Screening for, and treatment of, pre-cancerous cervical lesions has lead to dramatic reductions in cervical cancer in many countries. In all cases, cervical screening has been based on cytology, but that is beginning to change. Research studies, including randomised trials, clearly show that human papillomavirus (HPV) testing could be used to prevent a greater proportion of cervical cancer within a practical screening programme. Meanwhile, young adolescents are being vaccinated against HPV in developed countries, but cervical screening should continue for many years because it will take decades before most of those targeted by screening have been vaccinated. In the HPV vaccination era, the rate of cervical disease will decrease, and so will the positive predictive value of cytology. The screening characteristics of HPV testing make it the preferred choice for primary screening. However, questions regarding how to use HPV testing to screen vaccinated and unvaccinated women in the future remain unanswered.
Introduction
Cervical cancer is the third most common cancer in women worldwide and the most common in several developing countries. Cervical cytology has been the traditional screening modality for pre-cancerous cervical lesions. Cytology-based screening has been successful in countries in which adequate resources exist to ensure high quality and good coverage of the population at risk. It has had no major effect, however, on reducing incidence and mortality rates in the developing world where appropriate resources and infrastructure are not available or difficult to maintain.
The identification of certain oncogenic types of human papillomavirus (HPV) as a necessary cause of cervical cancer has provided a great opportunity to prevent disease on two fronts: by immunisation with HPV vaccines and by screening using HPV DNA assays. The implementation of universal HPV vaccination of adolescent girls is the best prospect to control cervical cancer; however, a reduction of cervical cancer burden is unlikely to be observed, even in young women, for at least several decades given the latency between HPV infection and cervical cancer. For older women (certainly those born before 1990), vaccination is currently not an option, so cervical screening will have to continue for many years after vaccine introduction.
It is expected that, as vaccinated cohorts reach the age of screening, the rate of screen-detected cervical disease will decrease, leading to a reduction of 40–60% of the current colposcopy referral rates in counties with high vaccine coverage. These reductions are likely to translate into initial savings to the healthcare system (partially compensating the cost of vaccination); however, the vaccine-induced decrease in cervical lesions may lead to a degradation of performance characteristics of cytology. In particular, the positive predictive value is expected to decline as lesions become less common. The screening performance characteristics of HPV testing make it an attractive screening test for primary screening, given its immediate high positive predictive value and long-term negative predictive value.
In this chapter, we describe the current and potential use of HPV testing for vaccinated populations who become eligible for screening in countries with cytology-based screening programmes and in those with limited resources for cervical cancer prevention.
Basis and limitations of cervical cytology
Cytological testing involves collection of exfoliated cells from the cervix and microscopic examination of these cells after staining. This technique was first developed by Papanicolaou and Babes in the 1920s as a way of identifying women with invasive cervical cancer. Subsequently, Papanicolaou and Traut also showed that cytology could be used to identify pre-cancerous lesions of the cervix. In the 1960s, cervical cytology began to be widely used in many developed countries as a screening technique, and it is now the main approach to cervical cancer prevention.
Cytology screening has been successful in countries with organised screening programmes and in those with high-coverage opportunistic screening. The quality of cytology, however, has been questioned, and several studies have shown that the technique has low sensitivity; even high-quality sample-taking and cytology can miss cervical lesions. This partly explains why cytology screening has failed to reduce cervical cancer in developing countries with high coverage, where, in addition to the low-quality sample-taking and cytology, adequate follow up and access to treatment are often poor.
Despite the low cost of consumables, cytology is labour intensive and requires continuous quality control and assurance. Thus, good cytology is expensive and may not necessarily be the most cost-effective option for screening, especially in many low- or middle-income countries with high burden of disease and little infrastructure for quality assurance.
Liquid-based cytology (LBC) was introduced as a way of improving the performance of the test. Instead of smearing the exfoliated cells onto a glass slide, samples are transferred to a liquid preservative solution that is transported to the laboratory where a slide is prepared. Several studies have compared the performance of the two most commonly used LBC methods (ThinPrep ® and SurePath ® ) with conventional cytology. A recent meta-analysis established that LBC has logistical and operational advantages (rapid interpretation, lower rate of unsatisfactory smears, and the possibility of ancillary molecular testing using remnant fluid) but it is more expensive (per sample), and is neither more sensitive nor more specific than conventional cytology in the detection of histologically confirmed high-grade cervical intraepithelial neoplasia (CIN). Despite these disadvantages, LBC has replaced conventional cytology in several countries, and it has the unproven advantage that it is harder for a tired or distracted cyto-screener to miss an abnormality.
Rationale for human papillomavirus testing and reliability for detecting cancer precursors
Research on the use of HPV DNA assays began in the 1980s. Although much of that research began with a focus on viral detection as an endpoint, attention soon turned to the potential clinical utility of HPV testing for identifying cervical cancer precursors. Techniques to detect the presence of HPV in cervical cell specimens have evolved considerably in the past 25 years. To date, three assays (Hybrid Capture 2 [HC2], Cervista ® and Cobas ® 4800) have been approved by the Food and Drug Administration (FDA) and several more CE-mark (European Conformity) assays (PreTect HPV-Proofer, careHPV™, Linear Array HPV Genotyping Test, RealT ime High Risk HPV, APTIMA HPV ® HPV assay) have proved to be more sensitive than cytology at detecting biopsy-confirmed high-grade cervical lesions. HPV testing is objective, highly reproducible (i.e. results are the same when assayed in different laboratories) and has the potential to be automated.
Detection of high-risk HPV DNA is potentially useful in four clinical applications: (1) as a primary screening test (solely or in combination with cytology) to detect cervical cancer precursors; (2) as a triage test to select which women who have minor cytological abnormalities are in need of referral for colposcopy; (3) in the continuing management of women referred for colposcopy for whom no high-grade lesion was found; and (4) as a follow-up ‘test of cure’ for women treated for high-grade intraepithelial lesion to identify women who required continued surveillance.
The use of HPV DNA testing has clear benefits in the triage of equivocal smears, low-grade smears in older women and in the post-treatment surveillance of women after treatment for CIN. Hence, HPV testing is widely used to triage mildly abnormal cytology in the USA, and is about to be rolled out nationally in England for such purpose.
In contrast, despite overwhelming evidence towards using HPV testing in primary screening, issues still remain about how best to implement it. HPV testing, however, is being used for primary screening nationally in Mexico, and it is approved to be used in combination with cytology (co-testing) in primary screening in the USA. Here, we will concentrate on the evidence for using HPV testing in primary screening.
Rationale for human papillomavirus testing and reliability for detecting cancer precursors
Research on the use of HPV DNA assays began in the 1980s. Although much of that research began with a focus on viral detection as an endpoint, attention soon turned to the potential clinical utility of HPV testing for identifying cervical cancer precursors. Techniques to detect the presence of HPV in cervical cell specimens have evolved considerably in the past 25 years. To date, three assays (Hybrid Capture 2 [HC2], Cervista ® and Cobas ® 4800) have been approved by the Food and Drug Administration (FDA) and several more CE-mark (European Conformity) assays (PreTect HPV-Proofer, careHPV™, Linear Array HPV Genotyping Test, RealT ime High Risk HPV, APTIMA HPV ® HPV assay) have proved to be more sensitive than cytology at detecting biopsy-confirmed high-grade cervical lesions. HPV testing is objective, highly reproducible (i.e. results are the same when assayed in different laboratories) and has the potential to be automated.
Detection of high-risk HPV DNA is potentially useful in four clinical applications: (1) as a primary screening test (solely or in combination with cytology) to detect cervical cancer precursors; (2) as a triage test to select which women who have minor cytological abnormalities are in need of referral for colposcopy; (3) in the continuing management of women referred for colposcopy for whom no high-grade lesion was found; and (4) as a follow-up ‘test of cure’ for women treated for high-grade intraepithelial lesion to identify women who required continued surveillance.
The use of HPV DNA testing has clear benefits in the triage of equivocal smears, low-grade smears in older women and in the post-treatment surveillance of women after treatment for CIN. Hence, HPV testing is widely used to triage mildly abnormal cytology in the USA, and is about to be rolled out nationally in England for such purpose.
In contrast, despite overwhelming evidence towards using HPV testing in primary screening, issues still remain about how best to implement it. HPV testing, however, is being used for primary screening nationally in Mexico, and it is approved to be used in combination with cytology (co-testing) in primary screening in the USA. Here, we will concentrate on the evidence for using HPV testing in primary screening.
Evidence for the use of human papillomavirus testing in primary screening
Published meta-analyses
In 2006, a pooled analysis of European and US studies, in which both cytology and HPV testing were carried out in parallel, clearly showed that HPV testing could be used as a sole primary screening test with cytology reserved for women who test HPV positive. Individual data from more than 60,000 women screened with cytology and HPV testing in 10 studies were included. HPV testing was done by HC2 in eight studies and by consensus polymerase chain reaction (PCR) with GP5+ and 6+ primers in the other two; and all studies used conventional cytology (see Cuzick et al., 2006 for details of included studies). HPV testing was substantially more sensitive for detecting CIN2+ than cytology (96.1% v 53.0%) but less specific (90.7% v 96.3%). The sensitivity of HPV testing was similar in all studies (85–100%), whereas the sensitivity of cytology was highly variable (18.6–76.7%) ( Fig. 1 ). HPV sensitivity was uniformly high at all ages, whereas the sensitivity of cytology was substantially better in women over the age of 50 years than in younger women (79.3% v 59.6%). The specificity of both tests increased with age.
Similar results were obtained 2 years later by a meta-analysis of published data, adding several new parallel testing studies and randomised trials. Because nine of the included studies used PCR methods for HPV DNA detection, comparisons between the performance of HPV testing with HC2 and with PCR methods were possible. The sensitivity of HC2 for CIN2+ was consistently high in eight studies conducted in Europe and the USA: pooled estimate of 98.1% (95% CI 96.8 to 99.4%; P for inter-study heterogeneity: 0.4) and the pooled specificity was: 91.7% (95% CI 90.3 to 93.1%; range: 85 to 95%). The pooled sensitivity of PCR assays for CIN2+ (84.2%; 95% CI: 77.0 to 91.5%) was lower, but its pooled specificity (95.1%; 95% CI 93.4 to 96.8%) was higher than HC2. Different studies, however, used different primers and different methods of detection of amplified sequences, and significant heterogeneity was found.
In a pooled analysis of 17 population-based cervical screening studies in China where 30,371 women were screened; HPV testing had a higher sensitivity of 97.5% (95% CI 95.7 to 98.7) for detection of CIN3+, and a lower specificity of 85.1% (82.3 to 87.9) compared with LBC (sensitivity 87.9%; 95% CI 84.7–90.7 and specificity 94.7%; 95%CI 93.5 to 96.0) and VIA (sensitivity 54.6%, 95% CI 48.0 to 61.2 and specificity 89.9%, 95%CI 86.8 to 93.0) .
A list of studies included in the two first meta-analyses is presented in Table 1 , along with studies or new analysis of ongoing trials published more recently.
| Cross-sectional studies | |
| A1 | Cuzick J, et al., Br J Cancer 1999; 81 :554–558 |
| A2 | Schneider A, et al., Int J Cancer 2000; 89 :529–534. |
| A3 | Ratnam S, et al., Cancer Epidemiol Biomarkers Prev 2000; 9 :945–951. |
| A4 | Clavel C, et al., Br J Cancer 2001; 12 :75–83. |
| A5 | Kulasingam SL, et al., JAMA 2002; 288 :1749–1757. |
| A6 | Petry KU, et al., Br J Cancer 2003; 88 :1570–1577. |
| B1 | Cuzick J, et al., Lancet 2003; 362 :1871–1876. |
| B2 | Bulkmans NW, et al., Int J Cancer 2004; 110 :94–110. |
| C1 | Cuzick J, et al., Lancet 1995; 345 :1533–1536 |
| C2 | Kuhn L, et al., J Natl Cancer Inst 2000; 92 :818–825. |
| C3 | Schiffman, M, et al., JAMA 2000; 283 :87–93. |
| C4 | Belinson J, et al., Gynecol Oncol 2001; 83 :439–444. |
| C5 | Blumenthal PD, et al., Int J Gynaecol Obstet 2001; 72 :47–53. |
| C6 | Oh YL, et al., Cytopathology 2001; 12 :75–83. |
| C7 | Paraskevaidis E, et al., Gynecol Oncol 2001; 82 :355–359. |
| C8 | Syrjanen S, et al., J Low Tract Dis 2002; 6 :97–110. |
| C9 | Belinson J, et al., Int J Gynecol Cancer 2003; 13 :819–826. |
| C10 | Coste J, et al., BMJ 2003; 326 :733. |
| C11 | Salmeron J, et al., Cancer Causes Control 2003; 14 :505–512 |
| C12 | Sankaranarayanan R, et al., Int J Cancer 2004; 112 :341–347. |
| C13 | Sankaranarayanan R, et al., J Med Screen 2004; 11 :77–84. |
| C14 | Agorastos T, et al., Gynecol Oncol 2005; 96 :714–720. |
| C15 | Bigras G, et al., Br J Cancer 2005; 93 :575–881. |
| C16 | Sarian LO, et al., J Med Screen 2005; 12 :142–149. |
| C17 | Mayrand MH, et al., N Engl J Med 2007; 357 :1579–1588. |
| C18 | Almonte M, et al., Int J Cancer 2007; 121 :796–802. |
| D19 | Luyten A, et al., J Clin Virol 2009; 46 (Suppl. 3):S5–10. |
| D20 | Gravitt PE, et al., PLoS One 2010; 5 :e13711. |
| D21 | Lazcano-Ponce, E et al., Cancer Causes Control 2010; 21 :1693–700. |
| D22 | Wu R, et al., Int J Gynecol Cancer 2010; 20 :1411–1414. |
| D23 | McAdam M, et al., PLoS One 2010; 5 :e13266. |
| D24 | Belinson J, et al., Am J Clin Pathol 2011; 135 :790–5. |
| D25 | Monsonego J, et al., Int J Cancer 2011; 129: 691–701. |
| D26 | Wright TC, et al., “High risk for CIN2+ in Cobas ® 4800 HPV 16/18 positive women ≥30 years with NILM cytology: ATHENA results”, abstract presented at the 26 th International Papillomavirus Conference, Montreal, 2010. |
| Randomised Trials | |
| E1 | Kotaniemi-Talonen L, et al., Br J Cancer 2005; 93 :862–867. |
| E2 | Sankaranarayanan R, et al., Int J Cancer 2005; 116 :617–623. |
| E3 | Ronco G, et al., J Natl Cancer Inst 2006; 98 :765–774. |
| E4 | Ronco G, et al., Lancet Oncol 2006; 7 :547–555. |
| E5 | Bulkmans, NW et al., Lancet 2007; 370:1764–1772. |
| E6 | Naucler P, et al., N Engl J Med 2007; 357 :1589–1597. |
| E7 | Kitchener, HC, et al., Lancet Oncol 2009; 10 :672–682. |
| E8 | Ogilvie, GS, et al., BMC Cancer 2010; 10 :111. |
Four new studies have been conducted in Germany, Vanuatu, India and Mexico comparing the performance of HPV testing (by HC2) with that of conventional cytology (and/or VIA). The studies confirmed that HPV testing is more sensitive than cytology and VIA. In three of the studies, HPV testing was less specific than cytology. In the Indian study, however, the specificity of HPV testing was 91%, higher than that of cytology and VIA (86% and 88%, respectively).
Randomised-controlled trials
Eight randomised-controlled trials have been carried out (or are ongoing) to evaluate the use of HPV testing compared with cytology in primary screening. The main characteristics of these trials are presented in Table 2 .
| Randomisation a | |||||
|---|---|---|---|---|---|
| Study | Country | Number of women | Age (years) | Intervention b /control | Management in the intervention arm |
| SWEDESCREEN | Sweden | 12,527 | 32–38 | HPV + CC v CC |
|
| POBASCAM | Netherlands | 44,102 | 30–60 | HPV + CC v CC |
|
| ARTISTIC | UK | 24,510 | 20–64 | HPV + LBC v LBC |
|
| NTCC Phase 1 | Italy | 45,174 | 25–60 | HPV + LBC v CC |
|
| NTCC Phase 2 | Italy | 49,196 | 25–60 | HPV v CC | 1.Refer to colposcopy if HPV positive at 0 months. |
| Public Health Trial | Finland | 61,149 | 30–60 | HPV v CC |
|
| Osmanabad Tria c d | India | 131,746 | 30–59 | HPV v RC | 1.Refer to colposcopy if any screening test positive at 0 months. |
| HPV FOCAL e f | Canada | 33,000 | 25–65 | HPV v LBC |
|
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
