Initial Accuracy of and Learning Curve for Transvaginal Ultrasound with Bowel Preparation for Deep Endometriosis in a US Tertiary Care Center

Abstract

Study objective: To evaluate the diagnostic accuracy and learning curve of a sonographic mapping protocol for deep endometriosis (DE).

Design: Retrospective cohort study (Canadian Task Force classification II-3).

Setting: Tertiary referral center in the United States.

Patients: 117 consecutive patients who presented to our gynecology clinic with complaints of significant noncyclic pelvic pain of at least 6 months' duration, and/or clinical findings concerning for deep endometriosis and who were referred for transvaginal ultrasound with bowel preparation.

Interventions: Patients underwent transvaginal ultrasound with bowel-preparation (TVUS-BP) performed by a single radiologist. Findings suspicious for DE were reported and correlated with surgical and histopathological findings. The duration of the examination and number of cases required to achieve proficiency were calculated for positive, equivocal, and negative findings.

Measurements and main results: Among 117 patients (median age, 35 years; range, 19-54 years) referred for TVUS-BP, 113 had complete examinations. Fifty-seven of these 113 patients underwent surgical exploration within 1 year, and DE was identified surgically in 23 of them. DE of the rectosigmoid colon and/or rectovaginal septum was detected with a sensitivity of 94% (95% confidence interval [CI], 70%-100%) and specificity of 100% (95% CI, 91%-100%). DE of the retrocervical region and/or uterosacral ligaments was detected with a sensitivity of 86% (95% CI, 65%-97%) and specificity of 94% (95% CI, 81%-99%). Proficiency, defined by a flattening of the learning curve, was achieved after 70 to 75 scans. The mean duration of the examination was 42 ± 4 minutes initially, but declined to 15 ± 4 minutes once proficiency was achieved. Cases of equivocal or minimal disease demonstrated the greatest decline in examination duration.

Conclusion: A newly applied TVUS-BP protocol for detection of pelvic DE is highly accurate and required only a modest learning curve to achieve procedural proficiency in a US tertiary referral center where physicians interpret but typically do not perform TVUS exams. Overcoming diagnostic uncertainty regarding minimal or equivocal disease appeared to be an important factor in the initial learning curve. With adequate training, TVUS-BP may be adapted as a primary diagnostic tool for detecting pelvic DE.

Keywords: Endometriosis; Laparoscopy; Learning; Transvaginal; Ultrasound.