Association of CD274 (PD-L1) Copy Number Changes with Immune Checkpoint Inhibitor Clinical Benefit in Non-Squamous Non-Small Cell Lung Cancer

Abstract Background We sought to characterize response to immune checkpoint inhibitor (ICI) in non-squamous non-small cell lung cancer (NSCLC) across various CD274 copy number gain and loss thresholds and identify an optimal cutoff. Materials and Methods A de-identified nationwide (US) real-world clinico-genomic database was leveraged to study 621 non-squamous NSCLC patients treated with ICI. All patients received second-line ICI monotherapy and underwent comprehensive genomic profiling as part of routine clinical care. Overall survival (OS) from start of ICI, for CD274 copy number gain and loss cohorts across varying copy number thresholds, were assessed. Results Among the 621 patients, patients with a CD274 CN greater than or equal to specimen ploidy +2 (N = 29) had a significantly higher median (m) OS when compared with the rest of the cohort (N = 592; 16.1 [8.9-37.3] vs 8.6 [7.1-10.9] months, hazard ratio (HR) = 0.6 [0.4-1.0], P-value = .05). Patients with a CD274 copy number less than specimen ploidy (N = 299) trended toward a lower mOS when compared to the rest of the cohort (N = 322; 7.5 [5.9-11.3] vs 9.6 [7.9-12.8] months, HR = 0.9 [0.7-1.1], P-value = .3). Conclusion This work shows that CD274 copy number gains at varying thresholds predict different response to ICI blockade in non-squamous NSCLC. Considering these data, prospective clinical trials should further validate these findings, specifically in the context of PD-L1 IHC test results.


Introduction
Immune checkpoint inhibitors (ICIs) have been approved for use in multiple tumor types and subsequently incorporated into the National Comprehensive Cancer Network (NCCN) guidelines, influencing real-world clinical management of patients with cancer. 1 Despite this, only an estimated 12.5% of eligible (based on PD-L1 positivity) patients are reported to respond to ICI, 2 while frequent immune-related adverse events are observed in ICI treated patients. 3,4 Hence, it is of the utmost importance to further develop both positive and negative predictive biomarkers for ICI response.
PD-L1 expression as detected by immunohistochemistry (IHC) has identified a subset of tumors more responsive to ICI 5 and is an US Food and Drug Administration (FDA) approved companion diagnostic (CDx) in multiple tumor types 6 ; however, PD-L1 IHC testing is complex and remains insufficient to consistently predict response to ICI. 3,[7][8][9] In addition, tumor mutational burden high (TMB-High defined as TMB greater than or equal to 10 mutations/Megabase [muts/ Mb]) and microsatellite instability-high (MSI-H) solid tumor patients are also eligible to receive ICI based on 2 pan solid tumor approvals. 10,11 However, the clinical outcomes of ICI treatments in these biomarker positive patients is varied. 9,12 Recently, interest has emerged in the study and development of composite biomarkers that incorporate both tumor cell intrinsic and tumor microenvironment derived predictors of ICI response. 13 Both CD274 (gene encoding PD-L1) gains and losses have been discussed in clinical studies as positive and negative predictive biomarkers for ICI in various tumor types. [14][15][16][17][18][19] Inoue et al 20 showed that CD274 amplified tumors (defined as ploidy times 2 as detected by Fluorescence in situ hybridization [FISH]) when compared with tumors with PD-L1 polysomy and PD-L1 disomy had better survival outcomes to nivolumab after progression on prior therapy, with the 1-year OS rate being 100% (N = 5), 46% (N = 27) and 57.6% (N = 162), respectively, in a cohort of 194 patients with NSCLC. Goodman et al 18 identified 9 CD274 amplified (using comprehensive genomic profiling (CGP) and at a cutoff of ploidy +4) solid-tumor patients treated with ICI and reported an ORR of 66.7% and a median progression-free survival of 15.2 months. However, different assays and CD274 copy number cutoffs were used in these different studies. Huang et al recently studied over 240 000 patient specimens across multiple tumor types 21 that underwent CGP and showed that CD274 copy number gains (defined as CD274 copy number of at least specimen ploidy +1) were more prevalent than CD274 amplifications (defined as CD274 copy number of at least specimen ploidy +4) and also correlated with increased PD-L1 expression. As previously shown 21 among 30 396 lung adenocarcinomas, we reported the prevalence of CD274 copy number gains defined as CD274 copy number of at least specimen ploidy +1, specimen ploidy +2, specimen ploidy +3, and specimen ploidy +4 as 15%, 5.1%, 1.8%, and 0.9%, respectively.
Due to the variable prevalence rates of positivity at different CD274 copy number cutoffs and given the varying responses based on different CD274 copy number cutoffs in the aforementioned clinical studies, it is imperative to find an optimal standardized copy number cutoff for CD274 that is correlated with patient response to ICI in specific tumor types. Here, we investigate the association of ICI response with CD274 copy number gains and losses at various cutoffs in a clinico-genomic cohort of 621 non-squamous patients with NSCLC.

Patients
This study used the nationwide (US-based) de-identified Flatiron Health-Foundation Medicine clinico-genomic database (CGDB). The de-identified data originated from approximately 280 cancer clinics (~800 sites of care). Retrospective longitudinal clinical data were derived from electronic health record data, comprising patient-level structured and unstructured data, curated via technology-enabled abstraction, and were linked to genomic data derived from FMI CGP tests in the CGDB by de-identified, deterministic matching. 22 Institutional Review Board approval of the study protocol was obtained prior to study conduct and included a waiver of informed consent.
This study included 621 patients satisfying the following cohort inclusion criteria: (1) chart-confirmed diagnosis of non-squamous NSCLC (data collected through December 31, 2020), (2) Had at least 2 documented clinical visits in the Flatiron Health network on or after January 1, 2011, (3) Underwent CGP testing on a pathologist-confirmed non-squamous NSCLC tumor specimen, at FMI, on or after date of chart-confirmed initial diagnosis of non-squamous NSCLC, on a sample collected no earlier than 30 days before the Flatiron Health diagnosis date. (4) Wild-type for any oncogenic EGFR and ALK genomic alteration as determined by the FoundationOne and FoundationOne CDx CGP test (5) Were treated with second-line ICI monotherapy, specifically, Atezolizumab, Durvalumab, Nivolumab, or Pembrolizumab (Patients who had already received any form of ICI in the first-line setting were excluded). Patients were observed to have received a second-line ICI monotherapy between May 2015 and November 2020.

Comprehensive Genomic Profiling
Clinical cases of non-squamous NSCLC (as diagnosed by the treating physician and confirmed on hematoxylin and eosin-stained slides) underwent CGP performed using the FoundationOne and FoundationOne CDx assays as described previously, in a Clinical Laboratory Improvement Amendments (CLIA) certified and College of American Pathologists (CAP) accredited laboratory. 23,24 All samples submitted for sequencing featured a minimum of 20% tumor cell nuclear area and yielded a minimum of 50 ng of extracted DNA. CGP was performed on hybridization-captured, adapter-ligation based libraries, to identify genomic alterations (base substitutions, small insertions/deletions, copy number alterations and rearrangements) in greater than 300 cancer-associated genes, tumor mutational burden (TMB) 25 and MSI. 26

CD274 Copy Number Calling
Copy number alterations were detected using a comparative genomic hybridization-like method applied to next generation sequencing data. 23,27 In the laboratory, each specimen was analyzed alongside a process-matched normal control (an internally validated mixture of 10 heterozygous diploid samples from the HapMap project), with custom algorithms to normalize the sequence coverage distribution across captured DNA regions. Log-ratios of normalized coverage data for exonic, intronic, and SNP targets accounting for stromal admixture, as well as genome-wide SNP frequencies, were used to generate the profiles. Using circular binary segmentation, custom algorithms further clustered groups of targets and SNP frequencies to define upper and lower bounds of genomic segments. Empirical Bayesian algorithms used a distribution of parameters including purity and base ploidy and probability matrices were derived using different statistical sampling methodologies to fit these data. Specimen-level ploidy was estimated as described by Sun et al 27 Computational models were reviewed by expert analysts for each sample. 23

PD-L1 Expression
PD-L1 IHC testing was run and interpreted by experienced board-certified pathologists according to the manufacturer instructions in a CLIA-certified and CAP-accredited laboratory (Foundation Medicine, Inc, Morrisville, NC) for a subset of specimens in this CGDB cohort. DAKO PD-L1 IHC 22C3 pharmDx's tumor proportion scoring (TPS) method was used to score the cases. 28

Outcomes and Statistical Analyses
The primary clinical endpoint was OS from start of second-line ICI monotherapy until death or loss of follow-up. To account for delayed entry into the real-world clinico-genomic cohort, risk set adjustment was performed to adjust for left truncation bias. The Kaplan-Meier method along with the log-rank test was used to estimate differences between outcome estimates. Categorical variables were compared using the 2-sided Fisher's exact test, while the 2-sided Wilcoxon rank sum test was used to compare continuous variables. All analyses were performed using the R software 29 version 4.0.3.

Patient Characteristics
Overall, 621 EGFR-and ALK-wild-type non-squamous patients with NSCLC treated with second-line ICI monotherapy that fit the predefined inclusion criteria were identified. Median (interquartile range) follow-up time was 10.9 (3.7-23.4) months and as of the CGDB data cutoff date, 73.3% had died. Among the 621 patients, majority were female (53.6%), self-reported race as White (73.2%), were stage IV at initial diagnosis (64.1%), 18.7% had an ECOG status over 2 at initiation of second-line ICI monotherapy and had a history of smoking (88.4%, Table 1). 59.1% and 33.5% of the patients had received either platinum-based chemotherapy or anti-VEGF combination therapy respectively, in the first-line setting (Table 1).

Association of CD274 Copy Number with Response to ICI Blockade
Across the overall cohort, 1.4%, 2.4%, 4.7%, and 15.0% patients had a CD274 copy number greater than or equal to specimen ploidy +4, greater than or equal to specimen ploidy +3, greater than or equal to specimen ploidy +2, greater than or equal to specimen ploidy +1, respectively, while 36.9% patients had a CD274 CN equal to specimen ploidy. Among patients with a CD274 loss, 48.1%, 11.8%, and 1.1% had a CD274 copy number lesser than or equal to specimen ploidy −1, lesser than or equal to specimen ploidy −2 and lesser than or equal to specimen ploidy −3, respectively. To examine the association of CD274 copy number (CN) to ICI blockade, we studied the OS of patients from the start of second-line ICI monotherapy, stratified by their CD274 CN relative to specimen ploidy, at various CD274 CN thresholds.
As the CD274 copy number threshold was increased from at least specimen ploidy +1 to at least specimen ploidy + 4, the 1-year OS rate amongst the patients with CD274 gains was observed to be 61.1%, 73.3%, 75%, and 66.7%, respectively. Given the significantly higher survival at a CD274 CN threshold of greater than or equal to specimen ploidy +2, we specifically examined the cohort using the ploidy +2 cutoff, and here we observed that there were no significant differences in the demographics and clinical characteristics of the gain group (CD274 CN threshold greater than or equal to specimen ploidy +2) vs the rest of the patients, (Supplementary  Table S1) but among the well-studied ICI biomarkers of PD-L1, TMB and MSI, TMB-High (at a threshold of 10 muts/ Mb), TMB-High was significantly enriched in the gain group (Table 2). Of note, although PD-L1 protein expression data were only available for a subset of cases, CD274 CN changes were overall correlated with PD-L1 protein expression (Supplementary Table S2). However, they were not entirely concordant and cases with CD274 CN gain with no PD-L1 protein expression, and CD274 CN loss with PD-L1 protein expression existed in this cohort.
At a CD274 copy number gain threshold of 2, when the OS from start of second-line ICI monotherapy was stratified by TMB-High, an additive pattern emerged. mOS of patients with CD274 CN less than ploidy +2 and TMB low (N = 330) was the lowest at 7.7 [6.3-10.9] months, mOS of patients with CD274 CN less than ploidy +2 and TMB-High (N = 262) was comparable with that of patients with CD274 CN greater than or equal to ploidy +2 and TMB low (N = 9) at 9.5 [7.1-13.2] months and 9.3 [1.3-NA] months, respectively, while mOS of patients with CD274 CN greater than or equal to ploidy +2 and TMB-High (N = 20) was the highest at 24.9 [11.1-NA] months, P = .04 (Fig. 2). As an exploratory analysis, we included the PD-L1 status where available in these different subgroups defined by CD274 CN and TMB (Supplementary Table 3), although the number of cases with available PD-L1 status is small to make any conclusions.

Discussion
While the importance of CD274 gains and losses as biomarkers of response to ICI has been increasingly emphasized, no data is available on the corresponding clinically relevant and optimal CD274 copy number thresholds. In this retrospective  Prospective clinical trials such as the phase II trial studying the efficacy of Nivolumab and Ipilimumab in patients with rare cancers (NCT02834013) are currently enrolling patients with CD274 amplifications, defined as at least 6 copies of CD274 detected through CGP. This on-going trial further emphasizes the importance to define and evaluate the clinical relevance of CD274 copy number gain thresholds used to enroll patients onto ICI-based clinical trials. In this manner, more patients can potentially be accrued and could benefit from such clinical studies. In addition, as previously described, 21 higher rates of CD274 gains have also been reported in a variety of tumors featuring squamous cell histology and hence it is important to identify disease specific clinically relevant CD274 copy number gain thresholds to predict ICI response.
The current study also identifies an additive effect of CD274 CN gain (at a threshold of at least 4 copies) and TMB on response to ICI inhibitors. The CD274 CN low and TMB low cohort had the lowest mOS at 7.7 months and the CD274 CN high and TMB high cohort had the highest mOS at 24.9 months, while the 2 mixed groups had a comparable mOS of approximately 9.5 months, right in between that of the 2 other cohorts. This parallels the independent and complimentary nature of PD-L1 IHC and TMB seen across multiple tumor types, including non-squamous NSCLC. 30 Interestingly, the gain in mOS between the TMB high and TMB low groups, was much higher in the CD274 CN high cohort (15.6 months) compared with that in the CD274 CN low group (1.8 months). We hypothesize that the tendency of immune evasion and hence response to ICI blockade is higher in the CD274 CN high group, specifically in the presence of a high neoantigen burden manifested in the TMB high cohort. Thus, further studies exploring the efficacy of chemotherapy, chemoimmunotherapy and immunotherapy across these 4 cohorts appears warranted and has the potential to add precision in the treatment of clinically advanced NSCLC patients.
This study has several limitations. Firstly, interpretability of the survival outcomes in the cohort of patients with a CD274 CN of at least ploidy +4 are limited because of the small cohort size. Second, since PD-L1 IHC data were not available for most of the cases, a head-to-head comparison on the predictive power of PD-L1 IHC vs CD274 CN gain could not be undertaken and should be considered in future studies to determine whether PD-L1 IHC or CD274 CN is a more predictive biomarker for ICI. It is important to note that in our previous study, 21 while CD274 CN gains with at least ploidy +2 was positively correlated with PD-L1 IHC in NSCLC, there was a subset of PD-L1-positive patients that were negative for CD274 CN gain and a subset of PD-L1-negative patients that were positive for CD274 gain at a threshold of at least ploidy + 2, indicating that CD274 CN positivity could be an independent predictive biomarker of ICPI response.

Conclusions
In this study, the survival outcomes with ICI monotherapy in non-squamous NSCLC varies with CD274 copy number gains defined at different cutoffs. In future validation studies, CD274 gains defined as at least 4 copies needs to be evaluated as a biomarker of ICI response in prospective large scale clinical studies.