Discussion Questions

• •
  

  What were the study’s objectives and design?

  
  
• •
  

  How have cervical cancer screening guidelines changed recently?

  
  
• •
  

  Why might colposcopy be used as a quantifiable measure of harm?

  
  
• •
  

  Which screening strategies seemed most promising?

  
  
• •
  

  What are the main strengths and limitations of this study?

  
  
• •
  

  Should the results of this study change our current clinical management?

Introduction

Screening programs based on cervical cytology have markedly reduced morbidity and mortality from cervical cancer, yet it remains an important public health problem; 12,170 diagnoses of the disease were estimated for 2012, along with 4200 deaths. The availability of molecular diagnostics that detect high-risk human papillomavirus (HPV) DNA, the cause of all cervical cancers, has fundamentally changed the screening landscape. Likewise, the HPV vaccine may alter the epidemiology of the disease. Screening algorithms will have to evolve concurrently in order to remain effective. Cox et al highlight this shifting paradigm, along with the important trade-offs between sensitivity and specificity inherent in any screening test.

Andrea R. Hagemann, MD, and George A. Macones, MD, MSCE, Associate Editor

Introduction

Screening programs based on cervical cytology have markedly reduced morbidity and mortality from cervical cancer, yet it remains an important public health problem; 12,170 diagnoses of the disease were estimated for 2012, along with 4200 deaths. The availability of molecular diagnostics that detect high-risk human papillomavirus (HPV) DNA, the cause of all cervical cancers, has fundamentally changed the screening landscape. Likewise, the HPV vaccine may alter the epidemiology of the disease. Screening algorithms will have to evolve concurrently in order to remain effective. Cox et al highlight this shifting paradigm, along with the important trade-offs between sensitivity and specificity inherent in any screening test.

Andrea R. Hagemann, MD, and George A. Macones, MD, MSCE, Associate Editor

Study Design

Hagemann: Can you summarize the study’s objectives and design?

Kuroki: The objective of this study was to compare the sensitivity and specificity of 9 cervical cancer screening strategies to the current screening standard—cytology with HPV testing of atypical squamous cells of undetermined significance (ASCUS) for detection of cervical intraepithelial neoplasia (CIN)2+ or CIN3+; women with HPV-positive ASCUS or low-grade squamous intraepithelial lesions or lesions of greater severity are referred for colposcopy.

In addition, the authors assessed the potential harm of each screening strategy, estimated by the number of tests and the number of colposcopies needed to detect each high-grade lesion at baseline. I would like to point out that this study is a post-hoc analysis of the ATHENA (Addressing THE Need for Advanced HPV Diagnostics) HPV trial. The original study evaluated the performance of the Cobas HPV test (Roche Molecular Diagnostics, Pleasanton, CA) in over 47,000 women and resulted in FDA approval for use of the test in the United States.

Squires: Yes, before focusing on the study at hand, it is important to consider the original design of the ATHENA HPV study. This was a prospective study that enrolled more than 47,000 women, aged 21 and older at initial presentation, for cervical cancer screening. These women then underwent both Papanicolaou testing (standard liquid-based cytology) and HPV genotyping with the Cobas HPV test. The primary objective was to clinically validate Cobas HPV testing, both alone and in combination with cervical cytology, as a viable option in cervical cancer screening. Cobas HPV testing identifies the presence of 11 high-risk HPV types and 1 possibly high-risk type in the form of a pooled result, and it specifically reports the presence or absence of HPV genotypes 16 and 18.

In contrast to the original study, the present design is a post-hoc analysis that focuses on women older than 30 years who were enrolled in the ATHENA HPV study. As noted, it compared 9 different combinations of screening strategies to the most widely used strategy in the United States—cytology with reflex HPV testing and when necessary, colposcopy—and examined the detection of CIN2 or CIN3 at enrollment. The goal was to determine the sensitivity and specificity of each strategy. The number of tests required at baseline and the number of initial colposcopic examinations needed to detect CIN2 or greater served as surrogates for harm.

Hagemann: Let’s put this study in context. What is the difference between traditional cervical cytology, HPV testing, and cotesting? How have cervical cancer screening guidelines changed recently?

Wilkinson-Ryan: Traditional cervical cytology involves looking at cells from the cervix under a microscope to detect changes associated with precancerous and cancerous lesions. HPV testing, which can be performed in a number of ways, is intended to detect the presence of the virus that causes cervical cancer. HPV 16 and HPV 18 are the types most strongly associated with cervical cancer, but there are at least 13 high-risk HPV genotypes. Cotesting uses cytology and HPV testing at the same time to determine both a patient’s risk for developing a precancerous or cancerous lesion and the next step in the evaluation.

The major changes in the most recent screening guidelines from the American Cancer Society, the American Society for Colposcopy and Cervical Pathology, and the American Society for Clinical Pathology (ACS/ASCCP/ASCP) have to do with the timing of screening, incorporation of routine cotesting for 30- to 65-year-olds, and use of HPV genotyping. Women aged 21–29 years with negative cytology or ASCUS with negative HPV can be screened at 3-year intervals by cytology alone. Women aged 30–65 years can now be followed with cytology alone every 3 years or preferably with HPV and cytology cotesting at 5-year intervals. Finally, HPV 16 and HPV 18 genotyping is one option for determining the next step for women with normal cytology who are HPV positive. This study focuses on additional alternative screening approaches for women aged 30–65 years; these could be endorsed in future guidelines.

Hagemann: Can you comment on the use of colposcopy as a quantifiable measure of harm? Along those lines, what are your thoughts regarding using CIN2+ vs CIN3+ as the most clinically meaningful outcome for screening?

Kuroki: The authors of this paper have followed the statement made by the recent ASC/ASCCP/ASCP cervical cancer screening guidelines, which acknowledge that colposcopies are associated with physical discomfort and are a prerequisite to more invasive treatments with greater short- and long-term risks of harm. “Since the number of subjects undergoing colposcopy is usually reported in controlled studies and more screening leads to more screen positives and therefore more colposcopy, it provides a surrogate for the potential harm.” Other considerations for possible harm are important, but are much more difficult to quantify. For example, consider the potential for a false-positive result to produce anxiety or a needless intervention, potential stigmatization from the diagnosis of a sexually transmitted disease, relationship disruption, and pain and bleeding. There could also be more long-term risks of pregnancy complications such as preterm labor and premature delivery as a consequence of diagnostic/therapeutic interventions. So it makes sense to use colposcopy as a surrogate marker for potential harm.

The most clinically meaningful outcome for screening would ideally be cervical cancer. However, given the natural history of cervical carcinogenesis, few studies have sufficient numbers of cancer cases to assess cancer risk directly, and they therefore use CIN2+ and/or CIN3+ as surrogate outcomes. However, the recent ASC/ASCCP/ASCP cervical cancer screening guidelines specifically designated CIN3 as their primary outcome, stating, “although CIN2 is the widely accepted threshold for treatment, to provide an additional margin of safety, we posited that CIN2 should not be the primary target of cervical cancer screening.”

Hagemann: Can you describe some of the important considerations in balancing sensitivity and specificity for cervical cancer screening? And, how does positive predictive value factor into these considerations?

Wilkinson-Ryan: Screening tests must have a low false-negative rate in order to maximize capture of patients with disease, even though we’ll never capture all patients at risk. Therefore a high sensitivity—where the patients who have the disease of interest test positive for it—is paramount for any screening test to be effective. However, in the case of cervical cancer screening, having an abnormal test requires the patient to have a colposcopy, which can be uncomfortable and may lead to more invasive procedures. Since most women with CIN2+ do not have cancer and those with CIN are often identified in subsequent screening rounds, maximizing sensitivity for a surrogate endpoint like CIN2+ may not be critical. Having an abnormal Pap test or positive HPV test can also have significant negative psychological and emotional effects on the patient. For these reasons, it is also important to have a low false positive rate (or high specificity, where those who don’t have the disease of interest actually have a negative test) in order to avoid unnecessary procedures and stress. As we lower the threshold of what is considered positive to increase sensitivity, we also reduce our specificity. This requires us to carefully balance what we consider a positive test with the risks for labeling a patient incorrectly.

Positive predictive value is related to test sensitivity and the prevalence of the disease for which the patient is being tested. When comparing tests within the same population, a more sensitive test will have a higher positive predictive value. For instance, a positive screening test in an area where the disease is highly prevalent is more likely to be a true positive than the same test done in an area where the disease is rare. Cervical cancer is a relatively rare disease in the United States. Although screening tests must be sensitive to capture all cases, we need tests with high specificity (low false-positive rates) to optimize the positive predicative value for cervical cancer screening.