Discussion Questions

• ■
  

  What is the research question?

  
  
• ■
  

  Can you briefly explain epigenetics?

  
  
• ■
  

  What do you think of the researchers’ choice of biomarkers?

  
  
• ■
  

  How would you describe the overall study design?

  
  
• ■
  

  What were the main results of the study?

  
  
• ■
  

  Can you comment on the strengths and weaknesses of this study?

Introduction

An estimated 12,200 new cases of cervical cancer and 4200 deaths were predicted for 2010. Pap testing has cut incidence and mortality in half over the last 35 years. Still, follow-up of vague results can be complex; 3 approaches are acceptable when a patient has atypical squamous cells of undetermined significance (ASC-US): 2 repeat tests at 6-month intervals, colposcopy, or human papillomavirus (HPV) testing. A minority of these patients will go on to develop cervical intraepithelial neoplasms; perhaps 0.1-0.2% already have invasive lesions despite ASC-US findings. Can a new type of analysis better predict serious lesions in this group of women?

George A. Macones, MD, MSCE, Associate Editor

Study Design

Macones: What is the research question?

Cahill: This is actually a very important topic. As we all know, Pap smears with ASC-US are common, occurring in approximately 5% of samples. The management of a patient with ASC-US is controversial, because a fraction of these have higher-grade lesions. One current strategy is to stratify who needs a colposcopy by doing “reflex” HPV typing on ASC-US Pap smears. However, this is an imperfect strategy. These authors take an epigenetic approach to refining the management of patients with ASC-US Pap tests. In this study, they assess and compare methylation tests for 3 tumor suppression genes: protocadherin 10 (Pcdh10), Wt1, And Pax1.

Macones: That sounds complicated. Can you briefly explain epigenetics?

Cahill: Sure. We are understanding more and more about the sequence of the human genome. However, there is much more to a phenotype than just a DNA sequence. This is where epigenetics comes in. I would define epigenetics as factors aside from gene sequence that control gene activity during development. Resulting changes in gene activity can be passed from 1 generation to the next without modifications to DNA. In maternal-fetal medicine, a classic example of epigenetics involves genomic imprinting. In genomic imprinting, only 1 parent’s copy of a specific gene is active in the child; the other is suppressed. Disease can result if the single active gene is flawed. That is the case with disorders such as Prader-Willi syndrome or Angelman syndrome. In cancer, the study of methylation is an important area of research, as we see in this study. Methylation of certain base pairs in a gene sequence can shut down a gene’s activity, and the degree to which methylation occurs can be shaped by factors, such as stress, age, or environment. This is an epigenetic change.

Macones: Thank you. Now that we understand a bit about epigenetics, let’s move on. What do you think of the researchers’ choice of biomarkers?

Odibo: They all seem very reasonable to me. The authors selected these specific biomarkers precisely because they had been suggested to be associated with cervical neoplasia in prior studies, making them biologically plausible candidates. I thought they did a nice job with this.

Macones: How would you describe the overall study design?

Odibo: Well, subjects with abnormal Pap smears have been enrolled and followed in a large multicenter cohort study in Taiwan—the T1899 cohort. The study by Lin et al includes a fraction of the T1899 cohort; specifically, women with an ASC-US smear who had a colposcopic-directed biopsy. If I had to give this a label, I would describe it as a cohort study.

Macones: Can you comment on the analytic approach?

Stamilio: The goal of the study was to look for 3 biomarkers—3 methylated tumor suppressor genes—and determine how their presence correlated with disease severity in comparison with traditional HPV testing among women with ASC-US Pap smears. Methylation-specific polymerase chain reaction (PCR) was used to examine cervical scrapings for the biomarkers. PCR-reverse line blot hybridization and Hybrid Capture 2 (Digene, Silver Spring, MD) were used to identify high-risk HPV DNA in cervical samples. Relationships between cervical lesions and biomarkers were examined with the Fisher’s exact test or χ ² test; logistic regression models were used to explore age-adjusted associations between positive biomarkers and high-grade cervical neoplasia. I think the approach was generally reasonable, although the real goal of this study was to look at characteristics of the different test options.

Macones: What about sample size? Is there a sample size estimate provided?

Stamilio: No, there is no calculation provided in the manuscript. My sense is that this is really a convenience sample. In other words, this is the sample of patients that the authors had available.