Molecular biomarker-based screening for early detection of cervical cancer

OBJECTIVE: To evaluate the effectiveness of a molecular biomarker-based screening method for early detection of cervical cancer. STUDY DESIGN: Fluorescent immunochemical labeling was used to classify cervical cytology specimens as probably normal or probably abnormal. These specimens were then Papanicolaou stained and evaluated twice by a cytotechnologist and by a pathologist when appropriate. The first evaluation was performed as conventional Pap screening to assign a reference per-specimen diagnosis. The second evaluation assigned a cytologic diagnosis to each cell in the specimen. The fluorescence results were correlated with those from each of the two morphologic evaluations to determine the sensitivity and specificity of the method on a per-specimen and a per-cell basis. Where available, in cases of positive morphologic analysis, biopsy was compared to fluorescence results. RESULTS: The per-specimen sensitivity and specificity were 87.5% and 81.8%, respectively, when using atypical squamous cells of undetermined significance and "above" as the decision threshold. For the same specimens evaluated on a cell-by-cell basis, the corresponding sensitivity and specificity were 71.4% and 66.3% when the same threshold was used. The per-specimen sensitivity for high grade squamous intraepithelial lesions was 100% under these conditions. CONCLUSION: This method appears to be a robust and reliable means of detecting cervical dysplasia and is now being evaluated in additional clinical studies.