番茄社区

The recent update to the Standardized Definitions for Efficacy End Points (STEEP) criteria for adjuvant breast cancer (BC) trials included some separately defined endpoints for neoadjuvant trials.

" identified a need to separately address end points for neoadjuvant clinical trials," Jennifer Litton, MD, of The University of Texas MD Anderson Cancer Center in Houston, and co-members of the working group explained in the .

"The multidisciplinary NeoSTEEP working group of experts was convened to critically evaluate and align neoadjuvant BC trial end points," the authors wrote. "Consistent prespecified definitions and interventions are paramount for clinically meaningful trial results and cross-trial comparison."

The group recommended a preferred definition for pathologic complete response (pCR) and residual cancer burden (RCB). Litton discussed those definitions as well as additional considerations in the following interview.

Why was it necessary to create the NeoSTEEP group and define some endpoints separately for neoadjuvant trials?

Litton: There are so many issues to consider when writing a meaningful clinical trial. During the STEEP 2.0 working group and subsequent publication, the authors recognized that there are endpoints that are very specific to neoadjuvant trials that STEEP does not address. The STEEP working group recommended a separate working group to evaluate and make recommendations on what endpoints would be important.

It has become clear that different trials may have chosen different endpoints, some even defining those endpoints differently, which may be confusing when looking at outcomes. NeoSTEEP looked at the history and evidence of currently used endpoints in neoadjuvant trials, and then the working group made recommendations for both what should be included and what the definitions should be for these endpoints.

It is important to define when the time point starts and what constitutes an event. For neoadjuvant trials, the initiation of the intervention happens before surgery, and so that clock should also start before surgery rather than, for example, at the time of surgery as for most adjuvant trials. Additionally, consideration of what happens if there is progression in the breast or at a distant site before surgery should also be included in these endpoints, which would be lost if surgery is the starting point.

What is your definition of pathologic complete response, and how did you arrive at it?

Litton: Pathologic complete response definitions were taken directly from the currently available FDA guidance referenced in the paper. We evaluated data for both definitions that differ only by the presence or absence of in situ cancer. While the working group did determine a definition preference, the group also acknowledged that either definition could be used as long as it was well defined before the study started, and that the pathologic evaluation and reporting was consistent throughout and at all participating sites.

What is your definition of residual cancer burden, and how did you determine that?

Litton: RCB is an index developed by Fraser Symmans and colleagues with both categorical and continuous variables that provides more granular information than just the binary answer of pCR yes/no. This is becoming more helpful when looking at potential survival endpoints with, for example, immunotherapy.

The working group did not feel the data were ready to make a recommendation that RCB should be included routinely as a primary endpoint, but the group did recommend to garner more information about it as well as to evaluate across multiple trials; RCB should be included, whenever possible, as a secondary endpoint.

What are one or two of the other key considerations you cover in your paper?

Litton: We discuss in the paper special considerations such as endpoints under investigation for hormone receptor positive breast cancer, nodal evaluation guidance, the importance of including patients with inflammatory breast cancer, and supporting quality-of-life measures in these clinical trials.

Read the NeoSTEEP document here.