PD-L1 testing in non-small cell lung carcinoma

Expires: November 9, 2023

  • Photo of Anja Roden, M.D.Presenter

    Anja Roden, M.D.
    Professor of Laboratory Medicine and Pathology
    Division of Anatomic Pathology
    Mayo Clinic, Rochester, Minnesota


    Contact us: mcleducation@mayo.edu.

    Transcript and references


    Welcome to today’s “Hot Topic.” We will be discussing PD-L1 testing in non-small cell lung carcinoma.

    I’m Anja Roden, M.D., one of the surgical pathologists at Mayo Clinic in Rochester with subspecialty in thoracic pathology, and I am also one of the medial directors of our Immunostains Laboratory.


    I have nothing to disclose.


    After a short introduction that will include the role of PD-1/PD-L1 interaction in the immune system, we will discuss the challenges of PD-L1 testing, and talk about the current PD-L1 testing in non-small cell lung carcinoma (NSCLC). And then I will leave you with a few take-home points.

    PD-L1 in the immune system1-3

    As you might know, PD-1 located on T-cells interacts with PD-L1, which is located on tumor cells. This interaction will actually inhibit T-cells, which then can lead to outgrow of the tumor. Therefore, anti-PD1 or anti-PD-L1 drugs have been developed that are supposed to block the PD1/PD-L1 interaction, and therefore, boost the host anti-tumor immune response, which then should inhibit tumor growth.

    And actually, responses to these drugs have not only been seen in lung adenocarcinoma (adenoCa) but also in squamous cell carcinomas (SQCC), which is really exciting because, until then, we did not have targeted therapy for squamous cell carcinomas.

    Anti-PD-1 versus chemotherapy

    The probably, seminal, clinical trial for using anti-PD-1 or anti-PD-L1 drugs in NSCLC was the one by Reck and colleagues, published in The New England Journal of Medicine in 2016. In this trial, 305 patients were included that had untreated stage 4 NSCLC, and all of these carcinomas had PD-L1 expression using clone 22C3 in at least 50% of the tumor cells.

    As you can see here, patients that were treated with anti-PD1, which is pembrolizumab, had a much better progression free survival, and also overall survival, than patients who were treated with chemotherapy.

    PD-L1 testing

    Now, in the subsequent years, multiple clones have been developed for PD-L1 testing. These clones are associated with certain drugs, as they had been used in corresponding clinical trials. In addition, the Food and Drug Administration (FDA) recommended, or introduced, companion tests. And the companion test is PD-L1 testing using an FDA-specified clone, which is required before the treatment is initiated. If it is a complementary test, that means that PD-L1 testing, again using an FDA-specified clone, is strongly encouraged before treatment. 

    PD-L1 testing in NSCLC

    In NSCLC, we review the tumor for the tumor proportion score (TPS). This score is the number of PD-L1 positive tumor cells, divided by all tumor cells, and then multiplied by 100. Basically, it is the percentage of PD-L1 positive tumor cells. We do look for membranous staining. Only membranous staining counts for the TPS, and we do need to have at least 100 viable tumor cells. In regards to the membranous staining, it doesn’t matter whether it is weak staining or strong staining, and whether the entire cell membrane stains or only part of the cell membrane. For immune cells, this is defined as the percent of PD-L1 positive tumor-associated immune cells. We also report those. If we look at a metastasis in the lung, we need to know what the primary tumor is because there are different interpretation guidelines for PD-L1 based on the primary tumor site.

    NSCLC therapy chart

    If we look at NSCLC, here you can see the companion tests that are currently FDA approved. Here we have Keytruda, which is pembrolizumab, or an anti-PD-1 drug. This requires testing with clone 22C3. And if the drug is used as first-line therapy — that means that the patient did not receive any other therapy before that — then a TPS of at least 50% is required for treatment. If the drug is used as a second-line therapy — that means the patient received probably chemotherapy already — then a TPS of 1% would suffice.

    We also have nivolumab, which is an anti-PD-1 drug, or Opdivo, requiring clone 28-8. And here, we look for a positive tumor cell percentage of 1% or above.

    The companion test is for atezolizumab, or Tecentriq, which is an anti-PD-L1 drug requiring clone SP142. And here, the clinician needs to look for at least 50% expression on tumor cells or 10% expression on immune cells.

    Here are some examples. This is a lung adenocarcinoma, and you can appreciate the tumor here, here, here, as well as here. On high power, you see the large tumor cells with the round nuclei and prominent nucleoli-forming glands typical of adenocarcinoma. PD-L1 expression on low power, using clone 22C3, shows clear expression over here, here, as well as here. And on high power, you can see that at least part of the cell membrane is expressing PD-L1 in all of the tumor cells. So the TPS for this case is 100%.

    Here is a case of lung adenocarcinoma metastatic to the axilla. On low power, you appreciate lymphoid tissues where we are likely in a lymph node. But tumor cells are actually only seen here as well as here, and you can appreciate on high power the tumor cells. If we look on low power for PD-L1, again using clone 22C3, we actually see PD-L1 expression elsewhere, which is not necessarily tumor, like over here. If we look at higher power, clearly these tumor cells are actually negative. They do not show membranous staining, but the staining is in accompanying immune cells. And also over here, the tumor cells are negative. They do not show membranous staining for PD-L1. Only the immune cells stain for PD-L1. So, in this case, the TPS is 0.

    Immune cells with PD-L1 expression

    And here is just an example of how PD-L1 expression on immune cells will look like.

    Macrophages express PD-L1

    We have to be careful, as macrophages also express PD-L1, and often macrophages actually express PD-L1 quite strongly. Therefore, we always have to look at our PD-L1 stain with the accompanying H&E. H&E on this case, you can appreciate the neoplastic glands. However, inside the neoplastic glands, there are clusters of macrophages. On low power, using PD-L1 stain 22C3, there is some expression of PD-L1. But it seems to be in the macrophages, which is also highlighted on high power. These macrophages are positive for PD-L1. If you look at the tumor cells, they are negative for PD-L1. So in this case, the TPS is 0.

    PD-L1 (SP263) control: Tonsil (placenta also possible)

    We use, as a positive control, tonsil, as it stains quite a few of the immune cells. However, placenta is also a good positive control for PD-L1.

    PD-L1 clones and expression on tumor cells4

    As I said, there are several clones for PD-L1 available for testing, and they have differences in sensitivity. Literature suggests that clone 22C3, 28-8, and SP263 are highly comparable in their sensitivities. However, clone SP142 has a lower sensitivity. And the more recently developed clone 73-10 has a higher sensitivity than these other three clones.

    Here is an example. On the left side, you see PD-L1 staining using clone SP263. While on the right side, you can appreciate the same tumor staining with SP142. On the left side, the TPS was 100%, as you can see on high power here. Pretty much all of the tumor cells have nice membranous staining.  On the right side, the TPS for SP142 was much lower, as only a few of the tumor cells show membranous staining.

    Heterogeneity of PD-L1 expression5

    There is also heterogeneity of PD-L1 expression. This heterogeneity can be seen within a single tumor where PD-L1 expression might be focal or patchy. It also can be seen between independent primary NSCLC. And actually, there is only an agreement of about 50% between independent primary tumors. There is a higher level of agreement between intrapulmonary metastases of almost 90%. However, overall, the data suggests that sampling might be an issue, specifically if you think about the focal or patchy expression within a single tumor. 

    Heterogeneous expression of PD-L1

    Here is an example where you can appreciate on H&E, the tumor pretty much involves the entire slide.


    If we look at PD-L1 expression, this is quite patchy. There is PD-L1 expression up here, or maybe here in these areas, but there are large areas of the tumor which are actually negative for PD-L1. And this really will make a difference if there is just a small biopsy, which might hit this area, and then the PD-L1 might be negative in that small biopsy.

    PD-L1 testing in NSCLC

    PD-L1 testing in NSCLC currently is FDA-approved for formalin-fixed paraffin embedded tissue (FFPE). However, we can run the PD-L1 testing on cell blocks, as well as decalcified tissue. However, if these cases are negative, then we would recommend testing another sample, ideally paraffin embedded tissue. Furthermore, it has been shown, and it is also our experience, that old FFPE tissue blocks show reduced to no staining. So even FFPE tissue blocks that are only three years old might show reduced or no staining. So overall, it is recommended to test the latest specimen, and ideally a recurrence or metastasis. However, if these results are negative or only just small biopsies, testing the original tumor in addition might be helpful.

    PD-L1 evaluation: Our approach

    Currently, we are offering the clones 22C3, SP263, and SP142 in our lab, and we do currently validate the clone 28-8. We require our oncologists to actually request PD-L1 staining and to request a certain clone based on their intended treatment. And then for NSCLC, we report the TPS and immune cells.

    Testing guidelines (NCCN)

    The NCCN testing guidelines currently call for PD-L1 testing in patients with advanced or metastatic NSCLC. And that includes adenocarcinomas, squamous cell carcinomas, large cell carcinomas, and non-small cell lung carcinomas not otherwise specified.  There is evidence that anti-PD-1/PD-L1 drugs will lead to a response of extensive stage or relapsed small cell lung carcinomas. However, this response appears to be regardless of PD-L1 expression, and therefore PD-L1 testing in small cell lung carcinoma is currently not necessary.

    Take-home message

    With that, I would like to summarize. As I talked, there are multiple clones for PD-L1 available for testing. Some of these clones are included in complementary or companion tests. Staining might be heterogeneous. Therefore, there might be sampling bias. And really, the oncologist needs to be involved in the test planning, including which clone should be tested and which specimen should be sent for testing.

    Thank you

    Thank you very much.



    1. Bang YJ, Im SA, Lee KW, et al: Randomized, double-bling phase II trial with prospective classification by ATM protein level to evaluate the efficacy and tolerability of olaparib plus paclitaxel in patients with recurrent or metastatic gastric cancer. J Clin Oncol 2015;33:2858-65.
    2. Kerr K, Nicolson M: Non-small cell lung cancer, PD-L1, and the pathologist. Arch Pathol Lab Med 2016;140:249-54.
    3. Ilie M, Hofman V, Dietel M, et al: Assessment of the PD-L1 status by immunohistochemistry: challenges and perspectives for therapeutic strategies in lung cancer patients. Virch Arch 2016;468:511-25.
    4. Tsao M, Kerr K, Kockx M, et al: PD-L1 immunohistochemistry comparability study in real-life clinical samples: results of blueprint phase 2 project. J Thorac Oncol. 2018;13:1302-11.
    5. Mansfield A, Murphy S, Peikert T, et al: Heterogeneity of programmed cell death ligand 1 expression in multifocal lung cancer. Clin Canc Res 2015;22(9):2177-81.

    MCL Education

    MCL Education

    This post was developed by our Education and Technical Publications Team.