Objective: To estimate the concordance rate in interpretation of cervicography. Methods: Three reviewers who practiced as a gynecologic oncologist over 5 years independently reviewed 392 cervicography images from patients with confirmed Cervical Intraepithelial Neoplasia (CIN) diagnosis (low grade squamous intraepithelial lesion (LSIL) in 47, high grade squamous intraepithelial lesion (HSIL) in 345). Reviewers were blinded to the diagnosis of each case and the composition of diagnosis in 392 cervicography images. After evaluating the quality of cervicography images, reviewers graded cervicography images as LSIL vs HSIL based on the first impression. Then, reviewers determined the presence or absence of findings: acetowhite lesion, mosaicism, punctuation, atypical vessels, cancerous mass, ulceration. Finally, reviewers graded again the cervicography images as LSIL vs HSIL. The concordance rate between reviewers was defined as portion of cases in which all three reviewers graded same diagnosis (LSIL vs HSIL). Results: Over two of three reviewers evaluated 67 cervicography images as not evaluable (67/392, 17%) and excluded from the further analysis. The sensitivity and specificity of reviewers were poor but compatible to numbers in literatures (sensitivity 0.48 0.56, specificity 0.55 0.76 by first impression; sensitivity 0.52 0.56, specificity 0.55 0.74 by final impression). However, the concordance rate is only 0.57 based on the first and final impression. Conclusion: In interpreting cervicography, the concordance rate between reviewers is poor. Novel technology with more Objective interpretation is necessary. Acknowledgements: This work was supported by the Technology Innovation Program funded By the Ministry of Trade, Industry and Energy (MOTIE) of Korea (10049785, Developm)