Diagnostic Accuracy of Transvaginal Ultrasound for Endometriosis Does Not Differ Between Women Undergoing Their First Versus Any Subsequent Surgical Treatment.

BACKGROUND: Australian endometriosis guidelines recommend transvaginal ultrasound (TVUS) as the first-line investigation for endometriosis. Where endometriosis has been surgically treated, there is a risk of recurrent disease and repeat surgery. It is unknown if the diagnostic accuracy of ultrasound for recurrent disease is equivalent to the surgery-naïve. AIMS: To compare the accuracy of TVUS to diagnose endometriosis in women undergoing a first versus any subsequent laparoscopy. MATERIALS AND METHODS: Retrospective analysis of the National Endometriosis Clinical and Scientific Trials registry included women aged 18-50 years who had undergone TVUS prior to gynaecological laparoscopy. Diagnostic accuracy of TVUS was compared between women with and without a history of surgical endometriosis treatment, with laparoscopy and histopathology as the reference standards. Subgroup analysis was undertaken by disease phenotype. RESULTS: There were 470 ultrasound-laparoscopy dyads included for analysis (first laparoscopy n = 245 vs. subsequent laparoscopy n = 225), with histopathology available for 443 cases (first laparoscopy n = 234 vs. subsequent laparoscopy n = 209). Diagnostic accuracy did not differ between the two groups when compared to either laparoscopy (first laparoscopy: AUC 0.58 [95% CI 0.31-0.85], sensitivity 40.7% [34.7%-47.0%], specificity 75.0% [29.0%-96.0%] vs. subsequent laparoscopy: AUC 0.60 [0.46-0.74], sensitivity 46.2% [39.6%-52.9%], specificity 73.3% [47.5%-89.3%]) or histopathology (first laparoscopy: AUC 0.54 [0.31-0.77], sensitivity 40.8% [34.6%-47.3%], specificity 66.7% [29.6%-90.4%] vs. subsequent laparoscopy: AUC 0.56 [0.45-0.67], sensitivity 47.5% [40.4%-54.8%], specificity 64.3% [45.7%-79.3%]). CONCLUSIONS: Diagnostic accuracy did not differ based on previous endometriosis surgery. Community-based TVUS performs considerably more poorly than research suggests, missing 55.5% of endometriosis overall and 63.2% of superficial disease, reinforcing that a negative ultrasound does not exclude endometriosis.