Objective: Visualization and digital documentation are critical to many cervical cancer programs worldwide. Practitioner competency remains a major limitation of visualization-based programs. High variability in image quality (cervix positioning and sharpness) is common among practitioners. To improve image quality on the smartphone-based Enhanced Visual Assessment (EVA) System, a framing ring designed to guide the user to properly position and scale the cervix within the image was inserted into the app. In this paper, the effects of the framing ring on cervix position and sharpness in digital cervicography images were quantitatively analyzed.
Methods: The framing ring was developed as a smartphone feature on the EVA System app. Anonymized images from MobileODT’s image portal acquired with and without the framing ring feature were selected. Comparisons were made in proper framing of the cervix, and in two measurements of image sharpness – a calculated Brenner score and manual labeling.
Results: Altogether, 911 images without the framing ring were analyzed, of which 695 (76%) were judged as well-framed. The mean Brenner and manual scores were 3.39 and 1.12, respectively. 174 images with the framing ring were analyzed, of which 159 (91%) were judged as well-framed. The mean Brenner and manual scores were 4.18 and 1.66, respectively. Conclusions: Cervix localization and sharpness remain challenges in digital cervicography. The framing ring has the potential to address cervix localization and sharpness in the images. In time this feature, in combination with other methods, could improve the performance of the EVA system as a tool for quality control in digital cervicography and developing machine learning algorithms of cervical pathology in images