Performance of the AutoPap primary screening system at Jefferson University Hospital

To evaluate the performance of the AutoPap System for primary screening of cervical/vaginal cytologic smears at the 25% no review rate at Jefferson University Hospital.

A total of 5,865 consecutive conventional and suitable cervical/vaginal smears, including high-risk cases, were processed by the AutoPap System. Slides designated "no further review" (NFR) were manually screened by a cytopathologist (rapid screening) and a cytotechnologist (detailed screening) to detect epithelial abnormalities and assess the slide adequacy. The presence of Candida or Trichomonas was noted. Three cytopathologists determined the final diagnosis of all epithelial abnormalities by majority agreement.

Of 5,865 slides, 5,120 (87%) qualified for scanning. The system classified 1,280 slides as NFR (25%) and 3,840 (75%) as review. Manual screening of 1,280 slides classified as NFR revealed 1,252 cases within normal limits, 10 atypical squamous cells of undetermined significance (4, favor low grade squamous intraepithelial lesion, 6 favor reactive) and 18 benign cellular changes. Adequacy assessment discrepancies were detected on 98 (19%) of 515 cases of satisfactory but limited by lack of endocervical component (SBLBLEC) and 15 (14%) of 105 cases on satisfactory but limited by inflammation/obscuration. Twenty-five percent of cases classified as SBLBLEC were vaginal smears. Trichomonas vaginalis was noted in 12 slides and Candida in 29.

These preliminary results show that in our patient population, the AutoPap primary screener selected as NFR mostly slides within normal limits. The adequacy assessment discrepancies were comparable to those found in routine laboratory practice.