An audit of liquid-based cervical cytology screening samples (ThinPrep and SurePath) reported as glandular neoplasia

Abstract

The aims of this study were to assess the number of cases diagnosed as glandular neoplasia (national report code 6) of cervical (6A) and non-cervical (6B) types on ThinPrep (TP) and SurePath (SP) liquid-based cytology (LBC) samples and to calculate the positive predictive value (PPV) of these diagnoses for significant glandular and/or squamous pathology for local audit and as a contribution to national data on glandular neoplasia. A computerized search identified all screening LBC samples reported as glandular neoplasia during the 24-month period from January 2006 to December 2007. Corresponding histology samples were identified, with a minimum follow-up period of 6 months for each case. A total of 70 samples, representing 70 patients, were reported as glandular neoplasia, 39 TP (55.7%) and 31 SP (44.3%), with 46 samples (31 TP, 15 SP) reported as 6A and 24 samples (eight TP, 16 SP) as 6B. PPV of glandular neoplasia was calculated for a biopsy diagnosis of cervical glandular intraepithelial neoplasia/adenocarcinoma and/or cervical intraepithelial neoplasia (CIN) 2 or worse. The PPV of 6A was 100% for both TP and SP. The PPV of 6B for adenocarcinoma was 62.5% for TP and 66.7% for SP. The combined PPV for 6A + 6B was 92.3% for TP, 83.3% for SP and 88.4% combined. The overall pick-up rates for the two methods were significantly different (TP 0.031%, SP 0.052%; P = 0.014). Histology showed only CIN3 with endocervical crypt involvement in nine TP cases and one SP case.