Obstetrics & Gynecology:
Clinical Examination and Dynamic Magnetic Resonance Imaging in Vaginal Vault Prolapse
Cortes, Eduard MD*; Reid, Wendy M. N. FRCOG*; Singh, Kavita MRCOG*; Berger, Leslie FRCR†
From the *Pelvic Floor Unit, Gynaecology Department, Royal Free and University, College Medical School, London, United Kingdom; and †MRI Unit, Radiology Department, Royal Free Hospital, London, United Kingdom.
Received June 4, 2003. Received in revised form September 16, 2003. Accepted September 18, 2003.
The authors thank S. Pickman, Senior Radiographer, G. Hoare, Radiographer; MRI unit, Royal Free Hospital, London, and J. Emberson, Statistician, Population Sciences Department, Royal Free Medical School, London.
Address reprint requests to: East Cortes, MD, Clinical Research Fellow, Royal Free Hospital, Academic Dept Obstetrics & Gynecology, 5th Floor, Pond Street, Hampstead, London NW3 2QG, United Kingdom; e-mail: firstname.lastname@example.org.
OBJECTIVE: To estimate the role of dynamic magnetic resonance imaging (MRI) as a diagnostic tool in the evaluation of vaginal apex prolapse in women with previous hysterectomy.
METHODS: Clinical examinations were performed on 51 women presenting with symptoms of prolapse. A preoperative dynamic MRI assessment was performed. The mid pubic line was the reference level used for prolapse grading. The parameters of analysis included 1) correlation by compartments of clinical and MRI grading of prolapse, 2) assessment of the accuracy of clinical examination of the middle compartment, and 3) identification of any additional information provided by MRI. All MRI films were analyzed and validated by the same two observers.
RESULTS: Analysis of each compartment separately revealed poor correlation between clinical and MRI assessment. Of the 51 cases with clinical vault prolapse, 27 (52.9%) cases were clinically overdiagnosed, 3 (6%) were underdiagnosed, and there was agreement in 21 (41.1%) when compared with MRI findings. Postoperative follow-up of the 18 (85%) patients who underwent colposacropexy after intraoperative assessment revealed the presence of cystocele in 4 (26.6%) occasions and rectocele in 3 (20%), which had been detected on MRI but not confirmed intraoperatively.
CONCLUSION: There is poor correlation between clinical and MRI findings when assessing vaginal apex prolapse. Magnetic resonance imaging allows the identification of other prolapsing compartments and may be a complementary diagnostic tool for the diagnosis of complex vaginal apex prolapse.
LEVEL OF EVIDENCE: III
Advances in magnetic resonance imaging (MRI) have made this technique an accurate imaging tool for the study of pelvic floor anatomy.1–6 Variations within MRI technology, such as the use of endovaginal coils, open-configuration MRI and, more recently, 3-dimensional MRI, have improved the understanding of pelvic floor pathology.7–12 As yet, few authors have studied the correlation between clinical examination and imaging findings and hence their usefulness as tools for preoperative assessment.13–15 Other imaging techniques used before MRI have shown a number of problems: some are poorly tolerated by patients or sometimes require the use of radiopaque contrast,16–22 or are unable to visualize the entire pelvic floor.23–28 The use of dynamic MRI does not involve ionizing radiation, is noninvasive, does not require patient preparation, and is a relatively rapid technique that provides high-quality images of soft tissues, clearly demonstrating pelvic viscera displacement. Arguments against the use of MRI as a diagnostic tool in pelvic floor pathology are the lack of an accepted standardized imaging grading protocol for prolapse and correlation with clinical assessment. MRI is relatively expensive and, as yet, in limited availability.
Most diagnostic studies performed with MRI to date have looked at pelvic floor anatomy and pathology in women presenting with symptoms in only one vaginal compartment.23–29 However, there is evidence that alterations in other vaginal compartments may play a role in the patient’s symptomatology. Law et al7 assessed 102 women with symptomatic and asymptomatic prolapse using open-configuration MRI and concluded that symptomatic women with symptoms in one compartment showed general weakness of the pelvic floor, suggesting the need to evaluate the entire pelvic floor before any surgery. The present study will assess the reliability of dynamic MRI grading as a valuable diagnostic tool for the assessment of the entire pelvic floor when compared with clinical findings in a population of women presenting with symptomatic vault prolapse (vaginal apex prolapse).
MATERIALS AND METHODS
Fifty-one women aged 40 to 95 years (mean 64 years) presenting to the pelvic floor unit at the Royal Free Hospital with symptomatic vaginal vault (apex) prolapse after hysterectomy (vaginal 42%, abdominal 58%), were recruited consecutively during the period of 1995 to 2001 and agreed to have a dynamic MRI evaluation of their pelvis. The study had local research ethics committee approval, and all women signed a consent form before enrollment. All women were given a clinical evaluation of their prolapse in the dorsal supine and standing position by using a quantified pelvic organ prolapse grading system, as described by Bump et al,30 by 3 gynecologists experienced in the assessment of pelvic floor dysfunction. After clinical assessment, all women received further evaluation of their prolapse by the use of dynamic MRI. Those women showing divergence between clinical and MRI findings in the middle compartment were clinically assessed intraoperatively. Only women who showed contraindications for MRI assessment were excluded from the test.
Imaging was performed in the dorsal supine position with both legs together and supported in a semiflexed position. Although this position may restrict mobility, it is the closest position to that of the clinical examination that can be adopted within a conventional MRI machine. All images were obtained with a 1.5 Tesla Philips Gyroscan (Philips Medical Systems, Best, The Netherlands) allocated in the X-ray department. Static T2 turbo spin echo images were obtained at 5 mm/5 mm on axial and sagittal sections with a field of view of 280 degrees, matrix 230 × 512, time repetition 6086/time echo 150, and 4.5 minutes acquisition time per plane. Dynamic imaging in the sagittal plane was performed by using a fast spin echo sequence. The slices were obtained on maximal straining (increasing Valsalva maneuver) at 5 mm/5 mm with field of view 350 and 17 seconds acquisition time per plane. The women had been instructed in the Valsalva maneuver before the imaging sequence and were asked to do the same at the time of clinical examination. To facilitate analysis of the films, sterile lubricating jelly (Aquagel, Adams Healthcare, Leeds, UK) was inserted vaginally with a vaginal applicator and rectally with a 10-mL syringe and a quill. This provides the vagina and rectum with a degree of opacity that enhances visualization. Women were imaged with a partially filled bladder and any vaginal support pessaries were removed.
Previous MRI studies have used different reference levels for grading prolapse, such as the pubosacral or pubococcygeal line,14,30 but these may overdiagnose or limit diagnostic accuracy to one specific compartment. In this study, the midpubic line in the sagittal plane, as described by Singh et al31, was used as the reference level for grading prolapse on MRI. The midpubic line is a line drawn across the midsagittal aspect of the pubic bone and is at the level of the vaginal hymen, hence coterminus with the clinical level, as described by Bump et al.30
A simplified scoring system was used to facilitate the description on MRI of the vaginal compartment involved: A = anterior compartment (equivalent to cystocele); M = middle compartment (vaginal apex or vault); P = posterior compartment (equivalent to rectocele). The different stages of prolapse were described as ranging from 1 to 4 as referenced by distance from the mid pubic line (Table 1). Images were analyzed and validated by the same 2 observers (E.C., K.S.), using a workstation (Philips Medical Systems), with zoom facility and electronic calipers. Both observers were blinded from the clinical findings at the time of MRI evaluation. As shown in Figure 1, a patient who was clinically diagnosed with a moderate anterior compartment descent (A2), severe middle compartment descent (M3), and mild posterior compartment descent (P1), when analyzed with dynamic MRI before and during the Valsalva maneuver, was described as having a severe anterior compartment prolapse A3, mild middle compartment prolapse M1, and nonexistent posterior prolapse P0.
An intraoperative assessment was performed by the authors in those patients who showed middle compartment disagreement between clinical and MRI diagnosis. Examination was performed in theater with the patient in a semiextended lithotomy position before reconstructive surgery. The vaginal apex was identified and held with a tenaculum at the same time gentle traction was applied. The distance the apex descended in relation to the hymen was documented.
The weighted κ test was used to compare values between both grading systems. A test for symmetry was used to assess the significance of over- and under-diagnosed results in each compartment. (Table 2). Statistical analysis of intraoperative assessment, to confirm either clinical or MRI findings preoperatively, was performed by using the McNemar’s χ2 test. Statistical significance was taken as a P value of less than .05.
Analysis of each compartment independently revealed poor correlation between clinical and MRI findings (Table 2). By using the mid pubic line on the MRI as the reference level for grading of prolapse, in only 17 (33%) of 51 was there agreement between clinical and MR scores in the anterior and posterior compartment. In the middle compartment, agreement was seen only in 21 (41%) patients of the 51. By using the same parameters, both in the anterior and posterior compartment, 19 (37%) cases of 51 were over-diagnosed clinically and 15 (29%) cases were under-diagnosed clinically when compared with MRI findings (Table 3).
When assessing the middle compartment with MRI as the reference diagnostic level, it would appear as if, in 27 (52.9%) cases, patients were clinically over-diagnosed and in only 3 (6%) cases they were under-diagnosed clinically. Symmetry test showed P = .002 and therefore a significant difference in this compartment when compared MRI staging and clinical assessment of prolapse. No significant differences were seen in the anterior and posterior compartment (Table 2). Looking in more detail at the middle compartment, we found that 35 (69%) of the 51 patients had been clinically diagnosed with a moderate-to-severe prolapse of this compartment (M2–M4). However, when we looked at these patients’ dynamic MRIs, 21 (60%) of 35 were reclassified as normal-to-mild prolapse (M0-M1). Interestingly, 20 of the 21 who were reclassified as grade 0–1 middle compartment prolapse also were found to show on MRI different degrees of prolapse in the anterior and posterior compartment (A2 to A4 and P2 to P4) (Table 4).
If we establish a division between grade 2 and 4 descent as significant prolapse and grade 0 to 1 as no significant in the anterior compartment, 7 (28%) of the 25 patients who were clinically diagnosed as significant prolapse were reclassified as nonsignificant with MRI. Likewise, in the middle compartment, this was seen in 21 (60%) of the 35 clinically significant prolapses and in 9 (28%) of the 32 clinically significant posterior compartment prolapse.
Three (14%) of the 21 patients mentioned above whose MRIs of the middle compartment reclassified them as nonsignificant prolapse were managed with ring pessaries. The remaining 18 (85.7%) patients were assessed clinically both intraoperatively and postoperatively. Clinical preoperative findings were confirmed intraoperatively in 15 (83%) of the 18 cases (P < .01), and a sacrocolpopexy procedure was performed. MRI findings reclassifying middle-compartment prolapse as nonsignificant were confirmed intraoperatively in 3 (17%) of the 18 cases, and vaginal surgery for correction of anterior and posterior compartment defects was performed instead after MRI findings. In 1 of the 15 cases, intraoperative findings confirmed both the clinical diagnosis in the middle compartment and concurred with the MRI findings in the anterior and posterior compartment. The patient underwent a combined abdominal and vaginal approach. Postoperative assessment at 3 months revealed no recurrent middle compartment prolapse in any patients. Four (26.6%) and 3 (20%) women of 15 had persistent anterior (cystocele) and posterior (rectocele) compartment prolapse, respectively. Although these had been detected preoperatively with MRI, intraoperative assessment was not conclusive enough to alter the initial surgical management.
Additional findings on MRI revealed peritoneum involvement in 12 (23%) cases, 1 anterior peritoneocele, and sigmoid colon descent in 2 (4%) cases. These were managed accordingly at the time of surgery. Two (4%) women of the 51 with grades 3 and 2 prolapse in the middle compartment found it impossible to reproduce their prolapse in the MRI environment, therefore invalidating this test for the purpose of this study.
Vaginal vault prolapse is the anatomical descent of the vaginal apex or middle vaginal compartment resulting from the disruption of its fascial supports. It can be the long-term consequence of both vaginal or total abdominal hysterectomy, perhaps where pre-existing, fascial, or muscle defects are present, or may also be the result of disruption of the endopelvic fascia at the time of hysterectomy. More recent surgical techniques have acknowledged this and advocated new surgical approaches.32–34 Complexity of vaginal vault prolapse may arise as the result of different compartments being involved within the anatomical herniation. Identification of these compartments preoperatively may facilitate both the surgical approach and repair.
Although previous studies have shown a correlation between grading prolapse when comparing the dorsal lithotomy with the upright position,35–37 these studies have been performed in open MRI settings. Other studies have looked at changes in prolapse grading between conventional and open-configuration MRI. Fielding et al28 found no significant difference on MRI grading when comparing the sitting and the supine position in 5 normal volunteers. Nevertheless, this study assessed only the pelvic floor laxity of the anterior compartment with reference to the pubococcygeal line.
Our results revealed poor correlation between clinical and MRI findings when defining compartment descent as well as identifying the descent clinically seen in the vaginal compartments separately. These differences may be the result of either mobility restrictions experienced by the patient at straining in an unconventional setting or the result of prolapse competition for space (compartment overlapping), when assessed in the dorsal supine position with the MRI machine.
Intraoperative assessment must also be carefully interpreted because pelvic floor descent under general anesthesia will not reproduce prolapse as seen under physiological conditions. Although subjects were assessed preoperatively at straining, abolishment of the active supports by muscular block of the pelvic floor muscles (Levator Ani) can make assessment of prolapse a reflection of its fascial supports only. Therefore, by removing any reflex contraction forces, clinical assessment may tend to accentuate the prolapse. In the current study, intraoperative assessment was used as a validating tool between the clinical diagnosis and MRI of the middle compartment, confirming clinical assessment of vaginal apex descent in 83% of cases. In those patients whose vault descent was not detected intraoperatively (3 of 18, 17%), the prolapse was repaired in accordance with the MRI findings in the anterior and posterior compartment. None of the 18 women presented with persistent or de novo vault prolapse at follow-up. Nevertheless, of the 83% (15 of 18) who underwent colposacropexy, some patients presented postoperatively with a cystocele (26.6%) or rectocele (20%) that had been detected on MRI preoperatively but had not been conclusively identified during either clinical or intraoperative assessment. These findings suggest that dynamic MRI may play a more reliable role in the assessment of the anterior and posterior compartment within the context of complex vault prolapse than clinical or intraoperative assessment.
Despite the poor correlation seen in our study, vaginal apex prolapse of the upper vagina can be the most complex type of prolapse to accurately diagnose because of the compartments involved and the etiologies for the varying presenting symptoms. The absence of anatomical landmarks, the presence of atrophic changes, and persistent long-term distension/pulsion forces, as well as the existence of previous pelvic surgery, may make it difficult to establish the boundaries between the 3 compartments. In women with recurrent vault prolapse or women who have undergone repetitive vaginal surgery for prolapse, the use of dynamic MRI may play a significant role in clarifying the prolapsed anatomy and identifying unexpected prolapsing viscera. In this population, the detection of unusual prolapsing structures, such as sigmoidoceles and peritoneoceles, may influence the surgical management by suggesting a combined abdominal and vaginal approach or by recommending surgery with a culdoplasty in cases were the peritoneum is seen to be involved.
Although previous studies have used the pubococcygeal line as a reference level for grading of prolapse, the midpubic line appears a more reliable reference level when assessing the anatomy of prolapse of all 3 compartments simultaneously. Although use of the pubococcygeal line may cause one to overdiagnose the degree of prolapse seen clinically, the lack of clinical correlation seen when using the midpubic line seems to be the result of difficulties experienced by patients at the time of reproducing their condition or prolapse competition for space rather than poor anatomical visualization or accuracy of definition of the compartments with reference to the level of the hymen. Studies using the midpubic line as the reference level in an open-configuration MRI setting may show more reliability because the positioning would allow assessment of prolapse under the effect of gravitational forces while minimizing the physical restrictions experienced in a conventional setting.
Dynamic MRI assessment of vaginal apex prolapse does not correlate with clinical findings and may reflect the compartment descent restrictions experienced by patients during straining. Clinical diagnosis of vaginal apex descent was confirmed intraoperatively in 83% of the cases not detected on MRI. The use of MRI may cause one to underdiagnose clinical assessment, particularly in the middle compartment. Dynamic MRI remains an accessory, complementary diagnostic tool in the clarification of the anatomy involved in complex vaginal prolapse and its associated symptomatology.
1. Tunn R, DeLancey JO, Quint EE. Visibility of pelvic organ support system structures in magnetic resonance images without an endovaginal coil. Am J Obstet Gynecol 2001;184:1156–63.
2. Goh V, Halligan S, Kaplan G, Healy JC, Bartram CI. Dynamic MR imaging of the pelvic floor in asymptomatic subjects. Am J Roentgenol 2000;174:661–6.
3. Schmeiser G, Putz R. The anatomy and function of the pelvic floor [review]. Radiologe 2000;40:429–36.
4. Strohbehn K, Ellis J H, Strohbehn JA, DeLancey JO. Magnetic resonance imaging of the levator ani with anatomical correlation. Obstet Gynecol 1996;87:277–85.
5. McCarthy S, Tauber C, Gore J. Female pelvic anatomy: MR assessment of variations during the menstrual cycle and with the use of oral contraceptives. Radiology 1986;160:119–23.
6. Barbaric ZL, Marumoto AK, Raz S. Magnetic resonance imaging of the perineum and pelvic floor. Top Magn Reson Imaging 2001;12:83–92.
7. Law PA, Danin JC, Lamb GM, Regan L, Darzi A, Gedroyc WM. Dynamic imaging of the pelvic floor using an open-configuration magnetic resonance scanner. J Magn Reson Imaging 2001;13:923–9.
8. Yang A, Mostwin JL, Rosenshein NB, Zerhouni EA. Pelvic floor descent in women: dynamic evaluation with fast MR imaging and cinematic display. Radiology 1991;179:25–33.
9. Tan IL, Stoker J, Lameris JS. Magnetic resonance imaging of the pelvic floor and urethra: body coil versus endovaginal coil. MAGMA 1997;5:59–63.
10. Hoyte L, Schierlitz L, Zou K, Flesh G, Fielding JR. Two- and 3-dimensional MRI comparison of Levator ani structure, volume, and integrity in women with stress incontinence and prolapse. Am J Obstet Gynecol 2001;185:11–9.
11. Fielding JR, Dumanli H, Schreyer AG, Okuda S, Gering DT, Zou KH, et al. MR-based three-dimensional modelling of the normal pelvic floor in women: quantification of muscle mass. AJR Am J Roentgenol 2000;174:657–60.
12. Hoyte L, Fielding JR, Versi E, Mamisch C, Kolvenbach C, Kikinis R. Variations in levator ani volume and geometry in women: the application of MR based 3D reconstruction in evaluating pelvic floor dysfunction. Arch Esp Urol 2001;54:532–9.
13. Kenton K, Shott S, Brubaker L. Vaginal topography does not correlate well with visceral position in women with pelvic organ prolapse. Int Urogynecol J Pelvic Floor Dysfunct 1997;8:336–9.
14. Comitter GV, Vasavada SP, Barbaric ZL, Gousse AE, Raz S. Grading pelvic prolapse and pelvic floor relaxation using dynamic magnetic resonance imaging. Urology 1999;54:454–457.
15. Rodriguez LV, Raz S. Diagnostic imaging of pelvic floor dysfunction [review]. Curr Opin Urol 2001;11:423–8.
16. Altringer WE, Saclarides TJ, Dominguez JM, Brubaker LT, Smith CS. Four-contrast defecography: pelvic “floor-oscopy”. Dis Colon Rectum 1995;38:695–9.
17. Kelvin FM, Maglinte DD, Benson JT, Brubaker LP, Smith C. Dynamic cystoproctography: a technique for assessing disorders of the pelvic floor in women. AJR Am J Roentgenol 1994;163:368–70.
18. Brubaker L, Retzky S, Smith C, Saclarides T. Pelvic floor evaluation with dynamic fluoroscopy. Obstet Gynecol 1993;82:863–8.
19. Dietz HP, Jarvis SK, Vancaillie TG. The assessment of levator muscle strength: a validation of three ultrasound techniques. Int Urogynaecol J Pelvic Floor Dysfunct 2002;13:156–9.
20. Mostwin JL, Genadry R, Saunders R, Yang A. Stress incontinence observed with real time sonography and dynamic fastcan magnetic resonance imaging. Scand J Urol Nephrol Suppl 2001;207:94–9.
21. Kelvin FM, Dean DT, Maglinte, Benson JT. Evacuation proctography (defecography): an aid to the investigation of pelvic floor disorders. Obstet Gynecol 1994;83:307–14.
22. Peschers UM, Gingelmaier A, Jundt K, Leib B, Dimpfl T. Evaluation of pelvic floor muscle strength using four different techniques. Int Urogynecol J Pelvic Floor Dysfunct 2001;12:27–30.
23. Lieneman A, Anthuber C, Baron A, Reiser M. Diagnosing enteroceles using dynamic magnetic resonance imaging. Dis Colon Rectum 2000;43:205–12; discussion 212––3.
24. Lienemann A, Anthuber C, Baron A, Kohz P, Reiser M. Dynamic MR colpocystorectography assessing pelvic-floor descent. Eur Radiol 1997;7(8):1309–17.
25. Paetzel C, Strotzer M, Furst A, Rentsch M, Lenhart M, Feuerbach S. Dynamic MR defecography for diagnosis of combined functional disorders of the pelvic floor in proctology. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 2001;173:410–15.
26. Armillotta M, Casillo A, Bonetti MG, Morcaldi M. Stress urinary incontinence in women: magnetic resonance assessment. Radiol Med 1995;89:112–6.
27. Tunn R, Paris S, Fischer W, Hamm B, Kuchinke J. Static magnetic resonance imaging of the pelvic floor muscle morphology in women with stress urinary incontinence and pelvic prolapse. Neurourol Urodyn 1998;17:579–89.
28. Fielding JR, Griffiths DJ, Versi E, Mulkern RV, Lee ML, Jolesz FA. MR imaging of pelvic floor continence mechanisms in the supine and sitting positions. AJR Am J Roentgenol 1998;171:1607–10.
29. Hedlund H, Bo K, Lilleas F, Talseth T, Tillung T. The clinical value of dynamic magnetic resonance imaging in normal and incontinent women: a preliminary study on micturition. Scand J Urol Nephrol Suppl 2001;207:87–91; discussion 106––25.
30. Bump RC, Mattiasson A, Bo K, Brubaker LP, DeLancey JO, Klarskov P, et al. The standardization of terminology of female pelvic organ prolapse and pelvic floor dysfunction. Am J Obstet Gynecol 1996;175:10–7.
31. K Singh, WMN Reid, LA Berger. Assessment and grading of pelvic organ prolapse by use of dynamic resonance imaging. Am J Obstet Gynecol 2001;185:71–7.
32. Morrison JE, Jacobs VR. 437 classic intrafascial supracervical hysterectomies in 8 years. J Am Assoc Gynecol Laparosc 2001;8:558–67.
33. Lyons TL, Winer WK, Adolph AJ. Classic intrafascial supraventricular hysterectomies. J Am Assoc Gynecol Laparosc 2002;9:401.
34. Conde-Agudelo A. Intrafascial abdominal hysterectomy: outcomes and complications of 867 operations. Int J Gynecol Obstet 2000;68:233–9.
35. Fielding JR, Versi E, Mulkern RV, Lerner MH, Griffiths DJ, Jolesz FA. MR imaging of the female pelvic floor in the supine and upright position. J Magn Reson Imaging 1996;6:961–3.
36. Barber MD, Lambers A, Visco AG, Bump RC. Effect of patient position on clinical evaluation of pelvic organ prolapse. Obstet Gynecol 2000;96:18–22.
37. Swift SE, Herring M. Comparison of pelvic organ prolapse in the dorsal lithotomy compared with the standing position. Obstet Gynecol 1998;91:961–4.
This article has been cited 6 time(s).
American Journal of RoentgenologyImaging Pelvic Floor Disorders: Trend Toward Comprehensive MRIAmerican Journal of Roentgenology
International Urogynecology JournalPOP-Q, dynamic MR imaging, and perineal ultrasonography: do they agree in the quantification of female pelvic organ prolapse?International Urogynecology Journal
International Urogynecology JournalA systematic review of clinical studies on dynamic magnetic resonance imaging of pelvic organ prolapse: the use of reference lines and anatomical landmarksInternational Urogynecology Journal
International Urogynecology JournalMagnetic resonance imaging of pelvic organ prolapse: comparing pubococcygeal and midpubic lines with clinical stagingInternational Urogynecology Journal
Neurourology and UrodynamicsDynamic magnetic resonance imaging for grading pelvic organ prolapse according to the international continence society classification: Which line should be used?Neurourology and Urodynamics
Abdominal ImagingThe urogynecological side of pelvic floor MRI: the clinician's needs and the radiologist's roleAbdominal Imaging
© 2004 The American College of Obstetricians and Gynecologists