The aim of this study was to evaluate in small cervical biopsies (non-cone, non-large loop excision of the transformation zone, LLETZ) the prognostic value of both routinely assessed and reviewed cervical intraepithelial neoplasia (CIN) grades 1 and 2, oncogenic human papillomavirus (onco-HPV) DNA (HPV status) and Ki-67 immuno-quantitative features for the prediction of progression. In biopsies from 44 CIN patients (the learning set), subjective CIN grade, onco-HPV by PCR, and Ki-67 immuno-quantitative features were assessed. We followed development of the lesions by colposcopy and cytology, but the final endpoint was the histological grade (again in small biopsies). The outcome was defined as progression (histological (CIN 1 to (CIN 2 or 3)) or CIN 2 to CIN 3) or not (all other cases). Single and multivariate (Cox regression) and survival analyses were applied. The resulting predictive combination of quantitative features was then applied to a new test set of 35 consecutive CIN 2 (small) biopsies followed by large (cone or LLETZ) biopsies. In the learning set, mean follow-up of non-progression cases was 18.8 months (range 4.7–35.9), and of progression cases 13.1 months (range 6.4–32.9) (p = 0.18). Five cases progressed (11%). Of the 16 CIN 1 and 28 CIN 2 lesions, 31 cases (70%) were onco-HPV positive (5 of the CIN 1 and 26 of the CIN 2). The age of women with progression or not did not differ (p = 0.68). All 5 progression cases were CIN 2 (on review, one of these was reclassified as CIN 1), and positive for onco-HPV. Cox regression analysis showed that the percentage of Ki-67-positive cells located in the middle third layer of the epithelium (MIDTHIRD) and the 90th percentile of the stratification index (SI90) was the best combination to predict progression (log rank = 5.1, p = 0.02). Furthermore, sensitivity (100%), specificity (56%), positive predictive value (23%), negative predictive value (100%), and overall percentage correctly classified cases (61%) of this Ki-67 combination were higher than that of subjective CIN grade or HPV status, either single or combined (both for routine and review CIN grades). Adding CIN grade or HPV status did not improve the Ki-67 prognostic results. Application of the prognostic Ki-67 combination to the test set of 35 small biopsies followed by large (cone or LLETZ biopsies) gave comparable results. Analyses on homogeneous subgroups (CIN 2 only, onco-HPV+ only, or CIN2/onco-HPV+ only) gave similar results. In conclusion, Ki-67 immuno-quantitation of small biopsies showing CIN 1 or CIN 2 has strong independent prognostic value for progression. Copyright © 2003 John Wiley & Sons, Ltd.