FC 11-08GENOTYPING AND CYTOLOGIC TRIAGE OF HPV POSITIVE WOMEN FOR THE DETECTION OF CERVICAL HIGH-GRADE LESIONS

09. HPV screening
M. El-Zein 1, S. Bouten 1, L. Sobhi Abdrabo 1, A. Siblini 1, K. Louvanto 2, E. Franco 1, A. Ferenczy 3.
1Division of Cancer Epidemiology, McGill University (Canada), 2Department of Obstetrics and Gynecology, University Hospital of Helsinki (Finland), 3Division of Gynecologic Oncology and Colposcopy, McGill University - Jewish General Hospital (Canada)

Background / Objectives

The VASCAR (viral testing alone with Pap triage for screening cervical cancer in routine practice) study was a single-center demonstration project initiated in Montreal, Canada in 2011 among more than 23,000 women attending routine cervical cancer screening. In a secondary phase of VASCAR, we determined genotype-specific risks of disease progression to biopsy-confirmed cervical intraepithelial neoplasia of grade 2 or worse (CIN2+) associated with HPV types 16, 18, and/or the other 12 (pooled) high-risk types, compared to Pap cytology. We also assessed the diagnostic performance of HPV genotyping compared to cytological triage.


Methods

Women aged 30-65 were originally screened for HPV using the Hybrid Capture® 2 (HC2) Test. Women with positive results were triaged using conventional cytology, and those with positive Pap cytology results (≥ASC-US; atypical squamous cells of undetermined significance) were referred to colposcopy. We retrospectively genotyped 1396 cervical specimens that were HPV+ with HC2 using the Roche’s cobas® 4800 HPV system, and extracted the women’s medical history. We evaluated diagnostic performance of triage tests in the first year of follow-up among women positive for: (1) HPV16; (2) HPV18; (3) HPV16 and/or HPV18 and; (4) one or more of the other 12 HPV types. Using hierarchical and exclusive categories of HPV positivity (any HPV16; else HPV18; else 12 other HPVs), we correlated HPV status at enrollment with detection of histologically confirmed CIN2+ by estimating hazards ratios (HR) with 95% confidence intervals (CI) using Cox proportional hazards regression.


Results

Of the 1396 women, 1092 (78%) were classified as normal, 136 (10%) had CIN1, 80 (6%) had CIN2, 81 (6%) had CIN3 and 7 women had cancer, throughout the entire follow-up period. Sensitivity of HPVs 16, 18, 16 and/or 18, and any high-risk HPV for prevalent CIN2+ (n=76) were 35.5% (CI:24.9-47.3), 9.2% (CI:3.8-18.1), 43.4% (CI:32.1-55.3), and 64.5% (CI:52.7-75.1), respectively. Conversely, cytology triage (ASC-US+) had a sensitivity of 92.0% (CI:83.4-97.0). Corresponding specificity values were 84.0% (CI:81.9-86.0), 95.0% (CI:93.7-96.1), 79.7% (CI:77.4-81.8), 33.4% (CI:30.9-36.0), and 73.6 (CI:71.1-76.0). Compared to cobas HPV- and HC2 HPV+ women, the HRs were 7.3 (CI:3.8-14.3), 3.9 (CI:1.5-10.2), and 2.7 (CI:1.4-5.2) for women with any HPV16, HPV18, and 12 other types, respectively. Compared to women with normal cytology, the HRs for AS-CUS, LSIL, and HSIL (SIL: squamous intraepithelial lesion) were 3.7 (CI:2.5-5.7), 5.0 (CI:3.1-8.0) and 16.5 (CI:11.0-24.7), respectively.


Conclusion

Cytology and genotyping seem to be comparable in triaging women with positive HPV results on screening.


References