Discussion Questions

• ■
  

  What novel idea is addressed in this study?

  
  
• ■
  

  What is the study design?

  
  
• ■
  

  What statistical tests were applied and why?

  
  
• ■
  

  What were the key findings of this study

  
  
• ■
  

  What do you think of the algorithm proposed in Figure 1?

  
  
• ■
  

  Are there limitations regarding the use of tumor diameter?

  
  
• ■
  

  What are the strengths and limitations of the study?

Introduction

Endometrial cancer, the most common gynecologic cancer, is usually confined to the uterus and associated with a good prognosis. Surgical staging includes hysterectomy, bilateral salpingo-oophorectomy, and pelvic- and para-aortic lymphadenectomy (per the International Federation of Gynecology and Obstetrics 2009). Staging is generally safe, but lymphadenectomy can be associated with bleeding, lymphocyst formation, and ureteral and vascular injury. Given that most patients with endometrial cancer do not have nodal metastases, it would be beneficial if cases requiring full staging could be accurately identified at the time of surgery. A new study evaluates the use of adding frozen section assessment of tumor diameter and recommends an algorithm that selects which patients should undergo lymphadenectomy.

Linda Van Le, MD and George A. Macones, MD, MSCE, Associate Editor

Study Design

Van Le: What novel idea is addressed in this study?

Rossi : The study proposes that tumor diameter measured at frozen section can assist in triaging patients to lymphadenectomy during surgical staging for endometrial cancer.

Ko : Triage would increase efficiency in the operating room, reduce potential morbidity and mortality associated with lymphadenectomy, and possibly, reduce costs.

Gehrig : I also think as the population’s body mass increases, there may be more technical limitations. If you have another piece of information at the time of the frozen section, that would reassure you against converting to a laparotomy to do the lymphadenectomy.

Van Le: What is the study design?

Ko : The study design is cross-sectional. It is a retrospective study, because all of the information has already been collected. In terms of measuring the outcome, they are trying to determine how many patients have lymph node metastases in this low-risk population. Their independent factor or variable would be tumor diameter. Other variables they are looking at include myometrial invasion and tumor grade and the association of these factors with positive lymph node metastases in low-risk early-stage endometrial cancer patients. They also looked at secondary outcomes, comparing myometrial invasion assessment on frozen section to permanent section. Basically they are asking whether the diagnoses match each other, how concordant they are, and similarly, for histologic grade, how concordant are the frozen and permanent sections?

Van Le: What statistical tests were applied and why?

Ko : They used the Jonckheere-Terpstra test. It is a nonparametric test to determine whether different categories are related to an outcome in an ascending or descending fashion. Among small tumors, medium tumors, and large tumors—the authors used <3 cm, 3-6 cm, and >6 cm—the researchers examined the concordance of tumor grade and myometrial invasion for small vs medium vs large tumors. They also looked at whether cases meeting low-risk criteria on frozen section were also low-risk on permanent sections among the 3 tumor size groups.

The Cohen’s kappa coefficient is a test to assess for interobserver concordance. This measure was used to assess how diagnoses between different observers agree with each other. There are different types of tests you can use for that but they chose the Cohen’s kappa coefficient. So it gives us a little bit of a sense of how much pathologists might agree with each other for findings related to different tumor sizes.

DiFurio : I would say there’s a bit of literature out there to suggest that unfortunately, we are not excellent when it comes to interobserver histology agreement.

Gehrig : Were the cutoffs of 3 cm, 3-6 cm, and >6 cm chosen a priori?

Ko : This is not completely clear. It appears that they had chosen 3 cm, 3-6 cm, and >6-cm cutoffs before they analyzed the data for myometrial invasion, histologic grade, and low-risk criteria. However, when they drew their final conclusion, they basically looked for the centimeter cutoff at which there were no lymph node metastases. This is demonstrated in their last table where they show you the frequency of cases with metastases at tumor diameter of ≤1 cm, ≤2 cm, ≤3 cm, ≤4 cm, and so on.