Determination of the PD-L1 status of a patient’s tumor has become increasingly important for informing the clinical decision whether to offer certain immunotherapeutic agents, making standardization of the tests and antibodies used to determine the PD-L1 status necessary to provide accurate and consistent results.
Toward this goal, investigators reported findings from an evaluation of several laboratory-developed tests (LDTs) as compared to PD-L1 assays used in clinical trials at the 17th World Lung Cancer Conference, at the Annual Meeting of the International Association for the Study of Lung Cancer (IASLC) in Vienna.
PD-L1 expression as assessed by immunohistochemistry (IHC) from fresh-frozen paraffin embedded tissue is the main biomarker currently used to decide whether a patient may benefit from anti-PD-1/PD-L1 agents. In clinical trials, Dako (22C3, 28-8) and Ventana (SP263) assays are most often used as diagnostic tests.
“Harmonization of assays and the development of laboratory-developed tests are urgently needed in France since Dako and Ventana platforms are not available in all pathology labs, and they are expensive; reimbursement for this testing remains insufficient,” commented Julien Adam, MD, of the Gustave Roussy Cancer Center in Villejuif, France. Adam and colleagues compared these 2 PD-L1 assays that are used in clinical trials with LDTs.
The investigators compared PD-L1 assays and LDTs with the 5 main PD-L+ antibodies in a 41-case series: For IHC analysis, 5 anti-PD-L1 clones, including 28-8, 22C3, E1L3N, SP142, and SP263 were used in Ventana BenchMark Ultra, Leica Bond or Dako Autostainer Link 48 IHC platforms. For matching platforms, Dako or Ventana assays were performed with clones 22C3, 28-8, and SP263.
LDTs were developed in each center in non-matching platforms and with other antibodies. A total of 35 PD-L1 stainings (8 with PD-L1 assays and 28 with LDTs) were performed across different platforms and antibodies for each case. Seven thoracic pathologists trained to PD-L1 scoring in expert courses scored tumor cell and immune cell staining. Each pathologist analyzed 6 cases, and compared the stainings obtained with the 5 antibodies on all platforms. Tumor cell and immune cell PD-L1 stainings were scored semi-quantitatively as recommended in PD-L1 Dako and Ventana assays.
The investigators found that the 22C3, 28-8, and SP263 assays were highly concordant as performed in several different centers regarding tumor cell staining. Similar concordance compared to those 3 assays was demonstrated by 14 of 27 (51.8%) LDTs. Specifically, LDTs using clone SP263 achieved the highest concordance rate across all platforms.
Immune cell staining produced poorer concordance rates, both for assays as well as LDTs.
“We found that 28-8, 22C3, and SP263 assays gave comparable results for tumor cell staining, as well as half (51.8%) of laboratory-developed tests using 28-8, 22C3, SP263, and E1L3N clones” said Adam. “These results indicate that LDTs can be an option for PD-L1 testing, but caution is required for validation and further use of these tests.”
Julie R. Brahmer, MD, MS, director of the Thoracic Oncology Program at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, who was a member of the panel of experts for this session, commented, “This study is unique in that it represents ‘real world’ testing and the type of results we receive in the clinic.”
The goal of the harmonization program is to ensure that the determination of PD-L1 levels for a single patient in a clinical laboratory will have the same accuracy seen in clinical trials.
Another attempt at standardizing PD-L1 testing is called the “Blueprint PD-L1 IHC Assay Comparison Project,” which represents a collaboration between the International Association for the Study of Lung Cancer (IASLC), the American Association for Cancer Research, and academic medical centers with pharmaceutical companies producing some of these drugs (including Merck, AstraZeneca, Genentech/Roche, and Bristol-Myers Squibb), together with the diagnostic companies Ventana and Dako that produce the PD-L1 assays.
This trial used 4 tests to evaluate 38 samples of human non-small cell lung cancer (NSCLC), and found that in all of the tests, half of the tumors were positive for PD-L1, and 5 of the tumors were negative. However, in 37% of cases, some tests considered the sample positive while others considered it negative. This disparity in results implies that the choice of test used to determine a tumor’s PD-L1 dependence may influence the decision of whether or not a patient is offered anti-PD-L1 therapy.
“A main issue is how to select patients for these immunotherapies. Each company is pursuing their own predictive PD-L1 assay in order to select patients. However, the PD-L1 assays are all different in terms of antibody used and cut-off values for positive/negative results,” explained Fred R. Hirsch, MD, PhD, CEO of the IASLC.