If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author. Address for correspondence: Shona Hendry, M.B.B.S., Peter MacCallum Cancer Centre, 305 Grattan St., Melbourne, Victoria 3000, Australia.
Department of Pathology, Peter MacCallum Cancer Centre, Melbourne, AustraliaSir Peter MacCallum Department of Oncology, University of Melbourne, Parkville, Australia
Department of Cardiothoracic Surgery, St. Vincent’s Hospital Melbourne, Fitzroy, AustraliaDepartment of Surgery, University of Melbourne, Parkville, Australia
Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Sydney, AustraliaSydney Medical School, University of Sydney, Sydney, AustraliaSchool of Medicine, University of Western Sydney, Sydney, Australia
Department of Pathology, Peter MacCallum Cancer Centre, Melbourne, AustraliaSir Peter MacCallum Department of Oncology, University of Melbourne, Parkville, Australia
Four different programmed death ligand 1 immunohistochemical assays are approved or in development as companion or complementary diagnostics to different immunotherapeutic agents in lung carcinoma. We sought to determine whether these assays are technically equivalent and whether one antibody can be used on an alternate staining platform.
Methods
Serial sections of tissue microarrays constructed from 368 cases of resected lung cancer were stained for 22C3 and 28-8 on the Dako Link 48 platform (Dako, Carpinteria, Ca) and for SP142 and SP263 on the Ventana Benchmark Ultra platform (Ventana Medical Systems, Tucson, AZ) strictly as per product insert. A protocol was developed to use the 22C3 antibody on the Ventana Benchmark Ultra platform.
Results
Differences in mean tumor cell and immune cell staining were observed between the four assays (p < 0.001). Differences between 22C3 and 28-8 were not statistically significant. Concordance of tumor cell scores was good (intraclass correlation coefficient [ICC] = 0.674), particularly when SP142 was excluded as an outlier (ICC = 0.755). The highest concordance was seen between 22C3 and 28-8 (ICC = 0.812). Concordance was poor for immune cell staining (ICC = 0.212). When dichotomized according to clinically relevant cutoffs, pairwise comparisons showed poor to moderate concordance (κ = 0.196–0.578), with positive percent agreement ranging from 15.1% to 90.0%. The 22C3 antibody performed comparably on the Dako Link 48 platform and the alternate Ventana Benchmark Ultra platform (ICC = 0.921, κ = 0.897).
Conclusions
Concordance between the four programmed death ligand 1 immunohistochemical assays when performed and scored as intended show that apart from 28-8 and 22C3, they cannot be used interchangeably in clinical practice. A protocol was successfully developed to use 22C3 on an alternate platform, which may help to overcome some barriers to implementation.
Immune checkpoint blockade through the use of monoclonal antibodies directed against cytotoxic T-lymphocyte associated protein 4 (CTLA4), programmed cell death 1 (PD-1), and programmed death ligand 1 (PD-L1) has profoundly altered the therapeutic approach in many cancer types with particularly impressive responses observed in a subset of patients with NSCLC.
Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial.
However, most patients do not respond, and there is a clear need for a predictive biomarker to guide treatment decisions. Recent emphasis on biomarkers for targeted therapies has resulted in the concept of the companion diagnostic, a test that provides information essential for the safe and effective use of a corresponding drug or biological product
Higher expression of PD-L1 correlates with higher rates of response to anti–PD-1 and anti–PD-L1 therapy, particularly in nonsquamous NSCLC, but this is not absolute, with objective response rates of 8% to 11% in patients with PD-L1–negative NSCLC.
Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial.
Different PD-L1 immunohistochemical (IHC) assays have been developed and approved in parallel with different therapeutic agents, with different cutoffs determined by clinical response. On the basis of data from the KEYNOTE-001 study,
U.S. Food and Drug Administration (FDA) approval for the anti–PD-1 agent pembrolizumab was limited to patients with NSCLC expressing PD-L1 in at least 50% of TCs, as determined by the companion diagnostic IHC assay, specifically, the 22C3 antibody clone, performed on the Dako Link 48 platform (Dako, Carpinteria, CA).
FDA approval summary: pembrolizumab for the treatment of patients with metastatic non-small cell lung cancer whose tumors express programmed death-ligand 1.
More recently, FDA approval has been updated to include treatment in the first-line setting for patients with NSCLC expressing PD-L1 in at least 50% of TCs and in the second-line setting for those expressing PD-L1 in at least 1% of TCs by the 22C3 PharmDx assay (Agilent, Technologies, Santa Clara, CA).
Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial.
studies. Nevertheless, the 28-8 Dako antibody was approved as a complementary diagnostic to nivolumab for nonsquamous NSCLC, a test that is not required to select patients for treatment but may be helpful in determining the likelihood of response to therapy.
Most recently, the PD-L1 inhibitor atezolizumab has been approved for the treatment of advanced NSCLC, alongside the SP142 antibody assay performed on the Ventana Benchmark Ultra staining platform (Ventana Medical Systems, Tucson, AZ) as a complementary diagnostic.
A phase III study of durvalumab (MEDI4736) with or without tremelimumab for previously treated patients with advanced NSCLC: rationale and protocol design of the ARCTIC study.
This situation has resulted in a number of challenges. Each drug, although targeting the same pathway, is different and equivalency cannot be assumed. Different antibody clones will target different epitopes, information about which may be proprietary and not publicly available. Different detection systems and automated staining platforms will also affect comparability. Importantly, most diagnostic laboratories are unlikely to run two different automated staining platforms, and use of four different antibodies on often very limited tissue samples is impractical, is expensive, and may not be the best use of precious tumor tissue.
International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary classification of lung adenocarcinoma.
Furthermore, funding approval in Australia for anti–PD1/PD-L1 therapies has been delayed because of the issues around identifying an agnostic robust reproducible test to accompany the drug.
Recent studies have raised concerns regarding the comparability of these four IHC assays.
The phase 1 Blueprint study, a collaboration between academic organizations and industry, compared the average scores from three pathologists in 39 cases and found similar analytic performance between the four assays, with the exception of SP142.
However, whether the appropriate instrumentation was used for the platform-specific assays is unclear, and the immune infiltrate was not scored. In another recent study of 90 cases by Rimm et al., statistically significant differences in mean TC staining were observed between 22C3, 28-8, and SP142, despite high concordance.
Our study aims to provide a technical comparison of the SP142, SP263, 22C3, and 28-8 PD-L1 IHC assays in a large cohort of NSCLC samples across two clinical laboratories in both TC and IC compartments, which will add to the evidence base being developed internationally. The staining has been performed as per the intended instructions for use and on the associated correct platform in a working diagnostic laboratory, without industry involvement or sponsorship, to reflect future practice as these tests and treatments are introduced. A further aim was to assess whether a protocol could be developed for one antibody (22C3) on an alternate platform (Ventana Benchmark Ultra) and therefore overcome one of the barriers to implementation of PD-L1 testing.
Methods
Sample Cohort
Tissue microarrays (TMAs) were constructed from two patient cohorts, as previously described.
The Melbourne cohort, from St. Vincent’s Hospital Melbourne and the Peter MacCallum Cancer Centre, comprised 368 cases of resected lung malignancies of all histological types (Supplementary Table 1). Cases were typed on the full resection specimen according to WHO 2015 criteria.
Travis W.D. Brambilla E. Burke A.P. Marx A. Nicholson A.G. WHO Classification of Tumours of the Lung, Pleura, Thymus and Heart. 4th ed. International Agency for Research on Cancer,
Lyon, France2015
Two cores were taken from each case, representing different areas of the same tumor and giving a total of 736 tissue samples. Data were complete for all four IHC assays in only 355 samples on account of missing tissue cores or lack of tumor tissue in the cores. The Sydney cohort, from Royal Prince Alfred Hospital, comprised one core from each of 68 cases of resected early-stage lung adenocarcinomas and squamous cell carcinomas. This study was approved by the institutional ethics committee (PMCC 03/90). No clinical or outcome data were included in this study. Supplementary Figure 1 describes the flow of samples through the analyses undertaken.
Immunohistochemistry
Serial sections (3 μm) of the Melbourne TMAs were stained on the Dako Link 48 platform for Dako pharmDx kits for clones 22C3 (SK006) and 28-8 (SK005) and on the Ventana Benchmark Ultra platform for SP263 (790-4905) and the predilute FDA-approved SP142 assay kit (740-4859) according to the package inserts (Supplementary Table 2). Formalin-fixed, paraffin-embedded agar cell pellets prepared from the PD-L1–positive NCI-H226 cell line and human placental and tonsil tissues were used as controls.
To compare the performance of the 22C3 antibody on the different automated staining platforms, predilute 22C3 antibody (using the stock concentration provided in the Dako pharmDx PD-L1 kit) was run on the Ventana Benchmark Ultra system according to the standard Ventana protocols recommended for SP263 using the OptiView HRP Multimer detection system (see Supplementary Table 2). All optimization protocols were performed simultaneously to limit machine, reagent, and run-to-run variation. TMA sections stained according to package instructions for 22C3 on the Dako Link 48 platform were compared with sections stained using the aforementioned protocol for 22C3 on the Ventana Benchmark Ultra system (hereafter referred to as 22C3-V), performed on the Melbourne TMAs at the Peter MacCallum Cancer Centre and on the Sydney TMAs at Royal Prince Alfred Hospital.
Pathologist Scoring
Scoring was performed by one pathologist between May 2016 and March 2017 according to the package insert instructions. A TC or IC was considered positive if there was any perceptible membrane staining, partial or complete, of any intensity that was perceived as distinct from cytoplasmic staining. Positive TCs were scored as the percentage of viable TCs present in the core (TC score) if a minimum of 50 viable TCs were present. The IC score represents the percentage of the tumor area covered by positive ICs. Alveolar macrophages, which often express PD-L1, were excluded, as were areas of necrosis and associated inflammation. Intensity of staining was assessed for the 22C3 assay and the 22C3-V modified assay as weak (score 1), moderate (score 2) or strong (score 3).
To compare the analytical performance of the different assays, raw TC and IC scores were compared. To compare the clinical diagnostic performance of the assays, scores were dichotomized into positive or negative according to the clinically relevant cutoffs defined for the corresponding assay by the package inserts, FDA safety and effectiveness data sheets, and associated clinical trials (i.e., 1% and 50% of TCs for 22C3,
Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial.
Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial.
Statistical analysis was performed using IBM SPSS Statistics software, version 24 (IBM Corp, Armonk, NY). Two-way mixed intraclass correlation coefficients (ICCs) with absolute agreement definition and single observer were calculated to evaluate concordance between the percentage of TCs and ICs stained across the four assays. The Friedman test and Wilcoxon signed rank test with Boneferroni correction were used to evaluate the differences in percent cell staining. To compare the clinical performance of the assays, overall percent agreement (OPA), positive percent agreement (PPA), negative percent agreement (NPA), and Cohen’s κ (unweighted) were calculated.
Results
Rates of PD-L1 Staining Vary according to Tumor Type
The different histological tumor types included in the TMAs showed varying positivity rates (see Supplementary Table 1), with the highest TC staining seen in the small numbers of large cell, adenosquamous, and sarcomatoid carcinomas included in the cohort. Adenocarcinomas and squamous cell carcinomas comprised most of the cohort (58.3% and 24.2%, respectively) and showed comparable mean TC scores and positivity rates when stained with 22C3, SP263, and SP142. Interestingly, the mean TC score when stained with 28-8 was significantly lower in the squamous cell carcinomas than in the adenocarcinomas (5.7% versus 10.2% [p = 0.007]); however, comparable percentages were above the cutoff of 1% (27.5% versus 23.3% [p = 0.448]). No carcinoids, small cell carcinomas, or large cell neuroendocrine carcinomas showed any TC staining, although the numbers were small.
Age of Sections Significantly Affects Staining Results
Staining with SP263 and 22C3 was performed on stored sections that had been cut from the Melbourne TMAs approximately 4 years before the current study and stored at room temperature, along with matching freshly cut sections. A significant increase in staining intensity was seen on the freshly cut sections compared with that of the stored sections for both antibodies (median intensity score 1 of 3 for archived sections and 2 of 3 for fresh sections [p < 0.001]) (illustrated in Supplementary Fig. 2). At a cutoff of 50% of TCs for the 22C3 assay, 3.7% of cases on the stored sections were positive, compared with 8.2% of cases on the fresh sections (p = 0.019). Similar results were seen for the SP263 assay: 3.3% of cases on the stored sections were positive at a cutoff of 25%, whereas 13.3% of cases on the fresh sections were positive (p < 0.001). Only the freshly cut sections were used for the remaining analyses.
Intratumor Heterogeneity Can Be Demonstrated in Approximately 10% of Cases
Comparison was made between different cores from the same tumor, which were available for 190 cases stained with 22C3 and SP263, 128 cases stained with 28-8, and 197 cases stained with SP142. Agreement between the different cores was fair to moderate for all four assays: an ICC of 0.355 for 22C3, 0.59 for 128-8, 0.600 for SP142, and 0.589 for SP263. When the clinically relevant cutoffs for each assay were used, 9.0% of cases overall showed concordant positive staining and 79.0% showed concordant negative staining results between the two different cores, with discordant results in 12.0%.
Analytical Performance Comparison
Figure 1 illustrates the patterns of staining in a representative NSCLC sample from the Melbourne TMA, and Figure 2 illustrates the distribution of TC (see Fig. 2A) and IC (see Fig. 2B) scores from the four IHC assays. The patterns of staining in the human placental and tonsillar control tissue are shown in Supplementary Figure 3. As has been previously described,
The SP142 assay typically showed prominent granular staining rather than crisp linear membrane staining because of the amplification components included in the assay kit, and 28-8 displayed significant cytoplasmic staining when the prescribed protocol was followed. The SP142 assay showed consistently lower TC and IC scores than the other three assays, whereas SP263 showed consistently higher TC scores. The mean TC percentage and the proportion of cases showing any TC staining were comparable for 22C3 and 28-8; however, they varied significantly across the remaining assay pairs (Table 1, Table 2, Table 3). The IC scores with 28-8 were consistently higher than with the other three assays. The differences in IC scoring mirrored those in TC scoring for 22C3, SP142, and SP263.
Figure 1Representative tumor microarray core stained with the programmed death ligand 1 immunohistochemical assays 22C3, 28-8, SP142, and SP263.
Figure 2(A) Tumor cell scores for each programmed death ligand 1 immunohistochemical assay. (B) Immune cell scores for each programmed death ligand 1 immunohistochemical assay. Cases with no staining by all four antibodies are not shown.
Table 1Comparison of the Mean TC IC Scores and the Proportion of Cases Showing TC (≥1% and ≥50%) or IC (≥1% and ≥10%) Staining across the Four PD-L1 Immunohistochemical Assays (N = 355)
Agreement between TC scores with the four assays on a continuous scale was good, with an ICC of 0.674 (95% confidence interval [CI]: 0.624–0.720). Excluding SP142 as an outlier increased the overall ICC to 0.755 (95% CI: 0.716–0.790), and agreement between TC scores with 22C3 and 28-8 was high (ICC = 0.812 [95% CI: 0.774–0.845]). Plots of interassay TC correlation are shown in Figure 3, and pairwise agreements between the TC scores using cutoffs of 1%, 25%, and 50% are shown in Supplementary Table 3. Overall and pairwise agreement between IC scores was poor (ICC=0.212 [95% CI: 0.159–0.269]).
Figure 3Scatter plots showing pairwise comparisons of the percentage tumor cell staining between the four programmed death ligand 1 immunohistochemical assays (N = 355), including regression line with 95% confidence intervals.
When the clinical trial positivity cutoffs allied with each assay/drug pair (as defined earlier in this article) were used, agreement between positive results and the four assays was moderate (κ = 0.43). OPA appeared high; however, this was due to the high number of concordant negative samples, as PPA was poor (Table 4).
Table 4Agreement between the Pairs of PD-L1 Immunohistochemical Assays Using the Clinically Relevant Cutoffs to Define Positive and Negative for Each Assay (N = 355)
Of the 355 cases, 14 (3.7%) were classified as positive by all four assays, and 254 cases (71.5%) were classified as negative by all four assays. A total of 88 cases (24.8%) showed discordance between the classifications provided by the different assays.
Cross-Platform Harmonization
There was excellent agreement between the 22C3 staining performed on the Dako Link 48 platform and that performed on the Ventana Benchmark Ultra platform (22C3-V), with an ICC of 0.921 for TC score and κ of 0.897 when scores were dichotomized at the clinical cutoff of 50%. Eleven of the 484 Melbourne cases (2.3%) and two of the 67 Sydney cases (3.0%) were positive on the Dako platform but negative on the Ventana platform. All cases scored as positive on the Ventana platform were also scored as positive on the Dako platform.
Discussion
The results of this study performed in a working diagnostic laboratory suggest that significant differences exist between the four PD-L1 IHC assays, and as such, they are not all comparable enough to use interchangeably. The assays were performed as intended, by following the prescribed protocol and using the appropriate automated staining platforms. Overall, the level of PD-L1 expression in the Melbourne TMA cohort was lower than has been previously reported,
Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial.
likely reflecting the composition of the cohort, which included early-stage resectable tumors and a variety of histological subtypes, some of which are known to rarely express PD-L1.
as demonstrated by our findings of discordant staining results in different cores from the same tumor. Although 88% of cases overall showed concordant results between the two cores, most of these (79%) were concordant negative results. Of the 21% of cases that showed a positive result in one or both cores, only 43% had a positive result in both cores. A recent study showed similarly high discordance between small biopsy specimens and resection specimens, with PD-L1 expression consistently underestimated by the small biopsy specimen.
Comparative study of the PD-L1 status between surgically resected specimens and matched biopsies of NSCLC patients reveal major discordances: a potential issue for anti-PD-L1 therapeutic strategies.
The use of TMAs in this study may therefore also have contributed to the low overall positivity rate as compared with the rates in studies that have used whole tissue sections.
However, the use of TMA sections allowed a large number of cases to be compared and may better reflect clinical practice in which advanced NSCLC is rarely resected and diagnosis relies on small tissue biopsy specimens or cytological specimens.
In agreement with previous studies, SP142 gave consistently lower TC and IC scores than the other three assays.
We found that 28-8 gave a consistently higher IC score than the other three assays, but the TC scores were comparable to those with 22C3 and lower than those with SP263. Significant cytoplasmic staining was noted with 28-8, which may contribute to the high IC scores, as differentiating cytoplasmic and membrane staining on small ICs is challenging and any staining is likely to be called positive in this scenario. In contrast, the high cytoplasmic staining can prevent recognition of any membrane accentuation in the larger TCs, resulting in a false-negative assessment. Subjectively, the cytoplasmic staining was noted to be most prominent in squamous cell carcinomas, which may have contributed to the lower mean TC scores for squamous cell carcinomas as compared with those for adenocarcinomas when stained with 28-8. This issue has not been previously reported but occurred when the prescribed protocol and staining platform were followed exactly. IC scoring is known to be more difficult than TC scoring and shows significant interobserver variability.
SP263 was subjectively the easiest assay to interpret with sharp, high-intensity membranous staining. Despite the good ICC of 0.755 between the assays when SP142 was excluded, there were still significant differences in the TC scores between the remaining assay pairs, with the exception of 22C3 and 28-8 (see Table 2, Table 3, Table 4).
Ventana’s SP263 assay has recently gained approval in Europe for identification of those patients with NSCLC most likely to benefit from nivolumab and pembrolizumab
we found that SP263 consistently stained a higher proportion of TCs than 28-8 and 22C3 and classified significantly more cases as positive (i.e., above the cutoff of 1%) (see Tables 2 and 3). This is in contrast to the results of the phase 1 Blueprint study, in which TC scores with 22C3, 28-8, and SP263 showed high correlation.
Further clinical validation data in the context of randomized trials would help to determine whether there are clinically meaningful differences between these assays and the potential consequences for patients.
Dichotomizing the raw scores into positive and negative on the basis of clinically relevant cutoffs for each antibody as determined in the respective clinical trials
Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial.
also showed significantly discordant results. The recent FDA approval for SP142 as a complementary diagnostic to atezolizumab for NSCLC specifies a cutoff of either 50% of TCs or 10% of ICs.
FDA approval summary: pembrolizumab for the treatment of patients with metastatic non-small cell lung cancer whose tumors express programmed death-ligand 1.
Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial.
The apparently high OPAs we report are likely skewed by the high number of concordant negative cases in our cohort. The PPAs are low, with only 3.7% of cases being scored as positive by all four assays and 24.8% of cases showing discordant results. The analytical performance of 22C3 and 28-8 was comparable with high agreement and no significant differences between TC scores. However, the PPA between this pair at the clinically relevant cutoffs was only 35.9%. These two assays may therefore be technically interchangeable, but the definition of positive will depend on the clinical situation and the intended treatment.
We were able to demonstrate comparable results when using the 22C3 antibody as prescribed on the Dako Link 48 automated staining platform and with a protocol modification on the Ventana Benchmark Ultra platform. Similar results were obtained when this procedure was performed in two different working diagnostic laboratories, showing that this could be rolled out to other laboratories. The discrepancies between two antibodies using the same platform (22C3 and 28-8) were much greater than those seen between the same antibody used on different platforms (22C3 and 22C3-V). A recent study, in which the 22C3 antibody protocol was calibrated on two different Ventana automated staining platforms to produce staining comparable to that seen on the Dako Link 48 platform, reported similar findings,.
Our study has a number of limitations. First, our patient cohort included mostly early-stage, resected NSCLC, in contrast to the advanced cases in which systemic treatment would be considered. In addition, as discussed earlier in this article, the use of small cores rather than whole sections may have contributed to the low overall expression of PD-L1 in our cohort. Furthermore, the Melbourne TMAs have been a valuable resource for a number of previous studies, and many of the cores were missing. It cannot be excluded that these missing cores may have biased the results. Only one observer scored all the samples and assays. This observer has undergone Dako 22C3 PharmDx training; however, training for the other assays was not available in Australia at the time of the study. It was beyond the scope of this study to investigate interobserver concordance, and previous studies involving a subset of these samples have demonstrated interobserver variability of less than 10%.
In summary, we have demonstrated significant differences between the four PD-L1 IHC assays approved or in development as complementary or companion diagnostics. The level of concordance generally accepted for IHC assays is 90%.
Principles of analytic validation of immunohistochemical assays: guideline from the College of American Pathologists Pathology and Laboratory Quality Center.
show that with the exception of 22C3 and 28-8, the concordance between these four assays is not acceptable for them to be used interchangeably. However, with appropriate validation steps, these antibodies may be able to be used on different automated staining platforms and produce comparable results. Nevertheless, if the laboratory-developed test were to be used as an approved test (which would constitute off-label use of the antibodies, as the manufacturers intended use statements specify the particular staining platform and methodology to be used with each assay kit), the usual careful validation would be needed. Like the Blueprint study,
our study was unable to include any clinical or outcome data and therefore cannot evaluate the clinical validity of these assays. The challenges in seeking a robust and reliable biomarker for immune checkpoint inhibitor therapy include both biological and technical issues.
The technical concerns regarding the comparability of the PD-L1 IHC assays that we have outlined here may further complicate our attempts to understand the biological determinants and significance of PD-L1 expression in NSCLC.
Acknowledgments
Dr. Fox is supported by an NHMRC Practitioner Fellowship. The authors would like to thank Prudence A. Russell for provision of tissue samples and histological subtype information, as well as Roche Diagnostics Australia for provision of the SP263 and SP142 assay kits, Merck and Co. for the 22C3 assay kit, and Bristol-Myers Squibb for the 28-8 assay kit.
Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial.
FDA approval summary: pembrolizumab for the treatment of patients with metastatic non-small cell lung cancer whose tumors express programmed death-ligand 1.
Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial.
A phase III study of durvalumab (MEDI4736) with or without tremelimumab for previously treated patients with advanced NSCLC: rationale and protocol design of the ARCTIC study.
International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary classification of lung adenocarcinoma.
Travis W.D. Brambilla E. Burke A.P. Marx A. Nicholson A.G. WHO Classification of Tumours of the Lung, Pleura, Thymus and Heart. 4th ed. International Agency for Research on Cancer,
Lyon, France2015
Comparative study of the PD-L1 status between surgically resected specimens and matched biopsies of NSCLC patients reveal major discordances: a potential issue for anti-PD-L1 therapeutic strategies.
Principles of analytic validation of immunohistochemical assays: guideline from the College of American Pathologists Pathology and Laboratory Quality Center.
Disclosure: Dr. Cooper is an advisory board member and has received honoraria for lectures from AstraZeneca, Bristol-Myers Squibb, and Merck and Co. The remaining authors declare no conflict of interest.
Drugs targeting programmed death 1 (PD-1) or its ligand, programmed death ligand 1 (PD-L1), have proved successful in treating advanced-stage NSCLC and revolutionized the way patients with this disease are now treated.1–6 By interrupting the interaction of PD-1 and PD-L1, the negative regulatory effects of this immune checkpoint on specific T-cell–driven immune responses are reversed. These treatments, however, work in only around 10% to 20% of unselected patients with advanced-stage NSCLC. Although there may be other generic immune inhibitory mechanisms, such as inhibitory tumor metabolism, or inhibitory cells or soluble factors in the tumor microenvironment that are active in the tumor,7 we presume that patients deriving benefit from such therapy are among those in whom PD1–PD-L1 interaction is inhibiting an available, specific immune response against the patient’s disease.
33054-X&id=gr1.jpg){kind=link}
33054-X&id=gr2.jpg){kind=link}
33054-X&id=gr3.jpg){kind=link}