Comparison of manual and automated methods of liquid-based cytology. A morphologic study

Abstract

To evaluate the morphologic characteristics of gynecologic samples prepared by 3 different methods of liquid-based cytology. Objective: To evaluate the morphologic characteristics of gynecologic samples prepared by 3 different methods of liquid-based cytology.

Study design: Cytologic samples from representative cases of each diagnostic category of squamous epithelial lesion, prepared by automated and manual liquid-based systems, were analyzed by 3 laboratories in the United States, Portugal and Brazil. The systems included: ThinPrep (automated, U.S. Food and Drug Administration approved; Cytyc Corp., Boxborough, Massachusetts, U.S.A.), Autocyte PREP (South American system, manual; TriPath Imaging, Inc., Burlington, North Carolina, U.S.A.) and DNACITOLIQ (manual; Digene Brazil, São Paulo, Brazil). A panel of 16 morphologic parameters was evaluated: cellularity, clean background, uniform distribution, artifacts, cellular overlapping, architectural and cytoplasmic distortion, cytoplasmic vacuolization, cellular elongation, imprecise cytoplasmic borders, folded cytoplasmic borders, nuclear hyperchromasia, coarse chromatin, prominent nucleolus, irregular nuclear borders, atypical mitosis and inflammatory infiltrate. Negative, atypical squamous cells of undetermined significance, low grade squamous intraepithelial lesion (LSIL) and high grade squamous intraepithelial lesion (HSIL) cases were included. Cases without biopsies were confirmed by consensus.

Results: Cellularity was adequate in all samples. Clean background was observed in the vast majority of samples with all liquid-based systems. Uniform distribution was frequently found in ThinPrep and Autocyte PREP samples but not in DNACITOLIQ. Artifacts were not present in DNACITOLIQ samples, rare in ThinPrep and observed in 8 (34.7%) Autocyte PREP. Cellular overlapping was observed in all 3 system samples: 11 (31.42%) cases in ThinPrep, 16 (69.56%) in Autocyte PREP and 17 (68%) in DNACITOLIQ System. Architectural and cytoplasmic distortion were present in 3 cases of HSIL (13%) and cytoplasmic vacuolization in 2 cases of LSIL and 1 HSIL of Autocyte PREP. Cellular elongation was found in 13 (56.5%) Autocyte PREP and in 5 (20%) DNACITOLIQ samples. Inflammatory infiltrate was found in all 3 system samples but with more frequency in the Autocyte PREP (69.56%) and DNACITOLIQ System (72%).

Conclusion: This study clearly indicated that in spite of the different methodologies, the 3 methods adequately preserved cellular structure for morphologic evaluation. The choice of method will depend on price, availability and procedures involved.