Computer grading of cervical intraepithelial neoplastic lesions. I. Cytologic indices.

In an attempt to derive a computer-assisted grading for cervical intraepithelial neoplastic (CIN) lesions, apparent discrepancies with the visual diagnostic evaluation could be explained from the data structure of the objective assessment. The 40 cases of dysplasia in this study appeared to fall into two subgroups having different curves of progression. A smaller subgroup, visually graded as milder, less-severe CIN, exhibited dysplastic cells with unexpectedly high abnormality indices. A larger subgroup showed dysplastic cells with abnormality indices in agreement with the tissue assessment.