Evaluation of endocervical curettage with conization in diagnosis of endocervical lesions.

Aim: Endocervical curettage (ECC) at the time of conization has been reported to be effective for diagnosing cervical intraepithelial neoplasia and/or early stage cervical cancer. We aimed to verify the accuracy of ECC with conization.

Methods: We retrospectively analyzed the records of 540 patients with suspected neoplastic cervical lesions who underwent conization at the Yokohama City University Hospital from January 2008 to December 2015. To validate the effectiveness of ECC for evaluating endocervical lesions, histopathologic findings from ECC samples were compared with those from endocervical specimens obtained by conization. In patients who subsequently underwent hysterectomy, specimens of residual endocervical stump lesions were compared with the specimens obtained by ECC.

Results: ECC was performed in 58.9% of patients who underwent conization. Positive findings were only observed in 7.9%, while negative findings were found in 67.3% of ECC samples; however, 24.8% of the samples were inadequate for diagnosis. None of the patients had an upgraded diagnosis according to ECC results. The sensitivity of ECC in predicting endocervical stump lesions that were identified by conization specimens was 25.0%, the specificity was 94.2% and the positive predictive value was 55.0% (κ = 0.238; P < 0.001). ECC samples yielded a sensitivity of 42.9%, a specificity of 83.9%, and positive predictive value of 54.5% (κ = 0.284; P = 0.053) in predicting residual endocervical lesions in the uterus.

Conclusions: As it offers low sensitivity and positive predictive value, ECC at the time of conization is of limited benefit for evaluating endocervical lesions.