Reliability and diagnostic performance of smartphone colposcopy

Abstract

Objective: To evaluate the interobserver and intraobserver reliability of smartphone colposcopy (SPC) versus conventional colposcopy and to determine diagnostic performance.Methods: A smartphone back camera was used to capture cervical images before and after application of acetic acid, and after application of lugol solution. Captured images were reviewed independently by two experienced colposcopists and findings were noted as per colposcopy. Smartphone-based diagnostic performance was calculated, and kappa statistics were used for measurement of agreement between SPC and conventional colposcopy findings.Results: A total of 114 women were included in the study. The kappa statistic for intraobserver reliability was 0.77 for both normal colposcopic findings and the transformation zone, indicating substantial agreement. Kappa values were 0.54 for acetowhite epithelium, 0.51 for lugol staining, and 0.51-0.60 for atypical vascularization.Kappa values for interobserver reliability were 0.76 for normal colposcopic findings, 0.56 for acetowhite epithelium, and 0.60 for lugol staining. The sensitivity, specificity, PPV, and NPV of SPC for CIN2+ were 88.2 (95% CI, 72.5-96.7), 48.7 (95% CI, 37.4-60.2), 0.42 (95% CI, 0.36-0.48), and 0.91 (95% CI, 0.79-0.96), respectively. Conclusion:SPC showed substantial agreement between the histologic diagnoses based on the captured images and conventional colposcopic findings.

Development of a prognostic prediction support system for cervical intraepithelial neoplasia using artificial intelligence-based diagnosis

“…As a result, we believe it is necessary to increase the number of low-resolution studies and to improve compatibility and feasibility in the future. If this type of deep learning is achieved, the algorithm could be applied in colposcopy using smartphones [ 40 ], which has been reported in recent years. Third, we considered only one prognostic factor in the pre-analysis.…”