Papnet-assisted, primary screening of cervico-vaginal smears

Pu,pose: The Papnet system was initially designed for rescreening negative Pap tests but may also be an effective primary screener. Methods: A set of 2,200 archival slides diagnosed by conventional, manual screening as 2,000 (90%) WNL, 47 (2.1%) carcinomas, 50 (2.3%), HSIL, 50 (2.3%) LSIL, and 53 (2 .4%) ASCUS/AGUS were compared to the results of Papnet-assisted, primary sc reening. Following Papnet scanning, the digitized images were triaged and c lassified as abnormal or negative. All abnormals had a full manual screenin g, whereas negatives had a limited screening. Results by each screening method were compared and discordant cases were pe er reviewed for a consensus result. Screening efficacy by each method was m easured against a standard result composed of the concordant and consensus results. Results: There were 101 concordant and 181 discordant abnormal res ults. The standard result for the slide set was 1,953 (88.9%) WNL, 87 (3.9% ) ASCUS/AGUS, 52 (2.4%) LSIL, 62 (2.8%) HSIL, 39 (1.8%) carcinomas, and 5 ( 0.2%) unsatisfactory. Papnet versus manual sensitivity rates were 87.6% vs 72.3% at the ASCUS/AGUS threshold, 85.6% vs 82.4% at the LSIL threshold, an d 89.1% vs 90.1% at the HSIL thereshold. Conclusions: Papnet-assisted, primary screening equals conventional, manual screening in the detection of a wide range of cell abnormalities and is mo re effective in the detection of abnormalities at the lower end of the abno rmal spectrum.