RESUMO
OBJECTIVES: Our aim was to evaluate the feasibility and performance of smartphone digital images for the detection of cervical intraepithelial neoplasia of grade 2 or worse (CIN2+) as an adjunct to a conventional visual inspection approach with acetic acid (VIA) and Lugol's iodine (VILI), in comparison with detection by histopathologic examination. METHODS: Three hundred women were primarily screened for human papillomavirus (HPV) using self-collected vaginal specimens. Human papillomavirus-positive women were then invited for VIA and VILI, which were interpreted as (i) nonpathological, (ii) pathological or, (iii) inconclusive. Cervical smearing, endocervical brushing, and cervical biopsies were performed. Digital images of the cervix were taken with a smartphone and evaluated offsite by experienced health care providers. Sensitivity and specificity for CIN2+ were compared between on-site and off-site observers, using histopathological diagnosis as the criterion standard. RESULTS: Eighty-eight HPV-positive women were screened for cervical cancer. Overall, 7 cases of CIN2+ (8.0%) were diagnosed using biopsy specimens. The on-site physician obtained a sensitivity of 28.6% (95% confidence interval [95% CI], 3.7-71) and a specificity of 87.2% (95% CI, 77.7- 93.7). The off-site physicians obtained a sensitivity ranging between 42.9% (95% CI, 9.9-81.6; p = 1) and 85.7% (95% CI, 42.1-99.6; p = .13) and a specificity between 48.1% (95% CI, 36.5- 59.7; p < .001) and 79.2% (95% CI, 68.5-87.6; p = .10). Comparison between observers did not reach significance. Observers assessed 95.6% of all images as very good or acceptable for interpretation purpose. CONCLUSION: Smartphone images may be a useful adjunct to conventional VIA and VILI for the detection of CIN2+ and improve cervical cancer screening in low-resource settings.