Search

To assess the concordance and performance of 2 different assays in detection of human papillomavirus (HPV) genotypes among women with abnormal Pap smear.A series of 575 women referred for colposcopy due to an abnormal Pap smear were analyzed with the Linear Array HPV Genotyping test detecting 37 HPV types and compared with Hybrid Capture II (HCII) assay for detection of carcinogenic HPV. Histologic outcomes of cervical intraepithelial neoplasia grade 2 (CIN2) or worse (CIN2+) and CIN3+ were the primary endpoints. Clinical performance, including receiver operating characteristics, was determined for both tests.HCII and linear array (LA) were concordant in 88.1% (433/491; 95% CI 85.3%-91.0%), having a substantial agreement with regular kappa (kappa = 0.70, 95% CI 0.62-0.77) and almost perfect agreement with weighted kappa (ICC = 0.82, 95% CI 0.7-0.85). In detecting CIN2+ and CIN3+, LA is 5% and 6% more sensitive but 9.5% and 8.7% less specific than HCII (area under ROC curve; P = 0.317 and P = 0.875, respectively).Performance of HCII and LA does not significantly differ in detecting CIN2+ or CIN3+.

Objectives. To assess the performance of a novel PCR-based assay (Roche AMPLICOR® HPV test) in detection of cervical pathology as a part of management for abnormal PAP smear (MAPS) and in women participating in cervical cancer screening. Study design. Altogether, 504 women comprising 270 patients referred for colposcopy due to an abnormal Pap smear and another 234 women participating in cervical cancer screening (tested for comparison) were analyzed for oncogenic (HR) Human papillomavirus (HPV) types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68 using the Roche AMPLICOR® HPV test in cervical samples collected in PreservCyt® liquid media. Colposcopic biopsy and/or LEEP cone biopsy was used as the gold standard in the triage group, while liquid-based cytology (LBC) was the reference test in the screening group. Results. The prevalence of HPV was significantly higher in the MAPS group (65.9%) than in the screening group (31.2%) ( P = 0.0001). There was a poor concordance between the referral PAP and the current LBC, being only moderate in the screening series, ICC (weighted kappa) = 0.291 (95%CI 0.070–0.459) ( P = 0.007), and almost poor in the MAPS Series, with ICC = 0.217 (95%CI 0.04–0.384) ( P = 0.023). AMPLICOR® HPV positivity increased linearly with the increasing grade of cervical lesions. In detecting high-grade (CIN2–3), colposcopy was the most sensitive test (96.5%), very similar to AMPLICOR (95.2%) ( P = 0.731), while LBC with HSIL cutoff was by far the most specific test (99.5%) and showed the highest PPV (96.1%). NPV of colposcopy (97.2%) and AMPLICOR (96.7%) were similar ( P = 0.839). Together with abnormal colposcopy and HSIL cytology, the AMPLICOR® HPV test is a powerful independent predictor of high-grade CIN2–3, and as such suitable to replace cervical cytology in management of women with abnormal PAP test (MAPS). Conclusions. The Roche AMPLICOR® HPV test is comparable to other HPV tests (HCII, PCR) in detecting CIN in MAPS. However, more data are clearly needed on the performance of AMPLICOR test in management of abnormal PAP and particularly as a screening tool.

Summary: The main objective of this study was to assess the feasibility of human papillomavirus (HPV) genotyping in women referred for colposcopy due to abnormal Papanicolaou (Pap) smear. A series of 248 women referred for colposcopy due to an abnormal Pap smear were analysed with the Roche Linear Array HPV genotyping test detecting 37 most frequent HPV types, and compared with hybrid capture II (HCII) assay for oncogenic (high-risk [HR] HPV) types as well as for p16INK4a expression using immunocytochemistry. All tests were performed in cervical samples collected in PreservCyt® liquid media for liquid-based cytology (ThinPrep), and colposcopic biopsy and/or loop electro excision procedure cone biopsy was used as the gold standard. HPV16 was the single most frequent genotype (29/258; 11.7%), followed by HPV51 (4.4%), HPV66 (3.6%), HPV42, 52 and 56 (3.2% for all). Linear array genotyping test significantly predicts both abnormal colposcopy (odds ratio [OR] = 9.0; 3.12-25.93), high-grade squamous intraepithelial lesions (OR = 9.6; 1.26-74.17) and cervical intraepithelial neoplasia (CIN) 3+ (OR = 29.3; 3.95-218.06). In detecting CIN3, linear array was equivalent (97.6%) to colposcopy in sensitivity (SE), both being superior to HCII (92.7%). Concordance between linear array and HCII was moderate (Cohen's kappa κ = 0.547; 95% confidence interval [CI]: 0.435-659). Specificity (SP) and positive predictive value (PPV) of linear array were significantly improved, if only HPV16 genotype was considered. Performance in the best balance is obtained, when linear array and colposcopy are combined, giving 82.9% SE, 93.9% SP, 73.9% PPV and 96.3% negative predictive value (NPV) as predictor of CIN3+ (OR 74.5; 95% CI: 27.36-202.72). In conclusion, linear array for HR-HPV is a highly sensitive test (97.6%) with high NPV (98.9%) in detecting CIN3+ lesions. HPV16 genotyping alone significantly improves SP and PPV of this test in management of women with abnormal cytology.