ª Springer Science+Business Media, LLC 2008 Abdom Imaging (2009) 34:121–133 Abdominal Published online: 15 August 2008 DOI: 10.1007/s00261-008-9432-4 Imaging

64-Row multidetector CT virtual hysterosalpingography Patricia Carrascosa,1 Carlos Capun˜ay,1 Mariano Baronio,2 Elba Martı´nLo´pez,1 Javier Vallejos,1 Mario Borghi,2 Carlos Sueldo,2 Sergio Papier2 1Department of Computed Tomography, Diagno´stico Maipu´, Av. Maipu´1668, Vicente Lo´pez (B1602ABQ), Buenos Aires, Argentina 2CEGYR, Ciudad Auto´noma de Buenos Aires, Buenos Aires, Argentina

Abstract Key words: Multidetector computed tomography—Hysterosalpingography—Virtual Objective: To illustrate the large variety of pathologies studies—Virtual hysterosalpingography— found on 64-row computed tomographic (CT) virtual Uterus—Fallopian tubes hysterosalpingography (CT-VHSG) in the evaluation of the female reproductive tract in infertile patients. Material and methods: We prospectively evaluated CT Virtual hysterosalpingography (VHSG) is a recent de- datasets from 209 patients with diagnosis of infertility. scribed, non-invasive modality to assess the uterus and CT-VHSG was performed with a 64-row CT scanner 2 female reproductive system based on volumetric com- using 64 9 0.625 mm collimation and 0.9 mm slice puted tomographic (CT) data. First evaluations per- thickness. A total volume of 20 mL of an iodine contrast formed using 16-row multidetector CT (MDCT) dilution was injected into the uterine cavity. The dura- scanners were promising, with good results for identifi- tion of the CT scan and the grade of patient discomfort cation of uterine pathologies but with limitations for the of the procedure were documented. Images were ana- fallopian tubes evaluation [1, 2]. The introduction of 64- lyzed on a workstation. The CT-VHSG exams were row CT scanners enables isotropic spatial resolution, divided in studies with normal or pathological findings; thinner collimation, better image quality, as well as pathologies were classified according to their locations temporal and contrast resolutions, ensuring a better (cervical, uterine, and fallopian tube pathology). voxel profile [3]. All these qualities give a significant Results: No complications occurred during the proce- improvement in the visualization and assessment of fal- dure. The mean scan time was 3.4 ± 0.4 s; the mean lopian tubes, one of the major goals in the evaluation of patient effective dose was 2.58 ± 0.75 mSv. In relation to the infertile patients [4]. the patient’s discomfort evaluation, 55.5% of the patients The purpose of the present study was to illustrate the referred no discomfort during the procedure. Cervical large variety of pathologies found on 64-row CT-VHSG pathology was found in 100/209 patients, uterine pathol- in the evaluation of the female reproductive tract in ogy in 93/209 patients, and fallopian tubes pathology in infertile patients. 37/209 patients. Conclusion: A 64-row CT-VHSG provides a reliable, Material and methods non-invasive alternative diagnostic technique in the infertility workup algorithm. Patients We prospectively evaluated CT datasets from 209 patients (mean age 34.9 ± 4.7 years old) with diagno- sis of primary or secondary infertility. All patients were studied in the follicular phase of the cycle. The study protocol was approved by the Institutional Review Board. Correspondence to: Patricia Carrascosa; email: investigacion@diag- nosticomaipu.com.ar 122 P. Carrascosa et al.: 64-MDCT Hysterosalpingography

Fig. 1. Cervical folds thickening. (A) Sagittal MIP image. The arrow marks the thick folds. (B) and (C) Virtual endoluminal views.

MDCT virtual hysteroscopic technique 2. Multiplanar reconstructions (MPR): this reconstruc- tions consisted in bi-dimensional images in different The CT-VHSG was performed with a 64-row CT scanner planes (coronal, sagittal, oblique), as well as curved- (Brilliance 64; Philips Medical Systems, Highland 2 MPRs unfolding the uterus, using soft tissue window. Heights, OH) using 64 9 0.625 mm collimation, 3. Maximum intensity projections (MIP): these projec- 0.9-mm slice thickness, a reconstruction interval of tions constituted the third step in the evaluation 0.45 mm, 120 kV, 200 mAs, and rotation time of 0.4 s. processes. This program calculates and displays two- The patient was positioned on the CT table in a supine and three-dimensional projections of the highest position. The perineum was cleansed with povidone– intensity pixels. iodine solution and draped with sterile towels. A sterile 4. Volume rendering: these three-dimensional recon- speculum was inserted into the vagina and the cervix was structions of the entire pelvis show the relation of the cleansed with povidone–iodine solution. A 5-French reproductive system with the pelvis. Complementary plastic cannula was positioned through the external cine images were created. cervical os. No cervical clamping was performed. A total 5. Virtual endoscopy display mode: this was the last step volume of 20 mL of an iodine contrast dilution [4 mL of in the evaluation. This algorithm permits to create water-soluble iodine contrast (iobitridol; Xenetix 350, images similar to those obtained with hysteroscopy, Laboratorios Temis Lostalo´, Buenos Aires, Argentina) giving an accurate delineation of the lesions. and 16 mL of saline solution] was injected into the uterine cavity. CT examinations included acquisition of an anteroposterior scout view followed by acquisition of Imaging findings axial images, ~5 s after the injection of the contrast solution. The CT-VHSG exams were divided in studies with nor- The duration of the CT scan and the grade of patient mal or pathological findings. Pathologies were classified discomfort of the procedure were documented. The patients according to their locations (cervical, uterine, and fallo- completed a questionnaire about the discomfort during the pian tube pathology). procedure, using a 4-step scale: no discomfort; mild dis- comfort; moderate discomfort; and severe discomfort. Results No complications occurred during the procedure. The Image analysis mean scan time was 3.4 ± 0.4 s; the mean patient effec- tive dose was 2.58 ± 0.75 mSv. In relation to the Images were sent to a workstation (Extended Brilliance patient’s discomfort evaluation, 116 patients (55.5%) Workspace; Philips Medical Systems) and reprocessed referred no discomfort during the procedure, 54 (25.8%) using different algorithms: mild discomfort, 33 (15.8%) moderate discomfort, and 1. Axial images: this was the first step in the evaluation. only 6 cases severe discomfort (2.9%). Original transverse CT images allowed an integral A total of 209 patients were studied. CT-VHSG was evaluation of the female reproductive organs. The interpreted as normal in 45/209 patients (21.5%). cervix, uterus, and fallopian tubes were analyzed with Cervical pathology was found in 100/209 patients a soft tissue window. (47.9%) including wall irregularities (n = 43), folds P. Carrascosa et al.: 64-MDCT Hysterosalpingography 123

Fig. 2. Cervical polyp. (A) and (B) Oblique coronal MIP and volume rendering images showing a polyp in the cervical canal (arrows). (C) and (D) Virtual endoluminal views. thickening (n = 22), cervical polyps (n = 16), glandular (n = 18), hydrosalpingx (n = 21), and intratubal syn- dilatations (n = 12), cervical stenosis (n = 10), and echiae (n = 2) (Figs. 12 and 13). synechiae (n = 3) (Figs. 1–3). Ovarian cysts were diagnosed in 27/209 patients Uterine findings were observed in 93/209 patients (12.9%). (44.5%) and included wall irregularities (n = 10), diver- ticulae (n = 9), cesarean scars (n = 6), adenomyosis Discussion (n = 11), submucosal, intramural, and subserosal leiomyomas (n = 30), synechiae (n = 9), endometrial Current imaging procedures that provide information polyps (n = 53), uterine anomalies (n = 12), and other of the female reproductive tract in the infertile patients findings (n = 12) (Figs. 4–11). include sonohysterography, hysteroscopy, and X-ray Fallopian tubes pathology was found in 37/209 HSG [5–9]. New volumetric, high-definition CT data patients (17.22%). They included total tube occlusion obtained with the 64-row CT scanners allow virtual 124 P. Carrascosa et al.: 64-MDCT Hysterosalpingography

Fig. 3. Cervical glandular dilatation and canal stenosis. (A) level of the cervical glandular dilatation (dotted arrow). (D) and (B) Coronal MIP and volume rendering images showing a Virtual endoluminal view at the level of the cervical canal cervical glandular dilatation (dotted arrows) and cervical canal stenosis (solid arrow). stenosis (solid arrows). (C) Virtual endoluminal view at the endoscopy evaluation of the female reproductive or- and endoluminal views give accurate characterization gans, and the CT-VHSG is proposed as a minimally of different kinds of female reproductive tract abnor- invasive diagnostic procedure for its evaluation. This malities with similar and even in several opportunities recent described non-invasive modality facilitates the better diagnostic accuracy of the conventional X-ray assessment of the female reproductive organs based on HSG [4]. volumetric CT data. The MPRs, combined with MIPs, Analysis of the CT data set requires a systematic and three-dimensional volume rendering reconstructions meticulous visualization of the different types of post- P. Carrascosa et al.: 64-MDCT Hysterosalpingography 125

Fig. 4. Cesarean scar. (A) and (B) Sagittal MIP and volume rendering images showing the uterine incision from a cesarean section (arrows). (C) Virtual endoluminal view. processing algorithms. Original transverse CT images This reprocess consists in the endoluminal view of the allow an integral evaluation of the female reproductive cervix, uterus, and even the fallopian tubes, demon- organs, facilitating the identification of myometrial strating the presence of endocavitary lesions. This algo- pathology (wall thickening, adenomyosis, intramural, or rithm permits to create images similar to those obtained subserosal myomas), intra-luminal pathology (endome- with hysteroscopy, giving an accurate delineation of the trial polyps, synechiae, submucosal myomas), and also lesions. the assessment of adnexal or other pelvic disorders. Advantages of the procedure are the possibility to Multiplanar reconstructions are extremely useful for a carry out the examination without using a tenaculum; the better delimitation and extension of the pathology, and facility to unfold and reconstruct the female reproductive for accurate measurements of the lesions. The MIP organs in any plane without detriment to the image reconstructions easily display two- and three-dimen- quality; a precise delimitation of submucous myomas sional projections that permit a conventional X-ray HSG and assessment of the entire uterine wall; the endolumi- like visualization of the cervix, uterus, and fallopian nal visualization of the fallopian tube lumen, when di- tubes as well as the determination of tube patency. The lated; and the evaluation of other pelvic organs. main limitation of this reprocess is the inadequate dis- Regarding the patient’s discomfort, this technique has crimination of intra-uterine pathology such as endome- been well received by the majority of the patients. Those trial polyps or synechiae. Three-dimensional volume who have ever carried out a conventional X-ray HSG rendering is a very user-friendly technique that shows the referred a better tolerance of the new examination mainly relationship of the female reproductive organs with the on the speed and the absence of clamping the cervix. rest pelvis structures and complements the analysis. It Another important argument for the CT-VHSG is the allows all the editing process within seconds and in a real significant dose reduction in comparison to conventional time. Overlying structures can be ease removed using X-ray HSG due to automatic tube current modulation clipping planes and the cervix; uterus, and fallopian during the scan. In our series, the mean patient effective tubes can be rotated into any orientation. Complemen- dose was 2.58 ± 0.75 mSv, inferior to the 5,13 ± tary cine images can be created to show them in different 0,24 mSv of the conventional X-ray HSG reported in our angles. The last step in the analysis is virtual endoscopy. previous series [4]. 126 P. Carrascosa et al.: 64-MDCT Hysterosalpingography

Fig. 5. Adenomyosis. (A–C) Coronal MIP and volume ren- adenomyosis. (C) Virtual endoluminal view. The arrows show dering images showing filling defects in the fundus with small the larger diverticula. contrast material-filled diverticula, findings that represent focal

Disadvantage of the CT-VHSG is the cost, threefolds the female reproductive tract and the other intrapelvic the cost of the HSG. Besides, in patients with proximal organs. Distinct advantages over conventional X-ray tubal occlusion, this technique does not allow to perform HSG, convert the 64-row CT-VHSG into a valuable an immediate tubal cannulation. alternative diagnostic technique in the infertility workup algorithm. Conclusion Acknowledgment. We would like to thank Dr. Graciela Fernandez The 64-row CT-VHSG provides a reliable, non-invasive Alonso for her assistance in editing this manuscript. diagnostic procedure for the simultaneous evaluation of P. Carrascosa et al.: 64-MDCT Hysterosalpingography 127

Fig. 6. Large leiomyoma. (A) Coronal MPR showing a volume rendering images showing distortion of the endome- round, large, soft tissue mass in the right myometrium dis- trial cavity. (D) Virtual endoluminal view. torting the endometrial cavity. (B) and (C) Coronal MIP and

References 4. Carrascosa P, Baronio M, Capun˜ay C, et al. (2008) Clinical use of 64-row multislice computed tomography hysterosalpingography in 1. Carrascosa P, Capun˜ay C, Baronio M, et al. (2007) Virtual hys- the evaluation of female factor infertility. Fertil Steril [Epub ahead teroscopy by multidetector computed tomography. Abdom Imag- of print] ing [Epub ahead of print] 5. Lindheim SR, Sprague C, Winter TC III (2006) Hysterosalpin- 2. Carrascosa P, Baronio M, Capun˜ay C, et al. (2007) Multidetector gography and sonohysterography: lessons in technique. AJR Am J computed tomography virtual hysterosalpingography in the inves- Roentgenol 186:24–29 tigation of the uterus and fallopian tubes. Eur J Radiol [Epub 6. Lindheim SR, Adsuar N, Kushner DM, et al. (2003) Sonohyste- ahead of print] rography: a valuable tool in evaluating the female pelvis. Obstet 3. Rydberg J, Liang Y, Teague SD (2004) Fundamentals of multi- Gynecol Surv 58:770–784 channel CT. Semin Musculoskelet Radiol 8:137–146 128 P. Carrascosa et al.: 64-MDCT Hysterosalpingography

Fig. 7. Large submucosal leiomyoma. (A) and (B) Axial and coronal thin slab MIP images showing a lobulated, large, soft tissue mass projecting into the endometrial cavity. (C) Sagittal volume rendering image. (D) Virtual endoluminal view.

7. Eng CW, Tang PH, Ong CL (2007) Hysterosalpingography: cur- 9. Pellicano M, Guida M, Zullo F, et al. (2003) Carbon dioxide versus rent applications. Singapore Med J 48:368–373 normal saline as a uterine distension medium for diagnostic vagi- 8. Simpson WL Jr, Beitia LG, Mester J (2006) Hysterosalpingogra- noscopic hysteroscopy in infertile patients: a prospective, ran- phy: a reemerging study. Radiographics 26:419–431 domized, multicenter study. Fertil Steril 79:418–421 P. Carrascosa et al.: 64-MDCT Hysterosalpingography 129

Fig. 8. Synechiae. (A) and (B) Coronal MIP and volume rendering images showing irregular filling defects in the endometrial cavity. (C) and (D) Virtual endoluminal views. 130 P. Carrascosa et al.: 64-MDCT Hysterosalpingography

Fig. 9. Endometrial polyp. (A) Coronal volume rendering showing a well-defined filling defect (arrow) in the posterior image showing a filling defect (arrow) in the anterior myo- myometrium. (D) Virtual endoluminal view. metrium. (B) Virtual endoluminal view. (C) Sagittal MIP image P. Carrascosa et al.: 64-MDCT Hysterosalpingography 131

Fig. 10. Endometrial polyp. (A) and (B) Coronal MIP and volume rendering images showing a filling defect (arrow) in the right anterior myometrium. (C) and (D) Virtual endoluminal views. 132 P. Carrascosa et al.: 64-MDCT Hysterosalpingography

Fig. 11. Unicornuate uterus. (A) and (B) Coronal volume rendering images demonstrating a single uterine horn. A single fallopian tube is also visualized. (C) and (D) Virtual endoluminal views. P. Carrascosa et al.: 64-MDCT Hysterosalpingography 133

Fig. 12. Right tubal occlusion. (A) and (B) Coronal MIP and volume rendering images showing absence of opacification of the right fallopian tube.

Fig. 13. Bilateral hydrosalpingx. (A) and (B) Coronal MIP and volume rendering images showing dilatation of the ampullary portion of the fallopian tubes. (C) Virtual endoluminal view of the left fallopian tube.