STUDY OBJECTIVE: To evaluate positive predictive value (PPV) of visual diagnosis at laparoscopy compared with biopsy findings according to severity of endometriosis. DESIGN: Retrospective study (Canadian Task Force classification II-2). SETTING: Academic referral center. PATIENTS: Women who underwent laparoscopic biopsies for suspected endometriosis. INTERVENTIONS: A total of 238 biopsy specimens (73 endometriomas and 165 peritoneal implants) were taken from 104 patients undergoing laparoscopy for evaluation of chronic pelvic pain thought to be caused by endometriosis. MEASUREMENTS AND MAIN RESULTS: Accuracy of laparoscopic findings compared with histology-proved endometriosis by severity of disease and location of endometriotic lesions. Overall PPV per patient was 86.5%, which was 75.8% for stage I disease compared with 89.7%, 100%, and 90.6%, respectively, for disease stages II to IV (p = .037). The PPV per biopsy specimen of stages I to IV endometriosis was 66.1%, 78.0%, 92.0%, and 81.1%, respectively (.049). When endometriomas and peritoneal biopsy specimens were analyzed separately, no difference in PPV existed (79% vs 77%; p = .67). CONCLUSION: High overall PPV existed in our study, especially in patients with advanced disease. The PPV per patient was higher than the PPV per biopsy specimen indicating that ability to diagnose endometriosis may be improved by performing multiple biopsies. This is particularly true in stage I where failure to confirm may be greatest.
STUDY OBJECTIVE: To evaluate positive predictive value (PPV) of visual diagnosis at laparoscopy compared with biopsy findings according to severity of endometriosis. DESIGN: Retrospective study (Canadian Task Force classification II-2). SETTING: Academic referral center. PATIENTS: Women who underwent laparoscopic biopsies for suspected endometriosis. INTERVENTIONS: A total of 238 biopsy specimens (73 endometriomas and 165 peritoneal implants) were taken from 104 patients undergoing laparoscopy for evaluation of chronic pelvic pain thought to be caused by endometriosis. MEASUREMENTS AND MAIN RESULTS: Accuracy of laparoscopic findings compared with histology-proved endometriosis by severity of disease and location of endometriotic lesions. Overall PPV per patient was 86.5%, which was 75.8% for stage I disease compared with 89.7%, 100%, and 90.6%, respectively, for disease stages II to IV (p = .037). The PPV per biopsy specimen of stages I to IV endometriosis was 66.1%, 78.0%, 92.0%, and 81.1%, respectively (.049). When endometriomas and peritoneal biopsy specimens were analyzed separately, no difference in PPV existed (79% vs 77%; p = .67). CONCLUSION: High overall PPV existed in our study, especially in patients with advanced disease. The PPV per patient was higher than the PPV per biopsy specimen indicating that ability to diagnose endometriosis may be improved by performing multiple biopsies. This is particularly true in stage I where failure to confirm may be greatest.