Letter to the Editor and Response from the Evidence-Based Practice Center: Human Papillomavirus Testing to Screen for Cervical Cancer
Cervical Cancer: Screening
February 15, 2012
Recommendations made by the USPSTF are independent of the U.S. government. They should not be construed as an official position of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.
Letter to the Editor and Response
Human Papillomavirus Testing to Screen for Cervical Cancer
First published as Letters to the Editor in Annals of Internal Medicine on November 17, 2011, and February 3, 2012.
To the Editor: The USPSTF report1 has evaluated the use of HPV testing and cytology in cervical cancer screening and concluded that more evidence is needed to determine the optimal approach. The report has considered the different possible strategies separately (HPV alone, HPV alone with cytology triage, cotesting, cotesting with cytology triage) in comparison to cytology only and considered solely results on prevention of invasive cancer.
It has been conclusively shown that HPV testing is able to detect about 50% more high-grade CIN, an established cancer precursor, than cytology2 at an initial screening. Furthermore, all RCTs conducted in industrialized countries3-6 have shown a similar reduction in CIN3+ after HPV testing (Table 3 of the report) indicating that HPV testing is able to detect non-regressive lesions earlier than cytology, thereby reducing precancerous lesions in subsequent screening rounds. This lead-time gain was similar for all strategies studied. In addition, a pooled analysis of phase 1 (cotesting) and phase 2 (HPV alone) of the NTCC6 trial also showed a major, clinically important and statistically significant reduction in invasive cancers. As different strategies using HPV testing have shown similar results, pooling data is not only reasonable but very much needed. As the protective effect is similar with all strategies, the least costly (HPV testing with cytology triage) should be chosen. As this strategy has the same need for colposcopy as cytology only3, lack of data on colposcopies needed for other strategies is not relevant for costs (and obviously not for effectiveness).
The report also contains factual mistakes. For example, it is stated that “ARTISTIC had the opposite result: more cases of cancer were found after 2 screening rounds in the cotesting group (8 total) compared with cytology (4 total).” In this trial5, there were proportionally more cancers with cytology because the cotesting:cytology randomization ratio was 3:1.
Further, the report considers quality issues used in pharmacological trials that are clearly irrelevant for large preventive trials. For example, masking patients is plausibly irrelevant when evaluating preventive programs where patient information is a vital part of the intervention itself and when the disease is asymptomatic and placebo effects unlikely. Observer bias is more plausible, but all trials had a blind assessment of the clinical endpoint.
Already in 2008, the EU guidelines7 on cervical cancer screening concluded that existing evidence is sufficient at least for the launch of controlled pilot implementation projects using HPV screening. This could provide additional evidence on cancer-protective effect, but only after many years. In our view, requiring new RCTs with cancer incidence as the endpoint to compare with cytology every small variation (like triage methods) in HPV-based screening strategies is unreasonable, unsustainable and, given the time-frame of acquiring evidence, also unethical.
1. Whitlock EP, Vesco KK, Eder M, Lin JS, Senger CA, Burda BU. Liquid-based cytology and human papillomavirus testing to screen for cervical cancer: a systematic review for the U.S. Preventive Services Task Force. Ann Intern Med. 2011. [PM:22006930]
2. Cuzick J, Arbyn M, Sankaranarayanan R, Tsu V, Ronco G, Mayrand MH, Dillner J, Meijer CJ. Overview of human papillomavirus-based and other novel options for cervical cancer screening in developed and developing countries. Vaccine. 2008;26S:K29-41.
3. Naucler P, Ryd W, Tornberg S, et al. Human papillomavirus and Papanicolau tests to screen for cervical cancer. N Engl J Med. 2007;357:1589-97.
4. Bulkmans N, Berkhof J, Rozendaal L, et al. Human papilllomavirus DNA testing for the detection of cervical intraepithelial neoplasia grade 3 and cancer: 5-year follow-up of a randomised controlled implementation trial. Lancet. 2007;370:1764-72.
5. Kitchener HC, Almonte M, Thomson C, et al. HPV testing in combination with liquid-based cytology in primary cervical screening (ARTISTIC): a randomised controlled trial. Lancet Oncol. 2009;10:672-82.
6. Ronco G, Giorgi-Rossi P, Carozzi F, Confortini M, Dalla Palma P, Del Mistro A, et al. Efficacy of human papillomavirus testing for the detection of invasive cervical cancers and cervical intraepithelial neoplasia: a randomised controlled trial. Lancet Oncol. 2010;11:249-57.
7. Arbyn M, Anttila A, Jordan J, Ronco G, Schenck U, Segnan N, Wiener H (eds). European Guidelines for Quality Assurance on Cervical Cancer Screening. 2nd ed. Luxembourg: Office for Official Publications of the European Communities; 2008.
To the Editor: We thank Dr. Ronco and colleagues for pointing out our error in contrasting a higher cumulative cancer detection after cotesting in ARTISTIC with the opposite pattern in other cotesting trials. Fortunately, this error was only in the manuscript and not in our draft report the USPSTF used for its deliberations. The error was introduced when we omitted proportional cancer results from tables as we were editing to fit within space constraints for dissemination products, including the manuscript. This error has now been corrected to indicate that cotesting trials tended to find fewer cancers cumulatively than cytology and, where reported, after the second screening round. Thus, judgment is required as to how trial and modeling results for cotesting apply in the United States outside of contexts1.
In response to Dr. Ronco and colleagues' other points: we separated the different types of HPV-based primary screening due to their heterogeneity in approaches to screening and abnormal test management protocols[[2, 3]]. This is especially important since compliance with recommended followup procedures was incompletely reported, but available data indicate variability4. Particularly without longer term follow-up, different screening strategies that result in differential opportunities for confirmatory testing with colposcopy, or that differ in compliance with recommended followup, may distort comparisons of screening sensitivity through asymmetry bias[[5, 6]]. Considering HPV-based primary screening strategies separately also facilitated indirect comparisons between different programmatic strategies; cotesting and primary HPV screening strategies differed little in benefit, with more potential harms with cotesting, which may be mitigated by adding a triage step (such as cytology) to primary HPV. However, we disagree with the assertion by Dr. Ronco and colleagues that colposcopy requirements for HPV with cytology triage will necessarily be similar to cytology alone. Recent reports from the FOCAL trial suggest they are 1.7-fold higher when the entire first screening round is considered7. These increases may be offset by relative decreases in colposcopy requirements in the second round (not yet reported). Thus, we still believe that determining the net impact of HPV with cytology triage will not be possible until we have more complete results from important ongoing trials[[8, 9]].
While we considered a hierarchy of potential outcomes, not just invasive cancers, no consensus exists on which outcomes are minimally sufficient to demonstrate a benefit over cytology screening in the absence of a clear impact on cervical cancer mortality or invasive cancer incidence. Demonstrated impacts on surrogates alone (e.g., reduced CIN3+ incidence) may require modeling to inform health policy changes (6). As such, while most cotesting evaluations within national screening programs in Europe found reduced incidence of CIN3+ in the second screening round (compared with cytology alone), relative CIN2+ detection was generally increased during the first round, and in some cases, cumulatively. An increased detection of CIN2+ cumulatively indicates a possible tradeoff between early disease detection and overdiagnosis10, a serious concern in a disease with very low prevalence of progressive cancers6.
Estimating both effectiveness and overdiagnosis requires longitudinal data from at least two screening rounds within a randomized trial10. At the time of our report, complete data from two screening rounds was available from only two of four cotesting trials (NTCC Phase I, ARTISTIC) and one primary HPV screening trial (NTCC Phase II). Incident cervical cancer cases after cotesting were sparsely reported and represented complete follow-up for the entire study population in one study only11. Since the publication of our report, complete second-round results from another trial (POBASCAM) suggest a similar pattern of reduced cancer incidence12, although data on cancer stages were not reported. The benefit of earlier detection might be considered more robust with a demonstrated reduction in cancer 1B+ incidence6.
Based on these new POBASCAM results, others have calculated that 3.2 additional CIN2/CIN3 would be treated to prevent 1 case of incident cancer as a means of estimating the potential overtreatment associated with cancer prevention13. To address the programmatic impact, however, one might consider the incremental difference in cumulative cancers due to higher cross-sectional sensitivity with HPV-based screening compare to cytology alone. Under these considerations, 8 CIN2/CIN3 would be treated to prevent 1 cancer diagnosis. Neither of these calculations, however, considers potential increased harms other than those stemming from overtreatment (i.e., not harms related to unnecessary retesting, colposcopy, biopsies).
By design, our systematic review was limited to reporting available empirical data to directly quantify benefits and harms. The tradeoffs between earlier disease detection and overdiagnosis still cannot be well quantified without more complete reporting from trials, unless supplemented by the addition of modeling exercises. To help further illuminate this issue, we highlighted the need for more complete reporting from available trials for all participants (complete second round screening and followup, perhaps with linkage to registries) and for a larger array of process measures and health outcomes (recommendations for and compliance with retesting for abnormal results, colposcopies, biopsies, and treatment, ideally with related harms). In doing so, we did not intend to suggest that randomized controlled trials are needed to evaluate every potential advance or nuance in cervical cancer screening, as we agree this would be inappropriate. Instead, comparative effectiveness evaluations to improve established cancer screening must rely on other evidence-based standards wherever appropriate14.
Modeling allows comparisons of different rescreening intervals for different tests, different starting and stopping ages, and other approaches not tested in trials. Modeled results from the United States suggest that a longer rescreening interval after cotesting (every five years) is required to achieve comparability with cytology in programmatic benefits and harms15. These programmatic benefits, however, require adherence with prolonged intervals after cotesting to avoid expected harms from detecting transient HPV infections, among other issues16. There is no national screening program in the United States and evidence suggests prolonged screening intervals up to three years (or longer) are not standard after cotesting in individual or small-group practices17. Thus, judgement is required as to how trial and modeling results for cotesting apply in the United States outside of contexts such as the Veterans Administration, health maintenance organizations, or others managing stable populations over time. These judgments about applicability and other inferential activities are clearly the role of the USPSTF18. Finally, we did not down-grade the quality of trials for issues such as participant blinding. Instead, we noted factors that could affect interpretation of results for these pragmatic trials, through mechanisms such as likelihood of seeking screening outside the program.
Evelyn Whitlock, MD, MPH
Kimberly K. Vesco, MD, MPH
Michelle Eder, PhD
Jennifer S. Lin, MD, MCR
Caitlyn A. Senger, MPH
Brittany U. Burda, MPH
Oregon Evidence-based Practice Center, Center for Health and Research, Kaiser Permanente Northwest, Portland, Oregon
1. Whitlock EP, Vesco KK, Eder M, Lin JL, Senger CA, Burda BU. Correction: Liquid-based cytology and human papillomavirus screening for cervical cancer. Ann Intern Med. 2012;156(1):71-72.
2. Whitlock EP, Vesco KK, Eder M, Lin JS, Senger CA, Burda BU. Liquid-based cytology and human papillomavirus testing to screen for cervical cancer: a systematic review for the U.S. Preventive Services Task Force. Ann Intern Med. 2011;155(10):687-5.
3. Vesco KK, Whitlock EP, Eder M, Lin J, Burda BU, Senger CA, et al. Screening for Cervical Cancer: A Systematic Evidence Review for the U.S. Preventive Services Task Force. Evidence Report No. 86. AHRQ Publication No. 11-05156-EF-1. Rockville, MD: Agency for Healthcare Research and Quality; 2011.
4. Rebolj M, Lynge E. Incomplete follow-up of positive HPV tests: overview of randomised controlled trials on primary cervical screening. Br J Cancer. 2010;103(3):310-314.
5. International Agency for Research on Cancer. IARC Handbooks of Cancer Prevention: Cervix Cancer Screening. Lyon, France: IARC Press; 2005.
6. Arbyn M, Ronco G, Cuzick J, Wentzensen N, Castle PE. How to evaluate emerging technologies in cervical cancer screening? Int J Cancer. 2009;125:2489-2496.
7. Ogilvie GS, van Niekerk D, Krajden M, Ceballos K, Ehlen TG, Martine RE, et al. HPV FOCAL: round one results of a cervical cancer screening trial. Berlin, Germany: International Papillomavirus Conference and Clinical Workshop; September 21, 2011.
8. Leinonen M, Nieminen P, Kotaniemi-Talonen L, Malila N, Tarkkanen J, Laurila P, et al. Age-specific evaluation of primary human papillomavirus screening vs conventional cytology in a randomized setting. J Natl Cancer Inst. 2009;101(23):1612-1623.
9. Ogilvie GS, van Niekerk DJ, Krajden M, Martin RE, Ehlen TG, Ceballos K, et al. A randomized controlled trial of human papillomavirus (HPV) testing for cervical cancer screening: trial design and preliminary results (HPV FOCAL Trial). BMC Cancer. 2010;10(1):111.
10. Ronco G, Miejer C. HPV screening: available data and recommendations for clinical practice. Curr Cancer Ther Rev. 2010;6(2):104-109.
11. Ronco G, Giorgi-Rossi P, Carozzi F, Confortini M, Dalla PP, Del MA, et al. Efficacy of human papillomavirus testing for the detection of invasive cervical cancers and cervical intraepithelial neoplasia: a randomised controlled trial. Lancet Oncology. 2010;11(3):249-257.
12. Rijkaart DC, Berkhof J, Rozendaal L, van Kemenade FJ, Bulkmans NW, Heideman DA, et al. Human papillomavirus testing for the detection of high-grade cervical intraepithelial neoplasia and cancer: final results of the POBASCAM randomised controlled trial. Lancet Oncol. 2012;13(1):78-88.
13. Katki HA, Wentzensen N. How might HPV testing be integrated into cervical screening? Lancet Oncol. 2012;13(1):8-10.
14. Lord SJ, Irwig L, Simes RJ. When is measuring sensitivity and specificity sufficient to evaluate a diagnostic test, and when do we need randomized trials? Ann Intern Med. 2006;144(11):850-855.
15. Kulasingam S, Havrilesky L, Ghebre R, Myers ER. Screening for Cervical Cancer: A Decision Analysis for the U.S. Preventive Services Task Force. AHRQ Publication No. 11-05157-EF-1. Rockville, MD: Agency for Healthcare Research and Quality; 2011.
16. Schiffman M, Wentzensen N, Wacholder S, Kinney W, Gage JC, Castle PE. Human papillomavirus testing in the prevention of cervical cancer. J Natl Cancer Inst. 2011;103(5):368-383.
17. Saraiya M, Berkowitz Z, Yabroff KR, Wideroff L, Kobrin S, Benard V. Cervical cancer screening with both human papillomavirus and Papanicolaou testing vs Papanicolaou testing alone: what screening intervals are physicians recommending? Arch Intern Med. 2010;170(11):977-985.
18. U.S. Preventive Services Task Force. USPSTF Procedure Manual. AHRQ Publication No. 08-05118-EF. Rockville, MD; Agency for Healthcare Research and Quality; 2008.