Technical Advance
Three-Dimensional Power Doppler Vascular Sampling A New Method for Predicting Ovarian Cancer in Vascularized Complex Adnexal Masses Juan Luis Alcázar, MD, Luis T. Mercé, MD, Manuel García Manero, MD Objective. The purpose of this study was to explore the role of a new concept (“vascular sampling”) as a third step to discriminate benign and malignant lesions in B-mode and color Doppler sonographically suggestive adnexal masses. Methods. Forty-five women (mean age, 52.3 years; range, 17–82 years) with the diagnosis of complex adnexal masses on B-mode sonography were evaluated using 3dimensional power Doppler sonography. Four women had bilateral masses. After a morphologic reevaluation was done, color pulsed Doppler sonography was used to obtain flow velocity waveforms, and velocimetric indices were calculated (resistive index, pulsatility index, and peak systolic velocity). Thereafter, 3-dimensional power Doppler sonography was used to assess vascularization of highly suggestive areas (gross papillary projections, solid areas, and thick septations), meaning a focused assessment (“sampling”) of a suggestive area of the tumor. With a virtual organ computer-aided analysis program, vascular indices (vascularization index, flow index, and vascular flow index) were automatically calculated. A definitive histologic diagnosis was obtained in each case. Results. Forty masses (82%) were malignant and 9 (18%) were benign. Morphologic evaluation revealed 10 (20%) unilocular solid masses, 20 (41%) multilocular solid masses, and 19 (39%) mostly solid masses. Blood flow was found in all cases. Median vascularization index (15.5% versus 8.2%; P = .002), flow index (33.6 versus 20.8; P = .007), and vascular flow index (5.2 versus 2.3; P = .001) were significantly higher in malignant tumors. No differences were found in resistive index (0.43 versus 0.45; P = .770), pulsatility index (0.62 versus 0.65; P = .694), and peak systolic velocity (15.6 versus 12 cm/s; P = .162). Conclusions. Threedimensional power Doppler vascular sampling seems to be a promising tool for predicting ovarian cancer in vascularized complex adnexal masses. It could be better than conventional color pulsed Doppler imaging. Key words: ovarian cancer; power Doppler sonography; 3-dimensional sonography. Abbreviations AUC, area under the curve; FI, flow index; PI, pulsatility index; PSV, peak systolic velocity; RI, resistive index; 3D, 3-dimensional; 2D, 2-dimensional; VFI, vascular flow index; VI, vascularization index; VOCAL, Virtual Organ Computer-Aided Analysis Received October 25, 2004, from the Department of Obstetrics and Gynecology, Clínica Universitaria de Navarra, School of Medicine, University of Navarra, Pamplona, Spain (J.L.A., M.G.M.); and Department of Obstetrics and Gynecology, Hospital Ruber Internacional, Madrid, Spain. (L.T.M.). Revision requested December 26, 2004. Revised manuscript accepted for publication December 30, 2004. Address correspondence to Juan Luis Alcázar, MD, Department of Obstetrics and Gynecology, Clínica Universitaria de Navarra, Avenida Pio XII 36, 31008 Pamplona, Spain. E-mail: [email protected]
T
he introduction of transvaginal sonography has greatly improved the ability to discriminate benign and malignant ovarian tumors.1 Although this technique has proved to be precise enough in the specific diagnosis of some benign ovarian lesions,2–4 it still has the problem of a relatively high false-positive rate in complex masses.5 Transvaginal color and pulsed Doppler sonography allows the assessment of tumor vascularization and was introduced into clinical practice in an attempt to decrease the false-positive rate of morphologic sonographic evaluation of adnexal masses. Although initial results, based on the use of velocimetric indices, were encouraging,6 the literature is full of conflicting results,7–9 which make this approach unpractical and poorly reproducible.
© 2005 by the American Institute of Ultrasound in Medicine • J Ultrasound Med 2005; 24:689–696 • 0278-4297/05/$3.50
Three-Dimensional Power Doppler Vascular Sampling
A simpler approach based on morphologic sonographic assessment and tumor vessel distribution using color or power Doppler sonography has been proposed.10 This approach has proved very useful for discriminating benign from malignant adnexal tumors because the probability of malignancy is high in complex adnexal masses with central vessel distribution, whereas complex masses without vessels or peripheral blood flow are usually benign.10 However, there are still some benign tumors that show a complex appearance with central vessel distribution such as cystadenofibromas, mucinous cystadenomas, tubo-ovarian abscesses, and solid ovarian benign tumors such as Brenner tumors, granulosa cell tumors, and ovarian fibromas. These tumors are actually very difficult to discriminate from ovarian malignancies. Recently, a new method to assess tumor vascularization was developed: 3-dimensional (3D) power Doppler sonography.11 The aim of this study was to evaluate the potential role of 3D power Doppler sonography as a new method in the evaluation of suggestive vascularized adnexal masses (complex morphologic appearance with central vessel distribution on conventional color Doppler transvaginal sonography) for discriminating benign from malignant lesions.
Materials and Methods From August 2002 to August 2004, 45 women with the diagnosis of complex adnexal mass with central vessel distribution were evaluated by transvaginal 3D power Doppler sonography. Patients’ mean age was 52.3 years (SD, 16.2 years; range, 17–82 years). Twenty-four patients (53%) were postmenopausal, and 21 (47%) were premenopausal. Four women had bilateral masses, giving a total number of 49 masses evaluated. All women had been evaluated first by transvaginal sonography with a Kretz Sonoace 9900 system (GE Medical Systems, Kretztechnik, Zipf, Austria) or a Voluson 730 Pro system (GE Medical Systems, Milwaukee, WI) with a 5- to 7.5MHz endovaginal probe and color, power, and pulsed Doppler as well as 3D sonographic capabilities. Transabdominal sonography (3.5–5 MHz) was also performed in large tumors. The scanning techniques have been described in detail elsewhere.12 On B-mode sonography, a complex mass was defined in the presence of thick papillary projec690
tions (>3 mm), solid areas, or mostly solid echogenicity (Figures 1 and 2). According to Granberg et al,1 lesions were classified as unilocular solid, multilocular solid, and mostly solid. After B-mode evaluation was done, the 2dimensional (2D) power Doppler gate was activated to assess tumor vascularization. Power Doppler settings were set to achieve maximum sensitivity to detect low-velocity flow without noise (frequency, 5 MHz; power Doppler gain, 20 [range, 1–30]; dynamic range, 20–40 dB; edge, 1; persistence, 2; color map, 1; gate, 2; filter, 3; and pulse repetition frequency, 0.6 kHz). Central vessel vascularization was defined in the presence of color spots within the suggestive areas of the tumor (thick papillary projections, solid areas, or central part of solid tumors; Figures 3 and 4).10 Pulsed Doppler sonography was used to interrogate color spots identified to obtain a flow velocity waveform and to confirm the arterial nature of the vessel. The pulsatility index (PI), resistive index (RI), and peak systolic velocity (PSV, centimeters per second) were recorded automatically. The lowest PI and RI and the highest PSV found in a given tumor were used for analysis. Then, a 3D volume was activated to obtain a 3D box from those suggestive areas, for example, thick papillary projections, a solid area, or, in the case of a mostly solid tumor, the whole tumor. Once a 3D volume was obtained, it was stored on a hard disk (Sonoview; GE Medical Systems, Kretztechnik). The volume acquisition time lasted from 2 to 6 seconds depending on the size of the volume box. Using the Virtual Organ Computer-Aided Analysis (VOCAL) program, we calculated 3 vascular indices of those Figure 1. Transvaginal sonogram from a mostly solid adnexal mass that corresponded to a primary serous ovarian carcinoma.
J Ultrasound Med 2005; 24:689–696
Alcázar et al
suggestive areas, excluding cystic areas in which there was no tissue and focusing only on “solid” areas (Figures 5–8). A 3D volume is composed of voxels (smallest unit of volume). Voxels contain all the information about gray scale and color according to an intensity scale ranging from 0 to 100. According to these values, this measurement system obtains the mean grayness and power Doppler indices to evaluate vessels and blood flow.13 The vascular indices calculated were as follows: The vascularization index (VI), expressed as a percentage, measures the number of color voxels in the studied volume, representing the blood vessels within the tissue; the flow index (FI) is the average color value of all color voxels, representing the average color intensity; and the vascular flow index (VFI) is the average color value of all gray and color voxels, which represents both blood flow and vascularization and might be regarded as “perfusion.” Measurements were undertaken with the use of the manual mode in plane A and a 9° rotation step. If there were more than 1 noncontiguous solid area, each was evaluated separately. In those tumors with more than 1 suggestive area, the highest VI, FI, and VFI values found were used for analysis. If the tumor showed areas of low echogenicity, these areas were not included in the VOCAL calculation. The same author (J.L.A.) performed all examinations. Intraobserver variability was estimated by performance of 2 different measures 1 week apart14 in the first 10 cases evaluated and calculation of the intraclass correlation coefficient (0.91). As indicated above, only complex adnexal masses with central vessel distribution were Figure 2. Transvaginal sonogram from a unilocular solid adnexal mass in which histologic examination revealed metastatic endometrial cancer to the ovary.
J Ultrasound Med 2005; 24:689–696
Figure 3. Transvaginal power Doppler sonogram from the case in Figure 1. Central vessel distribution is shown within solid areas.
included in this study. This means that no tumor with a clear benign appearance, such as simple cyst, endometrioma, mature teratoma, hydrosalpinx, or peritoneal cyst, even having central flow, or with solid areas or thick papillary projections but without vessels detected within these areas was included in this study. Patients were not selected except for sonographic findings and the availability of 3D sonography. All patients gave oral informed consent after the nature of the study was fully explained. Institutional Review Board approval was also obtained. All patients underwent surgery, and a definitive histologic diagnosis was obtained in every case. Tumors were classified according to World Health Organization criteria,15 and ovarian cancers were staged according to International Federation of Gynecology and Obstetrics criteria.16 Figure 4. Transvaginal power Doppler sonogram from the case in Figure 2. As in the previous case, central vessels are shown within solid areas.
691
Three-Dimensional Power Doppler Vascular Sampling
Results
Figure 5. Three-dimensional power Doppler sonogram from the case in Figure 1 showing how only solid portions of the tumor were analyzed by the VOCAL method.
Continuous variables were compared with the use of the Mann-Whitney U test. P < .05 was considered statistically significant for all tests. The SPSS version 11.0 for Windows statistical package was used (SPSS Inc, Chicago, IL). Receiver operating characteristic curves were plotted for each vascular index to assess its diagnostic performance.
Figure 6. Three-dimensional power-Doppler sonogram from the case in Figure 2 showing how only solid portions of the tumor were analyzed by the VOCAL method. Low-level echoes are disregarded.
692
After surgical removal, 40 (82%) of the masses proved to be malignant and 9 (18%) proved to be benign (Table 1). Primary ovarian cancer stages were as follows: stage Ia, 8 (32%); stage IIb, 2 (8%); stage IIIa, 1 (4%); stage IIIb, 2 (8%); stage IIIc, 8 (32%); and stage IV, 4 (16%). On B-mode sonography, adnexal mass appearances were as follows: unilocular solid cyst, 10 (20%); multilocular solid, 20 (41%); and mostly solid, 19 (39%). As stated, all lesions had blood flow detected within thick papillary projections, solid areas, or central areas of solid tumors. According to histologic findings, 8 (80%) of 10 unilocular solid masses were malignant; 14 (70%) of 20 multilocular solid tumors were malignant; and 18 (95%) of 19 solid tumors were malignant. Median VI, FI, and VFI were significantly higher in malignant tumors. No differences were found in median RI, PI, and PSV between benign and malignant tumors (Table 2). Receiver operating characteristic curves for VI, FI, and VFI are shown in Figures 9–11.
Discussion In this study, we explored the potential role of a new approach using 3D power Doppler sonography for predicting ovarian malignancy in complex adnexal masses that show central vessel distribution. We have termed this new approach “3D power Doppler vascular sampling.” It consists of assessing the vascularization of a given suggestive area in a given tumor. We have focused the use of this new technique on complex adnexal masses with central flow because we think that it will be useless in simple cystic masses or “complex” masses without flow, given that the probability of malignancy in the latter is known to be low.10,17,18 When we face a complex mass with detectable blood flow with solid areas or thick papillary projections, we should categorized this mass as “malignant” or “very highly suggestive.” However, a considerable number of benign tumors may have this appearance, for example, cystadenofibromas,19 tubo-ovarian abscesses,20 and solid benign ovarian tumors.21 Using conventional 2D color or power Doppler sonography, we have no means to differentiate these benign entities from true malignant tumors. Three-dimensional power Doppler J Ultrasound Med 2005; 24:689–696
Alcázar et al
sonography provides a new approach to assessing tumor vascularization. It has been shown that this technique is highly reliable with a very low intraobserver and interobserver variability.22–24 Therefore, this technique overcomes one of the main problems of conventional color Doppler sonography25 and provides an objective measure of vascularity. To our knowledge, the only previous study analyzing 3D power Doppler vascular indices in adnexal masses included only 6 surgically removed adnexal tumors, all benign.13 The authors did not report specific data on VI, FI, and VFI values. Therefore, no comparison can be made. However, there are several studies that have assessed the usefulness of 3D power Doppler sonography with a different approach than ours. Kurjak et al11 and Kupesic and Kurjak,26 from Zagreb, Croatia, compared the diagnostic performance of 3D power Doppler and conventional color Doppler sonography in different studies. They found that 3D power Doppler sonography was more sensitive than conventional color Doppler sonography,11 especially when a sonographic contrast agent was used.26 They based their diagnostic criteria for malignancy suggestion on vessel architecture as depicted by 3D sonography, such as branching pattern, vessel caliber, and presence of microaneurysms or vascular lakes. This was based on the chaos theory,27 which established that the architecture of a vascular network of newly formed vessels in malignant tumors is built in a chaotic distribution and not in a predetermined fashion. Although, this has been shown in corrosion studies,28 we think that the reproducibility of this approach would be low or at least difficult to achieve because basically it is based on a subjective analysis of a 3D reconstructed image. In fact, to the best of our knowledge, no further study has been published that tried to reproduce the excellent results reported by this group using this technique. Cohen and coworkers29 reported that the use of 3D power Doppler sonography in complex masses would reduce the false-positive rate of Bmode sonography. However, they used as a criterion for malignancy the presence of vessels in a central location, which is the same proposed for 2D sonography.10 Therefore, we wonder, as did Guerriero et al,30 whether this approach is superior to conventional 2D color or power Doppler J Ultrasound Med 2005; 24:689–696
Figure 7. Histogram showing VI, FI, and VFI values for solid areas of the case in Figure 1.
Figure 8. Histogram showing VI, FI, and VFI values for the case in Figure 2.
Table 1. Histologic Distribution of Adnexal Masses Histologic Finding
Mucinous cystadenoma Cystadenofibroma Tubo-ovarian abscess Low-malignant-potential tumor Primary carcinoma Metastatic carcinoma
n
%
3 3 3 3 25 12
6.1 6.1 6.1 6.1 53.0 24.6
693
Three-Dimensional Power Doppler Vascular Sampling
Table 2. Vascular and Velocimetric Indices in Malignant and Benign Adnexal Tumors Type
Malignant (n = 40) Benign (n = 9)
VI, %
15.5 (IQR, 8.2) (7.1–77) 8.2 (IQR, 11.5) (0.2–21)
FI
33.6 (IQR, 8.4) (16.2–51.9) 20.8 (IQR, 16) (12.6–36.9)
VFI
RI
5.2 0.43 (IQR, 3.2) (IQR, 0.15) (2.1–40.1) (0.23–0.74) 2.3 0.45 (IQR, 2.3) (IQR, 0.11) (0.04–11.5) (0.30–0.52)
PI
PSV, cm/s
0.62 (IQR, 0.2) (0.30–1.48) 0.65 (IQR, 0.3) (0.44–0.90)
15.6 (IQR, 9.8) (7.0–30.0) 12.0 (IQR, 4.8) (7.5–31.7)
Data are expressed as median with range in parentheses. IQR indicates interquartile range. VI, P = .002; FI, P = .007; VFI, P = .0001; RI, P = .770; PI, P = .694; PSV, P = .162.
694
sonography. In fact, a recent study from our group comparing 3D power Doppler and 2D power Doppler sonography, with the use of the same criteria for malignancy suggestion for both techniques, found that 3D power Doppler sonography was not superior to 2D power Doppler sonography for discriminating benign from malignant complex adnexal masses.31 The results of this study, although preliminary because of a relatively small series, indicate that 3D power Doppler vascular sampling may be a new tool for discriminating benign from malignant complex adnexal masses with central vessel distribution. We found that vascular indices (VI, FI, and VFI) from suggestive areas such as thick papillary projections, solid areas, and mostly solid tumors were significantly higher in malignant tumors compared with benign tumors, which may reflect different angiogenic patterns. Conversely, it has been shown that vas-
cularity as assessed by immunohistochemical microvessel density quantification correlates with both 2D power Doppler and 3D power Doppler findings,32,33 which would validate the findings obtained in vivo with 3D power Doppler technology. One interesting finding in our study is that 40% of ovarian cancers were stage II or I. This raises a question about the role of this technique for the early diagnosis of ovarian cancer. In fact, Kurjak et al34 were able to identify 97.7% of stage I ovarian cancers (42 of 43) in their study using this technique. Notwithstanding, we are aware that some potential sources of error could exist with this approach. One of them might be large vascularized tumors because it may be difficult to include the whole tumor in only a single 3D box; therefore, one should select a single area or use more than a single box for analysis. This may
Figure 9. Receiver operating characteristic curve for VI (area under the curve [AUC], 0.88; SE, 0.07).
Figure 10. Receiver operating characteristic curve for FI (AUC, 0.87; SE, 0.06).
J Ultrasound Med 2005; 24:689–696
Alcázar et al
introduce a subjective factor. We think that the simplest way is to use the area where the highest vascular indices are found. However, the best method for assessing these large tumors by using this technique needs to be established. Conversely, experience may be needed to appropriately select suggestive areas. However, this question needs to be evaluated. In fact, to the best of our knowledge, no study has formally assessed the learning curve for 3D sonography. We evaluated the diagnostic performance of this new approach in our series, and we found that each vascular index has acceptable diagnostic performance. However, we have to bear in mind that specificity and the negative predictive value are difficult to calculate precisely because the number of benign tumors, as could be expected, was low. Finally, we assessed only the intraobserver reproducibility, which might be another limitation in our study. Interobserver reproducibility must be assessed in future studies. In conclusion, we have defined and presented a new approach for predicting ovarian cancer in complex adnexal masses with central blood flow detected on conventional color or power Doppler transvaginal sonography. This new approach, termed 3D power Doppler vascular sampling, is based on 3D sonography and in our own experience seems to be reproducible. Studies in larger series are needed to establish its actual diagnostic performance.
References 1.
Granberg S, Norstrom A, Wikland M. Tumors of the lower pelvis as imaged by vaginal sonography. Gynecol Oncol 1990; 37:224–229.
2.
Alcázar JL, Laparte C, Jurado M, López-García G. The role of transvaginal ultrasonography combined with color velocity imaging and pulsed Doppler in the diagnosis of endometrioma. Fertil Steril 1997; 67:487– 491.
3.
Caspi B, Appelman Z, Rabinerson D, Elchalal U, Zalel Y, Katz Z. Pathognomonic echo patterns of benign cystic teratomas of the ovary: classification, incidence and accuracy rate of sonographic diagnosis. Ultrasound Obstet Gynecol 1996; 7:275–279.
4.
Guerriero S, Ajossa S, Lai MP, Mais V, Paoletti AM, Melis GB. Transvaginal ultrasonography associated with colour Doppler energy in the diagnosis of hydrosalpinx. Hum Reprod 2000; 15:1568–1572.
5.
Ferrazzi E, Zanetta G, Dordoni D, et al. Transvaginal ultrasonographic characterization of ovarian masses: comparison of five scoring systems in a multicenter study. Ultrasound Obstet Gynecol 1997; 10:192–197.
6.
Kurjak A, Zalud I, Alfirevic Z. Evaluation of adnexal masses with transvaginal color Doppler ultrasound. J Ultrasound Med 1991; 10:295–297.
7.
Kawai M, Kano T, Kikkawa F, Maeda O, Oguchi H, Tomoda Y. Transvaginal Doppler ultrasound with color flow imaging in the diagnosis of ovarian cancer. Obstet Gynecol 1992; 79:163–167.
8.
Tekay A, Jouppila P. Validity of pulsatility and resistance indices in classification of adnexal tumors with transvaginal color Doppler ultrasound. Ultrasound Obstet Gynecol 1992; 2:338–344.
9.
Levine D, Feldstein VA, Babcook CJ, et al. Sonography of ovarian masses: poor sensitivity of resistive index for identifying malignant lesions. AJR Am J Roentgenol 1994; 162:1355–1359.
Figure 11. Receiver operating characteristic curve for VFI (AUC, 0.86; SE, 0.1).
10. Guerriero S, Alcázar JL, Coccia ME, et al. Complex pelvic mass as a target of evaluation of vessel distribution by color Doppler for the diagnosis of adnexal malignancies: results of a multicenter European study. J Ultrasound Med 2002; 21:1105–1111. 11. Kurjak A, Kupesic S, Sparac V, Kosuta D. Threedimensional ultrasonographic and power Doppler characterization of ovarian lesions. Ultrasound Obstet Gynecol 2000; 16:365–371.
J Ultrasound Med 2005; 24:689–696
695
Three-Dimensional Power Doppler Vascular Sampling
12. Alcázar JL, Ruiz-Pérez ML, Errasti T. Transvaginal color Doppler sonography in adnexal masses: which parameter performs best? Ultrasound Obstet Gynecol 1996; 8:114–119. 13. Pairleitner H, Steiner H, Hasenoehrl G, Staudach A. Three-dimensional power Doppler sonography: imaging and quantifying blood flow and vascularization. Ultrasound Obstet Gynecol 1999; 14:139–143. 14. Food and Drug Administration, Center for Biologics Evaluation and Research. CBER guidances/ guidelines/points to consider. Food and Drug Administration website. Available at: http://www.fda. gov/cber/guidelines.htm.
24. Jarvela IY, Sladkevicius P, Tekay AH, Campbell S, Nargund G. Intraobserver and interobserver variability of ovarian volume, gray-scale and color flow indices obtained using transvaginal three-dimensional power Doppler ultrasonography. Ultrasound Obstet Gynecol 2003; 21:277–282. 25. Tekay A, Jouppila P. Controversies in assessment of ovarian tumors with transvaginal color Doppler ultrasound. Acta Obstet Gynecol Scand 1996; 75:316– 329.
15. Serov SF, Scully RE, Sobin LH. International Histological Classification of Tumors. No 9. Histological Typing of Ovarian Tumors. Geneva, Switzerland: World Health Organization; 1973.
26. Kupesic S, Kurjak A. Contrast-enhanced three-dimensional power Doppler sonography for differentiation of adnexal masses. Obstet Gynecol 2000; 96:452– 458.
16. Shepherd JH. Revised FIGO staging for gynecological cancer. Br J Obstet Gynaecol 1989; 96:889–892.
27. Breyer B, Kurjak A. Tumor vascularization, Doppler measurements and chaos: what to do? Ultrasound Obstet Gynecol 1995; 5:209–210.
17. Guerriero S, Alcázar JL, Ajossa S, et al. Comparison of conventional color Doppler imaging and power Doppler imaging for the diagnosis of ovarian cancer: results of a European study. Gynecol Oncol 2001; 83:299–304.
696
23. Merce LT, Barco MJ, Bau S. Reproducibility of the study of placental vascularization by three-dimensional power Doppler. J Perinat Med 2004; 32:228–233.
28. Konerding MA, Miodonski AJ, Lametschwandtner A. Microvascular corrosion casting in the study of tumor vascularity: a review. Scanning Microsc 1995; 9:1233–1243.
18. Castillo G, Alcazar JL, Jurado M. Natural history of sonographically detected simple unilocular adnexal cysts in asymptomatic postmenopausal women. Gynecol Oncol 2004; 92:965–969.
29. Cohen LS, Escobar PF, Scharm C, Glimco B, Fishman DA. Three-dimensional ultrasound power Doppler improves the diagnostic accuracy for ovarian cancer prediction. Gynecol Oncol 2001; 82:40–48.
19. Alcázar JL, Errasti T, Mínguez JA, Galán MJ, GarcíaManero M, Ceamanos C. Sonographic features of ovarian cystadenofibroma: spectrum of findings. J Ultrasound Med 2001; 20:915–919.
30. Guerriero S, Ajossa S, Melis GB. Is three-dimensional power Doppler ultrasound better than two-dimensional power Doppler? Gynecol Oncol 2002; 84: 352–353.
20. Varras M, Polyzos D, Perouli E, Noti P, Pantazis I, Akrivis C. Tubo-ovarian abscesses: spectrum of sonographic findings with surgical and pathological correlations. Clin Exp Obstet Gynecol 2003; 30:117–121.
31. Alcázar JL, Castillo G. Comparison of 2-dimensional and 3-dimensional power-Doppler imaging in complex adnexal masses for the prediction of ovarian cancer. Am J Obstet Gynecol 2005; 192:807–812.
21. Lee MS, Cho HC, Lee YH, Hong SR. Ovarian sclerosing stromal tumors: gray scale and color Doppler sonographic findings. J Ultrasound Med 2001; 20: 413–417.
32. Cheng WF, Lee CN, Chen CA, et al. Comparison between “in vivo” and “in vitro” methods for evaluating tumor angiogenesis using cervical carcinoma as a model. Angiogenesis 1999; 3:295–304.
22. Raine-Fenning NJ, Campbell BK, Clewes JS, Kendall NR, Johnson IR. The interobserver reliability of threedimensional power Doppler data acquisition within the female pelvis. Ultrasound Obstet Gynecol 2004; 23:501–508.
33. Sedelaar JP, van Leenders GJ, Hulsbergen-van de Kaa CA, et al. Microvessel density: correlation between contrast ultrasonography and histology of prostate cancer. Eur Urol 2001; 40:285–293. 34. Kurjak A, Kupesic S, Sparac V, Prka M, Bekavac I. The detection of stage I ovarian cancer by three-dimensional sonography and power Doppler. Gynecol Oncol 2003; 90:258–264. J Ultrasound Med 2005; 24:689–696
Three-Dimensional Power Doppler Vascular Sampling A New Method for Predicting Ovarian Cancer in Vascularized Complex Adnexal Masses Juan Luis Alcázar, MD, Luis T. Mercé, MD, Manuel García Manero, MD Objective. The purpose of this study was to explore the role of a new concept (“vascular sampling”) as a third step to discriminate benign and malignant lesions in B-mode and color Doppler sonographically suggestive adnexal masses. Methods. Forty-five women (mean age, 52.3 years; range, 17–82 years) with the diagnosis of complex adnexal masses on B-mode sonography were evaluated using 3dimensional power Doppler sonography. Four women had bilateral masses. After a morphologic reevaluation was done, color pulsed Doppler sonography was used to obtain flow velocity waveforms, and velocimetric indices were calculated (resistive index, pulsatility index, and peak systolic velocity). Thereafter, 3-dimensional power Doppler sonography was used to assess vascularization of highly suggestive areas (gross papillary projections, solid areas, and thick septations), meaning a focused assessment (“sampling”) of a suggestive area of the tumor. With a virtual organ computer-aided analysis program, vascular indices (vascularization index, flow index, and vascular flow index) were automatically calculated. A definitive histologic diagnosis was obtained in each case. Results. Forty masses (82%) were malignant and 9 (18%) were benign. Morphologic evaluation revealed 10 (20%) unilocular solid masses, 20 (41%) multilocular solid masses, and 19 (39%) mostly solid masses. Blood flow was found in all cases. Median vascularization index (15.5% versus 8.2%; P = .002), flow index (33.6 versus 20.8; P = .007), and vascular flow index (5.2 versus 2.3; P = .001) were significantly higher in malignant tumors. No differences were found in resistive index (0.43 versus 0.45; P = .770), pulsatility index (0.62 versus 0.65; P = .694), and peak systolic velocity (15.6 versus 12 cm/s; P = .162). Conclusions. Threedimensional power Doppler vascular sampling seems to be a promising tool for predicting ovarian cancer in vascularized complex adnexal masses. It could be better than conventional color pulsed Doppler imaging. Key words: ovarian cancer; power Doppler sonography; 3-dimensional sonography. Abbreviations AUC, area under the curve; FI, flow index; PI, pulsatility index; PSV, peak systolic velocity; RI, resistive index; 3D, 3-dimensional; 2D, 2-dimensional; VFI, vascular flow index; VI, vascularization index; VOCAL, Virtual Organ Computer-Aided Analysis Received October 25, 2004, from the Department of Obstetrics and Gynecology, Clínica Universitaria de Navarra, School of Medicine, University of Navarra, Pamplona, Spain (J.L.A., M.G.M.); and Department of Obstetrics and Gynecology, Hospital Ruber Internacional, Madrid, Spain. (L.T.M.). Revision requested December 26, 2004. Revised manuscript accepted for publication December 30, 2004. Address correspondence to Juan Luis Alcázar, MD, Department of Obstetrics and Gynecology, Clínica Universitaria de Navarra, Avenida Pio XII 36, 31008 Pamplona, Spain. E-mail: [email protected]
T
he introduction of transvaginal sonography has greatly improved the ability to discriminate benign and malignant ovarian tumors.1 Although this technique has proved to be precise enough in the specific diagnosis of some benign ovarian lesions,2–4 it still has the problem of a relatively high false-positive rate in complex masses.5 Transvaginal color and pulsed Doppler sonography allows the assessment of tumor vascularization and was introduced into clinical practice in an attempt to decrease the false-positive rate of morphologic sonographic evaluation of adnexal masses. Although initial results, based on the use of velocimetric indices, were encouraging,6 the literature is full of conflicting results,7–9 which make this approach unpractical and poorly reproducible.
© 2005 by the American Institute of Ultrasound in Medicine • J Ultrasound Med 2005; 24:689–696 • 0278-4297/05/$3.50
Three-Dimensional Power Doppler Vascular Sampling
A simpler approach based on morphologic sonographic assessment and tumor vessel distribution using color or power Doppler sonography has been proposed.10 This approach has proved very useful for discriminating benign from malignant adnexal tumors because the probability of malignancy is high in complex adnexal masses with central vessel distribution, whereas complex masses without vessels or peripheral blood flow are usually benign.10 However, there are still some benign tumors that show a complex appearance with central vessel distribution such as cystadenofibromas, mucinous cystadenomas, tubo-ovarian abscesses, and solid ovarian benign tumors such as Brenner tumors, granulosa cell tumors, and ovarian fibromas. These tumors are actually very difficult to discriminate from ovarian malignancies. Recently, a new method to assess tumor vascularization was developed: 3-dimensional (3D) power Doppler sonography.11 The aim of this study was to evaluate the potential role of 3D power Doppler sonography as a new method in the evaluation of suggestive vascularized adnexal masses (complex morphologic appearance with central vessel distribution on conventional color Doppler transvaginal sonography) for discriminating benign from malignant lesions.
Materials and Methods From August 2002 to August 2004, 45 women with the diagnosis of complex adnexal mass with central vessel distribution were evaluated by transvaginal 3D power Doppler sonography. Patients’ mean age was 52.3 years (SD, 16.2 years; range, 17–82 years). Twenty-four patients (53%) were postmenopausal, and 21 (47%) were premenopausal. Four women had bilateral masses, giving a total number of 49 masses evaluated. All women had been evaluated first by transvaginal sonography with a Kretz Sonoace 9900 system (GE Medical Systems, Kretztechnik, Zipf, Austria) or a Voluson 730 Pro system (GE Medical Systems, Milwaukee, WI) with a 5- to 7.5MHz endovaginal probe and color, power, and pulsed Doppler as well as 3D sonographic capabilities. Transabdominal sonography (3.5–5 MHz) was also performed in large tumors. The scanning techniques have been described in detail elsewhere.12 On B-mode sonography, a complex mass was defined in the presence of thick papillary projec690
tions (>3 mm), solid areas, or mostly solid echogenicity (Figures 1 and 2). According to Granberg et al,1 lesions were classified as unilocular solid, multilocular solid, and mostly solid. After B-mode evaluation was done, the 2dimensional (2D) power Doppler gate was activated to assess tumor vascularization. Power Doppler settings were set to achieve maximum sensitivity to detect low-velocity flow without noise (frequency, 5 MHz; power Doppler gain, 20 [range, 1–30]; dynamic range, 20–40 dB; edge, 1; persistence, 2; color map, 1; gate, 2; filter, 3; and pulse repetition frequency, 0.6 kHz). Central vessel vascularization was defined in the presence of color spots within the suggestive areas of the tumor (thick papillary projections, solid areas, or central part of solid tumors; Figures 3 and 4).10 Pulsed Doppler sonography was used to interrogate color spots identified to obtain a flow velocity waveform and to confirm the arterial nature of the vessel. The pulsatility index (PI), resistive index (RI), and peak systolic velocity (PSV, centimeters per second) were recorded automatically. The lowest PI and RI and the highest PSV found in a given tumor were used for analysis. Then, a 3D volume was activated to obtain a 3D box from those suggestive areas, for example, thick papillary projections, a solid area, or, in the case of a mostly solid tumor, the whole tumor. Once a 3D volume was obtained, it was stored on a hard disk (Sonoview; GE Medical Systems, Kretztechnik). The volume acquisition time lasted from 2 to 6 seconds depending on the size of the volume box. Using the Virtual Organ Computer-Aided Analysis (VOCAL) program, we calculated 3 vascular indices of those Figure 1. Transvaginal sonogram from a mostly solid adnexal mass that corresponded to a primary serous ovarian carcinoma.
J Ultrasound Med 2005; 24:689–696
Alcázar et al
suggestive areas, excluding cystic areas in which there was no tissue and focusing only on “solid” areas (Figures 5–8). A 3D volume is composed of voxels (smallest unit of volume). Voxels contain all the information about gray scale and color according to an intensity scale ranging from 0 to 100. According to these values, this measurement system obtains the mean grayness and power Doppler indices to evaluate vessels and blood flow.13 The vascular indices calculated were as follows: The vascularization index (VI), expressed as a percentage, measures the number of color voxels in the studied volume, representing the blood vessels within the tissue; the flow index (FI) is the average color value of all color voxels, representing the average color intensity; and the vascular flow index (VFI) is the average color value of all gray and color voxels, which represents both blood flow and vascularization and might be regarded as “perfusion.” Measurements were undertaken with the use of the manual mode in plane A and a 9° rotation step. If there were more than 1 noncontiguous solid area, each was evaluated separately. In those tumors with more than 1 suggestive area, the highest VI, FI, and VFI values found were used for analysis. If the tumor showed areas of low echogenicity, these areas were not included in the VOCAL calculation. The same author (J.L.A.) performed all examinations. Intraobserver variability was estimated by performance of 2 different measures 1 week apart14 in the first 10 cases evaluated and calculation of the intraclass correlation coefficient (0.91). As indicated above, only complex adnexal masses with central vessel distribution were Figure 2. Transvaginal sonogram from a unilocular solid adnexal mass in which histologic examination revealed metastatic endometrial cancer to the ovary.
J Ultrasound Med 2005; 24:689–696
Figure 3. Transvaginal power Doppler sonogram from the case in Figure 1. Central vessel distribution is shown within solid areas.
included in this study. This means that no tumor with a clear benign appearance, such as simple cyst, endometrioma, mature teratoma, hydrosalpinx, or peritoneal cyst, even having central flow, or with solid areas or thick papillary projections but without vessels detected within these areas was included in this study. Patients were not selected except for sonographic findings and the availability of 3D sonography. All patients gave oral informed consent after the nature of the study was fully explained. Institutional Review Board approval was also obtained. All patients underwent surgery, and a definitive histologic diagnosis was obtained in every case. Tumors were classified according to World Health Organization criteria,15 and ovarian cancers were staged according to International Federation of Gynecology and Obstetrics criteria.16 Figure 4. Transvaginal power Doppler sonogram from the case in Figure 2. As in the previous case, central vessels are shown within solid areas.
691
Three-Dimensional Power Doppler Vascular Sampling
Results
Figure 5. Three-dimensional power Doppler sonogram from the case in Figure 1 showing how only solid portions of the tumor were analyzed by the VOCAL method.
Continuous variables were compared with the use of the Mann-Whitney U test. P < .05 was considered statistically significant for all tests. The SPSS version 11.0 for Windows statistical package was used (SPSS Inc, Chicago, IL). Receiver operating characteristic curves were plotted for each vascular index to assess its diagnostic performance.
Figure 6. Three-dimensional power-Doppler sonogram from the case in Figure 2 showing how only solid portions of the tumor were analyzed by the VOCAL method. Low-level echoes are disregarded.
692
After surgical removal, 40 (82%) of the masses proved to be malignant and 9 (18%) proved to be benign (Table 1). Primary ovarian cancer stages were as follows: stage Ia, 8 (32%); stage IIb, 2 (8%); stage IIIa, 1 (4%); stage IIIb, 2 (8%); stage IIIc, 8 (32%); and stage IV, 4 (16%). On B-mode sonography, adnexal mass appearances were as follows: unilocular solid cyst, 10 (20%); multilocular solid, 20 (41%); and mostly solid, 19 (39%). As stated, all lesions had blood flow detected within thick papillary projections, solid areas, or central areas of solid tumors. According to histologic findings, 8 (80%) of 10 unilocular solid masses were malignant; 14 (70%) of 20 multilocular solid tumors were malignant; and 18 (95%) of 19 solid tumors were malignant. Median VI, FI, and VFI were significantly higher in malignant tumors. No differences were found in median RI, PI, and PSV between benign and malignant tumors (Table 2). Receiver operating characteristic curves for VI, FI, and VFI are shown in Figures 9–11.
Discussion In this study, we explored the potential role of a new approach using 3D power Doppler sonography for predicting ovarian malignancy in complex adnexal masses that show central vessel distribution. We have termed this new approach “3D power Doppler vascular sampling.” It consists of assessing the vascularization of a given suggestive area in a given tumor. We have focused the use of this new technique on complex adnexal masses with central flow because we think that it will be useless in simple cystic masses or “complex” masses without flow, given that the probability of malignancy in the latter is known to be low.10,17,18 When we face a complex mass with detectable blood flow with solid areas or thick papillary projections, we should categorized this mass as “malignant” or “very highly suggestive.” However, a considerable number of benign tumors may have this appearance, for example, cystadenofibromas,19 tubo-ovarian abscesses,20 and solid benign ovarian tumors.21 Using conventional 2D color or power Doppler sonography, we have no means to differentiate these benign entities from true malignant tumors. Three-dimensional power Doppler J Ultrasound Med 2005; 24:689–696
Alcázar et al
sonography provides a new approach to assessing tumor vascularization. It has been shown that this technique is highly reliable with a very low intraobserver and interobserver variability.22–24 Therefore, this technique overcomes one of the main problems of conventional color Doppler sonography25 and provides an objective measure of vascularity. To our knowledge, the only previous study analyzing 3D power Doppler vascular indices in adnexal masses included only 6 surgically removed adnexal tumors, all benign.13 The authors did not report specific data on VI, FI, and VFI values. Therefore, no comparison can be made. However, there are several studies that have assessed the usefulness of 3D power Doppler sonography with a different approach than ours. Kurjak et al11 and Kupesic and Kurjak,26 from Zagreb, Croatia, compared the diagnostic performance of 3D power Doppler and conventional color Doppler sonography in different studies. They found that 3D power Doppler sonography was more sensitive than conventional color Doppler sonography,11 especially when a sonographic contrast agent was used.26 They based their diagnostic criteria for malignancy suggestion on vessel architecture as depicted by 3D sonography, such as branching pattern, vessel caliber, and presence of microaneurysms or vascular lakes. This was based on the chaos theory,27 which established that the architecture of a vascular network of newly formed vessels in malignant tumors is built in a chaotic distribution and not in a predetermined fashion. Although, this has been shown in corrosion studies,28 we think that the reproducibility of this approach would be low or at least difficult to achieve because basically it is based on a subjective analysis of a 3D reconstructed image. In fact, to the best of our knowledge, no further study has been published that tried to reproduce the excellent results reported by this group using this technique. Cohen and coworkers29 reported that the use of 3D power Doppler sonography in complex masses would reduce the false-positive rate of Bmode sonography. However, they used as a criterion for malignancy the presence of vessels in a central location, which is the same proposed for 2D sonography.10 Therefore, we wonder, as did Guerriero et al,30 whether this approach is superior to conventional 2D color or power Doppler J Ultrasound Med 2005; 24:689–696
Figure 7. Histogram showing VI, FI, and VFI values for solid areas of the case in Figure 1.
Figure 8. Histogram showing VI, FI, and VFI values for the case in Figure 2.
Table 1. Histologic Distribution of Adnexal Masses Histologic Finding
Mucinous cystadenoma Cystadenofibroma Tubo-ovarian abscess Low-malignant-potential tumor Primary carcinoma Metastatic carcinoma
n
%
3 3 3 3 25 12
6.1 6.1 6.1 6.1 53.0 24.6
693
Three-Dimensional Power Doppler Vascular Sampling
Table 2. Vascular and Velocimetric Indices in Malignant and Benign Adnexal Tumors Type
Malignant (n = 40) Benign (n = 9)
VI, %
15.5 (IQR, 8.2) (7.1–77) 8.2 (IQR, 11.5) (0.2–21)
FI
33.6 (IQR, 8.4) (16.2–51.9) 20.8 (IQR, 16) (12.6–36.9)
VFI
RI
5.2 0.43 (IQR, 3.2) (IQR, 0.15) (2.1–40.1) (0.23–0.74) 2.3 0.45 (IQR, 2.3) (IQR, 0.11) (0.04–11.5) (0.30–0.52)
PI
PSV, cm/s
0.62 (IQR, 0.2) (0.30–1.48) 0.65 (IQR, 0.3) (0.44–0.90)
15.6 (IQR, 9.8) (7.0–30.0) 12.0 (IQR, 4.8) (7.5–31.7)
Data are expressed as median with range in parentheses. IQR indicates interquartile range. VI, P = .002; FI, P = .007; VFI, P = .0001; RI, P = .770; PI, P = .694; PSV, P = .162.
694
sonography. In fact, a recent study from our group comparing 3D power Doppler and 2D power Doppler sonography, with the use of the same criteria for malignancy suggestion for both techniques, found that 3D power Doppler sonography was not superior to 2D power Doppler sonography for discriminating benign from malignant complex adnexal masses.31 The results of this study, although preliminary because of a relatively small series, indicate that 3D power Doppler vascular sampling may be a new tool for discriminating benign from malignant complex adnexal masses with central vessel distribution. We found that vascular indices (VI, FI, and VFI) from suggestive areas such as thick papillary projections, solid areas, and mostly solid tumors were significantly higher in malignant tumors compared with benign tumors, which may reflect different angiogenic patterns. Conversely, it has been shown that vas-
cularity as assessed by immunohistochemical microvessel density quantification correlates with both 2D power Doppler and 3D power Doppler findings,32,33 which would validate the findings obtained in vivo with 3D power Doppler technology. One interesting finding in our study is that 40% of ovarian cancers were stage II or I. This raises a question about the role of this technique for the early diagnosis of ovarian cancer. In fact, Kurjak et al34 were able to identify 97.7% of stage I ovarian cancers (42 of 43) in their study using this technique. Notwithstanding, we are aware that some potential sources of error could exist with this approach. One of them might be large vascularized tumors because it may be difficult to include the whole tumor in only a single 3D box; therefore, one should select a single area or use more than a single box for analysis. This may
Figure 9. Receiver operating characteristic curve for VI (area under the curve [AUC], 0.88; SE, 0.07).
Figure 10. Receiver operating characteristic curve for FI (AUC, 0.87; SE, 0.06).
J Ultrasound Med 2005; 24:689–696
Alcázar et al
introduce a subjective factor. We think that the simplest way is to use the area where the highest vascular indices are found. However, the best method for assessing these large tumors by using this technique needs to be established. Conversely, experience may be needed to appropriately select suggestive areas. However, this question needs to be evaluated. In fact, to the best of our knowledge, no study has formally assessed the learning curve for 3D sonography. We evaluated the diagnostic performance of this new approach in our series, and we found that each vascular index has acceptable diagnostic performance. However, we have to bear in mind that specificity and the negative predictive value are difficult to calculate precisely because the number of benign tumors, as could be expected, was low. Finally, we assessed only the intraobserver reproducibility, which might be another limitation in our study. Interobserver reproducibility must be assessed in future studies. In conclusion, we have defined and presented a new approach for predicting ovarian cancer in complex adnexal masses with central blood flow detected on conventional color or power Doppler transvaginal sonography. This new approach, termed 3D power Doppler vascular sampling, is based on 3D sonography and in our own experience seems to be reproducible. Studies in larger series are needed to establish its actual diagnostic performance.
References 1.
Granberg S, Norstrom A, Wikland M. Tumors of the lower pelvis as imaged by vaginal sonography. Gynecol Oncol 1990; 37:224–229.
2.
Alcázar JL, Laparte C, Jurado M, López-García G. The role of transvaginal ultrasonography combined with color velocity imaging and pulsed Doppler in the diagnosis of endometrioma. Fertil Steril 1997; 67:487– 491.
3.
Caspi B, Appelman Z, Rabinerson D, Elchalal U, Zalel Y, Katz Z. Pathognomonic echo patterns of benign cystic teratomas of the ovary: classification, incidence and accuracy rate of sonographic diagnosis. Ultrasound Obstet Gynecol 1996; 7:275–279.
4.
Guerriero S, Ajossa S, Lai MP, Mais V, Paoletti AM, Melis GB. Transvaginal ultrasonography associated with colour Doppler energy in the diagnosis of hydrosalpinx. Hum Reprod 2000; 15:1568–1572.
5.
Ferrazzi E, Zanetta G, Dordoni D, et al. Transvaginal ultrasonographic characterization of ovarian masses: comparison of five scoring systems in a multicenter study. Ultrasound Obstet Gynecol 1997; 10:192–197.
6.
Kurjak A, Zalud I, Alfirevic Z. Evaluation of adnexal masses with transvaginal color Doppler ultrasound. J Ultrasound Med 1991; 10:295–297.
7.
Kawai M, Kano T, Kikkawa F, Maeda O, Oguchi H, Tomoda Y. Transvaginal Doppler ultrasound with color flow imaging in the diagnosis of ovarian cancer. Obstet Gynecol 1992; 79:163–167.
8.
Tekay A, Jouppila P. Validity of pulsatility and resistance indices in classification of adnexal tumors with transvaginal color Doppler ultrasound. Ultrasound Obstet Gynecol 1992; 2:338–344.
9.
Levine D, Feldstein VA, Babcook CJ, et al. Sonography of ovarian masses: poor sensitivity of resistive index for identifying malignant lesions. AJR Am J Roentgenol 1994; 162:1355–1359.
Figure 11. Receiver operating characteristic curve for VFI (AUC, 0.86; SE, 0.1).
10. Guerriero S, Alcázar JL, Coccia ME, et al. Complex pelvic mass as a target of evaluation of vessel distribution by color Doppler for the diagnosis of adnexal malignancies: results of a multicenter European study. J Ultrasound Med 2002; 21:1105–1111. 11. Kurjak A, Kupesic S, Sparac V, Kosuta D. Threedimensional ultrasonographic and power Doppler characterization of ovarian lesions. Ultrasound Obstet Gynecol 2000; 16:365–371.
J Ultrasound Med 2005; 24:689–696
695
Three-Dimensional Power Doppler Vascular Sampling
12. Alcázar JL, Ruiz-Pérez ML, Errasti T. Transvaginal color Doppler sonography in adnexal masses: which parameter performs best? Ultrasound Obstet Gynecol 1996; 8:114–119. 13. Pairleitner H, Steiner H, Hasenoehrl G, Staudach A. Three-dimensional power Doppler sonography: imaging and quantifying blood flow and vascularization. Ultrasound Obstet Gynecol 1999; 14:139–143. 14. Food and Drug Administration, Center for Biologics Evaluation and Research. CBER guidances/ guidelines/points to consider. Food and Drug Administration website. Available at: http://www.fda. gov/cber/guidelines.htm.
24. Jarvela IY, Sladkevicius P, Tekay AH, Campbell S, Nargund G. Intraobserver and interobserver variability of ovarian volume, gray-scale and color flow indices obtained using transvaginal three-dimensional power Doppler ultrasonography. Ultrasound Obstet Gynecol 2003; 21:277–282. 25. Tekay A, Jouppila P. Controversies in assessment of ovarian tumors with transvaginal color Doppler ultrasound. Acta Obstet Gynecol Scand 1996; 75:316– 329.
15. Serov SF, Scully RE, Sobin LH. International Histological Classification of Tumors. No 9. Histological Typing of Ovarian Tumors. Geneva, Switzerland: World Health Organization; 1973.
26. Kupesic S, Kurjak A. Contrast-enhanced three-dimensional power Doppler sonography for differentiation of adnexal masses. Obstet Gynecol 2000; 96:452– 458.
16. Shepherd JH. Revised FIGO staging for gynecological cancer. Br J Obstet Gynaecol 1989; 96:889–892.
27. Breyer B, Kurjak A. Tumor vascularization, Doppler measurements and chaos: what to do? Ultrasound Obstet Gynecol 1995; 5:209–210.
17. Guerriero S, Alcázar JL, Ajossa S, et al. Comparison of conventional color Doppler imaging and power Doppler imaging for the diagnosis of ovarian cancer: results of a European study. Gynecol Oncol 2001; 83:299–304.
696
23. Merce LT, Barco MJ, Bau S. Reproducibility of the study of placental vascularization by three-dimensional power Doppler. J Perinat Med 2004; 32:228–233.
28. Konerding MA, Miodonski AJ, Lametschwandtner A. Microvascular corrosion casting in the study of tumor vascularity: a review. Scanning Microsc 1995; 9:1233–1243.
18. Castillo G, Alcazar JL, Jurado M. Natural history of sonographically detected simple unilocular adnexal cysts in asymptomatic postmenopausal women. Gynecol Oncol 2004; 92:965–969.
29. Cohen LS, Escobar PF, Scharm C, Glimco B, Fishman DA. Three-dimensional ultrasound power Doppler improves the diagnostic accuracy for ovarian cancer prediction. Gynecol Oncol 2001; 82:40–48.
19. Alcázar JL, Errasti T, Mínguez JA, Galán MJ, GarcíaManero M, Ceamanos C. Sonographic features of ovarian cystadenofibroma: spectrum of findings. J Ultrasound Med 2001; 20:915–919.
30. Guerriero S, Ajossa S, Melis GB. Is three-dimensional power Doppler ultrasound better than two-dimensional power Doppler? Gynecol Oncol 2002; 84: 352–353.
20. Varras M, Polyzos D, Perouli E, Noti P, Pantazis I, Akrivis C. Tubo-ovarian abscesses: spectrum of sonographic findings with surgical and pathological correlations. Clin Exp Obstet Gynecol 2003; 30:117–121.
31. Alcázar JL, Castillo G. Comparison of 2-dimensional and 3-dimensional power-Doppler imaging in complex adnexal masses for the prediction of ovarian cancer. Am J Obstet Gynecol 2005; 192:807–812.
21. Lee MS, Cho HC, Lee YH, Hong SR. Ovarian sclerosing stromal tumors: gray scale and color Doppler sonographic findings. J Ultrasound Med 2001; 20: 413–417.
32. Cheng WF, Lee CN, Chen CA, et al. Comparison between “in vivo” and “in vitro” methods for evaluating tumor angiogenesis using cervical carcinoma as a model. Angiogenesis 1999; 3:295–304.
22. Raine-Fenning NJ, Campbell BK, Clewes JS, Kendall NR, Johnson IR. The interobserver reliability of threedimensional power Doppler data acquisition within the female pelvis. Ultrasound Obstet Gynecol 2004; 23:501–508.
33. Sedelaar JP, van Leenders GJ, Hulsbergen-van de Kaa CA, et al. Microvessel density: correlation between contrast ultrasonography and histology of prostate cancer. Eur Urol 2001; 40:285–293. 34. Kurjak A, Kupesic S, Sparac V, Prka M, Bekavac I. The detection of stage I ovarian cancer by three-dimensional sonography and power Doppler. Gynecol Oncol 2003; 90:258–264. J Ultrasound Med 2005; 24:689–696
