[Skip to Content]
[Skip to Content Landing]
Views 455
Citations 0
Comment & Response
August 27, 2019

PD-L1 Status and Survival in Patients With Lung Cancer—Reply

Author Affiliations
  • 1Department of Medical Oncology, Dana Farber Cancer Institute, Boston, Massachusetts
  • 2Foundation Medicine Inc, Cambridge, Massachusetts
JAMA. 2019;322(8):783-784. doi:10.1001/jama.2019.9196

In Reply Dr Yang and colleagues highlight some of the key challenges facing the field of biomarker discovery in general and in particular with the use of PD-L1 metrics to predict immunotherapy response in NSCLC. We agree that there are numerous challenges to interpreting PD-L1 pathological data, including the antibody clone used, intensity of staining, heterogeneity of expression within the tumor, and changes in expression over time and with interval treatments. Ultimately, controlling for all of these variables, even in the context of a clinical trial (eg, requiring evaluation of PD-L1 status from multiple biopsy sites), is very challenging and further highlights the importance of exploring these data in a real-world data set where community clinical practice precludes optimization of all of these parameters. Nevertheless, as noted in the Methods section of the article,1 we defined positive and negative PD-L1 results using previously published standards, including a requirement of 50% staining in a central laboratory using the 22C3 or SP142 antibodies, or annotation in the medical record of a positive result. Cases that did not adhere to these guidelines were excluded.