PD-L1: A Biomarker With Baggage

David Rimm, MD, PhD

David Rimm, MD, PhD

Although PD-L1 expression is the most established yardstick of response to immune checkpoint immunotherapy, its clinical utility as a biomarker remains an enigma further complicated by the variety of assays and algorithms for measuring it, experts say.

The implications of PD-L1 expression as a predictor of response to immune checkpoint inhibitors (ICIs) have been the subject of exploration and debate ever since the early development of therapies targeting the PD-1/ PD-L1 pathway.¹

Nevertheless, several immunohistochemistry (IHC) assays have been developed to detect PD-L1 expression levels for specific drugs; each biologic has its own IHC assay specific to a distinct anti—PD-L1 antibody clone and a distinct staining platform with specific tumor cell or immune cell thresholds.^2,3

Questions about the interchangeability of available assays have emerged, most recently in analyses of expression levels in patients with triple-negative breast cancer (TNBC) considered for treatment with atezolizumab (Tecentriq), a PD-L1 inhibitor.⁴ Other experts have expressed concern about the impact that using assays that have been developed so specifically for particular therapies could have on clinical practice.

“It’s unfortunately a mess that’s been conferred to us by vendors having different assays,” David Rimm, MD, PhD, director of both pathology tissue services and translational pathology at Yale School of Medicine in New Haven, Connecticut, said in an interview with OncologyLive^®.

“If they all had the same assay, we would see both tumor and immune cell staining in every tumor type. It’s crazy—we’re just trying to measure the same protein. The same assay should work on all the different cancer cells,” he said. “This is the first time in history that we’ve been told to use this company’s assay for this tumor and a second company’s assay for a second tumor. It’s a very bad idea.”

As it stands, PD-L1 expression testing using an FDA-approved companion diagnostic is required for several label indications for 2 ICIs, pembrolizumab (Keytruda) and atezolizumab (Table 1).^5-7

Even without a label indication, however, clinical studies of ICIs routinely evaluate responses by PD-L1 status as a stratification factor. “Careful interpretation of PD-L1 expression data may unmask poor risk groups who may not benefit from [immuno- oncology] drugs,” Kevin Harrington, PhD, a professor at The Institute of Cancer Research (ICR) in London, England, said in a presentation at the European Society for Medical Oncology 2019 Annual Congress.⁸

“This is really critical—that we don’t get swept away by the idea that everyone should get a PD-1 or a PD-L1 inhibitor first, when it is approved in the first line, because there may be patients who clearly don’t benefit from those therapies,” added Harrington, who also is the joint head of the Division of Radiotherapy and Imaging at ICR.

Overall, 5 IHC assays have been developed for pivotal clinical trials involving currently approved ICIs targeting the PD-1/PD-L1 pathway: Ventana PD-L1 (SP142) for atezolizumab, Dako 73-10 for avelumab (Bavencio), Ventana PD-L1 (SP263) for durvalumab (Imfinzi), IHC 28-8 pharmDx for nivolumab (Opdivo), and IHC 22C3 pharmDx for pembrolizumab (Table 2).^5-7,9-14 For the sixth approved ICI, cemiplimab-rwlc (Libtayo), a PD-1—directed antibody indicated in cutaneous squamous cell carcinoma settings, PD-L1 expression was evaluated in a small subset of patients in the pivotal trial but did not correlate with outcomes.¹⁵

Clinical decisions about which PD-L1 testing assay to order depend upon the patient’s cancer type and stage and the ICI under consideration, according to ARUP Consult, an online laboratory test selection tool sponsored by the University of Utah Department of Pathology.¹⁶

Impact of Assorted Assays

The availability of multiple assays is particularly noticeable in non—small cell lung cancer (NSCLC). “We’re in a bit of a conundrum at the moment in terms of the number of assays and the number of different technologies that are being used to assess PD-L1 staining,” Harrington said during his presentation. “For each of these different drugs with 4 different manufacturers, 4 different antibodies, there is the potential that there may be discrepancy and discordance.”

In a review article published in The Cancer Journal, Cottrell and Taube highlighted practical issues with having various PD-L1 assays. “The approval of multiple assays associated with unique interpretation thresholds has made it difficult to directly compare results across clinical trials. It has also created many practical concerns regarding which test to use and the interchangeability of the results,” they wrote.²

The authors go on to note that the assays 22C3, 28-8, and SP263 are comparable in their ability to detect PD-L1, whereas SP142 has a reduced sensitivity.² Of note, SP142 was used in the IMpassion130 trial for patients with TNBC.⁴

“It was shown that there is a greater degree of variability in scoring of immune cell PD-L1 expression across all 4 assays,” wrote the authors, “likely due to a lack of defined criteria and a lack of training, contributing to poor reproducibility amongst pathologists.”

The authors also described how IHC assay systems can differ in their ability to detect PD-L1 expression, and based on studies showing that different PD-L1 antibodies stain equally well in NSCLC and melanoma, they hypothesize other factors are at play, such as antigen retrieval, concentration of primary antibody, and amplification.²

Table 1. ICI Indications Linked to PD-L1 Expression^5-7 (Click to Enlarge)

Rationale for Multiple Tests

In an interview, FDA officials explained the agency’s support for approving multiple assays. “Different cancer types have different levels of PD-1/PD-L1, both on the surface of the tumor cells and on immune cells that surround the tumor cells,” an FDA spokesperson said in an email. “Each pharmaceutical company that develops PD-1/PD-L1-targeted therapeutics also determines the optimal scoring [level of PD-1/PDL1] based on patient response to their specific therapeutic.

“While the different FDA-approved tests may detect the same biomarker, the performance characteristics of these tests differ, which means they may identify different patient populations, as they are based on the data generated in the course of the corresponding therapeutic trial,” the spokesperson said. “Thus, a test is uniquely paired with a specific drug, and FDA approval of this drug-device pair has been shown to be safe and effective for the stated intended use, including the specific indication.”

A Roche spokesperson explained the rationale behind the development of the PD-L1 assay that Genentech, a member of the Roche Group, has used in studies of atezolizumab. “Immunotherapy biomarkers are fundamentally different from oncogene biomarkers in that they are continuous rather than categorical (binary)—spatially and temporally variable and reliant on multiple complex interactions rather than a single, dominant determinant,” the Roche spokesperson said.

Table 2. IHC PD-L1 Assays Developed for Clinical Trials of Approved ICI Therapies^5-7,9-14 (Click to Enlarge)

Tumor Versus Immune Cells

PD-L1, a ligand of the PD-1 receptor, is a dynamic marker that interacts with PD-1 to suppress the immune system’s response to cancer, but its expression can predict a favorable response to antibodies designed to unfetter antitumor activity. The predictive value of PD-L1 expression varies depending on the tumor type; in some NSCLC indications, for example, its expression on malignant cells is characterized as a robust predictor of response, whereas its impact in other cancers is less clear.¹⁸

Lymphocytes, including T cells, B cells, and natural killer cells, express the PD-1 receptor. On the other hand, PD-L1 is expressed on more cell types in addition to lymphocytes, such as tumor cells, macrophage-lineage cells, and endothelial cells.²

Testing immune cells for PD-L1 expression differs from traditional IHC tests that examine tumor cells, Rimm noted. “As pathologists, we look at the cell type that expresses PD-L1,” he said. “PD-L1 can be expressed by actual tumor cells because they have a certain morphologic and biologic appearance. It can also be expressed by immune cells, which is a basket term for lymphocytes, macrophages, dendritic cells, myeloid-derived stem cells, and other cells that are not tumor cells but are in the microenvironment of the tumor or stroma.” Grouped together, those less common cells are called immune cells, he said.

Importantly, different tumor types have been noted to exhibit different proportions of immune cell and tumor cell stain, Rimm said. For example, breast cancer and head and neck cancer tumors stain heavily for immune cells compared with tumor cells, whereas lung cancer stains more heavily for tumor cells.

The FDA explained its approach to scoring criteria: “Based on the PD-L1 device, scoring criteria may include the assessment of PD-L1 staining in tumor cells, immune cells, or both at specific percentage levels (cutoff). FDA requests and reviews the analytical performance of the IHC test across all critical steps to determine that the test is robust and informative for the specific use. An important part of the FDA review includes the review of the inter-reader precision where the performance of multiple pathologist readers [is] evaluated.”

Interchangeability Issues

Triple-Negative Breast Cancer

The availability of multiple PD-L1 assays and the appropriate choice for ICI therapy has surfaced in TNBC. In March 2019, atezolizumab became the first ICI to gain an indication for breast cancer when the FDA approved the drug in combination with nab-paclitaxel (Abraxane) for patients with unresectable locally advanced or metastatic PD-L1—positive TNBC.

The drug was approved based on findings from the phase III IMpassion130 trial, which showed an improvement in median progression-free survival (PFS) of 7.4 months with atezolizumab/nab-paclitaxel versus 4.8 months with chemotherapy (HR, 0.60; 95% CI, 0.48-0.77; P <.0001).⁵ In the study, PD-L1 positivity was defined as PD-L1 stained tumor-infiltrating immune cells of any intensity covering ≥1% of the tumor area using the SP142 assay.⁴

In an exploratory overall survival (OS) analysis, patients with PD-L1—positive tumors, median OS was 25.0 months (95% CI, 19.6-30.7) with atezolizumab versus 18.0 months (95% CI, 13.6-20.1) with placebo. The difference was not statistically significant according to the study’s prespecified parameters, but investigators said the benefit with atezolizumab therapy was clinically meaningful (stratified HR, 0.71; 95% CI, 0.54-0.94).¹⁹

Mathematical modeling has been unable to harmonize PD-L1 expression findings from the SP263 and 22C3 PD-L1 assays with the FDA-approved SP142 companion diagnostic, according to an exploratory analysis of the IMpassion130 trial presented at the 2019 San Antonio Breast Cancer Symposium.⁴ Accordingly, “the assays cannot be considered equivalent,” leaving the SP142 assay as “the only clinically validated companion assay to select patients with metastatic TNBC for treatment with atezolizumab plus nab-paclitaxel,” lead study author Hope S. Rugo, MD, FASCO, and coinvestigators wrote in the conclusion of their poster.⁴

In earlier data, Rugo and colleagues retrospectively scored tumor samples from participants in the study, with the 3 assays using central testing read by pathologists trained and qualified to read immune cells ≥1% (SP142 and SP263) and combined positive score (CPS) on tumor plus immune cells ≥1 (22C3).²⁰

“The different assays identified different percentages of cases that are positive for PD-L1,” said Rugo, professor of medicine and director of Breast Oncology and Clinical Trials Education at the University of California, San Francisco, Helen Diller Family Comprehensive Cancer Center. “More tumors were classified as positive for PD-L1 using the 2 alternative antibodies and assays [81% with 22C3 and 75% with SP263].

“Almost all SP142-positive cases are captured by either 22C3 CPS or SP263,” she said. “However, about one-third of patients’ tumors were positive for PD-L1 using only 1 of [these] 2 assays. The overall percentage of agreement between SP142 and 22C3 was 64% and between SP142 and SP263 was 69%.”

The biomarker-evaluable population (BEP) consisted of 614 patients, representing 68% of the intention-to-treat (ITT) IMpassion130 population. In this cohort, 46% were PD-L1- positive using SP142 versus 41% in the ITT population. The improvement in PFS with the addition of atezolizumab was slightly greater in the BEP than in the ITT group.²⁰

The median PFS was 8.3 months with the atezolizumab regimen versus 4.1 months with nab-paclitaxel plus placebo in the PD-L1—positive BEP population. In the PD-L1– positive ITT population, the median PFS was 7.5 months versus 5.3 months, respectively. Rugo et al noted that the benefit for the atezolizumab regimen was driven by the SP142 subgroup.²⁰

Although Roche understands the value of having fewer tests, the company said its goal with the SP142 assay was to best predict outcomes of therapy. “An amplification detection system was incorporated to visually enhance the assessment of PD-L1 particularly on immune cells. This is currently the only commercially available PD-L1 assay which is amplified and consequently has a different staining pattern to other approved PD-L1 assays,” the company spokesperson said. “This has created challenges for pathologists seeking to harmonize these assays so that they can use 1 PD-L1 assay to make therapeutic decisions for several different PD-1 and PD-L1 inhibitors across several different indications based on different scoring algorithms.

“Although the desire for streamlining of testing is understood, the role of a companion diagnostic assay is to discriminate between responders and nonresponders for that specific therapeutic product in a specific indication, with a cutoff based on clinical outcomes,” the Roche spokesperson stated.

Lung Cancer

Leading investigators into NSCLC were among the first cancer researchers to confront the question of interchangeability in PD-L1 assays. Starting in 2015, industry and academic leaders have been working on harmonizing tests to evaluate patients with the malignancy for ICI therapy. In 2018, Tsao et al published findings from the second phase of the Blueprint (BP) Programmed Death Ligand 1 (PD-L1) Immunohistochemistry Comparability Project (BP2).²¹

In BP2, pulmonary pathologists from 15 countries scored specimens with the 5 assays used in NSCLC studies (Table 2). Staining with the 22C3, 28-8, and SP263 assays was “highly comparable,” the investigators said. When detecting PD-L1 expression on tumor cells, they calculated lower sensitivity with SP142 and higher sensitivity with 73-10. Additionally, the results showed that scoring by pathologists was strongly reliable for tumor cells with all assays (overall intraclass correlation coefficient [ICC], 0.86-0.93) but poor reliability on immune cells (overall ICC, 0.18-0.19).²¹

Similarly, in a study sponsored by the National Comprehensive Cancer Network Oncology Research Program and Bristol- Myers Squibb, Rimm and colleagues found that in 4 commercial assays for NSCLC, the ICC for tumor cell testing was high: 0.83-0.88 for each antibody. However, ICC from immune cells for each antibody was 0.17-0.23. Again, this study involved multiple raters.²²

In the interview, Roche responded to questions about reproducibility with regard to its SP142 assay.

“The analytical characteristics of the Ventana PD-L1 (SP142) Assay, including inter-reader repeatability and reproducibility, have been published together with the effectiveness of pathologist training on the assessment of Ventana PD-L1 (SP142) Assay staining on both tumor cells (TCs) and tumor-infiltrating immune cells (ICs) in non— small cell lung cancer and urothelial cancer tissues, demonstrating that the assay reliably evaluated staining on both TC and IC across multiple expression levels/clinical cut-offs,” said the Roche spokesperson.

For its part, the FDA indicated that reproducibility with approved assays is not an issue. “PD-L1 devices and the associated scoring algorithms are developed based on data from clinical trials. For certain PD-L1 devices/drugs, data show that assessment of PD-L1 staining in [immune cells] is related to clinical outcome. Moreover, the assay reproducibility has been demonstrated in the course of FDA’s review to be sufficiently robust to identify a population of people who benefit from the drug in a clinical trial.”

Other Cancer Types

Differences in test results vary in other cancers. In a meta-analysis, Torlakovic et al looked for equivalence among commercially available assays across 22 studies and 376 assay comparisons, all with various thresholds. Most of the test comparisons (337) were for NSCLC; other tests were for urothelial carcinoma, mesothelioma, and thymic carcinoma.²³

Due to the nature of the studies, investigators were limited to 2 metrics used in most studies: 1% and 50% cutoffs for tumor proportion score. Comparisons were made between the 22C3, 28-8, and Ventana SP263 assay.

SP263 had the highest sensitivity and lowest specificity of tested assays, followed by 28-8, and 22C3. With respect to interchangeability, if this term were defined as 90% sensitivity and specificity for both cutoffs, no assay would be considered interchangeable. Nevertheless, the assays’ results are comparable, investigators said.²³

Notably, the investigators concluded that if an FDA-approved companion diagnostic is not available for PD-L1 expression analysis, “it is better to develop a properly validated laboratory-developed test for the same purpose(s)” than it is to substitute another companion diagnostic.²³

In findings presented at the 2019 American Society of Clinical Oncology Annual Meeting, Labriola et al compared the 28-8 and SP142 IHC assays in 32 patients with metastatic renal cell carcinoma (mRCC) and 18 with metastatic urothelial carcinoma (mUC). Findings showed that 91% of mRCC cases reflected agreement between the 2 assays, with 94% concordance for mUC based on tissue samples.²⁴

“Although PD-L1 status does not fully predict for response to ICIs, this suggests that PD-L1 testing could be used interchangeably for the majority of cases when selecting ICI treatment in mRCC and mUC,” investigators concluded.

Complexities of PD-L1 Expression

In addition to variability in assay performance and interpretation, there are other challenges with measuring PD-L1 expression.

First, no distinction between adaptive and constitutive patterns of PD-L1 expression are made in guidelines. Adaptive PD-L1 expression is a heterogeneous interferon gamma—driven mechanism of PD-L1 expression by tumor and immune cells at the host-tumor interface, whereas constitutive tumor cell PD-L1 expression refers to expression of PD-L1 on the cell surface of tumor cells regardless of the presence of an immune infiltrate and could be mediated by multiple mechanisms stemming from genetic mutations.²

The occurrence of constitutive PD-L1 expression apparently differs based on tumor type, with melanoma—in which adaptive expression was first discovered&mdash;displaying infrequent expression and cutaneous squamous cell carcinoma exhibiting more expression. Although adaptive and constitutive expression are not mutually exclusive, it could be that in the absence of a triggering immune response, constitutive expression does not predict response to therapy with anti-PD-1/PD-L1 inhibitors, as is the case with squamous cell cancers of the head and neck cancers, a tumor type for which response to anti-PD1/PD-L1 therapy appears to be predicted by immune cells.²

Second, the expression of PD-L1 over time could change due to its heterogeneous and focal nature. On a related note, tissue samples for evaluation could be too small.

Third, other resistance mechanisms could be present in the tumor that interfere with response to treatment, such as alternate PD-1/PD-L1 ligand/receptor activity (PD-L2, B7-1/CD80) or expression of other checkpoint molecules.²

Alternative Biomarkers

Investigators are on the hunt for alternative biomarkers that would indicate the potential efficacy of anti-PD-1/PD-L1 therapy. These approaches include measuring tumor infiltrating lymphocytes (TILs), assays that analyze PD-L1 expression plus TILs, tumor mutational burden, circulating tumor and immune cells, and microbiome-rich stool.²

Pharmaceutical companies also are moving forward with studies of emerging markers. “Two-thirds of our research and development projects are being developed with companion diagnostics, and Roche Tissue Diagnostics is a leader in CDx development,” the company spokesperson said. “The development of each targeted therapeutic candidate includes the assessment of biomarkers that may be associated with response to treatment. We continue to further our understanding of cancer and explore innovative diagnostic approaches, including new technologies in digital pathology and multiplex testing innovation.”

For Rimm, improvement of biomarker testing for immune-oncology agents is a pressing research priority. “Every pathologist doing their job right wants the best thing for the patients. We don’t interact with the patients; we only interact with the oncologists on a regular basis,” he said. “Our oncologists are completely at a loss for what they should do. They just want us to tell them if it’s positive or negative. We need to come up as quickly as possible with another approach to addressing this issue.”

“Different PD-1/PD-L1 inhibitors that have included a specific PD-L1 assay have assessed varying levels of PD-L1 expression in clinical trials for correlation with clinical outcomes. Consequently, there are different scoring algorithms and cutoffs for different assays as companion diagnostic assays for different drugs in different indications.”

Ionescu et al express similar concerns in a review article in Current Oncology. “Given the likely future availability of additional PD-L1 inhibitors, PD-L1 assays will become increasingly important in identifying patients most likely to benefit from treatment. However, the PD-L1 assays for use with durvalumab, pembrolizumab, nivolumab, and atezolizumab have all been developed independently, and they use different antibody clones, immunohistochemistry staining protocols, scoring algorithms, and cutoffs for determining PD-L1 status. Questions have therefore been raised about whether the tests can be used interchangeably to inform treatment decisions for the various PD-1 and PD-L1 inhibitors.”¹⁷

References

1. Lipson EJ, Sharfman WH, Drake CG, et al. Durable cancer regression off-treatment and effective reinduction therapy with an anti-PD-1 antibody. Clin Cancer Res. 2013;19(2):462-468. doi: 10.1158/1078-0432.CCR-12-2625.
2. Cottrell TR, Taube JM. PD-L1 and emerging biomarkers in immune checkpoint blockade therapy. Cancer J. 2018;24(1):41-46. doi: 10.1097/ppo.0000000000000301.
3. Kerr KM. The PD-L1 immunohistochemistry biomarker: two steps forward, one step back? J Thorac Oncol. 2018;13(3):291-294. doi: 10.1016/j.jtho.2018.01.020.
4. Rugo HS, Loi S, Adams S, et al. Exploratory analytical harmonization of PD-L1 immunohistochemistry assays in advanced triple-negative breast cancer: a retrospective substudy of IMpassion130. Presented at: 2019 San Antonio Breast Cancer Symposium; December 10-14, 2019; San Antonio, TX. Abstract PD1-07.
5. Tecentriq [prescribing information]. South San Francisco, CA: Genentech, Inc; 2019. www.accessdata.fda.gov/drugsatfda_docs/label/2019/761034s021lbl.pdf. Accessed January 13, 2020.
6. Keytruda [prescribing information]. Whitehouse Station, NJ: Merck & Co, Inc; 2020. www.accessdata.fda.gov/drugsatfda_docs/label/2020/125514s067lbl.pdf. Accessed January 13, 2020.
7. PD-L1 IHC 22C3 pharmDx. Premarket approval (PMA) database. FDA website. www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpma/pma.cfm?id=P150013S014. Updated January 13, 2020. Accessed January 13, 2020.
8. Harrington K. The value of PD-L1 as a biomarker beyond lung cancers. Presented at: European Society for Medical Oncology 2019 Congress; September 27-October 1, 2019; Barcelona, Spain. bit.ly/389B6Wl.
9. Ventana PD-L1 (SP142) Assay. Premarket approval (PMA) database. FDA website. www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpma/pma.cfm?id=P160002S009. Updated March 8, 2019. Accessed January 22, 2020.
10. Teixidó C, Vilariño N, Reyes R, Reguart N. PD-L1 expression testing in non-small cell lung cancer. Ther Adv Med Oncol. 2018;10:1758835918763493. eCollection 2018. doi: 10.1177/1758835918763493.
11. Kaufman HL, Russell J, Hamid O, et al. Avelumab in patients with chemotherapy-refractory metastatic Merkel cell carcinoma: a multicentre, single-group, open-label, phase 2 trial. Lancet Oncol. 2016;17(10):1374-1385. doi: 10.1016/S1470-2045(16)30364-3.
12. Antonia SJ, Balmanoukian A, Brahmer J, et al. Clinical activity, tolerability, and long-term follow-up of durvalumab in patients with advanced NSCLC. J Thorac Oncol. 2019;14(10):1794-1806. doi: 10.1016/j.jtho.2019.06.010.
13. Ventana PD-L1 (SP263) Assay. Premarket approval (PMA) database. FDA website. www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpma/pma.cfm?id=P160046. Accessed January 22, 2020.
14. PD-L1 IHC 28-8 pharmDX. Premarket approval (PMA) database. FDA website. www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpma/pma.cfm?id=P150025S003. Accessed January 22,2020.
15. Mutli-discipline review. FDA website. www.accessdata.fda.gov/drugsatfda_docs/nda/2018/761097Orig1s000MultidisciplineR.pdf. Accessed January 14, 2020.
16. PD-L1 testing. Arup Consult website. arupconsult.com/content/pd-l1-testing. Updated December 2019. Accessed January 18, 2020.
17. Ionescu DN, Downes MR, Christofides A, Tsao Ms. Harmonization of PD-L1 testing in oncology: a Canadian pathology perspective. Curr Oncol. 2018;25(3):e209-e216. doi: 10.3747/co.25.4031.
18. Galluzzi L, Chan TA, Kroemer G, Wolchok JD, López-Soto. The hallmarks of successful anticancer immunotherapy. Sci Transl Med. 2018;10(459): pii:eaat7807. doi: 10.1126/scitransmed.aat7807.
19. Schmid P, Rugo HS, Adams S, et al; IMpassion130 Investigators.
20. Atezolizumab plus nab-paclitaxel as first-line treatment for unresectable, locally advanced or metastatic triple-negative breast cancer (IMpassion130): updated efficacy results from a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol. 2020;21(1):44-59. doi: 10.1016/S1470-2045(19)30689-8.
21. Rugo HS, Loi S, Adams S, et al. Performance of PD-L1 immunohistochemistry (IHC) assays in unresectable locally advanced or metastatic triple-negative breast cancer (mTNBC): post-hoc analysis of IMpassion130. Presented at European Society for Medical Oncology Congress 2019; September 27-October 1, 2019; Barcelona, Spain. Abstract LBA20. bit.ly/30vdKI9.
22. Tsao MS, Kerr KM, 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(9):1302-1311. doi: 10.1016/j.jtho.2018.05.013.
23. Rimm DL, Han G, Taube JM, et al. A prospective, multi-institutional, pathologist-based assessment of 4 immunohistochemistry assays for PD-L1 expression in non-small cell lung cancer. JAMA Oncol. 2017;3(8):1051-1058. doi: 10.1001/jamaoncol.2017.0013.
24. Torlakovic E, Lim HJ, Adam J, et al. “Interchangeability" of PD-L1 immunohistochemistry assays: a meta-analysis of diagnostic accuracy. Mod Pathol. 2020;33(1):4-17. doi: 10.1038/s41379-019-0327-4.
25. Labriola M, Zhu J, Cheris S, et al. Concordance between PD-L1 assays for metastatic renal cell carcinoma (mRCC) and metastatic urothelial carcinoma (mUC). J Clin Oncol. 2019;37(suppl 15; abstr e14259). doi: 10.1200/JCO.2019.37.15_suppl.e14259.