Liquid-based cytology for primary cervical cancer screening: a multi-centre study.

The aim of this six-centre, split-sample study was to compare ThinPrep fluid-based cytology to the conventional Papanicolaou smear. Six cytopathology laboratories and 35 gynaecologists participated. 5428 patients met the inclusion criteria (age > 18 years old, intact cervix, informed consent). Each cervical sample was used first to prepare a conventional Pap smear, then the sampling device was rinsed into a PreservCyt vial, and a ThinPrep slide was made. Screening of slide pairs was blinded (n = 5428). All non-negative concordant cases (n = 101), all non-concordant cases (n = 206), and a 5% random sample of concordant negative cases (n = 272) underwent review by one independent pathologist then by the panel of 6 investigators. Initial (blinded) screening results for ThinPrep and conventional smears were correlated. Initial diagnoses were correlated with consensus cytological diagnoses. Differences in disease detection were evaluated using McNemar's test. On initial screening, 29% more ASCUS cases and 39% more low-grade squamous intraepithelial lesions (LSIL) and more severe lesions (LSIL+) were detected on the ThinPrep slides than on the conventional smears (P = 0.001), including 50% more LSIL and 18% more high-grade SIL (HSIL). The ASCUS:SIL ratio was lower for the ThinPrep method (115:132 = 0.87:1) than for the conventional smear method (89:94 = 0.95:1). The same trend was observed for the ASCUS/AGUS:LSIL ratio. Independent and consensus review confirmed 145 LSIL+ diagnoses; of these, 18% more had been detected initially on the ThinPrep slides than on the conventional smears (P = 0.041). The ThinPrep Pap Test is more accurate than the conventional Pap test and has the potential to optimize the effectiveness of primary cervical cancer screening.