BD focalpoint slide profiler performance with atypical glandular cells on SurePath Papanicolaou smears

Abstract

Background: The FocalPoint Slide Profiler is an automated cervical cytology screening system that is approved for primary screening. It identifies up to 25% of slides as requiring No Further Review. However, few studies have evaluated FocalPoint performance with glandular abnormalities.

Methods: Sixty-six SurePath Papanicolaou (Pap) tests with a diagnosis of atypical glandular cells were identified. A total of 172 Pap tests with a diagnosis of "endometrial cells present" were included as controls. Follow-up histology was abnormal if diagnosed as high-grade squamous intraepithelial lesions, adenocarcinoma in situ, carcinoma, or complex endometrial hyperplasia. The FocalPoint software ranked each case into 1 of 7 categories: quintiles 1 (high risk) through 5 (low risk), No Further Review, and Process Review.

Results: A total of 215 slides were qualified for review; 38 (57.6%) atypical glandular cells cases were abnormal on follow-up biopsy, and 27 (71.1%) atypical glandular cells with abnormal follow-up qualified for review; no cases were classified No Further Review, and 9 (33%) were ranked in quintile 1. Twenty-three (82.1%) atypical glandular cells with benign follow-up were qualified for review; 3 (11%) cases were classified No Further Review, and 4 (17%) were ranked in quintile 1. There was a statistically significant difference between the ranking of benign atypical glandular cells cases, abnormal atypical glandular cells cases, and control cases (P = .03). However, when collapsed into No Further Review versus all other quintiles, the differences were not significant (P = .20).

Conclusions: The FocalPoint Slide Profiler did not classify glandular lesions with abnormal follow-up in the No Further Review category. However, these cases were not preferentially ranked in quintile 1. FocalPoint-screened slides need to be carefully reviewed for glandular abnormalities, regardless of the quintile ranking.