141 Diagnostic Accuracy of Magnetic Resonance Imaging for the Evaluation of Acute Female Pelvic Pathology

In the evaluation of female pelvic etiologies, physicians often order both ultrasound and CT imaging within the diagnostic workup. In an effort to both avoid radiation and redundant imaging, this study evaluates the diagnostic accuracy of MRI for the evaluation of female pelvic pathology in emergency department (ED) patients presenting with lower abdominal pain. This study is a sub-analysis of a HIPAA-compliant and institutional review board-approved prospective single-center study of patients undergoing evaluation for appendicitis at an academic hospital’s ED. Patients were eligible if they were ≥12 years old and had a CT ordered for the evaluation of appendicitis. In this sub-analysis, only female patients who also underwent pelvic ultrasound were included. Participants underwent their research MRI scan after CT was performed. After the ED visit, 2 fellowship-trained abdominal radiologists and 1 fellowship-trained musculoskeletal radiologist independently interpreted each MRI examination in a randomized fashion, blinded to all clinical information. Images were evaluated according to a standardized interpretation form, which queried the presence of common gynecologic pathology including ovarian cyst, ovarian torsion, tubo-ovarian abscess, oophoritis/salpingitis, and endometrial/uterine mass. A reference standard of surgical/pathological reports was used if available, while a chart review and follow-up phone call was otherwise used. An expert panel of 2 senior radiologists who reviewed all 3 imaging sets (MRI, CT, and ultrasound) and available clinical information determined the “ground truth” disease state for each patient. If multiple diagnoses were present in one patient, the cases were analyzed as separate instances. Each case was evaluated against the presence or absence of the items listed on the interpretation sheet, resulting in 5 separate 2x2 tables. Inter-rater reliability was measured using the Fleiss kappa statistic. This study included 41 women, mean age 27.6 years (range 13-59 years). Of these patients, 26 (63.4%) had an acute finding seen on imaging. When evaluating for any of the 5 diagnoses of interest, cumulative sensitivity and specificity (95% confidence interval) was 66.2% (59.6-72.8%) and 96.5% (95.4-97.6%), respectively. Ovarian cysts were the most common pathology, observed in 21 cases. The sensitivity and specificity for detecting ovarian cysts was 69.8% (58.5-81.2%) and 100% (100-100%). Three cases of endometrial/uterine masses were identified and had a corresponding sensitivity and specificity of 66.7% (35.9-97.5%) and 94.7% (90.6-97.5%). For the 2 cases each of torsion and oophoritis/salpingitis, sensitivity was 33.3% (0-71.1%) and 50.5% (10-90%) while specificity was 99.1% (97.5-100%) and 99.1% (97.5-100%), respectively. There were no cases of tubo-ovarian abscess, yielding a sensitivity of 100% (100%) and specificity of 98.3% (96.1-100%). Inter-rater reliability kappa statistics varied from 0.225 (oophoritis/salphingitis) to 0.823 (ovarian cyst). MRI may be an acceptable single imaging alternative to the combination of ultrasound and CT in the emergent evaluation of pelvic pain.