Reproducibility of Linear Array for Human Papillomavirus Genotyping

Division of Cancer Epidemiology and Genetics, NCI/NIH/DHHS, Bethesda, MD, United States.
Journal of clinical microbiology (Impact Factor: 4.23). 11/2012; 51(2). DOI: 10.1128/JCM.03036-12
Source: PubMed

ABSTRACT We conducted a Linear Array® test/re-test analysis using cytologic specimens from 198 women. 67.2% of samples had the same HPV types detected in both tests [type-specific positive agreement 83.3% overall (Kappa=0.9), 86.8% for carcinogenic types (Kappa=0.92)]. Discordance was highest with low hybridization signal strength. Overall, Linear Array was highly reproducible.


Available from: Sophia Wang, Jun 27, 2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: Studies of the natural history of human papillomavirus (HPV) infection require reproducible, type-specific testing of the viral types that infect cervical tissue; Linear Array (LA) is one method that has been widely used. We sought to develop a cost-effective, high-throughput alternative using the same PGMY09/11 primer/probe system and offering comparable sensitivity and specificity to LA, to ensure continuity in longitudinal studies. We report here on a Luminex-based approach, PGMY-LX that offers type-specific detection of 33 oncogenic and non-oncogenic types. Detection of HPV type-specific plasmid DNA was highly specific, with high signal-to-noise ratios for all types except non-oncogenic type 40 and no cross-reactivity between types. Cohen's unweighted κ values for 378 clinical samples tested by both LA and PGMY-LX were greater than or equal to 0.8 (range 0.80 to 1.0) for 25 types, including the following oncogenic HPV (16, 31, 33, 39, 45, 58, 59), and possibly oncogenic types (53,66,73, and 82), and greater than 0.7 (range 0.74 to 0.79) for oncogenic (18, 35, 51, 56) and probable oncogenic type 68b, indicating substantial or better type-specific agreement between the two methods. Reproducibility by PGMY-LX of the types detected by LA varied from 94% when a single HPV type was present to 66% when multiple types were present. Inter-run reproducibility for PGMY-LX varied from 98% for single-type infections to 85% for multiple-type infections. The high reproducibility of PGMY-LX and type-specific agreement with LA allows PGMY-LX to be incorporated into longitudinal, cohort studies that have historically relied on LA. Copyright © 2015, American Society for Microbiology. All Rights Reserved.
    Journal of Clinical Microbiology 02/2015; 53(4). DOI:10.1128/JCM.03100-14 · 4.23 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The prevalence and genotype distribution of human papillomavirus (HPV) provides the basis for designing HPV prevention programs. The prevalence rates of type-specific HPV and coinfections in samples of Mexican women were investigated in 822 women aged 18-87 years. HPV detection was performed using a Linear Array(TM) genotyping test. HPV infection was found in 12.4% of controls, 46.3% of those with cervical intraepithelial neoplasia 1, and 100% of those with cervical intraepithelial neoplasia 3 or cervical cancer. HPV 16 was the most prevalent type in all diagnosis groups. The HPV types most frequently found in cervical cancers were 16, 18, 45, 52, 58, and 39; HPV types 16, 62, 51, 84, 18, 53, and CP6108 were the most prevalent in control women. Considering HPV-positive samples only, coinfections occurred most often in controls (63%) and were less frequent in those with cervical cancer (26%). The most frequent viral types in coinfections with HPV 16 in control women were HPV 62, 51, and 84; in women with cervical cancers, HPV 18, 39, and 70 were most common. In conclusion, in addition to HPV types 16 and 18, types 45, 39, 58, 52, and 71 were found in cervical cancers in Mexican women (78%); among them, only 65% were attributable to HPV types 16 and 18. Therefore, it is necessary to consider these viral types in the design of new vaccines, and to determine whether certain HPV types coinfecting with HPV 16 in precursor lesions determine tumor progression or regression. J. Med. Virol. © 2015 Wiley Periodicals, Inc. © 2015 Wiley Periodicals, Inc.
    Journal of Medical Virology 02/2015; 87(5). DOI:10.1002/jmv.24099 · 2.22 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Human papillomavirus (HPV) is a well-known pathogen for lower genital tract neoplasias, yet little is known regarding HPV prevalence in Bulgaria. The aim of this study was to investigate the prevalence of HPV DNA and to determine HPV types distribution among women with normal and abnormal cytology. Cervical smears with different cytological diagnoses were collected from 355 Bulgarian patients. The cohort of patients selected is the biggest ever studied in this country. Using the Roche Linear Array HPV Genotyping Test, papillomavirus DNA was found in 217 out of the 355 samples, 164 of which had only one and 53 had more than one HPV type. The distribution of the viruses tested in 355 samples was as follows: (i) the most common type was HPV 16, which was found in 61 samples; (ii) the next most frequent HPV type was HPV 33, found in 14 of the samples. A high prevalence of HPV infection was observed in this study. As HPV infection has a high correlation with cervical cancer, this study emphasizes the need for both primary prevention of cervical cancer with HPV vaccines as well as secondary prevention with screening. Currently, two HPV vaccines are included in the National immunization schedule in Bulgaria. Thus, new clinical studies will benefit from patient stratification by the presence or absence of HPV, and by designing separate clinical trials specifically for HPV associated cancers. J. Med. Virol. © 2014 Wiley Periodicals, Inc.
    Journal of Medical Virology 07/2014; 86(11). DOI:10.1002/jmv.24020 · 2.22 Impact Factor