Identification of progressive cervical epithelial cell abnormalities using DNA image cytometry

BACKGROUND: The objectives of the current study were to compare the capabilities of conventional cervical cytology and of DNA image cytometry (DNA-ICM) in the prediction of progressive or regressive behavior in atypical squamous cells (ASC), low-grade squamous intraepithelial lesions (LSIL), and atypical glandular cells (AGC). METHODS: One hundred ninety-six women with Papanicolaou (Pap) smears that yielded diagnoses of ASC, LSIL, or AGC were included in a prospective cohort study. Slides were classified according to the Bethesda system. DNA-ICM was performed according to the consensus reports of the European Society of Analytical Cellular Pathology. RESULTS: Reference standard verification was available in 108 patients. The rate of DNA aneuploidy in Pap smears increased significantly from cervical intraepithelial neoplasia 1 (CIN1) (54%) and CIN2 (64.3%) to CIN3 or greater (CIN3+) (83.3%) in subsequent biopsies (P < 0.05). Using ASC, LSIL, and AGC as input cytologic diagnoses and >/= CIN2 as the output histologic diagnosis, the positive predictive values (PPVs) for conventional cytology and DNA-ICM were 35.2% and 65.9%, respectively (P < 0.001). The negative predictive value (NPV) of DNA-ICM was 85.0%. When >/= CIN3 was used as the output histologic diagnosis, conventional cytology had a PPV of 22.2%. The PPV and NPV of DNA-ICM were 43.9% and 93.3%, respectively. CONCLUSIONS: The results of the current study confirmed the prognostic validity of DNA image cytometry for differentiation between progressive and regressive lesions in patients with ASC, LSIL, and AGC diagnoses.