Background: The Papanicolaou test has reduced mortality from cervical cancer but has not completely eradicated the disease. A reason for the false negative screening errors may be found in one of three steps of the test (obtaining, processing and interpretation of the cervicovaginal smears). The possible false negative primary screening errors may be detected with the help of rescreening. Therefore, an automated quality control by means of a system such as in the PAPNET is necessary. The PAPNET combines the use of algorithmic image analysis with neural networks and allows to conventionally reassess cervicovaginal smears. Nevertheless, the cytopathologist always decides whether to reassess the sample with the optic microscope as suggested by the presence of abnormal cells in the frames selected by the PAPNET system. Therefore the cytopathologist always formulates the conclusive diagnosis.
Methods: In this study, we rescreened 300 cervicovaginal smears.
Results: Only 122 smears have been reassessed by the optic microscope. The human papillomavirus infection and the SIL (squamous Intraepithelial Lesion) diagnoses formulated by means of rescreening with the PAPNET have coincided with the primary screening diagnoses. The only exception was a human papillomavirus infection not previously detected.
Conclusions: Our study indicates that the use of the PAPNET system facilitates the quality control measures and has reduced the false negatives.