, bearing in mind the fact that diagnostic performance was not affected by sphere volume, to use 5-cm 3 rather than smaller volume spherical sampling in most cases. However, our results confirm previous data that showed that 5-cm 3 spherical sampling cannot be used in all cases 4.Infact, only the 1-cm 3 volume could be applied successfully in all cases, making this sphere volume preferable for small solid tumors or tumors with small solid components in particular. Another limitation for the use of spherical sampling is when the most vascularized area of the tumor is close to the tumor border. In this case, extratumoral vessels such as an iliac vessel might be included within the spherical sampling volume, affecting the calculation of 3D power Doppler vascular indices. Care should be taken in these cases, particularly when using 5-cm 3 spherical sampling in small tumors. Ovarian tumors are easily characterized subjectively by examining their morphological and vascular features on B-mode and color Doppler imaging. In our opinion, most sonographers would consider ovarian cysts such as those analyzed in this study, i.e. with solid components and flow within them, as being highly suspicious for malignancy. However, about 15% of these tumors are benign 18,and the question is, how can these be identified? The IOTA group has put considerable effort into standardizing the description of adnexal masses 19. They use the color score, based on the examiner s subjective impression regarding the amount of flow present within the mass, in the assessment of malignancy. One of their most recently published papers presented data on 1066 adnexal masses 20, subclassifying tumors with solid components according to their color score into two groups: those with a color score of 1 2 and those with a score of 3 4, where a color score of 1 indicates no flow. If we were to group together their unilocular solid, multilocular solid and solid tumors with color scores of 3 4 (their cases that most closely resemble our present cases), the sensitivity for diagnosing ovarian cancer would be 74.5% and the specificity 82.5%. Including cases with a color score of 2, to give a group with color score of 2 4, would probably increase the sensitivity but also the falsepositive rate. We believe that including in this assessment a quantification of flow within the tumor by means of 3D- PDA vascular indices could increase both the sensitivity and particularly the specificity: using the conventional approach all histologically benign masses with centrally vascularized solid or cystic-solid appearance on B-mode and 2D power Doppler analysis would be classifed as malignant (100% false positives), while, according to our data, the use of 3D spherical sampling could reduce the number of false-positive cases by 77%, for a sensitivity of 91% (Table 6). In conclusion, based on our results we would recommend the use of different cut-off values for 3D-PDA indices for discriminating between benign and malignant adnexal masses, depending on the spherical sampling volume. Since VI from the most vascularized area of the tumor was more specific than were FI and VFI for predicting malignancy in solid and cystic-solid adnexal masses, this index should be used in preference. Since there was no difference in the diagnostic performance for VI between 1-cm 3 and 5-cm 3 sample volumes, both can be used. However, we would recommend the 1-cm 3 sphere volume since it can be used in all tumors, even small ones. REFERENCES 1. Pairleitner H, Steiner H, Hasenoehrl G, Staudach A. Threedimensional power Doppler sonography: imaging and quantifying blood flow and vascularization. Ultrasound Obstet Gynecol 1999; 14: 139 143. 2. Alcazar JL, Merce LT, Garcia Manero M. Three-dimensional power Doppler vascular sampling: a new method for predicting ovarian cancer in vascularized complex adnexal masses. J Ultrasound Med 2005; 24: 689 696. 3. Jokubkiene L, Sladkevicius P, Valentin L. Does threedimensional power Doppler ultrasound help in discrimination between benign and malignant ovarian masses? Ultrasound Obstet Gynecol 2007; 29: 215 225. 4. Alcazar JL, Prka M. Evaluation of two different methods for vascular sampling by three-dimensional power Doppler angiography in solid and cystic-solid adnexal masses. Ultrasound Obstet Gynecol 2009; 33: 349 354. 5. Kudla MJ, Timor-Tritsch IE, Hope JM, Monteagudo A, Popiolek D, Monda S, Lee CJ, Arslan AA. Spherical tissue sampling in 3-dimensional power Doppler angiography: a new approach for evaluation of ovarian tumors. J Ultrasound Med 2008; 27: 425 433. 6. Alcazar JL, Rodriguez D. Three-dimensional power Doppler vascular sonographic sampling for predicting ovarian cancer in cystic-solid and solid vascularized masses. J Ultrasound Med 2009; 28: 275 281. 7. Kudla MJ, Alcazar JL. Does the size of three-dimensional power Doppler spherical sampling affect the interobserver reproducibility of measurements of vascular indexes in adnexal masses? Ultrasound Obstet Gynecol 2009; 34: 732 734. 8. Serov SF, Scully RE in collaboration with Sobin LH. Histological typing of ovarian tumors. International Histological Classification of Tumors. No 9. World Health Organization: Geneva, 1973. 9. Sheperd JH. Revised FIGO staging for gynecological cancer. Br J Obstet Gynaecol 1989; 96: 889 892. 10. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983; 148: 839 843. 11. Orre M, Lotfi-Miri M, Mamers P, Rogers PA. Increased microvessel density in mucinous compared with malignant serous and benign tumours of the ovary. Br J Cancer 1998; 77: 2204 2209.

608 Kudla and Alcázar 12. Alcázar JL. Three-dimensional power Doppler derived vascular indices: what are we measuring and how are we doing it? Ultrasound Obstet Gynecol 2008; 32: 485 487. 13. Geomini PM, Kluivers KB, Moret E, Bremer GL, Kruitwagen RF, Mol BW. Evaluation of adnexal masses with threedimensional ultrasonography. Obstet Gynecol 2006; 108: 1167 1175. 14. Czekierdowski A, Stachowicz N, Smolen A, Kotarski J. The use of 2-dimensional and 3-dimensional color and power Doppler in the diagnosis of blood flow indices of adnexal tumors. Ginekol Pol 2006; 77: 296 306. 15. Geomini PM, Coppus SF, Kluivers KB, Bremer GL, Kruitwagen RF, Mol BW. Is three-dimensional ultrasonography of additional value in the assessment of adnexal masses? Gynecol Oncol 2007; 106: 153 159. 16. Raine-Fenning NJ, Nordin NM, Ramnarine KV, Campbell BK, Clewes JS, Perkins A, Johnson IR. Evaluation of the effect of machine settings on quantitative three-dimensional power Doppler angiography: an in-vitro flow phantom experiment. Ultrasound Obstet Gynecol 2008; 32: 551 559. 17. Kudla M, Alcazar JL. 3DPD imaging of ovarian pathology: advantages and limitations of the method. How can we standardise the results? Donald School J Ultrasound Obstet Gynecol 2009; 3: 47 53. 18. Alcázar JL, Royo P, Jurado M, Mínguez JA, García-Manero M, Laparte C, Galván R, López-García G. Triage for surgical management of ovarian tumors in asymptomatic women: assessment of an ultrasound-based scoring system. Ultrasound Obstet Gynecol 2008; 32: 220 225. 19. Timmerman D, Valentin L, Bourne TH, Collins WP, Verrelst H, Vergote I; International Ovarian Tumor Analysis (IOTA) Group. 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: 500 505. 20. 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: 681 690.