To determine the usefulness of computerized rescreening for detecting high-grade cervical intraepithelial neoplasia (CIN) in patients with cytological evidence of atypical squamous cells of undertermined significance (ASCUS).Methods.
One hundred and eight patients with ASCUS cytology who had undergone colposcopy had their Papanicolaou smears reviewed using PAPNET, a neural network-based instrument.Results.
Twenty (18.5%) of the 108 slides were reclassified to a higher-grade lesion. Analysis showed sensitivity (77.8%), specificity (86.9%), and positive (35%) and negative (97.7%) predictive values for detection of high-grade CIN.Conclusions.
PAPNET neural network-based rescreening of ASCUS smears identifies patients with high-grade cancer precursors. It may be a triage test for ASCUS cytology.