BACKGROUND Screening programs for the early detection of cervical carcinoma are criticized because of the problem of false‐negative diagnoses. A successful approach for solving this problem is applying neural network technology (PAPNET®) to assist the cytotechnologist (CT) in finding the (few) abnormal cells in the smear. METHODS In 3 consecutive years (1992, 1993, and 1994), 25,767 smears were screened conventionally and 65,527 with the aid of PAPNET by 7 CTs. For each CT, the scores for atypias of undetermined significance, squamous or glandular (ASCUS/AGUS according to the Bethesda classification system), indicated by Positive I, for low grade precursor lesions, by Positive II, for high grade lesions and invasive carcinoma, by Positive III, were calculated for both screening methods. The histologic scores were also calculated. RESULTS The mean positive scores of the seven CTs were higher for PAPNET than for conventional screening, and the coefficients of variability were lower. For Positive III smears, the consistency in screening was significantly higher for PAPNET than for conventional screening. The higher histologically positive scores for carcinoma in situ and invasive carcinoma indicated an increased screening sensitivity. CONCLUSIONS As demonstrated by the improvement in the performances of all CTs involved, screening efficacy was enhanced by the use of neural network technology. Cancer 1996;78:112‐7.

No abstract

Neural network-based screening (NNS) of cervical smears can be performed as a so-called "hybrid screening method," in which parts of the cases are additionally studied by light microscope, and it can also be used as "pure" NNS, in which the cytological diagnosis is based only on the digital images, generated by the NNS system. A random enriched sample of 985 cases, in a previous study diagnosed by hybrid NNS, was drawn to be screened by pure NNS. This study population comprised 192 women with (pre)neoplasia of the cervix, and 793 negative cases. With pure NNS, more cases were recognized as severely abnormal; with hybrid NNS, more cases were cytologically diagnosed as low-grade. For a threshold value > or = HSIL (high-grade squamous intraepithelial lesions), the areas under the receiver operating characteristic (ROC) curves (AUC) were 81% (95% CI, 75-88%) for pure NNS vs. 78% (95% CI, 75-81%) for hybrid NNS. For low-grade squamous intraepithelial lesions (LSIL), the AUC was significantly higher for hybrid NNS (81%; 95% CI, 77-85%) than for pure NNS (75%; 95% CI, 70-80%). Pure NNS provides optimized prediction of HSIL cases or negative outcome. For the detection of LSIL, light microscopy has additional value.

No abstract