RESUMO
PURPOSE: To evaluate the rate of atypical hyperplasia (AH) underestimating endometrial cancer (EC) comparing endometrial biopsy (EB) accomplished by hysteroscopic biopsy with dilatation and curettage (D&C). Second, to compare the two techniques to foresee EC grading. METHODS: This trial was based on the findings of two Gynecological Departments within the same Public Utility, sharing pathological service and database but routinely performing EB under hysteroscopic visualization (group A) or hysteroscopy followed by D&C (group B). We retrieved the clinical data of patients showing EC on hysterectomy throughout a 10-year period. The accuracy of hysteroscopic-view diagnosis and EB pathology were compared, having the pathologic findings of hysterectomy as reference. RESULTS: A total of 161 patients met the inclusion criteria. Among these, 109 and 52 were included in groups A and B, respectively. In group A, 32.1% of patients underwent EB in an out-patient setting. To foresee EC, hysteroscopic view showed a sensitivity of 82.5% and 70.2% in groups A and B, respectively (P = 0.019). An underestimation of EC diagnosed as AH on EB was found in 20 patients (12.4%). Among these, 18 (16.5%) and 2 (3.8%) were included in groups A and B, respectively (P = 0.022). In group A, a fault diagnosis of AH resulted higher when EB was performed as out-patient setting (P = 0.006). EB allowed the grading of EC in 73.3% and 90.3% of patients in groups A and B, respectively. The agreement was 73.7% and 85.1%, leading to moderate (κ = 0.56) and good (κ = 0.77) "κ" coefficient of concordance for groups A and B, respectively. CONCLUSIONS: EB performed by D&C lowers the rate of AH underestimating concurrent EC and improves the grading agreement when compared with hysteroscopic sampling.