Superior performance of liquid-based versus conventional cytology in a population-based cervical cancer screening program

Objective. Liquid-based cytology may offer improvements over conventional cytology for cervical cancer screening. The two cytology techniques were compared in a group of 86,469 women who participated in a population-based screening program. Using a nation-wide pathology database containing both cervical cytology and histology records for all patients, we compared the outcome of the two screenings methods with regard to the detection rate of histological proven abnormalities and the determination of the true false-negative rates for both methods. Methods. Two cohorts of women living in the same geographical region were used. Cohort 1 (n = 51,154 women) was analysed using conventional cytology (conventional cohort) and cohort 2 (liquid cohort) (n=35,315 women) was analysed using liquid-based cytology (SurePath (R)). The samples were processed in one laboratory. The results of histological follow up were available via a central database. Results. The rate of unsatisfactory slides was significantly lower using liquid-based cytology (0.13% vs. 0.89%, p<0.0001). Detection of ASCUS+ (Atypical squamous cells of unknown significance OF higher abnormalities) was significantly higher using liquid-based cytology (2.97% vs. 1.64%, p<0.0001), mainly due to the increase in the ASCUS category. The percentage of histological abnormalities within the ASCUS samples was approximately equal in both cohorts, indicating that more true abnormal cases were detected using liquid-based cytology. The sensitivity for detection of a histological proven lesion is significantly higher in the liquid cohort compared to the conventional cohort (96.2% vs. 92.0%), with only a slight difference in specificity (97.8% vs. 98.2%). Conclusion. This population study confirmed previous institution-based reports of decreased numbers of unsatisfactory samples based on liquid-based cytology and showed an increased sensitivity for the detection of cytological abnormalities that was validated by subsequent histological investigation. (C) 2008 Elsevier Inc. All rights reserved.