Three-Dimensional Sonography of the Endometrium and Adjacent Myometrium

Transcription

1 Technical dvance Three-Dimensional Sonography of the Endometrium and djacent Myometrium Preliminary Observations Rochelle F. ndreotti, MD, rthur C. Fleischer, MD, Lawrence E. Mason, Jr, MD Objective. y evaluating a series of patients undergoing pelvic sonography with routine 2-dimensional (2D) as well as 3-dimensional (3D) reconstructed images in the coronal plane, we attempted to characterize the types of additional information that can be obtained. Methods. Ninety randomly selected patients undergoing transvaginal pelvic sonography were imaged according to a standard 2D protocol. 3D uterine volume was then acquired in the sagittal plane and reconstructed in the coronal plane. The endometrium and surrounding myometrium were evaluated for architecture, masses, the relationship of masses to the endometrial cavity, and the anatomic configuration of the cavity. Results. Ninety-one studies were obtained. dditional findings were obtained on the coronal view in 28 studies (30.8%). No additional findings were obtained in 63 studies (69.2%). Normal endometrial and myometrial findings were obtained by conventional 2D imaging in 42 of 91 patients. Of this group, additional findings were shown in 2 (5%) patients. Forty-nine of the 91 patients had abnormal findings by 2D imaging. dditional information was obtained in 26 (53%) of these patients. dded information included uterine anomalies, better definition of the endometrium, more accurate delineation and location of endometrial polyps, location of leiomyomas, visualization of cystic areas within the myometrium, and confirmation of the location of intrauterine devices. Conclusions. The 3D reconstructed view of the endometrium and adjacent myometrium appears to be most helpful after a conventional transvaginal study, showing abnormalities within the endometrium and myometrium but being of little added benefit if the conventional findings are normal. Key words: coronal plane; endometrium; 3-dimensional sonography; uterus. bbreviations IUD, intrauterine device; 3D, 3-dimensional; 2D, 2-dimensional Received pril 26, 2006, from the Departments of Radiology and Radiological Sciences (R.F..,.C.F., L.E.M.) and Obstetrics and Gynecology (R.F..,.C.F.), Vanderbilt University Medical Center, Nashville, Tennessee US. Revision requested May 4, Revised manuscript accepted for publication May 11, We thank the sonographers at Vanderbilt University Medical Center for their technical contributions and Vera. Merriweather for editorial assistance. ddress correspondence to Rochelle F. ndreotti, MD, Department of Radiology and Radiological Sciences, Vanderbilt University Medical Center, st ve S, Nashville, TN US. rochelle.f.andreotti@vanderbilt.edu Traditional views of the uterus and endometrium on 2-dimensional (2D) sonography are usually limited to sagittal and transverse images of the uterus. This is due to inherent limitations of the scan planes that can be obtained with transabdominal and transvaginal transducers and the limited mobility of the transvaginal probe during transvaginal sonography. With volume imaging, any desired plane through the uterus can be obtained regardless of the original scan plane of acquisition so that the uterus and endometrium can now be imaged in the coronal plane. Investigators have shown additional findings on the coronal images that cannot be appreciated by traditional views. This view is essential for evaluating the uterine 2006 by the merican Institute of Ultrasound in Medicine J Ultrasound Med 2006; 25: /06/$3.50

2 3D Sonography of the Endometrium and djacent Myometrium contour for anomalies. Raga et al 1 correctly characterized the external configuration of the uterus with 3-dimensional (3D) sonography as determined at laparoscopy in 41 of 42 patients. Wu et al 2 showed sensitivity and specificity for 3D sonography of 100% for diagnosing an anomalous uterus using laparoscopy and hysteroscopy for final diagnosis. Visualization of intrauterine devices (IUDs) within the endometrial cavity in the coronal plane has been evaluated. Lee et al 3 reported the complete visualization of all parts of Tcu380 IUDs in 95% of 96 cases using the coronal plane. The relationship of abnormalities to the endometrial cavity and the assessment of the endometrium for focal changes in echo texture have also been under investigation. 4 Most of these investigators have focused on the use of volume imaging of the endometrial cavity during sonohysterography. 5 7 lthough 3D volume imaging techniques have become routine in computed tomography and magnetic resonance imaging, and their clinical value is appreciated, it has only been more recently that these techniques have become available in the field of sonography. With 3D sonography, a volume rather than a single image plane is acquired, stored, reformatted, and analyzed, showing any arbitrary plane. Using a multiplanar display, 3 perpendicular planes are displayed simultaneously, and correlation among these planes can be obtained. y evaluating a series of patients undergoing pelvic sonography with routine 2D as well as 3D reconstructed images of the endometrium in the coronal plane, we attempted to further characterize the types of additional information shown in this plane and the frequency with which one may expect to obtain additional findings. Given this information, the answer to the question of whether this should be a routine addition to the transvaginal sonographic examination may become more apparent. Materials and Methods Ninety consecutively referred menstrual or postmenopausal patients undergoing transvaginal pelvic sonography at the ultrasound department of Vanderbilt University Medical Center who were randomly assigned to systems with 3D imaging capabilities were scanned with standard 2D and newly developed 3D protocols by experienced sonographers using either a 3D 9-3 MHz broadband transvaginal transducer (IU-22; Philips Medical Systems, othell, W) or an 8-4 MHz broadband transvaginal transducer (HDI 5000; Philips Medical Systems). fter the 2D examination, a 3D volume was obtained through the uterus by either an automated or a manual sweep in the sagittal plane. The manual sweep requires freehand movement of the probe as opposed to the automated method, which, when activated, automatically sweeps through the selected area of interest while the transducer is held stationary. The digitally stored volume is then presented in a multiplanar display showing 3 perpendicular planes through the volume: the sagittal, transverse, and reconstructed coronal views. y scrolling through the reconstructed coronal view, multiple images were obtained by the sonographer through the anteroposterior thickness of the endometrium. Images were reviewed retrospectively by experienced sonologists. The endometrium and surrounding myometrium were evaluated for architecture, masses, the relationship of masses to the endometrial cavity, and the anatomic configuration of the cavity. The images were subjectively analyzed and separated into those that showed better delineation of findings versus those in which findings were similar. dditional information obtained in the 3D reconstructed image was divided into multiple categories, which included uterine anomalies, definition of the endometrium, delineation and location of endometrial polyps, location of leiomyomas, cystic areas within the myometrium, and location of IUDs. To accurately assess the uterine contour and shape of the endometrial cavity, careful evaluation of the perpendicular planes was performed by the sonographer to ensure that a true midcoronal view of the uterus was obtained. Otherwise, a false impression of an arcuate uterus may have been given. Results Of the 90 patients imaged by transvaginal sonography, 91 studies were obtained. One patient who had menorrhagia was reevaluated at the request of the referring physician 3 weeks after the initial evaluation by 2D and 3D imaging yielded normal findings. dditional findings were obtained in 28 studies (30.8%). Two additional findings were shown in 3 of the cases. No additional findings were obtained in 63 studies (69.2%). Normal find J Ultrasound Med 2006; 25:

3 ndreotti et al ings relating to the endometrium and myometrium were obtained by conventional 2D imaging in 42 of 91 patients. Of this group, additional findings were shown in 2 (5%) patients. In 1 of these cases, an IUD location was confirmed, and in the other case, an arcuate uterus was identified. Forty-nine of the 91 patients had abnormal findings by 2D imaging. bnormalities seen on conventional images included leiomyomas (12 cases), a heterogeneous myometrium (10 cases), suggested endometrial polyps (5 cases), a poorly delineated endometrium (4 cases), thickness of the endometrium greater than 14 mm in a menstruating patient and 5 mm in a postmenopausal patient (4 cases), fluid within the endometrial cavity (4 cases), and uterine anomalies (4 cases). It was common for patients to have more than 1 of these abnormalities on the same study. dditional findings were obtained with the 3D reconstructed image in 26 (53%) of these patients. dded information included uterine anomalies, better definition of the endometrium, more accurate delineation and location of endometrial polyps, location of leiomyomas, visualization of cystic areas within the myometrium, and confirmation of the location of IUDs (Table 1). On reevaluation of the patient who underwent 2 studies in this series, the initial study showing normal findings, the additional finding of a polyp was suggested on the 2D images and confirmed on the coronal reconstruction. Table 1. Characterization of dditional Information by 3D Reconstructed Images (Total n = 91) No. of % of dded Information Studies Total Uterine anomaly 8 9 etter delineation of endometrium 7 8 Delineation and location of polyps 6 7 Location of leiomyomas 5 5 Cystic areas in myometrium 3 3 Confirmation of IUD location 2 2 Uterine anomalies were shown in 8 studies. Of these 8 studies, 7 were considered to represent arcuate uteri because the thickness of the myometrial septum projecting within the cavity was measured as less than 1 cm (Figure 1). In only 1 case was a larger septum shown, consistent with a true septate uterus (Figure 1). Differentiation among a septate, an arcuate, or a bicornuate uterus by 2D imaging was inconclusive in this patient, and it was only on the coronal images that the endometrial cavity and surrounding myometrium could be delineated. etter definition of the endometrium was observed in 7 cases. These cases were divided into several categories. There were 3 studies in patients with a poorly defined endometrium by 2D imaging, likely due to adenomyosis, in which better delineation of the endometrium was shown on the coronal reconstruction. In 1 study, Figure 1. Three-dimensional reconstructed images in the coronal plane showing a normal uterine contour., Small septum extending into the fundal endometrial cavity, consistent with an arcuate uterus., Larger septum measuring greater than 1 cm, consistent with a septate uterus. J Ultrasound Med 2006; 25:

4 3D Sonography of the Endometrium and djacent Myometrium the location of cysts in either the endometrium or the adjacent myometrium could not be determined on the 2D images but was confirmed within the endometrium on the coronal reconstructed image. This was confirmed on a surgical pathologic specimen. nother patient had a history of surgical removal of an endometrial polyp that had been diagnosed on a prior transvaginal sonogram (Figure 2). The study obtained in our series showed a thin endometrium on routine 2D sonography. However, the reconstructed coronal image clearly showed an oval hypoechoic defect in the region of the previously noted endometrial polyp, consistent with a scar. One case showed a minimal quantity of fluid within the endometrial cavity on the coronal reconstructed view not seen on 2D views. The coronal view was able to better delineate and show the location of endometrial polyps in 6 studies. In 2 cases, only a heterogeneous, thickened endometrium was seen on 2D sonography. mass was delineated within the cavity on coronal reconstructions. One of these studies showed fluid surrounding the polyp. In 2 cases, the presence of a polyp was questioned on the 2D images and confirmed on the coronal reconstructions (Figure 3). More numerous polyps were shown on the coronal views in 2 studies. In 1 of these studies, the location of 2 other polyps was confirmed. Pathologic correlation by curettage, polypectomy, or hysterectomy was obtained in 3 of the 6 cases. The location of 1 or more leiomyomas was better appreciated on the coronal reconstructed image in 5 studies. In 4 of these studies, a submucosal component was shown on the 3D reconstructed image in the coronal plane, which was suggested by conventional imaging because of the proximity of the mass to the endometrium but not confirmed. In 1 case, sonohysterography was performed to confirm the submucosal location of the leiomyoma. The coronal view was able to show an intramural location with no submucosal component of a uterine leiomyoma in 1 case in which the 2D images were inconclusive. In 3 studies, cystic areas were shown within the myometrium on the 3D reconstructed coronal images that were not appreciated or were poorly Figure 2. Endometrial scar after polypectomy., Thin endometrial bilayer on 2D sagittal image of the uterus., Three-dimensional coronal reconstruction showing a hypoechoic defect in the fundal endometrium, consistent with a scar. C, Conventional sagittal 2D sonography of the endometrium before polypectomy showing a polyp within the fundal cavity (arrow). C 1316 J Ultrasound Med 2006; 25:

5 ndreotti et al Figure 3. etter delineation and demonstration of the location of an endometrial polyp., Two-dimensional imaging in the sagittal plane showing the questionable presence of an endometrial polyp (arrow)., n endometrial polyp is welldelineated in the left corneal area on the coronal reconstruction. visualized on the 2D images. ll 3 of these patients had other myometrial findings by 2D imaging, which were suggestive of adenomyosis. Surgical pathologic specimens were not obtained in any of these 3 patients. The confirmation of IUD location was shown on the coronal view in 2 cases (Figure 4). oth IUDs had the configuration of a copper T IUD and could be seen in their entirety only on the reconstructed coronal image and not on traditional 2D sagittal and transverse images. lthough the long arm of the IUD could be seen within the mid cavity by 2D imaging, the short arms of the IUD within the uterine cornu were only seen on the coronal view, ruling out any myometrial involvement in the patients, who both had pelvic pain. Discussion The coronal image of the endometrium and adjacent myometrium is an easily obtainable adjunct to any conventional transvaginal pelvic sonogram. Within less than 1 minute, a sweep through the midsagittal plane of the uterus can be obtained, producing the multiplanar display. The resulting coronal plane is reconstructed at the level of the endometrium. One is then able to Figure 4. Demonstration of IUD location within the endometrial cavity., Two-dimensional sagittal image showing obscuration of the endometrium by the presence of the IUD., Three-dimensional coronal reconstructed image showing the entire IUD within the cavity, confirming its location. J Ultrasound Med 2006; 25:

6 3D Sonography of the Endometrium and djacent Myometrium quickly scroll through the volume, finding the appropriate images that show endometrial and myometrial findings. In our series, the sonographer rather than the radiologist had the responsibility of scrolling through the 3D volume and choosing pertinent images. We thought that this would be the more realistic scenario in most institutions because a separate workstation where the radiologist can manipulate data is often financially impractical, and the volume is usually reconstructed on the ultrasound system. This is most conveniently performed at the time of the examination by the sonographer. dditional information was obtained on the reconstructed coronal image in 31% of the total studies and 53% of the studies with abnormal uterine or endometrial findings. The most common additional finding was a uterine anomaly in 9% of cases, with most being arcuate uteri. ecause this finding is widely considered a variation of normal, whether this adds anything clinically is debatable. However, a recent study by Woelfer et al 8 has shown that women with an arcuate uterus had a significantly greater risk of second-trimester loss and preterm labor than those with a normal uterine configuration. This information may then be important in the management of subsequent pregnancies. Given these results, the answer to the question of whether the coronal view should be a routine addition to the transvaginal sonogram obviously would seem positive. However, there were additional findings on the reconstructed coronal image in only 5% of studies with normal findings. In view of the paucity of additional information obtained in this subgroup of patients, it seems highly unlikely that the supplementary reconstructed image would be of added benefit after normal 2D findings. Technical limitations of this study included the lack of radiologist participation in the selection of reconstructed images produced by scrolling through the volume of data, the loss of spatial resolution in the coronal plane, and the lack of control of the phase of the menstrual cycle. lthough the additional time and expertise of the radiologist may add even more findings to the study, this is not a cost- or time-efficient approach unless there is a separate workstation to manipulate the data volume. Furthermore, there is a substantial loss of spatial resolution in the coronally reconstructed image because of its distance from the original sagittal plane. s a result, some of the information seen on the 2D sagittal and transverse images may be lost in the reconstructed image. The random nature of patient selection in this series made it difficult to control for the phase of the menstrual cycle. ecause of contrasting echogenicity of the endometrium and associated abnormalities that may be observed during a specific phase of the menstrual cycle, there is a potential advantage of selecting the more beneficial time period. For example, an endometrial polyp is echogenic and will be best visualized against a more hypoechoic endometrium seen in the proliferative phase, whereas uterine leiomyomas, which tend to be hypoechoic, will be seen more clearly adjacent to an echogenic, secretory endometrium. dditionally, there were a limited number of pathologic or radiologic correlations of findings. In cases without pathologic or radiologic correlations, we cannot be entirely sure of the correctness of our 3D findings. However, 3D transvaginal sonography with reconstructions in the coronal plane has been substantiated in the literature as a confirmatory method of evaluating IUD location and uterine anomalies so that other types of correlations are probably unnecessary. 1 4 Pathologic correlation was obtained in 3 of 6 cases of endometrial polyps, and a sonohysterogram was obtained, confirming a submucosal location in 1 of 5 cases of leiomyomas associated with the endometrial cavity. Correlating the distinctness of the endometrium is probably not feasible in most cases. Cysts within the endometrium were confirmed on a surgical pathologic specimen in 1 study. What is most likely a scar within the endometrium after polypectomy is also supported by a prior transvaginal sonogram that showed the location of the polyp. dmittedly, our conclusions are preliminary and await larger series that have further pathologic proof. Three-dimensional sonography of the endometrium and adjacent myometrium using reconstructions in the coronal plane is easily obtainable on a routine basis. The extra view appears to be most helpful after a conventional transvaginal study that shows abnormalities within the endometrium and myometrium, and additional findings were seen in more than 50% of the patients in this study. Conversely, if the findings from the conventional study of the endometrium and myometrium are normal, it is unlikely that the 3D study will be of any added benefit J Ultrasound Med 2006; 25:

7 ndreotti et al References 1. Raga F, onilla-musoles F, lanes J, Osborne NG. Congenital mullerian anomalies: diagnostic accuracy of three-dimensional ultrasound. Fertil Steril 1996; 65: Wu MH, Hsu CC, Huang KE. Detection of congenital mullerian duct anomalies using three-dimensional ultrasound. J Clin Ultrasound 1997; 25: Lee, Eppel W, Sam C, Kratochwil, Deutinger J, ernaschek G. Intrauterine device localization by threedimensional transvaginal sonography. Ultrasound Obstet Gynecol 1997; 10: ega G, Lev-Toaff S, O Kane P, ecker E, Kurtz. Threedimensional ultrasonography in gynecology: technical aspects and clinical applications. J Ultrasound Med 2003; 22: Sylvestre C, Child TJ, Tulandi T, Tan SL. prospective study to evaluate the efficacy of two- and three-dimensional sonohysterography in women with intrauterine lesions. Fertil Steril 2003; 79: Lev-Toaff S, Pinheiro LW, ega G, Kurtz, Goldberg. Three dimensional multiplanar sonohysteropgraphy and x- ray hysterosalpingography. J Ultrasound Med 2001; 20: La Torre R, De Felice C, De ngelis C, Coacci F, Mastrone M, Cosmi EV. Transvaginal sonographic evaluation of endometrial polyps: a comparison with two dimensional and three dimensional contrast sonography. Clin Exp Obstet Gynecol 1999; 26: Woelfer, Salim R, anerjee S, Elson J, Regan L, Jurkovic D. Reproductive outcomes in women with congenital uterine anomalies detected by three-dimensional ultrasound screening. Obstet Gynecol 2001; 98: J Ultrasound Med 2006; 25: