Diagnostic Accuracy of Fluorescein Sodium for Targeted Cervical Biopsies

Abstract

Background: Visual inspection methods for cervical cancer screening are widely used in low resource settings. Fluorescent sodium could improve accuracy of cancer screening. This study aimed to assess diagnostic accuracy of fluorescein sodium (FNa) to detect cervical neoplasia. Methods: Seventy consecutive patients referred for colposcopy were enrolled prospectively. Acetic acid, Lugol’s iodine, and FNa were used sequentially. Biopsies were taken from all abnormal areas. If there was no obvious abnormality, two random biopsies and endocervical curettage were done. Reference standard was the highest grade lesion on cervical biopsy with a threshold of CIN2+. The patterns of each staining agent were recorded as absent, faint, or distinct. Diagnostic accuracy estimates with 95% confidence intervals were calculated. Correlation between the various tests were also determined using the kappa statistic. Results: There were 27 cases of CIN2+ (38.6%). The sensitivity of any fluorescence for CIN2+ was 82% (62, 94) and for distinct fluorescence was 59% (39, 78). The specificity was 65% (49, 79) for any fluorescence and 95% (84, 99) for distinct fluorescence, the same as for Swede score > 7. For any fluorescence, the positive likelihood ratio was 2.34 (1.5, 3.65) and the negative likelihood ratio was 0.28 (0.13, 0.65). For distinct fluorescence, the positive likelihood ratio was 12.74 (3.18, 51.1) and the negative likelihood ratio was 0.43 (0.27, 0.68). There was moderate correlation between FNa and the other tests. Conclusion: Distinct fluorescence with FNa was very specific, low cost, and easy to perform and may contribute to confirm CIN2+ disease.