Three-dimensional ultrasonographic evaluation of ovarian tumours: a preliminary study

Toshiyuki Hata1,3, Toshihiro Yanagihara1, Keiji Hayashi1, Chizu Yamashiro1, Yohichi Ohnishi1, Masashi Akiyama1, Atsushi Manabe2 and Kohji Miyazaki2

1 Department of Perinatology, Kagawa Medical University, 1750–1 Ikenobe, Miki, Kagawa 761-0793, and 2 Department of Obstetrics and Gynecology, Shimane Medical University, Izumo 693, Japan


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Our objective was to determine the accuracy of three dimensional (3D) ultrasonography in the differentiation of ovarian tumours. Women (n = 20) with ovarian tumours (13 benign and seven malignant) were first examined using two-dimensional (2D) ultrasonography and were then evaluated with three-dimensional (3D) ultrasonography. All patients underwent surgery within 7 days of 2D and 3D ultrasonographic evaluations, and a histopathological diagnosis was made. A 2D ultrasonographic diagnosis of malignancy was done using an established scoring system. Diagnostic criteria for ovarian malignancy by 3D ultrasonography were: irregular thick septa, irregular papillary projection, mostly solid tumour, high echogenicity, and irregular inner wall. Compared with 2D ultrasound, 3D ultrasound had a significantly higher specificity (P < 0.005) and accuracy (P < 0.01), and a significantly lower false positive rate (P < 0.005). Our results suggest that 3D ultrasonography might be a better means of differentiating between malignant and benign ovarian tumours. However, in view of the small number of ovarian tumours, these observations must be considered preliminary.

Key words: 2D ultrasonography/3D ultrasonography/ovarian cancer/tumours


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Data have been reported regarding the effectiveness of transvaginal sonography for ovarian cancer screening (van Nagell et al., 1991Go). Transvaginal sonography is more sensitive than a pelvic examination in detecting ovarian abnormalities but lacks specificity in distinguishing between benign and malignant lesions (Sparks and Varner, 1991Go). Moreover, a major concern about transvaginal sonography screening for ovarian cancer is its relatively low positive predictive value. Recent studies indicate that transvaginal sonography screening has a positive predictive value of only 10% in post-menopausal women (van Nagell et al., 1991Go; DePriest et al., 1993aGo). As a result, there has been increased interest in the evaluation of several adjunctive diagnostic methods as a means to increase the specificity and positive predictive value of transvaginal sonography screening (van Nagell, 1994Go). These include morphology indexing of tumours based on sonographic findings (Sassone et al., 1991Go; DePriest et al., 1993bGo), colour Doppler assessment of intratumoral blood flow (Hata et al., 1992Go, 1995Go), and serum CA 125 (Gadducci et al., 1992Go).

Three-dimensional (3D) ultrasonography has been introduced into clinical practice in gynaecology. This technique overcomes anatomical limitations which restrict the number and orientation of the scanning planes on transvaginal sonography (Jurkovik et al., 1994Go). This ultrasonic innovation has diverse applications, e.g. detection of congenital uterine abnormalities (Jurkovik et al., 1995), evaluation of ovarian masses (Bonilla-Musoles et al., 1995Go), assessment of uterine cavity pathology (Weinraub et al., 1996Go), and ovarian and endometrial volume measurements (Kyei-Mensah et al., 1996Go). Recently, further technical development of 3D technology has led to a self-contained imaging system that can not only produce conventional two-dimensional (2D) images but can generate within seconds a high-quality 3D image with no need for an external workstation or other additional, costly equipment (Baba et al., 1996Go; Hata et al., 1997Go). The objectives of our study were to demonstrate the feasibility of intratumoral surface visualization of ovarian tumours using 3D surface rendering and to determine whether 3D ultrasonography is valid in differentiating malignant from benign ovarian tumours and whether this method has a higher diagnostic accuracy than 2D ultrasonography.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Japanese women (n = 20) with ovarian masses participated in this study. These women were referred to our department for sonographic evaluation because of suspected pelvic masses. They were a consecutive series of patients with an ovarian mass. The age range of the patients was 19–78 years with a mean of 45.5; eight of the patients were post-menopausal. All patients underwent surgery within 7 days of 2D and 3D ultrasonographic evaluations, and histopathological findings were classified according to the criteria of the World Health Organization (Serov et al., 1973Go). The patients signed an informed consent form as approved through the Research Ethics Committee of Shimane Medical University Hospital, Japan.

All patients were studied with a specially developed abdominal 3D ultrasonography transducer (Aloka ASU-1000B, 3.5 MHz; Aloka, Tokyo, Japan). This ultrasonic transducer is connected to an ultrasonography device (Aloka SSD-1700; Aloka). This imaging system can provide conventional 2D ultrasonographic images and can also generate within seconds high-quality 3D images in the surface and transparent mode with no need for an external workstation. A 3D image is built by selecting a region of interest from a 2D image and superimposing on it a volume box defined by the examiner. The crystal array of the transducer then sweeps mechanically over the 2D region selected through a 60° angle. Within 5 s, the outlined volume is automatically scanned and a sculpture-like 3D image is displayed simultaneously on the screen. In this system (Vol-mode), the ultrasound beam is regarded as a projection ray in volume rendering, and ray tracing is conducted in real time. The procedure was not as complex as that of conventional 3D ultrasonography, and images can be obtained immediately (Baba et al., 1996Go; Hata et al., 1997Go). At present we use a 128 MB removable hard disk drive for the permanent storage of 3D images.

First one examiner (A.M.) recorded the 2D ultrasonographic images with transabdominal and transvaginal approach, then another examiner (T.H.) recorded the 3D ultrasonographic images without knowing the results of the 2D study. A 2D ultrasonographic diagnosis of malignancy was done using the scoring system by Sassone et al. (1991). The scores of nine or more were regarded as malignancy. The surface rendering mode of the 3D ultrasonography allows the study of the inner wall surfaces of the tumour. Diagnostic criteria for ovarian malignancy by 3D ultrasonography were irregular thick septa (Figure 1Go), irregular papillary projection (Figure 2Go), high echogenicity (Figure 2Go), irregular inner wall (Figure 2Go), and mostly solid irregular tumour (Figure 3Go). Each variable was assigned as one point in 3D scoring. Ovarian tumours with two and more 3D diagnostic criteria (2 or greater in scoring) were defined as malignancy. Unfortunately, in this study, other indices such as Doppler evaluation of these tumours or serum screening, e.g. CA 125 were not carried out in these patients.



View larger version (60K):
[in this window]
[in a new window]
 
Figure 1. Endometrioid cancer in a 52 year old woman (case 3). (A) Two-dimensional and (B) three-dimensional ultrasonography showed papillary projection and irregular septum (arrows). Tumour size is 12x6x5 cm.

 


View larger version (92K):
[in this window]
[in a new window]
 
Figure 2. Endometrioid cancer in a 40 year old woman (case 10). (A) Two-dimensional and (B) three-dimensional (3D) ultrasonography showed papillary projection (arrows). An irregular inner wall was clearly depicted by 3D ultrasound. Tumour size is 16x13x10 cm.

 


View larger version (100K):
[in this window]
[in a new window]
 
Figure 3. Serous cystadenocarcinoma in a 67 year old woman (case 17). (A) Two-dimensional and (B) three-dimensional ultrasonography showed mostly solid tumour. Tumour size is 14x10x10 cm.

 
Statistical analysis
Data were analysed using the {chi}2 test. P < 0.05 was considered to be significantly different.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
There were 13 patients with benign ovarian tumours, one with carcinoma of low malignant potential, and 6 with malignant ovarian tumours (Table IGo). Table IIGo lists the diagnostic values of each modality. The 3D ultrasound specificity (92.3%) and accuracy (95%) were significantly higher than those with 2D ultrasound (specificity 38.4% and accuracy 60%) (P < 0.005 and P < 0.01 respectively). The 3D ultrasound false positive rate (7.7%) was significantly lower than that with 2D ultrasound (61.5%) (P < 0.005). No significant differences were noted between sensitivity, positive predictive value, negative predictive value, and false negative rate characterized by 2D and 3D ultrasound diagnosis in detecting malignant ovarian tumours.


View this table:
[in this window]
[in a new window]
 
Table I. Demographic data, two-dimensional (2D) and three-dimensional (3D) morphological scores, and final diagnosis
 

View this table:
[in this window]
[in a new window]
 
Table II. Diagnostic value of two-dimensional (2D) and three-dimensional (3D) ultrasonography for the diagnosis of ovarian tumours
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Using a commercially available 3D ultrasound machine, the scanning time was 4–8 s and was dependent on the size of the volume imaged (Chan et al., 1997Go). The volumes of images were then analysed with the patient off the examination table. The time required to review the volumes of images depended on the complexity of the adnexal mass. A simple cyst took only 1–3 min, whereas a complex mass required as long as 4–8 min. In the current study we used a specially developed abdominal 3D transducer; this imaging system proved to be capable of providing conventional 2D ultrasonography images and, at the same time, having the capacity to generate (within seconds) high-quality fine 3D images on the screen without an external workstation being required (Baba et al., 1996Go; Hata et al., 1997Go). With respect to the advantages and disadvantages of our 3D system used in this study, the advantages might include: (i) rapid time for acquisition; (ii) rapid time for display, both of these allowing for rapid repeat of image acquisition if necessary; (iii) clear pictures; and (iv) reduced cost. Disadvantages might include: (i) lack of planar data, (ii) lack of rotation, and (iii) inability to `edit' data to eliminate echoes, e.g. intestines. Therefore, our 3D system (rapid surface rendering and volume rendering) has advantages over multi planar 3D technology.

In this investigation, we compared 3D ultrasonographic findings with 2D ultrasonographic findings for detecting ovarian malignancy. There were no significant differences for sensitivity, negative predictive value, and false negative rate between 3D and 2D ultrasonography. The positive predictive value (46.6%) of 2D ultrasound is lower than that (87.5%) of 3D ultrasound, but there was no significant difference. Compared with 2D ultrasound, 3D ultrasound had significantly higher specificity and accuracy, and significantly lower false positive rate. Bonilla-Musoles et al. (1995) reported that 2D ultrasound detected four out of five ovarian malignancies, and one additional ovarian carcinoma was diagnosed by 3D scanning. In this study, 2D ultrasonographic diagnosis of malignancy was carried out using the scoring system by Sassone et al. (1991). Lerner et al. (1994) pointed out that Sassone's scoring system is plagued by the existence of complex and high scoring lesions that are associated with benign lesions, such as teratoma and fibroma. In particular, the specificity and positive predictive value were hampered by masses that obviously were benign but nonetheless scored high, such as benign cystic teratoma and fibroma–thecoma. In this study, five of the 20 patients studied had benign cystic teratoma. The low specificity and low positive predictive value using 2D ultrasound in this study may be due to this population study bias. On the other hand, there was only one false positive diagnosis with 3D ultrasound in a patient with seromucinous cystadenoma. In order to evaluate the 3D imaging system in the diagnosis and differentiation of benign/malignant ovarian tumours, the same scoring system should be used for both the 2D and the 3D images. However, in this study, we used only the surface rendering mode for detection of ovarian tumour structures, and did not use planar images. Hence, we could not use the same scoring system for both the 2D and the 3D images. There are various scoring systems for 2D images, and different diagnostic criteria are used for these scoring systems, Therefore, there are primary differences between the various 2D scoring systems, as there are between the 3D and 2D scoring systems used in this study. Moreover, the scoring system of Sassone et al. (1991) is the most useful standard scoring system for 2D imaging in the diagnosis and differentiation of benign/malignant ovarian tumours, and we did not try to compare other scoring systems for 2D images with our 3D scoring system. We believe that 3D ultrasonography provides novel information on the visualization of intratumoral structures of the ovarian masses. These results suggest that 3D ultrasonography has the potential to be a supplement to 2D ultrasonography and might be useful in identifying malignant ovarian tumours. However, in view of the small number of ovarian tumours, these observations must be considered preliminary.


    Notes
 
3 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Baba, K., Okai, T. and Kozuma, S. (1996) Real-time processable three-dimensional fetal ultrasound. Lancet, 348, 1307.[ISI][Medline]

Bonilla-Musoles, F., Raga, F. and Osborne, N.G. (1995) Three-dimensional ultrasound evaluation of ovarian masses. Gynecol. Oncol., 59, 129–135.[ISI][Medline]

Chan, L., Lin, W.M., Uerpairojkit, B. et al. (1997) Evaluation of adnexal masses using three-dimensional ultrasonographic technology: Preliminary report. J. Ultrasound Med., 16, 349–354.[Abstract]

DePriest, P.D., van Nagell, J.R., Gallion, H.H. et al. (1993a) Ovarian cancer screening in asymptomatic postmenopausal women. Gynecol. Oncol., 51, 205–209.[ISI][Medline]

DePriest, P.D., Shenson, D.S., Fried, A. et al. (1993b) A morphology index based on sonographic findings in ovarian cancer. Gynecol. Oncol., 51, 7–11.[ISI][Medline]

Gadducci, A., Ferdeghini, M., Prontera, C. et al. (1992) The concomitant determination of different tumor markers in patients with epithelial ovarian cancer and benign ovarian masses. Gynecol. Oncol., 44, 147–154.[ISI][Medline]

Hata, K., Hata, T., Manabe, A. et al. (1992) A critical evaluation of transvaginal Doppler studies, transvaginal sonography, magnetic resonance imaging, and CA 125 in detecting ovarian cancer. Obstet. Gynecol., 80, 922–926.[Abstract]

Hata, K., Hata, T. and Kitao, M. (1995) Intratumoral peak systolic velocity as a new possible predictor for detection of adnexal malignancy. Am. J. Obstet. Gynecol., 172, 1496–500.[ISI][Medline]

Hata, T., Aoki, S., Manabe, A. et al. (1997) Three-dimensional ultrasonography in the first trimester of human pregnancy. Hum. Reprod., 12, 1800–1804.[Abstract]

Jurkovik, D., Jauniaux, E. and Campbell, S. (1994) Three-dimensional ultrasound in obstetrics and gynecology. In Kurjak, A. and Chervenak, F.A. (eds), The Fetus as a Patient. Parthenon Publishing Group, Carnforth, UK, pp. 135–140.

Jurkovic, D., Geipel, A., Gruboek, K. et al. (1995) Three-dimensional ultrasound for the assessment of uterine anatomy and detection of congenital anomalies: a comparison with hysterosalpingography and two-dimensional sonography. Ultrasound Obstet. Gynecol., 5, 233–237.[ISI][Medline]

Kyei-Mensah, A., Pittrof, R., Maconochie, N. et al. (1996) Transvaginal three-dimensional ultrasound: reproducibility of ovarian and endometrial volume measurements. Fertil. Steril., 66, 718–722.[ISI][Medline]

Lerner, J.P., Timor-Tritsch, I.E., Federman, A. and Abramovich, G. (1994) Transvaginal ultrasonographic characterization of ovarian masses with an improved weighted system. Am. J. Obstet. Gynecol., 170, 81–85.[ISI][Medline]

Sassone, M., Timor-Tritsch, I., Arter, A. et al. (1991) Transvaginal sonographic characterization of ovarian disease. Evaluation of a new scoring system to predict ovarian malignancy. Obstet. Gynecol., 78, 70–76.[Abstract]

Serov, S.F., Scully, R.E. and Sobin, L.H. (1973) International Histological Classification of Tumors. No. 9. Histological Typing of Ovarian Tumors. World Health Organization, Geneva, Switzerland.

Sparks, J.M. and Varner, R.E. (1991) Ovarian cancer screening. Obstet. Gynecol., 77, 787–792.[Abstract]

van Nagell, J.R. (1994) Methods to improve the diagnostic accuracy of ultrasound in the detection of ovarian cancer. Gynecol. Oncol., 54, 115–116.[ISI][Medline]

van Nagell, J.R., DePriest, P.D., Puls, L.E. et al. (1991) Ovarian cancer screening in asymptomatic postmenopausal women by transvaginal sonography. Cancer, 68, 458–462.[ISI][Medline]

Weinraub, Z., Maymon, R., Shiman, A. et al. (1996) Three dimensional saline contrast hysterosonography and surface rendering of uterine cavity pathology. Ultrasound Obstet. Gynecol., 8, 277–282.[ISI][Medline]

Submitted on July 9, 1998; accepted on November 30, 1998.