High concordance observed among PD-L1 expression diagnostic assays

NEW ORLEANS — Three commercially available diagnostic tests demonstrated similar efficacy when measuring PD-L1 protein expression in non–small cell lung cancer, according to data presented at the American Association for Cancer Research Annual Meeting.

These results suggest that clinicians may be able to use these tests interchangeably to determine which patients will benefit from anti–PD-L1/PD-1 immunotherapy drugs, according to Marianne J. Ratcliffe, MA, PhD, diagnostic associate director at AstraZeneca.

Marianne J. Ratcliffe

“We know that PD-L1 is informative for treating lung cancer in the monotherapy setting,” Ratcliffe told HemOnc Today. “We think it is going to be even more important in the future, as combinations begin to come to market. So, PD-L1 testing is going to be very important, and we want to provide a bit of clarity around the unusual situation of having a number of targeted tests.”

Patients with NSCLC and high PD-L1 expression are more likely to respond to targeted therapies than those with low or no PD-L1 expression, according to the results of several clinical trials.

Ratcliffe and colleagues evaluated data from approximately 500 NSCLC tumor biopsy samples, representing squamous and nonsquamous histologies.

They used these samples to compare the diagnostic efficacy of three assays:

Ventana SP263 (Ventana PD-L1 Rabbit Monoclonal Primary Antibody; Ventana/Roche and AstraZeneca/MedImmune), currently used to evaluate PD-L1 expression in clinical trials of durvalumab (MEDI4736; AstraZeneca/MedImmune) with PD-L1 positivity cutoff of 25% of tumor cells with membrane staining;

Dako 22C3 (PD-L1 IHC 22C3 pharmDx; Dako), an FDA–approved companion diagnostic for pembrolizumab (Keytruda, Merck) that was used in the KEYNOTE-001 trial, which validated PD-L1 expression with 1% and 50% positivity cutoffs; and

Dako 28-8 (PD-L1 IHC 28-8 pharmDx; Dako), an FDA–approved companion diagnostic for nivolumab (Opdivo, Bristol-Myers Squibb) used in the Checkmate 057 trial with 1%, 5% and 10% tumor membrane positivity cutoffs.

Ratcliffe reported data from nearly 500 NSCLC patients. The researchers observed good concordance (>90% agreement) between all three assays at multiple cut offs.

The researchers observed good concordance between Ventana SP263 at the 25% cutoff point and Dako 28-8 at the 10% cutoff, with an overall percent agreement of 96%, a positive percent agreement of 91% and a negative percent agreement of 98%.

“There is really good agreement among the three commercially available PD-L1 assays,” Ratcliffe said. “That builds confidence that we may be able to use them interchangeably in the future. But I think the real take-home message is that, instead of thinking of PD-L1 in simple binary terms — positive and negative — there are going to be multiple cutoff points that are going to be clinically relevant when making treatment decisions in the future.”

Ratcliffe noted that PD-L1 assays are likely to become a common factor in pathologic determinations of NSCLC, as well as in treatment decision-making.

“We think of this study as a starting point for debate with pathologists in the clinical community,” Ratcliffe said. “We are figuring out what kind of information we need to assure that patients with get a high quality PD-L1 test, so the oncologist will know the right treatment option for that patient based on that result. In the meantime, we plan to get an independent pathologist to review our data and really validate the quality.”– by Cameron Kelsall

For more information:

Marianne J. Ratcliffe, MA, PhD, can be reached at marianne.ratcliffe@astrazeneca.com.

Reference:

Ratcliffe MJ, et al. Abstract LB-094. Presented at: American Association for Cancer Research Annual Meeting; April 16-20, 2016; New Orleans.

Disclosure: AstraZeneca funded this study. All study researchers are employed by AstraZeneca. Ratcliffe reports that she is an AstraZeneca shareholder.