Evaluating the use of virtual microscopy in cytology education.

Abstract

Whole slide imaging is a promising tool for cytology. In the present study, we assessed the performance of and user experience with virtual microscopy (VM) to determine its feasibility and usage in an educational setting. From January 1 through August 31, 2022, 46 Papanicolaou slides, of which 22 (48%) were abnormal, 23 (50%) were negative, and 1 (2%) was unsatisfactory, were reviewed by the students using both VM and light microscopy (LM) platforms. In addition to assessing VM's overall performance, the SurePath imaged slides' accuracy was reviewed as a potential alternative to ThinPrep because of its cloud storage advantage. Finally, the students' weekly feedback logs were analyzed to gain insights for improving the digital screening experience. The overall diagnostic concordance difference was significant between the 2 screening platforms (Z= 5.38; P < 0.001), favoring LM (86% correct diagnosis) over VM (70% correct diagnosis). The overall sensitivity of VM and LM was 54.0% and 89.6%, respectively. VM also had an overall higher specificity (91.8%) compared with LM (81.3%). LM performed better than whole slide imaging for the correct identification of an organism when one was present, with 77.6% sensitivity compared with 58.9% for the digital platform. The rate of agreement for the SurePath imaged slides with the reference diagnosis was 74.3% compared with 65.7% for the ThinPrep slides. Finally, 4 themes were discerned from reviewing the user logs, with issues about image quality and the lack of fine focus functionality most frequently mentioned, followed by themes associated with a higher learning curve and novelty associated with the digital screening. Although the VM results were poorer than the LM results in our validation, its use in an educational setting is promising considering the continued technological improvements and the renewed focus on improving the digital user experience.