Background: We aimed to identify the number of histologically verified cervical intraepithelial neoplasia (CIN)/cervical cancer lesions detected by endocervical cytobrush (EC) which have been missed by repeat PAP smear and colposcopy before large loop excision of the transformation zone (LLETZ).
Patients and methods: A retrospective cohort study of 1,676 consecutive patients in a colposcopy clinic undergoing repeat PAP smear, colposcopy, biopsy, and subsequent LLETZ.
Results: Data were available for 1,421 patients. EC identified 31/1,367 (2.2%) cases of CIN and/or cervical cancer missed by repeat PAP smear and colposcopy. Compared to repeat PAP smear and colposcopy, the combination of repeat PAP smear, colposcopy and EC increased the positive predictive value (PPV) (89.6% vs. 98.9%, p=0.07), but this difference was not statistically significant. Sensitivity (93.1% vs. 93.9%; p=0.8), specificity (27.7% vs. 27.7%; p=1.0), and negative predictive value (NPV) (11.9% vs. 13.1%, p=0.9) were also not significantly different. The number needed to screen (NNS) for identifying one additional case of CIN and/or invasive cancer by EC was 45. In a multivariate analysis, presence of CIN/cervical cancer in the LLETZ specimen and human papilloma virus infection, but not age, visibility of the transformation zone or presence of an endovcervical lesion were independently associated with the likelihood of CIN and/or invasive cancer detected by EC.
Conclusion: Adding EC to repeat PAP smear and colposcopy identifies 2.2% more cases of CIN and/or cervical cancer, but does not significantly increase the sensitivity, specificity, PPV or NPV of repeat PAP smear and colposcopy.