Objective: To estimate the sensitivity and specificity of visual inspection using acetic acid as a primary screen for cervical intraepithelial neoplasia (CIN).
Methods: Visual inspection was done on 1997 women aged 35-45 years in a screening trial in rural China. Each women had colposcopy and at least five cervical biopsies (directed biopsy of lesions, one biopsy at 2, 4, 8, or 10 o'clock at the squamocolumnar junction in each normal quadrant, and an endocervical curettage).
Results: Forty-three women had biopsy-proven CIN II, 31 had CIN III, and 12 had invasive cancer. In two women only the endocervix was positive (one with CIN II and one with CIN III). Visual inspection yielded normal results in 1445 women (72%), low-grade intraepithelial neoplasia in 525 (26%), high-grade in 21 (1%), and cancer in six (0.3%). With abnormal visual inspection defined as low-grade intraepithelial neoplasia or worse, the sensitivity for detecting biopsy proven CIN II or worse was 71% (61 of 86, 95% confidence interval [CI] 60%, 80%); the specificity was 74% (1420 of 1911, 95% CI 72%, 76%); the sensitivity was 65% for smaller lesions (37 of 57, 95% CI 51%, 77%), and 89% for larger lesions (24 of 27, 95% CI 71%, 98%) (P =.03).
Conclusion: The sensitivity of visual inspection equaled or exceeded reported rates for conventional cervical cytology. Visual inspection and colposcopy have similar specificity profiles for CIN II and greater. The benefit of an inexpensive point-of-care diagnosis and treatment algorithm will be a powerful incentive to pursue visual inspection for cervical cancer screening in developing countries.