Using virtual reality for evaluation of fetal ambiguous genitalia
Abstract
Objective
The utility of a virtual reality system was examined in the visualization of three-dimensional (3D) ultrasound images of fetal ambiguous genitalia.
Methods
In 2005, fetal ambiguous genitalia were diagnosed in four patients referred to our department for prenatal ultrasound assessment. The patients were examined by two-dimensional (2D) and 3D ultrasound and, subsequently, the volumes obtained on 3D ultrasound were visualized in the BARCO I-Space virtual reality system. This system projects stereoscopic images on three walls and the floor of a small ‘room’, allowing several viewers to see a 3D ‘hologram’ of the data being visualized. The results of 2D and 3D ultrasound examination and the virtual reality images of the I-Space were compared with diagnoses made postpartum.
Results
In all cases, prenatal diagnosis was unclear based on 2D ultrasound alone. Surface rendering of 3D data provided an impression of ambiguity, but diagnosis based on these data proved incorrect at birth in three cases. Conclusions based on the evaluation of 3D volumes in virtual reality best fitted the postpartum diagnosis in all cases.
Conclusions
This study suggests that by evaluation of the genitals in the I-Space, a better impression of genital ambiguity can be established. Binocular depth perception appeared particularly useful in distinguishing either a micropenis or enlarged clitoris from labia minora, since it helps in the estimation of size and position. Therefore, we see potential for the application of virtual reality not only for the evaluation of fetal ambiguous genitalia, but in all those cases where depth perception would improve the visualization of anatomical structures. Copyright © 2008 ISUOG. Published by John Wiley & Sons, Ltd.
Introduction
Determination of fetal sex is an important aspect of prenatal ultrasound assessment. Advances in two- and three-dimensional (2D and 3D) ultrasound technology have improved the visualization of fetal genitalia, allowing identification of fetal sex as early as the end of the first trimester1, 2. Any uncertainty or doubt about a baby's sex is extremely worrying and unsettling for its parents and family, but the assessment and diagnosis of fetal genital anomalies remain challenging and difficult tasks for ultrasonographers.
The objective of this study was to assess the clinical potential of a virtual reality technique by visualizing fetal genitalia that were considered abnormal or ambiguous on routine ultrasound scan using an immersive virtual reality system, the BARCO I-Space (Barco, Kortrijk, Belgium). In the I-Space, binocular depth perception provides the investigator with a realistic 3D illusion that allows much better assessment of 3D images such as fetal genitalia as compared to conventional 3D rendering in surface mode on a workstation.
Materials and Methods
I-Space: volume visualization in a virtual environment
The BARCO I-Space installed at the department of bioinformatics of the Erasmus MC is a four-walled CAVE™-like virtual reality system3 that surrounds investigators with computer-generated stereoscopic images (Videoclip S1 online). These images are projected on three walls and the floor of a small ‘room’, by eight high-quality Digital Light Processing (DLP) projectors. The images need to be viewed through glasses with polarizing lenses to perceive depth. The CAVORE4 volume rendering application is used to create a ‘hologram’ of the ultrasound volume that is being investigated, which can then be manipulated by means of a virtual pointer, controlled by a wireless joystick. Wireless tracking of the viewer's head allows the computer to provide the correct perspective and motion parallax, which in addition to the stereoscopic images helps in discerning fine details and understanding of 3D structures in the volumes. The CAVORE software was originally developed as a general-purpose volume rendering application for use in immersive virtual environments such as the CAVE. It uses a combination of direct manipulation of the data set with a pointing device with six degrees of freedom and a simple graphical user interface with a drop-down menu and a ‘widget’ (a graphical object that can be manipulated by the user) to control the transfer function that assigns gray-scale and opacity values to the data.
Methods
Between January 2005 and May 2006 ambiguous genitalia were diagnosed as an isolated or additional finding in the fetuses of four women referred to our department of obstetrics and prenatal medicine for prenatal assessment. 2D ultrasound examination of all suspected fetal abnormalities was performed on a GE Voluson 730 Expert system (GE Medical Systems, Zipf, Austria). For the evaluation of the genital area we also performed 3D ultrasound examinations, but we used only 3D ultrasound rendering in surface mode (not the sectional plane mode). All fetal malformations in our department are discussed by a team of doctors and obstetricians, including ultrasonographers, but after reviewing the data from examination of the four cases under consideration, analysis of the genital area was still unsatisfactory. To demonstrate the potential of the Barco I-Space in analyzing genital abnormalities, the 3D volumes were transferred to a personal computer for off-line evaluation using specialized 3D software (4D view, GE Medical Systems), without any image processing. These data were then saved as Cartesian (rectangular) volumes and transferred to the BARCO I-Space. In the I-Space, the volumes were screened for quality, completeness and visibility of the genital area. Volumes were resized and turned around in space, and gray scale and opacity were adjusted. To obtain the best view of the genital area the hologram was directly manipulated by cutting away part of it using the virtual pointer.
Results
Patient 1
A 27-year-old woman, gravida 2 para 1, was referred at 20 weeks' amenorrhea because of a suspected fetal cardiac anomaly. Amniocentesis was performed at that time and karyotyping revealed 46,XY, although ultrasound examinations at 20 and 27 weeks' gestational age had revealed normal female genitalia. Knowing the karyotype, the genitals were considered ambiguous on ultrasound examination at 34 weeks. On 3D surface rendering a bifid scrotum seemed most likely, with a small central structure in between, possibly a micropenis. In the I-Space large labia majora were clearly recognizable with a prominent structure between them, which most resembled labia minora. An enlarged clitoris could be neither excluded nor confirmed (Figure 1, Videoclip S2 online).
Three-dimensional (3D) image (a) and I-Space image (b) of the genital area of Patient 1. On the 3D image a bifid scrotum with a central micropenis was diagnosed, whereas in the I-Space the labia minora are clearly recognizable. (A comparison of the two techniques from these images is not valid, since a picture of the I-Space is nothing more than a 3D picture on a two-dimensional medium.).
At 38 weeks the infant was delivered by Cesarean section. Examination of the external genitals revealed normally appearing female genitalia: the labia majora were normal, there were no palpable gonads, and the clitoris seemed enlarged (1.5 cm) with a small palpable invagination. The introitus vaginae appeared normal. Ultrasound examination of the internal genitalia revealed normal aspects of the uterus and ovaries, with follicles visualized on both sides. The XY karyotype was confirmed and the infant was diagnosed as an XY-female5.
Furthermore, the child was diagnosed with a hypertrophic cardiomyopathy, ventricular septal defect, atrial septal defect and hypoplasia of the right lung. Several cerebral malformations (hypomyelinization) were also found. The infant died at the age of eight months. Several syndromes and metabolic diseases were considered, but a definite diagnosis was not possible and the parents refused autopsy.
Patient 2
A 23-year-old woman, gravida 3 para 1, was referred because of suspected fetal pyelectasia, polyhydramnios and ambiguous genitalia at 30 weeks' amenorrhea. Amniocentesis had already been performed and karyotyping revealed 46,XX. A series of ultrasound examinations confirmed polyhydramnios, a single-artery umbilical cord, two large cysts in the fetal pelvis and ambiguous genitalia; there was a prominent structure in between two skin creases. In the I-Space two small skin creases were visible in the genital area, with a prominent, protruding, blunt structure on top resembling a penis with hypospadias (Figure 2).
Detailed I-Space image of the genital area of Patient 2, showing the blunt protruding structure.
At 42 weeks the infant was born and several congenital anomalies were observed, including esophageal atresia with a trachea-esophageal fistula, anomalies of the S1–S2 vertebrae, urogenital sinus, hydronephrosis on both sides and ambiguous genitalia: labia majora, labia minora and a micropenis were observed. The anatomy of the urinary tract could not be distinguished as male or female and therefore it was concluded that it was probably an unusual variation of a cloacal malformation. The infant was diagnosed with VACTERL syndrome.
Patient 3
A 30-year-old woman, gravida 2 para 1, was referred because of a suspicion of hypospadias at 32 weeks' amenorrhea. 2D ultrasound examination revealed ambiguous genitalia. Severe hypospadias with micropenis and a bifid scrotum seemed most likely, but hypertrophy of the clitoris could not be excluded. Karyotyping following amniocentesis revealed 46,XX. On 3D surface rendering mode two skin creases were visualized with a structure in between, the most likely diagnosis being an enlarged clitoris between the labia majora. In the I-Space, normal labia majora were found, with somewhat enlarged labia minora (Figure 3).
(a) Three-dimensional (3D) ultrasound surface-rendered image of the genital area of Patient 3. The skin creases were interpreted as labia majora and the clitoris was thought to be enlarged. (b) I-Space image in which the labia minora are quite obvious. The operator (C.M.V.-D.) is visible holding a joystick, while manipulating the fully rendered 3D image.
At 41 weeks a healthy infant was delivered. Postpartum, the genitals were confirmed as female, with no evident abnormalities.
Patient 4
A 25-year-old woman, gravida 2 para 0, was referred because of suspected genital abnormalities at 32 weeks' amenorrhea. 2D ultrasound examination revealed ambiguous genitalia. On 3D surface-rendering mode ultrasound labia majora with an enlarged clitoris seemed most likely. The adrenal glands were not enlarged. The parents refused amniocentesis for karyotyping. In the I-Space normal labia majora were visualized, with markedly enlarged labia minora. The clitoris did not seem enlarged (Figure 4).
(a) Three-dimensional ultrasound image of the genital area of Patient 4, showing the labia majora with an enlarged clitoris. (b) I-Space rendering image, showing the enlarged appearance of the labia minora; in virtual reality this was even more clearly recognizable than it is in this two-dimensional representation.
At 40 weeks the patient delivered a healthy infant. The genital area appeared female, with somewhat enlarged labia minora, though the clitoris was not markedly enlarged. Ultrasound examination showed normal internal genital organs. A normal 46,XX karyotype was found.
Table 1 shows the diagnoses of the genitalia following the different examinations.
| Patient | Referral diagnosis | Classification on: | Karyotype | Postpartum diagnosis | ||
|---|---|---|---|---|---|---|
| 2D US | 3D US | I-Space | ||||
| 1 | Female/ambiguous | Female | Ambiguous | Female | 46,XY | XY female, normal female genitalia |
| 2 | Ambiguous | Ambiguous | Ambiguous | Ambiguous | 46,XX | VACTERL, ambiguous genitalia, cloacal malformation |
| 3 | Male hypospadias | Ambiguous | Clitoral hypertrophy | Normal female | 46,XX | Normal female |
| 4 | Ambiguous | Ambiguous | Clitoral hypertrophy | Normal female | 46,XX | Normal female |
- 2D, two-dimensional; 3D, three-dimensional; US, ultrasonography.
Discussion
Prenatal diagnosis of fetal gender by ultrasound generally has a high rate of accuracy and is a well-established part of routine ultrasonography6. However, when fetal genitalia appear malformed or ambiguous, or gender assigned by ultrasound does not match gender by karyotype, a plethora of syndromes must be considered and the limitations of ultrasound imaging in observing the external contours of the genitals become obvious.
In this study we tried to establish the clinical potential of a virtual reality technique in the assessment of fetal ambiguous genitalia after unsatisfying analysis of these malformations in four cases using 2D and 3D ultrasound. Although these scans were performed by senior ultrasonographers with more than 10 years of experience in the detection of fetal abnormalities, definite diagnoses of the genital malformations remained impossible. This study shows that by evaluation of the genitals in the I-Space, a better impression of ambiguity can be established. Depth perception appeared particularly useful in distinguishing either micropenis or enlarged clitoris from labia minora, since depth helps estimation of size and exact position. Although the numbers are small, we decided to present our results because the findings in the I-Space turned out to be extremely helpful in prenatal diagnosis. We learned that a protruding structure in the genital area can very well represent (normal) labia minora instead of malformations such as an enlarged clitoris or micropenis. This has already benefited the counseling of several patients seen after we conducted this study.
Using 3D ultrasound with conventional imaging systems still has the disadvantage that evaluation of the 3D ultrasound image is in fact performed in a 2D manner either on a print or on a computer screen. Therefore, the third dimension is not used optimally, since depth perception is not possible. In the I-Space the investigator views the equivalent of a hologram floating in space, thus benefiting from depth perception. When comparing the images of the 3D volumes with the images of the volumes in the I-Space it is important to remember the limitations of such comparisons, since the third dimension cannot be put on paper. Hence, the image of the volume in the I-Space shows neither all the information gathered from the volume in the I-Space itself nor all its beneficial aspects.
Of course, this study has the limitation that we only used surface-mode rendering. Evaluating the 3D volumes in various modes, such as the sectional plane mode, enhances the information acquired from the volume. However, in this study we provided a valid comparison, since we evaluated only the surfaces of both 3D and I-Space volumes.
The CAVORE software can be used to visualize many different 3D imaging modalities, such as computed tomography, magnetic resonance imaging, single photon emission computed tomography, positron emission tomography and, as in this study, 3D ultrasound. It can also animate time series of 3D data, also known as 4D data. This application has already been found effective and to have additional value in visualizing a fetal meningomyelocele7 and dynamic 3D adult echocardiographic data8, 9. In order to be integrated into daily clinical practice, there is a need to implement semi-immersive (such as GeoWall http://geowall.geo.lsa.umich.edu]) or desktop (such as the Personal Space Station (http://www.ps-tech.com)) virtual reality systems in the hospital environment, for instance in consulting rooms and operating theaters. The user interface of CAVORE has already been adapted to allow it to be run on this type of virtual reality system.
While a complete I-Space system at this time may cost in the region of $ 500 000, for this type of application a virtual reality system with a single projection surface would be sufficient. These can be put together for around $ 20 000–50 000, depending on projector and tracking choices. A potential side-effect of virtual reality might be that the systems using head-mounted displays (virtual reality ‘helmets’) are associated with ‘simulation sickness’, i.e. nausea and/or dizziness similar to motion sickness. However, users of projection-based systems, and in particular the operator who views the correct perspective, suffer far less from this problem.
We conclude that there is potential for the use of the virtual reality technique in evaluating difficult anatomical structures where depth perception would improve visualization, as is the case with fetal ambiguous genitalia.
SUPPORTING INFORMATION ON THE INTERNET
The following supporting information may be found in the online version of this article:
Videoclip S1 Overview of the Barco I-Space virtual reality system. The operator (A.H.J.K.) demonstrates a fetal face of 33 weeks' amenorrhea in the I-Space.
Videoclip S2 I-Space rendering of the genital area of Patient 1.
