In view of these findings, the girl was diagnosed with a congenital mllerian anomaly, consistent with a uterus didelphys with an obstructed left hemivagina resulting in a hematocolpos (collection of blood in the vagina) and left (ipsilateral) renal agenesis. She subsequently had an intravenous pyelogram (IVP) that confirmed the unilateral renal agenesis. MRI and CT scans did not add any new information to what was obtained with 3D U/S. The patient underwent surgery to relieve the obstruction of the left hemivagina.

The bladder (B) is visible and a uterus to each side. The hematocolpos is slightly to the right of the midline.

The complexities of this case This case was challenging in many ways. The patient was young and in pain. In addition, she had a history of normal periods since the onset of her menarche, 3 months before presentation. This information at first was confusing, because when a hematocolpos is seen, the initial differential diagnosis typically is imperforate hymen. Because of the child's age, transvaginal sonography was not an option. Both transrectal and translabial sonography were attempted, but the only structure visible was the hematocolpos. Transabdominal U/S using both two-dimensional (2D) and 3D U/S techniques with a full bladder provided the correct diagnosis: didelphic uterus with an obstructed left hemivagina and ipsilateral renal agenesis.

When scanning a gynecology patient, especially if a uterine anomaly is imaged, the kidneys need to be assessed; this adds only a few minutes to the scan and provides valuable information. The MRI confirmed the sonographic diagnosis of the uterus didelphys. The renal agenesis was confirmed with an IVP and the obstruction was confirmed and treated surgically.

Perspective on the anomaly Uterine malformations occur in approximately 0.1% to 3% of all women.^1,2 The incidence is significantly higher in the subgroups that have a history of recurrent pregnancy loss and approaches 10% to 15%. A population study of both fertile and infertile women found that 37% had bicornuate uterus, 15% had arcuate uterus, 13% had incomplete septate uterus, 11% had uterus didelphys, 9% had a complete septate uterus, and 4.4% had a unicornuate uterus.³

Renal anomalies can occur in 20% to 30% of women with mllerian defects.⁴ Ipsilateral renal agenesis is common in women who have obstructive defects, as was the case with our patient.

Congenital uterine anomalies result from failure of the paramesonephric ducts to fuse. The fused segments of the paramesonephric ducts give rise to the uterus and part of the vagina and the unfused portion gives rise to the uterine (fallopian) tubes. In normal development, this process is completed by the 10th week postconception.⁵ Lateral fusion defects are the most common type of mllerian anomaly. The other two common types of developmental defects are agenesis (as in the case of the Mayer-Rokitansky-Kster-Hauser syndrome) and vertical fusion defects that give rise to anomalies of the vagina (complete or incomplete septae) and of the cervix.

At times, when an anomaly involving a transverse vertical septum is diagnosed remote from menarche, not only is the vagina filled with stagnant blood but the uterine cavity also is dilated (hematocolpos with hematometria). Our patient had only three previous menstrual cycles, so her uterine cavity was not yet expanded by accumulating blood in the vagina.

The American Fertility Society's classification of mllerian anomalies is divided into six major types, each of which has several subtypes that are not discussed here⁶:

Type I: Mllerian agenesis or hypoplasia

Type II: Unicornuate uterus

Type III: Didelphic uterus

Type IV: Bicornuate uterus

Type V: Septate uterus

Type VI: Diethylstilbestrol-related anomalies

Spotting imaging clues On 2D and 3D ultrasonography, several clues can help you spot and classify congenital uterine anomalies. The most important ones are the shape of the uterine cavity represented by the endometrial echo or stripe, the outer and inner fundal contours, and the cervix.

A: A normal uterus showing a single central endometrial echo. B: A septate uterus showing two endometrial echoes. C: A bicornuate uterus during saline infusion, which shows two endometrial cavities filled with saline.

First, focus on the endometrial echo or stripe. Under normal circumstances, a uterus in a transverse plane on abdominal U/S in a woman with a full or partly full bladder has a single endometrial echo (Figure 3A). In a patient with uterine anomalies, however, a scan in the transverse plane will show two endometrial echoes (Figure 3B and C). The echoes may be separate, with tissue visible between them (as in the case of septum), or they can be further apart (as in the case of a bicornuate uterus). In some instances—such as with a uterus didelphys—the echoes may be even further apart and separated by bowel. On 2D ultrasound in the transverse plane, a uterus with a mllerian anomaly appears wider than a normal uterus. In these cases, the ratio of the anterior-posterior diameter to the transverse diameter is usually less than 0.6.

The outer and inner fundal contours and the cervix.

Next, evaluate the outer fundal contour of the uterus (Figure 4). This is best achieved with a 3D transvaginal scan. To get the best image of the outer fundal contour, manipulate the saved 3D volume (three orthogonal planes) to create a perfect coronal section. Barring the presence of fibroids, the outer fundal contour of a normal uterus should be smooth and slightly convex (Figure 4). Abnormalities of the outer uterine contour can range from a flat or minimal indentation (≤ 1 cm) (Figure 5) to an indentation greater than 1 cm (Figure 6). The former is sometimes seen in patients with a complete septate uterus, whereas the latter is seen with bicornuate uterus. In our patient, the abnormality was even more extreme, resulting in two widely separate uterine horns (uterus didelphys).

The outer fundal contour has a slight indentation. The two endometrial cavities are obvious.

The third step is to look at the shape of the endometrial cavity in the coronal plane. The normal inner fundal contour of the endometrial cavity is straight (Figure 4). A hypoplastic uterus may have a very small cavity, whereas an arcuate uterus will show a slightly curved bulge into the endometrial cavity. In a septate uterus, the septum runs all the way to the cervix; if the condition is partial, it runs halfway down the cavity. A bicornuate uterus has a deep cleft of the outer fundal contour, as well as two endometrial cavities that merge in the cervix (Figure 6). A didelphic uterus, such as in our patient, is actually a double uterus with two uterine bodies, two endometrial cavities, and two cervices. In most such cases, you will see two vaginal tracts with or without transverse septae at various levels. Not all mllerian anomalies fall into the categories we've just described, but the information should serve as a foundation for evaluation of uterine anomalies.

Finally, evaluate the cervix. The type of anomaly a woman has is determined by the number of cervical canals and the area of the external os.

This bicornuate uterus shows a deeply indented fundal contour. The two endometrial cavities are widely separated superiorly and merge together in the area of the cervix.

Choosing an imaging technique Several imaging modalities have been used to assess mllerian anomalies. The first technique, hysterosalpingography, reportedly has a false-positive rate of 38% and a false-negative rate of 28%.⁷ Another major drawback is its inability to assess the outer fundal contour, making it impossible to differentiate between a septate and a bicornuate uterus. As a result, the patient has to undergo laparoscopy to get a correct diagnosis. Magnetic resonance imaging (MRI) allows imaging of the uterus in the coronal plane and shows the endometrial cavity as well as the fundal contours. It therefore has become the gold standard for noninvasive diagnosis of mllerian anomalies.⁸ But it, too, has drawbacks: high cost and lack of access. MRI requires referral to an imaging facility, which delays the diagnosis. More recently, 3D U/S has moved to the forefront and is gaining a prominent—if not dominant—role in diagnosis of uterine malformations.⁹ Three-dimensional imaging of the uterus provides information that is the same or better than that obtained with MRI. It is significantly less expensive and readily available in most centers, which enables expedited diagnosis. When combined with saline infusion, 3D U/S yields information similar to that obtained with conventional hysterosalpingography.

Conclusions It's crucial that patients with a congenital uterine anomaly be correctly defined and classified, since their reproductive potential depends on the type of anomaly they have. Women with an arcuate uterus have reproductive potential similar to those whose uterus is normal. Women with a septate uterus, in contrast, have increased rates of miscarriage.¹⁰ The rate of spontaneous abortion reportedly is 40% in those with a bicornuate uterus, with a preterm delivery rate approaching 40%¹¹ and a live-birth rate of about 57%.¹ Reproductive outcomes in women with didelphic uterus usually are good. All women who have mllerian anomalies need to be evaluated for the presence of renal anomalies. Early diagnosis is critical in the subgroups that have obstructive anomalies, since they consistently will have ipsilateral renal agenesis.

With the rapid increase in U/S machines equipped with 3D imaging capability, diagnosing uterine anomalies is becoming easier, faster, and more reliable. Most importantly, this diagnostic process can be achieved as "one-stop shopping" in ob/gyn or infertility offices.

REFERENCES1. Rock JA, Schlaff WD. The obstetric consequences of uterovaginal anomalies. Fertil Steril. 1985;43:681-692.

2. Sanfilippo JS, Wakim NG, Schikler KN, et al. Endometriosis in association with uterine anomaly. Am J Obstet Gynecol. 1986;154:39-43.

3. Acien P. Incidence of mllerian defects in fertile and infertile women. Hum Reprod. 1997;12:1372-1376.

4. O'Rahilly R, Muller F. Human Embryology & Teratology. New York, NY: Wiley-Liss; 1992.

5. Lin PC, Bhatnagar KP, Nettleton GS, et al. Female genital anomalies affecting reproduction. Fertil Steril. 2002;78:899-915.

6. The American Fertility Society classifications of adnexal adhesions, distal tubal occlusion, tubal occlusion secondary to tubal ligation, tubal pregnancies, mllerian anomalies and intrauterine adhesions. Fertil Steril. 1988;49:944-955.

7. Raziel A, Arieli S, Bukovsky I, et al. Investigation of the uterine cavity in recurrent aborters. Fertil Steril. 1994;62:1080-1082.

8. Pellerito JS, McCarthy SM, Doyle MB, et al. Diagnosis of uterine anomalies: relative accuracy of MR imaging, endovaginal sonography, and hysterosalpingography. Radiology. 1992;183:795-800.

9. Woelfer B, Salim R, Banerjee S, et al. Reproductive outcomes in women with congenital uterine anomalies detected by three-dimensional ultrasound screening. Obstet Gynecol. 2001;98:1099-1103.

10. Lin PC. Reproductive outcomes in women with uterine anomalies. J Womens Health (Larchmt). 2004;13:33-39,

11. Daya S. Ultrasonographic evaluation of uterine anomalies. In: Jaffe R, Pierson RA, Abramowicz JS, eds. Imaging in Infertility and Reproductive Endocrinology. Philadelphia, Pa: Lippincott; 1994:63-92.