Background Vaginal agenesis may be associated with anomalies of the upper mullerian duct system such as rudimentary uterine horns or absence of a uterus. In the Mayer-Rokitansky-Kuster-Hauser Syndrome, approximately 90% of patients will have some mullerian structures. It has been demonstrated that rudimentary uterine horns have the potential for reproductive function with the utilization of assisted reproductive technologies, and therefore it is important to diagnose uterine structures both to control retrograde menses and for preservation of future fertility. The evaluation of a patient with vaginal agenesis has classically consisted of a history. physical exam, ultrasound, and/or MRI. MRI is particularly useful in the setting of vaginal atresia with hematocolpos or other obstructive utero-vaginal anomalies; however, its use in detecting uterine remnants is less well established. Traditionally, laparoscopy has not been considered a necessary step in evaluating patients with vaginal agenesis who are asymptomatic and have a negative MRI. We hypothesize that in cases of vaginal agenesis with or without symptoms of pelvic pain, surgical evaluation is a necessary tool for the detection of rudimentary uterine structures so that menstrual suppressive therapy can be initiated to avoid retrograde menses, endometriosis, adhesions, and possible future infenility. Methods All cases of vaginal agenesis referred to one gynecologist (MRL) at Children's Hospital and Brigham and Women's Hospital between 1993–1997 were reviewed. Subjects who had an MRI to assess mullerian structures were identified and their clinical charts reviewed to correlate physical exam, MRI reports, and operative findings. Results 22 cases of vaginal agenesis were identified. On physical exam, all of the women had absence of the vagina and only two (9%) had a midline abdominal mass on recto-abdominal exam. Of the 22 women, 14 underwent laparoscopy or laparotomy and 12 (86%) had some uterine structure(s). 13 of the 22 women (59%) had both MRI and an operative procedure to further delineate mullerain structures. In this subgroup. MRI successfully predicted uterine anomalies in four cases (31%) and lack of any uterine structures in one case (8%). In the remaining eight cases (75%), MRI diagnosis did not correlate with diagnosis on laparoscopy. Three had an MRI diagnosis of a hypoplastic midline uterus but on laparoscopy were found to have bilateral rudimentary uterine horns. Five women (42%) had an MRI showing no uterine structures, with the subsequent laparoscopy showing either unilateral or bilateral rudimentary uterine horns. Of these 5 women, 2 (40%) had symptoms of pelvic pain, and 3 (60%) had no symptoms of pelvic pain. Using laparoscopy as the gold standard. MRI was found to have a sensitivity of 31% a specificity of 100%, a positive predictive value of 100% and a negative predictive value of 11%. Conclusions Laparoscopy is an excellent method for diagnosing uterine structures associated with vaginal agenesis in women with or without symptoms of pelvic pain. In this small series, MRI had a low sensitivity (31%) for detecting uterine structures, panicularly rudimentary uterine horns in cases of vaginal agenesis. Laparoscopy is useful in the routine assessment of patients with vaginal agenesis either in conjunction with MRI or alone so as to accurately define the anatomy and preserve options for reproductive function.