Contribution of tumour and immune cells to PD-L1 expression as a predictive biomarker in metastatic triple-negative breast cancer: exploratory analysis from KEYNOTE-119.

The efficacy of pembrolizumab monotherapy versus chemotherapy increased with increasing programmed death ligand 1 (PD-L1) expression, as quantified by combined positive score (CPS; PD-L1 expression on both tumour cells and immune cells) in patients with previously treated metastatic triple-negative breast cancer (mTNBC) in the phase 3 KEYNOTE-119 study. This exploratory analysis was conducted to determine whether the expression of PD-L1 on tumour cells contributes to the predictive value of PD-L1 CPS in mTNBC. PD-L1 expression in tumour samples was assessed using PD-L1 IHC 22C3 pharmDx and quantified using both CPS and tumour proportion score (TPS; PD-L1 expression on tumour cells alone). Calculated immune cell density (CID) was defined as CPS minus TPS. The ability of each scoring method (CPS, TPS, and CID) to predict clinical outcomes with pembrolizumab was evaluated. With pembrolizumab, the area under the receiver operating characteristic curve was 0.69 (95% CI = 0.58-0.80) for CPS, 0.55 (95% CI = 0.46-0.64) for TPS, and 0.67 (95% CI = 0.56-0.77) for CID. After correction for cutoff prevalence, CPS performed as well as, if not better than, CID with respect to predicting objective response rate, progression-free survival, and overall survival. Data from this exploratory analysis suggest that, although PD-L1 expression on immune cells alone is predictive of response to programmed death 1 blockade in mTNBC, adding tumour PD-L1 expression assessment (i.e. CPS, which combines immune cell and tumour cell PD-L1 expression) may improve prediction. PD-L1 CPS thus remains an effective and broadly applicable uniform scoring system for enriching response to programmed death 1 blockade with pembrolizumab in mTNBC as well as other tumour types.

Equivalence of laboratory-developed test and PD-L1 IHC 22C3 pharmDx across all combined positive score indications.

We conducted an analysis across multiple PD-L1 combined positive score (CPS) indications to establish concordance of a 22C3 antibody-based laboratory-developed test (LDT) on the Ventana BenchMark XT or BenchMark ULTRA platform and the regulatory-approved PD-L1 IHC 22C3 pharmDx in cervical cancer (CC), esophageal squamous cell carcinoma (ESCC), head and neck squamous cell carcinoma (HNSCC), triple-negative breast cancer (TNBC), and urothelial carcinoma (UC). Tumor specimens from each tumor type were stained with 22C3 antibody and scored using the 22C3 antibody-based LDT, and scores were compared with those using PD-L1 IHC 22C3 pharmDx. PD-L1 status was measured by the pathologist using CPS as a continuous score and using clinically relevant cutoffs (CC, ≥1 and ≥10; HNSCC, ≥1 and ≥20; ESCC, TNBC, and UC, ≥10). The agreement between the BenchMark platforms and PD-L1 IHC 22C3 pharmDx was assessed by intraclass correlation coefficient (ICC) and a contingency table for clinical interpretation. A total of 522 samples were evaluated for the pan-tumor analysis (CC, n = 77; ESCC, n = 80; HNSCC, n = 126; TNBC, n = 118, UC, n = 121). Most clinical interpretations of PD-L1 status were concordant between the BenchMark XT and PD-L1 IHC 22C3 pharmDx for all five tumor types with regard to negative percentage agreement (NPA; 83-97%), positive percentage agreement (PPA; 86-100%), and overall percentage agreement (OPA; 90-97%); the ICC by tumor type was high (≥0.88). Importantly, the pan-tumor ICC was 0.95 (95% CI 0.94-0.96). Thirty additional TNBC samples were evaluated using the BenchMark ULTRA and PD-L1 IHC 22C3 pharmDx; the NPA, PPA, and OPA were 100%. The 22C3 antibody-based LDT on Ventana BenchMark XT and BenchMark ULTRA platforms demonstrated high concordance with the regulatory-approved PD-L1 IHC 22C3 pharmDx across multiple tumor types. These findings suggest the comparability of PD-L1 IHC 22C3 pharmDx with an LDT based on the 22C3 antibody.

Association Between Biomarkers and Clinical Outcomes of Pembrolizumab Monotherapy in Patients With Metastatic Triple-Negative Breast Cancer: KEYNOTE-086 Exploratory Analysis.

PURPOSE: In the two-cohort phase II KEYNOTE-086 study (ClinicalTrials.gov identifier: NCT02447003), first-line and second-line or later pembrolizumab monotherapy demonstrated antitumor activity in metastatic triple-negative breast cancer (mTNBC; N = 254). This exploratory analysis evaluates the association between prespecified molecular biomarkers and clinical outcomes. METHODS: Cohort A enrolled patients with disease progression after one or more systemic therapies for metastatic disease irrespective of PD-L1 status; Cohort B enrolled patients with previously untreated PD-L1-positive (combined positive score [CPS] ≥ 1) metastatic disease. The association between the following biomarkers as continuous variables and clinical outcomes (objective response rate [ORR], progression-free survival [PFS], and overall survival [OS]) was evaluated: PD-L1 CPS (immunohistochemistry), cluster of differentiation 8 (CD8; immunohistochemistry), stromal tumor-infiltrating lymphocyte (sTIL; hematoxylin and eosin staining), tumor mutational burden (TMB; whole-exome sequencing [WES]), homologous recombination deficiency-loss of heterozygosity, mutational signature 3 (WES), mutational signature 2 (apolipoprotein B mRNA editing catalytic polypeptide-like; WES), T-cell-inflamed gene expression profile (TcellinfGEP; RNA sequencing), and 10 non-TcellinfGEP signatures (RNA sequencing); Wald test P values were calculated, and significance was prespecified at α = 0.05. RESULTS: In the combined cohorts (A and B), PD-L1 (P = .040), CD8 (P < .001), sTILs (P = .012), TMB (P = .007), and TcellinfGEP (P = .011) were significantly associated with ORR; CD8 (P < .001), TMB (P = .034), Signature 3 (P = .009), and TcellinfGEP (P = .002) with PFS; and CD8 (P < .001), sTILs (P = .004), TMB (P = .025), and TcellinfGEP (P = .001) with OS. None of the non-TcellinfGEP signatures were associated with outcomes of pembrolizumab after adjusting for the TcellinfGEP. CONCLUSION: In this exploratory biomarker analysis from KEYNOTE-086, baseline tumor PD-L1, CD8, sTILs, TMB, and TcellinfGEP were associated with improved clinical outcomes of pembrolizumab and may help identify patients with mTNBC who are most likely to respond to pembrolizumab monotherapy.

Biomarkers predictive of response to pembrolizumab in head and neck cancer.

BACKGROUND: We performed an integrated biomarker evaluation in pembrolizumab-treated patients with R/M HNSCC enrolled in KEYNOTE-012 or KEYNOTE-055. The relationship between biomarkers and HPV status was explored. METHODS: We evaluated PD-L1 (combined positive score [CPS]), TMB, T-cell-inflamed gene expression profile (Tcellinf GEP), and HPV status. Associations between biomarkers were evaluated by logistic regression (ORR) and Cox regression (PFS, OS). RESULTS: Two hundred and fifty-seven patients (KEYNOTE-012, n = 106; KEYNOTE-055, n = 151) had TMB data available; of these, 254 had PD-L1 and 236 had Tcellinf GEP. TMB, PD-L1, and Tcellinf GEP were each significantly associated with ORR (p < 0.01). Kaplan-Meier curves at prespecified cutoffs generally showed PFS and OS separation in the anticipated direction for these biomarkers, except for OS and TMB. TMB did not correlate with PD-L1 or Tcellinf GEP (Spearman ρ = -0.03 and ρ = -0.13, respectively); PD-L1 and Tcellinf GEP were moderately correlated (Spearman ρ = 0.47). In multivariate models, TMB, PD-L1, and Tcellinf GEP were each independently predictive for ORR (p < 0.001). ORR was higher in patients with high versus low levels of biomarkers when dichotomized using prespecified cutoffs; patients with higher versus lower levels of TMB and PD-L1 or TMB and Tcellinf GEP had the highest ORRs. Within HPV subgroups, higher versus lower distributions of biomarkers (PD-L1, TMB, and Tcellinf GEP) were associated with response. HPV detection by p16-immunohistochemistry and WES showed good concordance (81%); results were generally similar by HPV status, regardless of the detection method. CONCLUSIONS: TMB and the inflammatory biomarkers PD-L1 and Tcellinf GEP, assessed alone or together, may be useful for characterizing clinical response to pembrolizumab in R/M HNSCC.

Indirect Clinical Validation of a Programmed Death-Ligand 1 Laboratory-Developed Test for Gastric/Gastroesophageal Junction Adenocarcinoma with 22C3 Antibody Concentrate.

BACKGROUND: The PD-L1 IHC 22C3 pharmDx used on the Dako Autostainer Link 48 (ASL48) staining platform is an established method for assessing programmed death-ligand 1 (PD-L1) expression in tumor tissue and determining patient eligibility for pembrolizumab treatment; however, the availability of this platform is limited in Europe and Asia. OBJECTIVES: The aims of this study were to develop and optimize protocols for the PD-L1 22C3 antibody concentrate with multiple immunohistochemistry staining platforms and to validate these protocols using PD-L1 combined positive score (CPS) with a cut-off of ≥ 1 in gastric or gastroesophageal junction adenocarcinoma. DESIGN: The 22C3 antibody concentrate was tested and optimized protocols were developed for use with three staining platforms: Dako ASL48, Ventana BenchMark ULTRA, and Leica BOND-MAX. Tumor specimens (N = 120) from patients with gastric or gastroesophageal junction adenocarcinoma were used for the validation study; these specimens were evaluated independently by three pathologists for PD-L1 CPS as a continuous variable and using a cut-off of ≥ 1. PD-L1 IHC 22C3 pharmDx used on the Dako ASL48 platform served as the reference or gold standard. RESULTS: The intraclass correlation coefficient of CPS as a continuous variable between the gold standard and each staining platform assessed was 0.910-0.989. When CPS was dichotomized based on a cut-off of ≥ 1, depending on the pathologist and the platform used, positive percentage agreement was 81-99% and negative percentage agreement was 90-100%. Interobserver agreement using the gold standard showed substantial agreement (κ = 0.779). CONCLUSION: The PD-L1 22C3 antibody concentrate can potentially be used with the laboratory-developed test on three commercially available immunohistochemistry staining platforms to determine PD-L1 expression in tumor samples from patients with gastric or gastroesophageal junction adenocarcinoma.

CAPITAL: Optimal subgroup identification via constrained policy tree search.

Personalized medicine, a paradigm of medicine tailored to a patient's characteristics, is an increasingly attractive field in health care. An important goal of personalized medicine is to identify a subgroup of patients, based on baseline covariates, that benefits more from the targeted treatment than other comparative treatments. Most of the current subgroup identification methods only focus on obtaining a subgroup with an enhanced treatment effect without paying attention to subgroup size. Yet, a clinically meaningful subgroup learning approach should identify the maximum number of patients who can benefit from the better treatment. In this article, we present an optimal subgroup selection rule (SSR) that maximizes the number of selected patients, and in the meantime, achieves the pre-specified clinically meaningful mean outcome, such as the average treatment effect. We derive two equivalent theoretical forms of the optimal SSR based on the contrast function that describes the treatment-covariates interaction in the outcome. We further propose a constrained policy tree search algorithm (CAPITAL) to find the optimal SSR within the interpretable decision tree class. The proposed method is flexible to handle multiple constraints that penalize the inclusion of patients with negative treatment effects, and to address time to event data using the restricted mean survival time as the clinically interesting mean outcome. Extensive simulations, comparison studies, and real data applications are conducted to demonstrate the validity and utility of our method.

Influence of tumor mutational burden, inflammatory gene expression profile, and PD-L1 expression on response to pembrolizumab in head and neck squamous cell carcinoma.

BACKGROUND: To characterize genomic determinants of response to pembrolizumab in recurrent/metastatic (R/M) head and neck squamous cell carcinoma (HNSCC) in the KEYNOTE-012 study. METHODS: Associations between biomarkers (tumor mutational burden (TMB), neoantigen load (NL), 18-gene T-cell-inflamed gene expression profile (TcellinfGEP), and PD-L1 combined positive score (CPS)) and clinical outcomes with pembrolizumab were assessed in patients with R/M HNSCC (n=192). Tumor human papillomavirus (HPV) status was also evaluated with the use of p16 immunohistochemistry and whole exome sequencing (WES; HPV+, mapping >20 HPV reads) in pretreatment tumor samples (n=106). RESULTS: TMB, clonality-weighted TMB, and TcellinfGEP were significantly associated with objective response (p=0.0276, p=0.0201, and p=0.006, respectively), and a positive trend was observed between NL and PD-L1 CPS and clinical response (p=0.0550 and p=0.0682, respectively). No correlation was observed between TMB and TcellinfGEP (Spearman ρ=-0.026) or TMB and PD-L1 (Spearman ρ=0.009); a correlation was observed between TcellinfGEP and PD-L1 (Spearman ρ=0.511). HPV status by WES and p16 immunohistochemistry showed concordance (84% Ò¡=0.573) among patients whose HPV results were available using both methods. CONCLUSIONS: TMB and inflammatory biomarkers (TcellinfGEP and PD-L1) may represent distinct and complementary biomarkers predicting response to anti-programmed death 1 therapies in HNSCC; further study of these relationships in randomized clinical trials is needed. TRIAL REGISTRATION NUMBER: NCT01848834.

Germline HLA landscape does not predict efficacy of pembrolizumab monotherapy across solid tumor types.

We evaluated the impact of class I and class II human leukocyte antigen (HLA) genotypes, heterozygosity, and diversity on the efficacy of pembrolizumab. Seventeen pembrolizumab clinical trials across eight tumor types and one basket trial in patients with advanced solid tumors were included (n > 3,500 analyzed). Germline DNA was genotyped using a custom genotyping array. HLA diversity (measured by heterozygosity and evolutionary divergence) across class I loci was not associated with improved response to pembrolizumab, either within each tumor type evaluated or across all patients. Similarly, HLA heterozygosity at each class I and class II gene was not associated with response to pembrolizumab after accounting for the number of tests conducted. No conclusive association between HLA genotype and response to pembrolizumab was identified in this dataset. Germline HLA genotype or diversity alone is not an important independent determinant of response to pembrolizumab and should not be used for clinical decision-making in patients treated with pembrolizumab.

Transcriptomic Determinants of Response to Pembrolizumab Monotherapy across Solid Tumor Types.

PURPOSE: To explore relationships between biological gene expression signatures and pembrolizumab response. EXPERIMENTAL DESIGN: RNA-sequencing data on baseline tumor tissue from 1,188 patients across seven tumor types treated with pembrolizumab monotherapy in nine clinical trials were used. A total of 11 prespecified gene expression signatures [18-gene T-cell-inflamed gene expression profile (TcellinfGEP), angiogenesis, hypoxia, glycolysis, proliferation, MYC, RAS, granulocytic myeloid-derived suppressor cell (gMDSC), monocytic myeloid-derived suppressor cell (mMDSC), stroma/epithelial-to-mesenchymal transition (EMT)/TGFß, and WNT] were evaluated for their relationship to objective response rate (per RECIST, version 1.1). Logistic regression analysis of response for consensus signatures was adjusted for tumor type, Eastern Cooperative Oncology Group performance status, and TcellinfGEP, an approach equivalent to evaluating the association between response and the residuals of consensus signatures after detrending them for their relationship with the TcellinfGEP (previously identified as a determinant of pembrolizumab response) and tumor type. Testing of the 10 prespecified non-TcellinfGEP consensus signatures for negative association [except proliferation (hypothesized positive association)] with response was adjusted for multiplicity. RESULTS: Covariance patterns of the 11 signatures (including TcellinfGEP) identified in Merck-Moffitt and The Cancer Genome Atlas datasets showed highly concordant coexpression patterns in the RNA-sequencing data from pembrolizumab trials. TcellinfGEP was positively associated with response; signatures for angiogenesis, mMDSC, and stroma/EMT/TGFß were negatively associated with response to pembrolizumab monotherapy. CONCLUSIONS: These findings suggest that features beyond IFNγ-related T-cell inflammation may be relevant to anti-programmed death 1 monotherapy response and may define other axes of tumor biology as candidates for pembrolizumab combinations. See related commentary by Cho et al., p. 1479.

Comparing programmed death ligand 1 scores for predicting pembrolizumab efficacy in head and neck cancer.

Tumor proportion score (TPS) and combined positive score ([CPS] includes immune cells), 2 methods for scoring programmed death ligand 1 (PD-L1) expression, have been used in clinical trials investigating the immune checkpoint inhibitor pembrolizumab in head and neck squamous cell carcinoma (HNSCC). These trials resulted in regulatory approval for pembrolizumab in the first- and second-line setting outside the United States. We performed a post hoc analysis of the KEYNOTE-040 study (NCT02252042) to determine whether CPS is a practical and suitable alternative scoring method to TPS. In KEYNOTE-040, patients with metastatic HNSCC received pembrolizumab or investigator choice of standard of care (SOC). The relative utility and equivalence of CPS ≥ 50 and TPS ≥ 50% for defining PD-L1 expression status in patients with HNSCC and comparability of scoring methods by tandem receiver operating characteristic (ROC) analysis were analyzed. The cutoff for each method was also evaluated. CPS ≥ 50 appeared equivalent to TPS ≥ 50% for predicting objective response rate (ORR), overall survival, and progression-free survival. ORR for pembrolizumab versus SOC was 26.2 versus 8.5% for TPS ≥ 50%, 28.1 versus 7.7% for CPS ≥ 50, 10.6 versus 11.6% for TPS < 50%, and 10.0 versus 12.0% for CPS < 50. Tandem ROC analysis showed that TPS 50% and CPS 50 maximized delta Youden index and suggested that CPS is more sensitive than TPS at lower cutoffs (i.e., CPS ≥ 1). In conclusion, CPS 50 can be used interchangeably with TPS 50% to determine PD-L1 status in patients with HNSCC. CPS may be more sensitive than TPS at lower cutoffs.

Prevalence of High Tumor Mutational Burden and Association With Survival in Patients With Less Common Solid Tumors.

Importance: Tumor mutational burden (TMB) is a potential biomarker associated with response to immune checkpoint inhibitor therapies. The prognostic value associated with TMB in the absence of immunotherapy is uncertain. Objective: To assess the prevalence of high TMB (TMB-H) and its association with overall survival (OS) among patients not treated with immunotherapy with the same 10 tumor types from the KEYNOTE-158 study. Design, Setting, and Participants: This retrospective cohort study evaluated the prognostic value of TMB-H, assessed by Foundation Medicine (FMI) and defined as at least 10 mutations/megabase (mut/Mb) in the absence of immunotherapy. Data were sourced from the deidentified Flatiron Health-FMI clinicogenomic database collected up to July 31, 2018. Eligible patients were aged 18 years or older with any of the following solid cancer types: anal, biliary, endometrial, cervical, vulvar, small cell lung, thyroid, salivary gland, mesothelioma, or neuroendocrine tumor. Patients with microsatellite instability-high tumors were excluded from primary analysis. For OS analysis, patients were excluded if immunotherapy started on the FMI report date or earlier or if patients died before January 1, 2012, and patients were censored if immunotherapy was started later than the FMI report date. Data were analyzed from November 2018 to February 2019. Main Outcomes and Measures: Overall survival was analyzed using the Kaplan-Meier method and Cox proportional hazards model, adjusting for age, sex, cancer types, practice type, and albumin level. Results: Of 2589 eligible patients, 1671 (64.5%) were women, and the mean (SD) age was 63.7 (11.7) years. Median (interquartile range) TMB was 2.6 (1.7-6.1) mut/Mb, and 332 patients (12.8%) had TMB-H (≥10 mut/Mb). Prevalence of TMB-H was highest among patients with small cell lung cancer (40.0%; 95% CI, 34.7%-45.6%) and neuroendocrine tumor (29.3%; 95% CI, 22.8%-36.6%) and lowest was among patients with mesothelioma (1.2%; 95% CI, 0.3%-4.4%) and thyroid cancer (2.7%; 95% CI, 1.2%-5.7%). Adjusted hazard ratio for OS of patients not treated with immunotherapy with TMB-H vs those without TMB-H was 0.94 (95% CI, 0.77-1.13). Comparable results were observed when including patients with high microsatellite instability tumors and calculating OS from first observed antineoplastic treatment date. Conclusions and Relevance: These findings suggest that prevalence of TMB-H varies widely depending on tumor type and TMB-H does not appear to be a factor associated with OS among patients across these cancer types treated in the absence of immunotherapy.

Impact of PD-L1 and T-cell inflamed gene expression profile on survival in advanced ovarian cancer.

OBJECTIVE: Programmed death ligand 1 (PD-L1) expression affects tumor evasion of immune surveillance. The prognostic value and relationship of PD-L1 expression to T-cell-inflamed immune signatures in ovarian cancer are unclear. The purpose of this study is to evaluate the impact of PD-L1 on overall survival and its correlation with an immune-mediated gene expression profile in patients with advanced ovarian cancer. METHODS: PD-L1 expression in tumor and immune cells was assessed by immunohistochemistry, and PD-L1-positive expression was defined as a combined positive score ≥1; a T-cell-inflamed gene expression profile containing interferon γ response genes was evaluated using extracted RNA from surgical samples. Associations between PD-L1 expression, gene expression profile status, and overall survival were analyzed using the Kaplan-Meier method, log-rank test, and multivariate Cox proportional hazards regression models. RESULTS: A total of 376 patients with advanced epithelial ovarian, primary peritoneal, or fallopian tube cancer treated by cytoreductive surgery and platinum-based therapy were included. PD-L1-positive expression was observed in 50.5% of patients and associated with more advanced stage (p=0.047), more aggressive histologic subtype (p=0.001), and platinum sensitivity defined by increasing treatment-free interval from first platinum-based chemotherapy to next systemic treatment (p=0.027). PD-L1-positive expression was associated with longer overall survival in multivariate analyses (adjusted HR 0.72, 95% CI 0.56 to 0.93). In subgroup analyses, this association was most pronounced in patients with partially platinum-sensitive disease (treatment-free interval ≥6 to <12 months). T-cell-inflamed gene expression profile status correlated with PD-L1 expression (Spearman, ρ=0.712) but was not an independent predictor of overall survival. CONCLUSION: PD-L1 expression is associated with longer overall survival among advanced ovarian cancer patients. PD-L1 expression may be an independent prognostic biomarker.

Use of the 22C3 anti-programmed death-ligand 1 antibody to determine programmed death-ligand 1 expression in cytology samples obtained from non-small cell lung cancer patients.

BACKGROUND: Pembrolizumab monotherapy is a standard-of-care treatment for the first- and second-line treatment of advanced non-small cell lung cancer with programmed death-ligand 1 (PD-L1) tumor proportion score (TPS) values ≥ 50% and ≥ 1%, respectively. PD-L1 testing with the PD-L1 immunohistochemistry (IHC) 22C3 pharmDx companion assay has been validated on tumor tissue with the Dako Autostainer Link 48 (ASL48). 22C3 anti-PD-L1 antibody-based laboratory-developed tests (LDTs) compatible with other autostainers and cytology samples are essential to support pembrolizumab treatment decisions across institutions globally. METHODS: ASL48 and BenchMark Ultra LDTs were optimized for the evaluation of cytology samples through comparisons with cell lines with known PD-L1 expression levels (strong, moderate, and negative). The PD-L1 TPS was then evaluated for 70 paired biopsy and cytology samples (bronchial washes, n = 40; pleural effusions, n = 30) with these LDTs. Biopsy and cytology LDT TPS values were also compared with a subset of biopsy samples (n = 37) evaluated with the PD-L1 IHC 22C3 pharmDx assay on the ASL48. RESULTS: Intraclass correlation coefficients of 0.884 to 0.898 were observed for biopsy samples versus cytology samples with the ASL48 and BenchMark Ultra LDTs. Concordance was high, regardless of the TPS cut point (<1% vs ≥ 1% and <50% vs ≥ 50%), sample type (pleural effusion vs bronchial wash), or tumor histology (adenocarcinoma vs squamous cell carcinoma). Concordance was high for each LDT versus the PD-L1 IHC 22C3 pharmDx assay. CONCLUSIONS: ASL48 and BenchMark Ultra 22C3 antibody concentrate-based LDTs have been validated for PD-L1 testing in cytology samples, and they will support reliable, high-quality PD-L1 testing across regions globally. Cancer Cytopathol 2018;126:264-74. © 2018 American Cancer Society.

Correction: Use of the 22C3 anti-PD-L1 antibody to determine PD-L1 expression in multiple automated immunohistochemistry platforms.

[This corrects the article DOI: 10.1371/journal.pone.0183023.].

Use of the 22C3 anti-PD-L1 antibody to determine PD-L1 expression in multiple automated immunohistochemistry platforms.

BACKGROUND: For non-small cell lung cancer (NSCLC), treatment with pembrolizumab is limited to patients with tumours expressing PD-L1 assessed by immunohistochemistry (IHC) using the PD-L1 IHC 22C3 pharmDx (Dako, Inc.) companion diagnostic test, on the Dako Autostainer Link 48 (ASL48) platform. Optimised protocols are urgently needed for use of the 22C3 antibody concentrate to test PD-L1 expression on more widely available IHC autostainers. METHODS: We evaluated PD-L1 expression using the 22C3 antibody concentrate in the three main commercially available autostainers Dako ASL48, BenchMark ULTRA (Ventana Medical Systems, Inc.), and Bond-III (Leica Biosystems) and compared the staining results with the PD-L1 IHC 22C3 pharmDx kit on the Dako ASL48 platform. Several technical conditions for laboratory-developed tests (LDTs) were evaluated in tonsil specimens and a training set of three NSCLC samples. Optimised protocols were then validated in 120 NSCLC specimens. RESULTS: Optimised protocols were obtained on both the VENTANA BenchMark ULTRA and Dako ASL48 platforms. Significant expression of PD-L1 was obtained on tissue controls with the Leica Bond-III autostainer when high concentrations of the 22C3 antibody were used. It therefore was not tested on the 120 NSCLC specimens. An almost 100% concordance rate for dichotomized tumour proportion score (TPS) results was observed between TPS ratings using the 22C3 antibody concentrate on the Dako ASL48 and VENTANA BenchMark ULTRA platforms relative to the PD-L1 IHC 22C3 pharmDx kit on the Dako ASL48 platform. Interpathologist agreement was high on both LDTs and the PD-L1 IHC 22C3 pharmDx kit on the Dako ASL48 platform. CONCLUSION: Availability of standardized protocols for determining PD-L1 expression using the 22C3 antibody concentrate on the widely available Dako ASL48 and VENTANA BenchMark ULTRA IHC platforms will expand the number of laboratories able to determine eligibility of patients with NSCLC for treatment with pembrolizumab in a reliable and concordant manner.

Genetic effects on treatment response of umeclidinium/vilanterol in chronic obstructive pulmonary disease.

BACKGROUND: Treatment with long-acting ß2-agonists (LABAs) and long-acting muscarinic antagonists (LAMAs) for chronic obstructive pulmonary disease (COPD) is standard, but response varies. We investigated genetic association with treatment response to umeclidinium (UMEC, a LAMA), vilanterol (VI, a LABA), and combination therapy. METHODS: Data from 17 clinical trials (N = 6075) in patients with COPD receiving once-daily UMEC/VI (125/25mcg or 62.5/25mcg), UMEC (125 or 62.5mcg), VI (25mcg) or placebo were used. Genetic association with change from baseline in trough forced expiratory volume in 1 s (FEV1) ∼24 h post-dosing was assessed for: (i) 3 ß2-adrenoceptor (ADRB2) gene variants; (ii) 298 single-nucleotide polymorphisms (SNPs) with prior evidence of associations; (iii) human leukocyte antigen (HLA) alleles and (iv) genome-wide association study (GWAS) SNPs. Other endpoints were (i) reversibility at screening; and at baseline: (ii) FEV1; (iii) forced vital capacity (FVC), and (iv) FEV1/FVC ratio. Using linear regression, the inverse normal transformed residuals were pooled together, first across treatment group, then across studies for each monotherapy, then combination therapy and finally for every treated patient. RESULTS: Of 6075 patients, 1849 received UMEC/VI, 1390 received UMEC, 1795 received VI, and 1041 received placebo. None of the ADRB2 variants, HLA alleles or GWAS variants tested were associated with treatment response or baseline endpoints. Four SNPs in FAM13A (rs7671167, rs2869967, rs1964516, rs1903003) were significantly associated with baseline FEV1/FVC ratio (p < 3 × 10(-5)) after adjusting for multiple testing. CONCLUSIONS: No genetic association was found with treatment response to UMEC or VI when administered as monotherapies or in combination.

Improved Sparse Multi-Class SVM and Its Application for Gene Selection in Cancer Classification.

BACKGROUND: Microarray techniques provide promising tools for cancer diagnosis using gene expression profiles. However, molecular diagnosis based on high-throughput platforms presents great challenges due to the overwhelming number of variables versus the small sample size and the complex nature of multi-type tumors. Support vector machines (SVMs) have shown superior performance in cancer classification due to their ability to handle high dimensional low sample size data. The multi-class SVM algorithm of Crammer and Singer provides a natural framework for multi-class learning. Despite its effective performance, the procedure utilizes all variables without selection. In this paper, we propose to improve the procedure by imposing shrinkage penalties in learning to enforce solution sparsity. RESULTS: The original multi-class SVM of Crammer and Singer is effective for multi-class classification but does not conduct variable selection. We improved the method by introducing soft-thresholding type penalties to incorporate variable selection into multi-class classification for high dimensional data. The new methods were applied to simulated data and two cancer gene expression data sets. The results demonstrate that the new methods can select a small number of genes for building accurate multi-class classification rules. Furthermore, the important genes selected by the methods overlap significantly, suggesting general agreement among different variable selection schemes. CONCLUSIONS: High accuracy and sparsity make the new methods attractive for cancer diagnostics with gene expression data and defining targets of therapeutic intervention. AVAILABILITY: The source MATLAB code are available from http://math.arizona.edu/~hzhang/software.html.

Genetic analysis of asthma exacerbations.

BACKGROUND: Identifying genetic markers of susceptibility to exacerbations may improve patient management, decrease morbidity, and lead to drug development. OBJECTIVES: To assess whether genetic markers associated with severe asthma exacerbations in previous reports are associated with less severe events that do not require intensive care and intubation and to identify additional markers in candidate genes and throughout the genome. METHODS: A total of 199 patients and 502 controls (individuals without an exacerbation) were identified from 4 clinical trials. We genotyped 51 markers from 17 genes previously reported to be associated with exacerbations; a whole genome scan was used to identify additional markers. Admixture analysis was conducted to characterize the presence of ancestral groups. The genetic marker effects were assessed by logistic regression for each study followed by a meta-analysis. RESULTS: Several coding variants in the IL4R gene had a genetic effect across 3 studies, including rs1805011 in IL4R (P < .0006). In addition, 3 markers in the IFNB1 gene showed evidence of association (P < .002) but only in the study with African Americans. Because these markers did not meet the prespecified multiplicity-adjusted significance level of P = .0002, we were unable to confirm previously published results for less severe events. The whole genome scan identified genes related to mast cell mediator release. The admixture analysis suggests that ancestry was best characterized by the presence of 3 ancestral groups. CONCLUSION: We were unable to confirm previously reported associations of genetic markers with asthma exacerbations. Although, in general, the patients studied had less severe asthma than patients in earlier reports, these results suggest involvement of similar pathways. TRIAL REGISTRATION: clinicaltrials.gov Identifiers: NTC00452699, NCT00452348, NTC00102765, NCT00843193.

Association of the hemochromatosis gene with pazopanib-induced transaminase elevation in renal cell carcinoma.

BACKGROUND & AIMS: Pazopanib has demonstrated clinical benefit in patients with advanced renal cell carcinoma (RCC) and is generally well tolerated. However, transaminase elevations have commonly been observed. This 2-stage study sought to identify genetic determinants of alanine transaminase (ALT) elevations in pazopanib-treated white patients with RCC. METHODS: Data from two separate clinical studies were used to examine the association of genetic polymorphisms with maximum on-treatment ALT levels. RESULTS: Of 6852 polymorphisms in 282 candidate genes examined in an exploratory dataset of 115 patients, 92 polymorphisms in 40 genes were significantly associated with ALT elevation (p<0.01). Two markers (rs2858996 and rs707889) in the HFE gene, which are not yet known to be associated with hemochromatosis, showed evidence for replication. Because of multiple comparisons, there was a 12% likelihood the replication occurred by chance. These two markers demonstrated strong linkage disequilibrium (r(2)=0.99). In the combined dataset, median (25-75th percentile) maximum ALT values were 1.2 (0.7-1.9), 1.1 (0.8-2.5), and 5.4 (1.9-7.6)×ULN for rs2858996 GG (n=148), GT (n=82), and TT (n=1 2) genotypes, respectively. All 12 TT patients had a maximum ALT>ULN, and 8 (67%) had ALT≥3×ULN. The odds ratio (95% CI) for ALT≥3×ULN for TT genotype was 39.7 (2.2-703.7) compared with other genotypes. As a predictor of ALT≥3×ULN, the TT genotype had a negative predictive value of 0.83 and positive predictive value of 0.67. No TT patients developed liver failure. CONCLUSIONS: The rs2858996/rs707889 polymorphisms in the HFE gene may be associated with reversible ALT elevation in pazo-panib-treated patients with RCC.

Genes related to apoptosis predict necrosis of the liver as a phenotype observed in rats exposed to a compendium of hepatotoxicants.

BACKGROUND: Some of the biochemical events that lead to necrosis of the liver are well-known. However, the pathogenesis of necrosis of the liver from exposure to hepatotoxicants is a complex biological response to the injury. We hypothesize that gene expression profiles can serve as a signature to predict the level of necrosis elicited by acute exposure of rats to a variety of hepatotoxicants and postulate that the expression profiles of the predictor genes in the signature can provide insight to some of the biological processes and molecular pathways that may be involved in the manifestation of necrosis of the rat liver. RESULTS: Rats were treated individually with one of seven known hepatotoxicants and were analyzed for gene expression by microarray. Liver samples were grouped by the level of necrosis exhibited in the tissue. Analysis of significantly differentially expressed genes between adjacent necrosis levels revealed that inflammation follows programmed cell death in response to the agents. Using a Random Forest classifier with feature selection, 21 informative genes were identified which achieved 90%, 80% and 60% prediction accuracies of necrosis against independent test data derived from the livers of rats exposed to acetaminophen, carbon tetrachloride, and allyl alcohol, respectively. Pathway and gene network analyses of the genes in the signature revealed several gene interactions suggestive of apoptosis as a process possibly involved in the manifestation of necrosis of the liver from exposure to the hepatotoxicants. Cytotoxic effects of TNF-alpha, as well as transcriptional regulation by JUN and TP53, and apoptosis-related genes possibly lead to necrosis. CONCLUSION: The data analysis, gene selection and prediction approaches permitted grouping of the classes of rat liver samples exhibiting necrosis to improve the accuracy of predicting the level of necrosis as a phenotypic end-point observed from the exposure. The strategy, along with pathway analysis and gene network reconstruction, led to the identification of 1) expression profiles of genes as a signature of necrosis and 2) perturbed regulatory processes that exhibited biological relevance to the manifestation of necrosis from exposure of rat livers to the compendium of hepatotoxicants.