Page 190 - Strategies for non-invasive managementof high-grade cervical intraepithelial neoplasia - prognostic biomarkers and immunotherapy Margot Maria Koeneman

Chapter 10
[11-21] Eleven such studies were identified and included a total of 678 women with CIN2/3 lesions. Follow-up terms ranged from six weeks to two years, during which colposcopies were performed every two to six months. Three cases of progression to micro-invasive disease were described and one case of invasive disease.[13, 15] All three cases of micro-invasive disease occurred in a study with a follow-up term of 20 weeks, in which a control colposcopy with biopsies was performed at 8 weeks. The author states that presence of micro-invasive disease at baseline cannot be excluded in these cases.[15] The only case of invasive disease occurred in an RCT with a two year follow-up term, in which control colposcopies were performed every 3 months and biopsies every six months. Overall, four cases of (micro-)invasive disease were diagnosed in 678 women (0.6%), of which one occurred in a study with a relatively long follow-up term. Based on these study results, short-term prospective observational studies in high-grade CIN do seem feasible, whilst histological follow- up is provided at limited intervals. This strategy provides an opportunity for prognostic biomarker research. Short term follow-up, however, does imply that those lesions that would regress after the follow-up limit, will be classified as non-regressing lesions. Indeed, increasing regression rates with longer follow-up have been reported. Munk et al. reported that regression of CIN2/3 was 5% in women who were treated within 8 weeks, compared to 38% in women who were treated after 8 weeks follow-up.[22] Godfrey et al. followed 100 women with a CIN2 lesion, of whom 62 women showed spontaneous regression. The time to cytological normalization in these women varied from 3 to 43 months, indicating that spontaneous regression can occur even after longer periods. [23] The inability to conduct long-term studies with associated underestimation of spontaneous regression seems to be an unavoidable limitation to prognostic biomarker research in high-grade CIN. Despite the feasibility of short-term biomarker research, motivating women with a CIN3 lesion for participation in a natural history study proved to be difficult, as we experienced in the first version of the TOPIC trial. This version included an observational arm, with the aim to develop a prognostic biomarker panel to predict spontaneous regression of high-grade CIN. Patients in the observational arm underwent no treatment for a period of maximum 20 weeks. Histological assessment of disease development was performed after 10 and 20 weeks by colposcopy with diagnostic biopsies. Inclusion of patients into the study was hampered by this observational arm: patients declined the study because they wished to be treated, rather than undergo observational management. The observational arm was removed from the study.
We studied Human Leucocyte Antigen (HLA) types and gain of the human telomerase RNA gene (hTERC) as potential prognostic biomarkers in high-grade CIN. Both were identified in our review as promising candidate markers. We found no effect of the HLA repertoire on spontaneous regression of high-grade CIN and could therefore not identify HLA alleles as a prognostic biomarker in high-grade CIN. The prognostic properties of hTERC gain were studied by means of 3q26 locus, which contains the hTERC gene. A pilot study in 19 women showed that 3q26 gain had a high negative predictive value: all lesions without 3q26 gain showed disease regression. As such, the absence of 3q26 gain could potentially serve as a prognostic biomarker for the identification of CIN lesions with a high probability of disease regression. Nevertheless, the positive predictive value of 3q26 gain was limited, leading to incorrect classification as non-regressing lesions in women whose lesion would actually regress. This would mean that a significant subset of women would still be overtreated.
188