Switching from Neural Networks (PAPNET) to the Imager (Hologic) for Computer-Assisted Screening

Abstract

Objective: The large set of ThinPrep slides prepared in the Leiden Cytology and Pathology Laboratory is exploited for calculating the impact of the transition from PAPNET neural network scanning to the Imager technology. Study Design: All cervical samples were suspended and fixed in the coagulant fixative BoonFix. We compared 57,541 ThinPrep slides which were scanned by PAPNET and 64,273 ThinPrep slides processed with the Imager: 99,157 cases originated from the Dutch population screening program of asymptomatic women (screenees) and the remaining 22,657 samples were of symptomatic women. In the PAPNET series, 23% were diagnosed by additional light microscopy; in the Imager method, all slides were studied light microscopically. The cytoscores (positive cytology per 1,000 samples) were calculated for normal, atypical squamous cells of undetermined significance (ASC-US), cervical intraepithelial neoplasia (CIN) grades I–II, and for CIN III+. The odds ratios (ORs) for the positive cytoscores were assessed for both the screenees and the symptomatic women. Results: The cytoscores, per 1,000 cases, for ASC-US varied from 17.77 to 40.59, for CIN I–II from 7.17 to 33.35, and for CIN III+ from 2.81 to 8.8. These 6 cytoscores were higher for symptomatic women than for screenees. We observe significantly elevated ORs for the Imager for ASC-US (1.26 and 1.23), CIN I–II (1.45) and for CIN III+ (1.58 and 1.45). These 3 ORs are higher for screenees than for symptomatic women. Conclusion: The Imager technology is more efficacious, particularly for handling screenee slides.

Inter‐ and intraobserver agreement in whole‐slide digital ThinPrep samples of low‐grade squamous lesions of the cervix uteri with known high‐risk HPV status: A multicentric international study

“…In fact, whole‐slide digital cytology will profit enormously from artificial intelligence–based diagnostic algorithms. In the past, several attempts were made with automated screening systems, 44 , 45 , 46 which were based on glass slides. Applying artificial intelligence to virtual slides may open new perspectives, reducing the strain of the screening activity and helping in the interpretation of slides in the context of large population‐based screening programs.…”

Artificial Neural Networks as Decision Support Tools in Cytopathology: Past, Present, and Future

“…Finally, PAPNET was used for comparative studies involving human papillomavirus-related molecular tests 31 with colposcopy techniques 71 and has been compared with other non-ANN-based semiautomated screening systems, such as the Imager® by Hologic® 72 or the AutoPap® Guided System. 52 Cervix is not the sole organ of the female reproductive system.…”

BestCyte® primary screening of 500 ThinPrep Pap Test thin-layers: 3 Cytologists’ Interobserver diagnostic concordance with predicate manual microscopy relative to Truth Reference diagnoses defining NILM, ASCUS+, LSIL+, and ASCH+ thresholds for specificity, sensitivity, and equivalency grading

“…Thereafter, the ThinPrep Imaging System (TIS) (Hologic, Bedford, MD, USA) and the FocalPoint GS Imaging System (Becton Dickinson, Burlington, NC, USA) were introduced for primary screening in 2004 and 2008, respectively, to complement proprietary (i.e., ThinPrep and SurePath) thin-layers and to facilitate final case reporting using Bethesda nomenclature. 2 Comparative studies by Boon et al 11 in 2010 reported significantly higher odds ratios for detecting abnormal cells in the ASCUS, CIN I-II, and CIN III+ diagnostic categories, and notably in asymptomatic routinely screened women, thus they concluded improved abnormal cell detection rates may be realized when applying TIS on ThinPrep thin-layers relative to PAPNET.…”