Cervical Cancer Screening: How Do We Make Sense of It All and What Is the Right Balance?

It was not so long ago when providers were screening women for cervical cancer with annual cytology. We kept it simple: we told women to come in every year, and women acknowledged that this was a paradigm that they needed to follow. Most would agree that this has contributed to at least a 60% reduction in invasive cervical cancer since the widespread implementation of screening.¹ Roll forward to 2016, and providers now practice under completely different recommendations from the American Cancer Society, the American Society for Colposcopy and Cervical Pathology, the American Society for Clinical Pathology, the US Preventive Services Task Force, and the American College of Obstetricians and Gynecologists. Although for the first time we saw consistency in the recommendations between the US Preventive Services Task Force and the major professional societies, we witnessed a controversial and perhaps dramatic leap from annual screening to screening every 3 years with cytology alone and every 5 years with cotesting (cytology and high-risk human papillomavirus [HPV] testing) in women aged ≥30 years.^2–5 In addition, to add to these changes, a third option for screening, with high-risk HPV testing alone, was introduced in 2014.⁶

In this issue of Cancer, Silver et al have put forth clinically significant information regarding the use of a 3-year cotesting interval from the Kaiser Permanente Northern California (KPNC) health system.⁷ Their study provides valuable data regarding the ratio of the number of biopsies to detect cervical intraepithelial neoplasia 3/adenocarcinoma in situ (CIN3/AIS) in a long-term screening group. It is important to note that the study by Silver et al provides meaningful information for providers and groups who initially might find a 3-year screening interval more appropriate and less of an initial “quantum leap” as they eventually work toward a 5-year interval. More importantly, however, for the first time based on US data, it is clear that we need to view screening tests as a continuum and not as individual tests that function independently.

The most useful reported data by Silver et al⁷ are the ratios of biopsy and CIN3/AIS detection, and the rate of adenocarcinoma and squamous cell carcinoma (SCC) by 3-year interval in the long-term screening cohort. The ratio of SCC to adenocarcinoma at the beginning of the study period was approximately 3.6:1, a finding that is consistent with national database information. However, the ratio of SCC to adenocarcinoma at the 3-year and 5-year intervals was approximately 1:1 and the number of adenocarcinoma cases detected at the 3-year interval was not dissimilar from that at the beginning. This finding suggests that the 3-year cotest screening interval is not sufficient for preventing adenocarcinoma because the same number of new cases continues to be found at the next testing opportunity. Furthermore, although the rate of SCC was found to be reduced at the second and third rounds of cotesting in the long-term screening cohort, there still were 4 to 5 cases per 100,000 women screened.⁷ This raises the question: Can we do better? Should the target for long-term cervical cancer screening and prevention be zero cases of cancer?

As the authors describe, it is unclear how many women underwent a colposcopic evaluation; Silver et al only knew whether a cervical biopsy was submitted for histopathology.⁷ Therefore, the number of colposcopic examinations per screened subject was not known and this represents a limitation of the analysis. Because the clinical protocol intended 3-year interval cotesting, and because this was achieved (with a median screening interval of 2.9 years), these results do not inform a comparison with 1-year screening or 5-year screening. When the rate of all cancers in the study was 13 per 100,000 women screened, the rate of cervical biopsy to CIN3/AIS was 20:1.⁷ If we could reduce the cancer rate from 13 per 100,000 women screened to 1 per 100,000 women screened but the biopsy to CIN3/AIS rate rose to 200:1, is that acceptable? Who should decide and set the “appropriate” threshold?

All judgments regarding patient care recommendations consider the balance between benefits and risks. The individual benefits and risks are never of equal importance; another judgment must be made concerning the weight assigned to each outcome, factored with the likelihood of that outcome. The aggregate of those judgments results in a perceived balance that drives the decision making. Who should assign those weights and make those judgments? An individual woman? Her individual physician? A collective guideline panel? Under the principles of patient autonomy and physician beneficence, the woman should decide. Before we can support that decision making we must be able to provide all the necessary information in a way that each woman can understand and use to make a wise decision. That is an aspirational goal, and to date, we are not there-not even close. The same could be said for the experienced physician; we do not currently have the necessary information at the point of care. If a guideline panel assigns those weights and makes those judgments, they should include patient preference data. Future research should determine the average comparative weight assigned by a woman associated with the risk (ie, adverse experience) of a screening test, colposcopy, biopsy, false-positive result, treatment, and so forth. Future research also should determine the average comparative risk assigned by a woman for preventing cervical cancer. Finally, a guideline panel could use that research to inform their judgments regarding the balance of benefits and risks.

Can the results of the study by Silver et al be readily applied to other populations? There may be unique characteristics of the women in the KPNC study population that differ from those of women in the United States overall, or a different subpopulation, or an individual woman. These could include socioeconomic status, race, health behaviors, and other factors. There may be unique features of the KPNC health delivery system that may not be experienced in other health systems. There may be less continuity of care, more variability in care, and other factors at play when trying to apply the KPNC results to other settings. The long-term screening cohort was a selected subgroup of women who had an average of 2.9 visits for cancer screening over a 9-year period. The results from that group may not be applicable to women who move location, or change provider, or change insurance status.

As practicing clinicians, we have seen firsthand the continued consternation and struggle providers have with changes in screening guidelines, particularly with increased screening intervals. It is important to recognize that the adoption of new guidelines is rarely swift and complete. The use of cotesting first was recommended in 2004, and it has taken >10 years to achieve >60% use in the screened population (unpublished data). Modifications to screening recommendations should be gradual, with an emphasis on education and an in-depth understanding of barriers and poor uptake. Future challenges include, but are not limited to, the following: 1) educating providers and patients regarding balancing the benefits and harms of screening; 2) educating providers regarding the concepts of risk and the justification for certain thresholds; 3) educating providers and patients regarding the continuum of screening and how one round of screening might influence the risk assessments in the next round; and 4) understanding and protecting the autonomy of women when making screening choices and balancing this autonomy with guidelines and provider experience and understanding. In the end, our understanding and knowledge regarding screening become ever more data-driven and complex, but we have a specific responsibility to keep it simple and easy to understand. Otherwise, we risk undermining the importance of screening altogether.