DENVER – In patients with non–small cell lung cancer (NSCLC), the programmed death-ligand 1 (PD-L1) status of the primary tumor is generally a reliable predictor of the status of tumor elsewhere the body, suggest a pair of retrospective cohort studies presented at a world conference on lung cancer.
Results showed a high rate of concordance of PD-L1 immunohistochemical (IHC) staining, whether comparing primary with nodes (81%-89%) or primary with metastasis (77%), investigators reported.
Dr. David Rimm
These findings are relevant in that some studies have suggested that high PD-L1 expression is a biomarker for benefit from agents that inhibit the programmed cell death 1 (PD-1) signaling pathway. Thus, being able to use archival primary tumor to assess PD-L1 status might help guide decisions about treatment options in the metastatic setting.
“This is pretty good concordance, especially considering that our assays only have about 75% concordance [among them], depending on how you measure it and what cutpoints you use and which antibodies you use,” said invited discussant Dr. David Rimm, professor of pathology and of medicine, director of pathology tissue services, and director of translational pathology at Yale University in New Haven, Conn.
“The take-home message is, is the glass half full or is the glass half empty? And the real message is that maybe the glass is twice as big as it needs to be,” he said. “That is, what we really need to do is come up with a uniform assay here that’s standardized so that we can actually do studies like this a little more carefully.”
In the first study, Dr. Brandon S. Sheffield, of the pathology office at BC Cancer Agency, Vancouver, studied 78 patients who underwent resection of a primary nonsquamous NSCLC and were found to have nodal involvement.
Dr. Brandon Sheffield
For each patient, they compared PD-L1 staining between the primary tumor and the matched nodal tumor, simultaneously testing various antibodies; Bristol-Myers Squibb Canada performed some of the IHC. Tumors were considered positive if at least 1% of cells stained with any intensity.
Results showed an 81% rate of concordance of PD-L1 staining between the primary and node; in about 8% of cases, only the primary was positive, and in about 9%, only the node was positive.
Study results also showed good concordance across the three different antibodies: SP142 (Spring Bioscience), E1L3N (Cell Signaling Technology), and 28-8 (Dako). In 76% of cases, all three antibodies were in agreement.
In some cases, clusters of infiltrating macrophages stained for PD-L1. “This could represent a possible pitfall, especially using a cutoff as low as 1%, although with some practice, one can appreciate that the staining of macrophages is somewhat different than the crisp membranous staining seen in tumor cells,” Dr. Sheffield.
“PD-L1 IHC is feasible and it can be done in your laboratory. There are small but very relevant differences in testing primary tumor tissue and lymph node metastasis, and that will need to be explored with a bias toward testing more, not less, tissue,” he concluded. “Multiple methods for PD-L1 IHC appear to be equivalent, and we should be able to have some freedom in choosing the best PD-L1 IHC assay for our own laboratories.”
Dr. Paul Mitchell
In the second study, Dr. Paul Mitchell of Austin Health, Melbourne, and colleagues assessed PD-L1 staining among 433 sequential patients who underwent resection of primary NSCLC between 1992 and 2010. Bristol-Myers Squibb performed the IHC staining and some of the scoring.
In one set of assays, the investigators used the 28.8 antibody and considered tumors to be positive if at least 5% of cells showed membranous staining of any intensity.
Results here showed that 28% of primaries were PD-L1 positive. The rate was similar for men and women, but higher in squamous tumors than in adenocarcinomas (39% vs. 18%). It was merely 14% in patients having an epidermal growth factor receptor (EGFR) mutation. The rate of positivity increased with the number of pack-years of smoking and, starting 5 years after cessation, decreased over time.
A total of 57 patients had paired primary tumor and metastatic tumor (most commonly from brain metastases). The median time from primary to metastasis was 1.3 years; in eight patients, metastasis was synchronous.
In this cohort, there was a 77% rate of concordance of PD-L1 staining between the primary and metastasis (r = 0.37, P = .0049); in 11% of cases, only the primary was positive, and in 12%, only the metastasis was positive. Also, among the eight patients having multiple metastases, all samples were concordant in six patients.