Hysteroscopically targeted biopsies compared with blind samplings in endometrial assessment of menopausal women taking tamoxifen for breast cancer

STUDY OBJECTIVE: To determine the validity of tissue sampling accomplished by hysteroscopically targeted or blind biopsies in the assessment of endometrial morbidity associated with tamoxifen treatment. DESIGN: Retrospective, unrandomized study (Canadian Task Force classification II-2). SETTING: Public hospital. PATIENTS: One hundred seventy-six menopausal women who had an endometrial stripe of more than 4 mm on transvaginal ultrasonography. INTERVENTION: Review of histopathologic reports of patients undergoing hysteroscopy followed by targeted (94 samplings) or blind (82 samplings) endometrial biopsies. MEASUREMENTS AND MAIN RESULTS: Histopathology was considered the reference test to assess endometrial morbidity, and correlates with hysteroscopic findings were made to evaluate the validity of the two sampling procedures. Overall, in 23 women (13.0%) tissue samples were insufficient for pathologic evaluation. Functional or atrophic endometrium and cystic atrophy were found in 51 (28.8%) and 37 patients (21.0%), respectively. Polyps, hyperplasias, and carcinomas were found in 38 (21.5%), 19 (10.7%), and 6 (3.3%), respectively. Blind biopsies failed to detect 5 of 5 polyps and 33 of 37 cystic atrophies, and in 34.1% of cases provided insufficient tissue for diagnosis; however, no hyperplasias or carcinomas were undetected. All specimens collected under vision were pathologically evaluable; 34 of 38 hysteroscopic reports of cystic atrophy were confirmed, and neither endometrial polyps nor hyperplasias and carcinomas were undetected. In distinguishing between normal and abnormal endometrium, hysteroscopy showed sensitivity and negative predictive value of 100% regardless of sampling modality. We found better specificity (80.0% vs 68.9%) and positive predictive value (68.9% vs 43.7%) for hysteroscopic diagnosis when tissue was collected under vision compared with blind sampling. CONCLUSION: In women taking tamoxifen, endometrial evaluation performed by blind sampling is safe in excluding hyperplasias or carcinomas. For complete understanding of tamoxifen-associated morbidity, hysteroscopy with sampling under vision has better diagnostic compliance.