Related Experiment Videos
Rescreening of atypical cervicovaginal smears using PAPNET
E Lerma1, L Colomo, A Carreras
1Department of Pathology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain.
Background:
Atypical squamous cells of undetermined significance (ASCUS) is a cytopathologic term used to describe cases without specific pathologic substratum. Between 10-60% of ASCUS cases correspond to squamous intraepithelial lesions (SIL).
Methods:
The objectives of this study were: 1) to detect the pathologic significance of ASCUS in study patients, 2) to determine whether PAPNET identifies these cases, and 3) to compare the results of PAPNET with those of a second conventional screening. One hundred and sixty-three consecutive patients with the cytologic diagnosis of ASCUS and adequate follow-up were selected. Of these, 111 patients had colposcopic lesions and biopsies were performed; in the remaining 52 cases colposcopy was negative, as were 3 consecutive annual Papanicolaou smears. In a blind review, all 163 cases were rescreened using PAPNET. A second manual screening was performed for comparison.
Results:
One hundred and twenty-six of the 163 cases (77.3%) showed no SIL on biopsy or follow-up. Of the 37 pathologic cases, the diagnosis was koilocytosis (flat condyloma) in 13 cases (8%), cervical intraepithelial neoplasia (CIN) type I in 11 cases (6.8%) low grade SIL [LSIL] in a total of 24 cases [14.8%]), and CIN II-III or high grade SIL (HSIL) in 11 cases (6.8%). In the review using PAPNET, 57 previous ASCUS cases were classified correctly as negative, and 7 of 13 koilocytosis cases (54%), 9 of 11 CIN I cases (82%), and 7 of 11 CIN II-III cases (64%) were diagnosed correctly. In the second conventional screening, 74 cases were negative and 77 cases were ASCUS; only 3 of 13 koilocytosis cases (23%), 4 of 11 CIN I cases (36.4%) and 5 of 11 CIN II-III cases (45.5%) were reclassified correctly.
Conclusions:
Among 163 patients with ASCUS, 77.3% had no precancerous squamous lesions. Concordance with definitive diagnosis was more accurate in our study using PAPNET analysis (Kappa index [K] = 0.7158) than by second conventional screening (K = 0.4537). Furthermore, we reclassified 35% of smears as negative and 15% as SIL.