Adaptation of CytoProcessor for cervical cancer screening of challenging slides. Issue 9 (20th May 2019)

• Record Type: Journal Article
• Authors: Crowell, Elizabeth F.; Bazin, Cyril; Thurotte, Vianney; Elie, Hubert; Jitaru, Laurette; Olivier, Grégoire; Caillot, Yann; Brixtel, Romain; Lesner, Boris; Toutain, Matthieu; Renouf, Arnaud
• Is Part Of: Diagnostic cytopathology. Volume 47:Issue 9(2019)
• Page Start: 890
• Page End: 897
• Publication Date: 2019-05-20
• Subjects: artificial intelligence -- cervical screening -- digital pathology -- gynecologic cytology -- Papanicolaou test; Cytodiagnosis -- Periodicals; Pathology, Cellular -- Periodicals; Cytology -- Periodicals; 616.07582
• DOI: 10.1002/dc.24213
• Languages: English
• ISSNs: 8755-1039

Abstract

Abstract: Background: Current automated cervical cytology screening systems require purchase of a dedicated preparation machine and use of a specific staining protocol. CytoProcessor (DATEXIM, Caen, France) is a new automated system, designed to integrate seamlessly into the laboratory's existing workflow. We previously demonstrated the superior performance of CytoProcessor for diagnosis of ThinPrep slides compared to the ThinPrep Imaging System (HOLOGIC, Marlborough, MA). Next, we analyzed whether CytoProcessor technology can be adapted for use on Novaprep slides. Methods: Using artificial intelligence, we developed a new algorithm in CytoProcessor for the analysis of slides prepared using the NOVAPREP Processor System NPS50 (Novacyt, Vélizy‐Villacoublay, France). A representative population of 309 cases was selected from the routine workflow in a public hospital. We compared the diagnoses made using CytoProcessor or conventional screening with a microscope. All discordances were resolved by a consensus committee. Results: The performance of CytoProcessor in terms of diagnostic accuracy on Novaprep slides was very similar to that observed previously on ThinPrep slides. Compared to conventional screening, CytoProcessor slightly improves diagnostic sensitivity while maintaining a statistically equivalent specificity. Diagnosis was reached 1.6 times faster with CytoProcessor compared to using a microscope. Conclusion: CytoProcessor is a robust automated cervical cytology screening system that can be used successfully with samples having very different characteristics. As previously shown, CytoProcessor confers significant gains in processing time and diagnostic precision. CytoProcessor is accessible through a secured internet connection, making remote diagnosis of Papanicolaou tests possible.