Don’t Forget HPV-45 in Cervical Cancer ScreeningThe Authors’ Reply

To the Editor The ATHENA trial is focusing only on human papillomavirus (HPV)-16 and HPV-18 to identify women at high risk for high-grade cervical disease or worse.1 Type-specific screening tests should focus not only on HPV-16 and HPV-18 but also on HPV-45. HPV types 16, 18, and 45 are the 3 most common types in cervical cancer. Together, they account for 75% of squamous cell carcinoma and 94% of adenocarcinoma.2,3 HPV-18 and HPV-45 together are significantly more common in cases of adenocarcinoma than in cases of squamous cell carcinoma (44% vs 14%; P < .0001).3 Women younger than 30 years have twice (4%) as many invasive cervical cancers related to HPV-45 than to HPV types other than 16, 18, or 45 (2%).3 The mean age of patients with cervical cancer is the lowest for HPV-45 (46.8 years), followed by HPV-18 (48.2 years) and HPV-16 (50.0 years). The average age of patients with invasive cancer caused by HPV-16, HPV-18, and HPV-45 is 47 years, and the average age of women with this cancer related to any other HPV type is 56 years.2,3 A possible explanation for this younger age is the integration of the viral DNA in the human genome. The percentage of integration increases from progression of a low-grade lesion to an invasive carcinoma. Patients with invasive cancer with integrated HPV DNA have a statistically significant younger median age at diagnosis (43 years) compared with patients without integration (52.5 years; P = .017).4 Integration frequency in invasive cancer is the highest for HPV-18, which varies between 100% and 92%, followed by HPV-45 (83%) and HPV-16 (which varies between 55% and 80%).4–6 Other types, such as HPV-33 (37% integration) and HPV-31 (14% integration), are less likely to integrate. HPV-45 and HPV-18 are rare in women with normal cytology and high-grade lesions (0.4% and 2.3% and 0.9 % and 6.9%, respectively) compared with HPV-16 (2.6% and 45.4%, respectively).7 The earlier occurrence of invasive cervical cancers that are positive for HPV-45 and HPV-18 is indicative of a short time of progression to invasive cancer, related to the higher early integration rate.3 The younger age of women with invasive cervical cancers that are positive for HPV types 16, 18, and 45 is relevant for primary and secondary cervical cancer prevention programs. The type-specific genotype screening protocols should therefore at least include HPV-16, HPV-18, and HPV-45. Vaccination with human papillomavirus bivalent (types 16 and 18) vaccine, recombinant (Cervarix) is highly effective (>98%) against high-risk (HR)-HPV types 16, 18, and 45 and in part also against types 31 and 35, whereas vaccination with human papillomavirus quadrivalent (types 6, 11, 16, and 18) vaccine, recombinant (Gardasil) is only highly effective (>98%) against HR-HPV types 16 and 18 and in part against type 31. One could consider delaying cervical screening in HPV-vaccinated cohorts owing to the late occurrence of cancer with HPV types that are not related to HPV types 16, 18, or 45. The ATHENA trial uses only genotyping for HPV-16 and HPV-18.1 Based on the published results, it is only worthwhile to genotype for HPV-16.1 The absolute risks of cervical intraepithelial neoplasia (CIN) 2 or worse and of CIN 3 or worse among women who were HPV-16+ are 31.5% and 20.0%, respectively. For HPV-18+ women, these rates are 4.3% and 4.3%, respectively, and for women positive for a non–HPV-16/HPV-18 HR-HPV (12 other HR-HPV genotypes), the rates are 8.6% and 4.4%, respectively. The relative risks of CIN 2 or worse and CIN 3 or worse for women who were HPV-16+ vs women positive for a non–HPV-16/HPV-18 HR-HPV are 3.7 and 4.5, respectively. HPV-18 positivity at baseline alone did not confer a significantly increased relative risk for CIN 2 or worse or CIN 3 or worse compared with the combination of all 12 other HR-HPV genotypes. Prospective studies like the ATHENA trial or the ALTS study have an underrepresentation of cervical precancerous lesions caused by HPV types 18 and/or 45 compared with the proportion of cervical cancers attributed to these HPV types. This is at least strange for prospective trials, given the fact that these cancers occur earlier than the ones related to HPV-16. Possible explanations could be the fast progression or that these lesions are missed. The latter could be because they are often higher up in the endocervical canal, especially the glandular lesions. This is supported by the fact that screening has a major impact on squamous lesions but not on adenocarcinomas. Another explanation could be that the HPV test or HPV genotyping is not adequate to detect integrated HPV-18 or HPV-45. HPV testing can be done with primers in the L1 or E6/E7 regions. During integration, there is never a loss of E6/E7, but L1 expression can be lost. The latter could therefore lead to a false-negative result if only an L1 test is used. Comparison of primers for the different regions revealed a difference of 10% in HR-HPV detection.8 In the ATHENA trial, the cobas 4800 HPV DNA Test (Roche Molecular Systems, Pleasanton, CA) was used for HPV detection, and this is a solely L1 test. Cervical cancer screening should be based on HPV testing owing to the superior sensitivity over cytology. Current HPV tests should focus on the top 3 HPV types: 16, 18, and 45. For the future, type-specific genotyping for all HR-HPV types is essential to identify whether type replacement is occurring after vaccination. The use of primers located in regions that can be absent in HPV-integrated tumors (eg, L1 and E2) should be avoided. Primers detecting the oncogenic E6/E7 regions, which are always present, even in integrated cervical cancers (15%), should be considered in order to not miss integrated HPV infections. 2012 We thank Drs Tjalma and Depuydt for their interest in our study and the cobas HPV Test and would like to assure them that we did not forget HPV-45 in cervical cancer screening. Our assay was designed to detect the presence of the HR oncogenic HPV genotypes, including HPV-45, in a pooled manner, while reporting HPV-16 and HPV-18 separately. Although many studies have demonstrated the dominance of HPV-16- and HPV-18-associated cervical cancer,1,2 there seems to be no clear consensus on the ranking of the next most oncogenic genotypes, as evidenced by data from several investigators. For example, HPV-45 was ranked as the third most oncogenic type globally by de Sanjose et al1 but ranked 11th for CIN 3 or worse after long-term follow-up.3 Moreover, in the ARTISTIC trial, infection with HPV-16/HPV-18 conferred a 3-fold greater rate of CIN 3 or worse at baseline compared with HPV types 31, 33, 45, 52, and 58 combined.4 In the baseline phase of our ATHENA trial, HPV-45 (detected by the Roche LINEAR ARRAY High Risk HPV Genotyping Test) ranked ninth for risk of CIN 3 or worse and 10th for prevalence, whereas HPV-16 is ranked first for risk of CIN 3 or worse and prevalence, and HPV-18 is ranked fifth for risk and second for prevalence. A more detailed analysis of the prevalence and the risks attributable to the various HR-HPV genotypes in the ATHENA trial will be the subject of a future manuscript. Tjalma and Depuydt also rightly point out that HPV-18 did not confer an absolute risk for CIN 2/CIN 3 or worse that was greater than that for the 12 other HR-HPV genotypes in the ATHENA trial. To this point, it should be noted that in contrast with the absolute risks of CIN 2 or worse and CIN 3 or worse observed in women positive for 12 other HR-HPVs (8.6% and 4.4%, respectively), the absolute risk in women positive for HPV-18 is identical for CIN 2 or worse and CIN 3 or worse (4.3% and 4.3%, respectively), indicating that HPV-18 disease is associated with higher-grade histology and is often glandular in its presentation. In addition, it should be noted that the risks reported in this analysis are only for the baseline phase of the ATHENA study. This is relevant because it is widely accepted that HPV-18-associated disease is identified during longitudinal follow-up, whereas HPV-16 is more evident at initial evaluation.5,6 Explanations for this difference include intrinsic difficulties in detection and/or identification of HPV-18-associated disease (because glandular lesions occur higher up in the cervical canal than squamous cell lesions) or a requirement for longer periods of persistence before HPV-18-associated disease develops. For these reasons, the follow-up phase of the ATHENA trial will be critical to fully understand the risk for high-grade cervical disease that infection with HPV genotype 18 poses. The follow-up phase will also allow evaluation of the risk over time of HPV-45 at baseline. We would also like to address the other concern of Tjalma and Depuydt about the possibility of missing disease by using an HPV test with primers directed against the L1 region. It is widely accepted that sufficient full-length episomal copies of HR-HPV are present in the disease state to allow its detection with L1-directed primers.7,8 In fact, the study by de Sanjose et al,1 which Tjalma and Depuydt quoted extensively, used the SPF-10 assay, a polymerase chain reaction–based L1 test. In addition, the ATHENA data presented in our article include a comparison of performance of the cobas HPV Test with the Hybrid Capture 2 test (QIAGEN, Gaithersburg, MD) (which uses HPV RNA probes not directed against the L1 region of HPV DNA to create an RNA-DNA hybrid), and we demonstrate that there are no significant differences in sensitivity or specificity for the detection of high-grade cervical disease between the 2 tests.9 Again, we thank Tjalma and Depuydt for their interest and insights. We presented the first data set for the cobas HPV Test that was designed to detect all oncogenic HPV genotypes (including HPV-45) as a pooled result and additionally to identify separately HPV-16 and HPV-18, which are responsible for 70% of cervical cancers worldwide. We demonstrated comparable sensitivity and specificity for detecting high-grade disease (CIN 2 or worse and CIN 3 or worse) to Hybrid Capture 2, which has been the standard of HPV testing for the past decade. We hope we have provided additional information to reassure Tjalma and Depuydt that we have conducted a well-designed study to support the clinical validation of a new pooled HR-HPV test that includes integrated HPV-16 and HPV-18 identification.