Use of the ThinPrep Imaging System for internal quality control of cervical cytology


T. Heard, Anatomical Pathology & Cytopathology, Pathology Queensland – Central, Level 2 Block 7, Royal Brisbane & Women's Hospital. Herston QLD 4029
Tel.: +61 7 3646 8241; Fax: +61 7 3646 8908


T. Heard, A. Chandra, G. Culora, S.S. Gupta, A. Herbert and M. Morgan
Use of the ThinPrep Imaging System for internal quality control of cervical cytology

Objective:  To audit the use of the ThinPrep Imaging System (TIS) for internal quality control (IQC) in the place of rapid review (RR), and to compare its performance with routine primary screening.

Method:  During 9 months, 16 462 ThinPrep slides were processed by TIS. Slides were initially reviewed using the TIS review scope, as recommended by the manufacturer: 22 fields of view were observed and, if considered abnormal, a full microscopic review was conducted using the review scope. Different biomedical scientists (BMSs), working on each procedure in rotation, performed batches of TIS-assisted quality control and routine primary screening independently on unmarked slides. Any slides with abnormalities detected by either method were referred to a consultant pathologist or advanced BMS practitioner for a final report. TIS results were compared with both previous records of RR and routine primary screening carried out on the same slides. We used the UK terminology in which ‘dyskaryosis’ is equivalent to squamous intraepithelial lesion (SIL) and borderline to atypical (including squamous and glandular cells).

Results:  TIS preview detected significantly more high-grade dyskaryosis compared with RR during the previous 4 years: 2.0–4.2 compared with 0.1–1.8 detected per 1000 slides (= 0.0001). TIS and routine screening were equivalent in sensitivity and specificity for the final cytology result, but BMSs were significantly more likely to classify slides as dyskaryotic rather than borderline when using TIS compared with routine screening. Referrals for potentially high-grade abnormalities detected by TIS-assisted IQC alone found 28 biopsies of at least cervical intraepithelial neoplasia grade 2 (CIN2+), whereas 15 CIN2+ biopsies were found on routine screening but missed using TIS. There was no significant change in the rates of inadequate tests, high- or low-grade cytological abnormalities, or positive predictive value for CIN2+ when TIS was in use.

Conclusions:  Screening with TIS was more sensitive than RR for IQC, providing a rescreening method equivalent to routine primary screening in overall accuracy.