Rescreening of Cervical Papanicolaou Smears Using PAPNET

To the Editor.— The article by Dr O'Leary and colleagues¹ and the Editorial by Dr Garber² present a misleading view of automated cervical cytology smear rescreening. Although the authors briefly suggest that PAPNET rescreening might prove more effective in a higher-risk population, neither article adequately points out how unusual a setting was studied. The women of the armed forces have outstanding access to health care services and are encouraged to avail themselves of them liberally. The low prevalence of lesions expected in this population is confirmed by the unusually low number of abnormal Pap smears found in the study by O'Leary et al. Furthermore, PAPNET was used "off label" in this study and was applied to Pap smears already twice manually rescreened; the indicated use of PAPNET is as the first rescreening after examination by a cytologist. Because the smears actually examined with PAPNET had already been twice filtered for false negatives, the initially small number of false-negative smears in the sample was dramatically reduced. Accordingly, it is remarkable that PAPNET found as many as it did. The comparison with 100% manual rescreening was, for related reasons, biased: the manual rescreening results were based on Pap smears that had been rescreened only once, leaving a greater number of false negatives to be detected.