Background

The incorporation of immune-checkpoint inhibitors (ICIs) into the multimodal treatment of stage III NSCLC is likely to change future treatment standards. PD-L1 protein expression on tumor and/or immune cells has emerged as the first predictive biomarker for sensitivity to ICIs targeting the PD-1/PD-L1 axis. Little is known on the impact of treatment modalities such as chemotherapy (CT), radiotherapy (RT) and/or combinations on PD-L1.

Methods

We collected tissue samples from patients enrolled in three consecutive SAKK trials (16/96, 16/00, 16/01) and analyzed PD-L1 expression (SP263 assay) before and after neoadjuvant (neo-adj) treatment (NAT). The SAKK trials were designed for operable stage III NSCLC and conducted in a pre-immunotherapy era. All patients were treated with 3 cycles of induction chemotherapy (CT) (cisplatin/docetaxel), followed in some patients by radiotherapy (RT) (44Gy, 22 fractions) (16/00, 16/01). Thus, the cohort included a bimodal (CT/S) and a trimodal treatment approach (CT/RT/S).

Results

From 368 patients enrolled in the three trials, we obtained matched pre- and post-neo-adj samples from 100 patients, that were PD-L1 evaluable. Distribution of PD-L1 expression on tumor cells grouped into categories <1%, 1-49%, and ≥50% was 46%, 35% and 19%, respectively in the pre-neo-adj tissue samples. Overall, 58 patients were in the bimodal, 42 patients in the trimodal group. Median PD-L1 values were 1-4% for both the pre-neo-adj and post-neo-adj samples in the trimodal group. In the bimodal group, median pre-neo-adj PD-L1 value was <1% and post-neo-adj 1-4%. Statistically there was no significant difference in either of the groups (trimodal and bimodal group) for the median PD-L1 expression between the pre-neo-adj and the post-neo-adj samples. In the comparison of the treatment arms there was no statistically significant difference for the median PD-L1 value after NAT (p = 0.77).

Conclusions

In our analysis NAT did not impact on PD-L1. This was the case for both neo-adj CT as well as neo-adj CT/RT. If the observation that RT can sensitize to treatment with ICIs holds true (Shaverdian N, Lancet Oncol. 2017), it is unlikely to be mediated through an upregulation of PD-L1.

Background

Pembrolizumab + chemotherapy (pembro–chemo) have shown improved OS, PFS and ORR vs chemo alone regardless of PD-L1 tumor proportion score (TPS) in patients (pts) with advanced NSCLC. This exploratory analysis presents pooled results of pembro–chemo vs chemo in East Asian pts with advanced/metastatic PD-L1–negative (ie, TPS <1%) NSCLC.

Methods

Pts enrolled in China, Japan, Korea, Thailand, and Taiwan in KEYNOTE-021 cohort G (NCT02039674), KEYNOTE-189 (NCT02578680), KEYNOTE-189 Japan extension (NCT03950674), KEYNOTE-407 (NCT02775435) and KEYNOTE-407 China extension (NCT03875092) were included. Pts were randomized to platinum-pemetrexed ± pembro (nonsquamous) or carboplatin-paclitaxel/nab-paclitaxel ± pembro (squamous). Efficacy was assessed in the pooled ITT population, safety in all treated pts. Tumor response was assessed by BICR per RECIST v1.1. Tumor PD-L1 expression was assessed centrally using the PD-L1 IHC 22C3 pharmDx assay. Analyses are descriptive.

Results

107 Asian pts with PD-L1 TPS <1% were included (pembro–chemo, n = 56; chemo, n = 51). Median time from randomization to data cutoff was 33.4 (range, 25.3–49.2) mo. OS, PFS, PFS2, and ORR were improved for pembro–chemo vs chemo alone (Table). Grade ≥3 AEs occurred in 80% of pts receiving pembro–chemo and 82% receiving chemo. Among 18 pts who crossed over to pembro on study from the chemo group, median OS from pembro initiation was 11.7 (95% CI, 6.0–18.9) mo. 9 pts (16%) in the pembro–chemo group completed 35 cycles of pembro, all had PR (median DOR 31.1 [range, 9.1–34.5+] mo) and were alive at data cutoff.

Conclusions

Pembro–chemo provided clinically meaningful benefit vs chemo alone with manageable safety in East Asian pts with PD-L1–negative advanced/metastatic NSCLC. Results are consistent with global phase 3 studies and support continued use of pembro–chemo as standard-of-care therapy in pts with NSCLC, regardless of PD-L1 expression.

Table

Clinical trial identification

NCT02039674, NCT02578680, NCT03950674, NCT02775435, NCT03875092

Editorial acknowledgement

Medical writing assistance was provided by Kathleen Estes, PhD, of ICON plc (North Wales, PA, USA). This assistance was funded by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, USA.

Background

Our understanding of the immunopathology of early-stage NSCLC is still limited. While neoadjuvant immunotherapeutic strategies have recently shown anti-tumor effects in resectable NSCLC, their mechanisms remain inadequately understood. Here, we explore immune programs that inform of tumor immunity and response to neoadjuvant chemotherapy and chemoimmunotherapy in localized NSCLC.

Methods

Targeted immune gene sequencing using the HTG Precision Immuno-Oncology panel was performed in localized NSCLCs from three cohorts based on treatment: naïve (n=190), neoadjuvant chemotherapy (n=38) and neoadjuvant chemoimmunotherapy (n=21). Three tumor immune microenvironment (TIME) phenotypes (inflamed, cold, excluded) were derived based on CD8⁺ T cell infiltration. Signatures of immune cell abundance and immune genes were statistically compared based on tumoral PD-L1 expression, immune phenotypes, associated with pathological response, and were cross-compared across the three cohorts.

Results

PD-L1 positive tumors exhibited increased signature scores for various lymphoid and myeloid cell subsets (both, p<0.05). TIME phenotypes exhibited disparate frequencies by stage, PD-L1 expression, and mutational burden. Inflamed NSCLCs displayed overall significantly heightened levels of immune signatures with the excluded group representing an intermediate state. A signature of cytotoxic T cells was associated with favorable survival in neoadjuvant chemotherapy-treated NSCLCs (p<0.05). Major pathological response to chemoimmunotherapy was positively associated with CD8 T cells (p<0.05) and Th1 cells were significantly reduced post-chemoimmunotherapy (p<0.001). Among the three cohorts, chemoimmunotherapy-treated NSCLCs exhibited highest scores for various immune cell subsets including T effector and B cells (both, p<0.05).

Conclusions

Our findings highlight immune gene programs that may underlie host tumor immunity and response to neoadjuvant chemotherapy and chemoimmunotherapy in early-stage NSCLC.

Background

Metastatic non-small-cell lung cancers (NSCLC) with PD-L1 tumor proportion score (TPS) ≥ 50% are immune reactive neoplasms with sensitivity to first-line pembrolizumab monotherapy. However, reliable biomarkers to identify patients who benefit from this line of treatment are still lacking.

Methods

We retrospectively analyzed baseline blood samples of newly diagnosed stage IV NSCLC patients with PD-L1 TPS ≥50% who received pembrolizumab monotherapy and experienced rapid progression (RP, progression-free survival [PFS] <3 months, n=17), or a durable benefit under treatment (DB, [PFS] >6 months, n=27). The expression for 13 genes (PD1, FOXP3, CD247, PRF1, GZMB, TBX21, MKI67, HLA-DRA, FAS, FASLG, GATA3, RORγt, IFNγ) of special immunologic interest was quantified in absolute terms using real-time PCR from reverse-transcribed whole blood RNA.

Results

Patients in the DB group had a significantly higher percentage of lymphocytes (Ly%, 15.7% vs 11.8%; p= 0.004), a lower neutrophil-to-lymphocyte ratio (NLR, 5.0 vs. 7.8, p=0.011), a higher absolute lymphocyte count (ALC, 1.49 vs 1.22 x10³/µl; p =0.043) and a higher overall expression of PD1 (2,204 vs. 1,671 x10³ copies/μg blood RNA, p=0.016) in peripheral blood compared to the RP group. Adjustment for the Ly% of each sample revealed that patients in the RP had a higher relative expression of FOXP3 and TBX21 compared to these in the DB group (6,150 vs. 3,531 x10³ copies/μg RNA, p= 0.003; and 80,776 vs. 61,361 x10³ copies/μg RNA, p= 0.023, respectively).

Conclusions

Durable clinical response to pembrolizumab monotherapy in NSCLC is associated with higher absolute counts and percentages of lymphocytes, a higher expression of PD-1, and a relative downregulation of FOXP3 and TBX21 in the peripheral blood prior to therapy start. Ongoing work aims to validate these findings in larger patient cohorts, evaluate longitudinal changes under treatment, and analyze the potential clinical utility in other NSCLC patient subsets.

Background

Immune checkpoint inhibitors have revolutionised the management of NSCLC. However, the development of irAEs is associated with morbidity that can be lifelong and even fatal. While it has been suggested that human leukocyte antigen (HLA-I) homozygosity is associated with worse overall survival among NSCLC treated with single agent immunotherapy, its association with toxicity has not been examined.

Methods

We collected blood from 193 NSCLC patients treated with anti-PD1/L1 in the first- or second-line setting. Blood cells DNA was extracted and high-quality HLA typing performed. Toxicity data was collected and graded as per common terminology criteria for adverse event (CTCAE) V5.0. Univariate analysis using GraphPad Prism was used to correlate between HLA-I/II heterozygosity with toxicity. We investigated the relationship between toxicity, overall response rate (ORR), progression free survival (PFS) and overall survival (OS). In addition, we investigated the association between irAEs and different HLA-I/II supertypes and genotypes.

Results

We recorded 103 irAE among 179 patients suitable for analysis. Homozygosity at one or more HLA-I loci, but not HLA-II, was associated with reduced risk of irAE (RR=0.57, P=0.025). None of the patients with homozygosity at one or more HLA-I loci developed pneumonitis of any grade (RR=0.00, P=0.037) or grade 3 toxicity (RR=0.00, P=0.023). In addition, the HLA-A03 supertype or the HLA-DRB1*0401 allele were associated with increased risk of developing irAE, (RR=1.40, P=0.039) and (RR=1.51, P=0.023), respectively. None of the patients with HLA-DQB1*0301 experienced gastrointestinal toxicity (RR=0.00, P=0.048). The occurrence of any irAE was associated with improved ORR (RR=1.94, P<0.0001), PFS (HR=0.37, P<0.0001) and OS (HR=0.26, P<0.0001), respectively. The use of steroids to treat irAE did not diminish this association.

Conclusions

Maximal HLA-I loci heterozygosity may be a useful biomarker to predict irAEs, especially pneumonitis among NSCLC patients treated with anti-PD1/PD-L1 in the first- or second-line setting. However, the occurrence of any irAE is correlated with favourable clinical outcome.

Background

Germline DDR genes mutations are the most frequently involved mechanism for inherited cancer susceptibility and we reported that high burden of family history of cancer (FHC), namely FHC-high, was associated with prolonged OS/PFS to PD-1/PD-L1 inhibitors in a multiple-cancer cohort.

Methods

We present the outcome analysis according to FHC from two large multicenter cohorts of patients with metastatic NSCLC receiving either first-line pembrolizumab (PD-L1 expression ≥ 50%) or chemotherapy. Patients were categorized as FHC-high and FHC-low/negative, as previously reported. To explore the association between somatic DDR genes alteration and FHC, we gathered targeted DNA tumour sequencing (FoundationOne CDx assay), from a parallel cohort of patients with NSCLC identifying 24 genes of interest (MLH1-6, PMS2, ATM, ATR, CHEK1-2, BAP1, BARD1, BRCA1-2, BRIP1, PALB2, RAD51-52-51C, FANCA-C-G-L, POLD1, POLE, ERCC4, XRCC2).

Results

728 and 652 patients were included in the pembrolizumab/chemotherapy cohorts. We performed a perfect random case-control matching between the two cohorts and 607 patients from each cohort were randomly paired on the basis of the FHC, age, ECOG-PS, and burden of disease. As compared to FHC-low/negative patients, FHC high achieved longer OS (HR=0.67, 95%CI: 0.46-0.95; p = 0.028), longer PFS (HR=0.65, 95%CI: 0.48-0.89; p = 0.007) and higher DCR (86.4 vs 67.5, p = 0.009), within the pembrolizumab cohort. On the contrary, no significant associations were found between FHC and OS (p = 0.075), PFS (p = 0.001), and DCR (p = 0.120), within the chemotherapy cohort. 118 patients were included in the DDR cohort, of which 20 FHC-high (16.9%) and 98 FHC-low/negative (83.1%). The prevalence of at least one DDR genes somatic mutations was 20% and 24.5% for FHC-low/negative and FHC-high patients, respectively (p = 0.668). The median TMBs were 6 and 7.6 Mut/Mb for FHC-high and low/negative (p = 0.6018), no association with PD-L1 was reported.

Conclusions

FHC-high identifies NSCLC with improved outcomes to pembrolizumab but not chemotherapy, suggesting its role as a surrogate marker. Somatic DDR genes are not associated with FHC and further investigations with broader germline testing are warranted.

Background

Immunosenescence is a progressive remodeling of immune functions with a multifactorial etiology (aging, chronic inflammation, persistent infections, cancer). CMV has been shown to act as chronic antigenic stressor and to accelerate immune ageing by affecting peripheral blood T cell phenotypes, including loss of CD28 or overexpression of CD57. Latent viral infections were shown to be associated with chronic type I IFN signature that might promote lymphocyte senescence. We defined SIP as the proportion of CD28^–CD57⁺KLRG1⁺CD8⁺ circulating T cells. We showed that a high pretreatment SIP (>39.5%, SIP+) was associated with resistance to ICB in patients with aNSCLC. We aimed to assess the role of SIP combined to CMV status on outcomes in aNSCLC patients and the association between SIP and chronic type I IFN signature.

Methods

Baseline SIP status was assessed by flow cytometry on fresh blood samples from ICB-treated and polychemotherapy-treated (PCT) aNSCLC patients. Type I IFN score was calculated by the sum of relative expression of 5 IFN-stimulated genes (IFI44, IFIT1, IFITM1, LY6E, MX1), determined by RT-qPCR. Soluble factors associated to interferons (IFNα, IFNβ, IFNλ, IP-10, PD-L1) were quantified using the MSD assay.

Results

203 aNSCLC patients (142 ICB-treated, 61 PCT-treated) were evaluable for SIP (19.7% SIP+) and 180 patients (89%) for CMV status. CMV+ patient’s rate was significantly higher in SIP+ compared to SIP- patients (91.4% vs 50%, p<0.0001). Among CMV+ patients, SIP+ represents only 30%. In ICB-treated patients, CMV status was not sufficient to predict outcomes (median PFS 6.21 in CMV- vs 3 months in CMV+ patients, p=0.087). Among CMV+ patients, SIP+ patients had lower PFS (1.9 vs 3.3 months, p=0.006) and OS (6.2 vs 22.8 months, p=0.008) than SIP-. Otherwise, at baseline, no association between SIP and type I IFN signature nor plasmatic levels of IFN-related factors was found.

Conclusions

SIP+ patients are CMV+, SIP+ identifies patients with poorer outcomes in CMV+ ICB-treated aNSCLC. No association between type I IFN signature or IFN-related plasmatic factors and SIP was found.

Clinical trial identification

NCT02666612, NCT02105168, NCT03984318

Background

This study aimed to investigate the feasibility of using circulating tumor cells (CTCs) and peripheral blood cells (PBCs) as biomarkers for predicting the efficacy of concurrent chemoradiotherapy (CCRT) and durvalumab consolidation (DC) in patients with locally advanced non-small cell lung cancer (NSCLC).

Methods

We recruited patients diagnosed with unresectable stage III NSCLC who received definitive CCRT between March 2020 and March 2021. DC was permitted in patients who did not progress after CCRT and tumor PD-L1 ≥1%. Blood samples were collected before (C0) and after CCRT (C1) to calculate PBC counts and analyze CTCs. CTCs, isolated using CD-PRIME^TM system, exhibited EpCAM/CK+/CD45− phenotype in BioViewCCBS^TM.

Results

A total of 50 patients were enrolled and 23 patients received DC. The median progression-free survival (PFS) was not significantly different between patients with and without DC (12.7 vs. 11.7 months; p=0.532). In overall, patients with higher platelets (PLT^hi, >252ⅹ10³/uL) at C1 had worse median PFS than those with lower platelets (PLT^lo, ≤252ⅹ10³/uL) (5.9 vs. 13.5 months; p<0.001). In DC group, patients with residual CTC clusters after CCRT (C1) had worse median PFS than those without detectable CTC cluster (9.5 vs. 13.1 months; p=0.099). In multivariate analysis, PLT^hi at C1 was an independent factor for PFS (hazard ratio [HR] 10.86, 95% confidence interval [CI] 2.83-41.77; p=0.001), and patients with DC who had PLT^hi and residual CTC clusters at C1 showed the worst PFS (2.6 months, HR 33.30; p=0.001), even worse than those without DC who had PLT^hi (5.9 months, HR 13.37; p=0.004).

Conclusions

Comprehensive analysis of CTC and PBC counts before and after CCRT, especially CTC clusters and platelets at C1, demonstrated they might be biomarkers for predicting survival. This finding could aid in risk stratification of patients with unresectable stage III NSCLC who are eligible for DC after definitive CCRT.

Background

Systemic inflammatory response has independent prognostic value, across tumour types and geographical locations, in patients with advanced cancer: an elevated ratio of peripheral neutrophils-to-lymphocytes (NLR) has been recognized as a poor prognostic indicator in various cancers. This study aimed to determine whether NLR and platelets-to-lymphocytes ratio (PLR) reflect poor treatment benefits in patients suffering from metastatic Non Small Cell Lung Cancer (mNSCLC) who underwent first-line pembrolizumab monotherapy.

Methods

We retrospectively analyzed 87 mNSCLC patients with programmed cell-death ligand 1 (PD-L1) TPS ≥50%, who received pembrolizumab monotherapy in our institution between 2017 and 2020. NLR and PLR were calculated from pre-treatment complete blood counts. We evaluated univariate Kaplan Meier estimated overall survival and multivariate Cox Regression Hazard Ratio: cut-point of NLR and PLR were obtained with a Survival Analysis for Continuous Explanatory Variable

Results

Patients with NLR lower than 8.59 had a median OS of 19 months vs 2,8 months of NLR high counterpart (p-value 0,0005); patients with PLR higher and lower than 207 have a median OS of 18,8 vs 12,1 months, respectively (p-value 0,173). Multivariate analysis revealed that high NLR, bone and liver metastases significantly and independently correlated to poor overall survival.

Conclusions

High NLR is associated to significantly shorter overall survival in previously untreated patients with NSCLC with high expression of PD-L1 treated with pembrolizumab monotherapy.

Background

Sintilimab + docetaxel showed encouraging efficacy and tolerable safety profile in advanced Chinese NSCLC pts who had failed first-line therapy. This study aims to explore potential circulating biomarker(s), especially dynamics of ctDNA, predicting therapeutic response and long-term outcome.

Methods

Advanced NSCLC pts who had failed standard platinum doublet or EGFR-TKI without receiving any ICIs before would be enrolled in this single-arm phase II study. Docetaxel (75mg/m2, day 1) + Sintilimab (200mg, day 3) every 3 weeks for 4-6 cycles followed by Sintilimab maintenance were prescribed until disease progression, unacceptable toxicity, or up to 2 years. 10 mL blood would be drawn from each patient immediately after enrollment (baseline, C0D0) and after two cycles of treatment. ctDNA was detected via a 448-gene panel. bTMB was calculated as the number of mutations divided by sequence length (1.16Mb). Positivity was defined as ≥ 2 somatic variants, ≤ 1 somatic variant within a sample was considered negative. Clonality was defined as AF/AFmax ≥ 50%.

Results

Pts with positive ctDNA after 2 courses of treatment (at 6^th week) displayed inferior prognosis (mPFS, 91d vs NR, ctDNA (+) vs ctDNA (-) after 2 treatment courses respectively, P < 0.0001; mOS, 147d vs 467d, P = 0.0039). Irrespective of driver mutations (EGFR, ERBB2 in this study), ctDNA residual at 6^th week was an independent risk factor for progression or death (HR = 100 [4.10-2503.00], P = 0.005, Cox Regression). Furtherly, clearance of clonal mutations and no clonal formation at 6^th week were associated with longer PFS (mPFS 89d vs 266d, HR = 4.40 [0.84-23.16], P = 0.003); Clearance of clonal mutations was associated with longer OS (mOS 147d vs 467d, HR = 7.01 [1.30-37.69], P = 0.0048). Baseline bTMB was not associated with the efficacy of Sintilimab+ Docetaxel treatment.

Conclusions

ctDNA residual, as well as dynamics of clonality in ctDNA could predict long-term clinical benefit of sintilimab + docetaxel in pts with previously treated advanced NSCLC, irrespective of driver mutations as well as bTMB. Further validation of ctDNA residual and dynamics of clonality as robust predictive biomarkers is warranted.

Clinical trial identification

The trial was registered on chictr.org.cn with identifier ChiCTR1900027634.

Background

There is limited data on the predictive biomarkers of response to immune checkpoint blockade (ICB) treatment of non-small cell lung cancer (NSCLC) patients. The main aim of this prospective study is to understand the utility of pre-treatment soluble immune checkpoint and tumor markers as early predictors of response in locally advanced/metastatic NSCLC patients treated with ICBs.

Methods

The study was conducted at the National Center for Cancer Care and Research (NCCCR), HMC, Qatar. A total of 26 patients on anti-PD-1/PDL-1 treatment were enrolled, and pre-treatment blood samples were collected. Multiplex Magnetic Bead Panel kits were utilized to measure concentrations of soluble immune checkpoint and tumor markers including BTLA, GITR, HVEM, IDO, LAG-3, PD-1, PDL-1, PDL-2, TIM-3, CD28, CD80, 4-1BB, CD27, CTLA-4, ICOS Ligand, CD276, VISTA, B7-H6; CD47 (IAP), BLAST-1, Galectin-9, TIMD-4; OX40, S100A8/A9, E-cadherin, MICB, Nectin 2, NTSE, PVR, Singlec 7, Singlec 9,CEA, CA-19-9, CA-125 CYFRA21-1 and CA-15-3. Mann-Whitney test was used to evaluate difference in medians in responders and non-responders. Response to treatment was assessed 4 months after initiation of treatment via PET CT imaging data.

Results

Clinical response to ICB treatment was determined based on RECIST criteria and PET CT imaging data. 12/26 (46%) patients showed clinical response to the treatment while 14/26 (54%) were identified as non-responders. Interestingly, in clinically responding patients, significant upregulation of the immune inhibitory soluble programmed death-ligand 1 (sPD-L1) molecule (<0.002**) and the immune stimulatory biomarkers glucocorticoid-induced TNFR-related protein (GITR) (<0.0005***), and Herpes virus entry mediator (HVEM) (<0.006**) were recorded. However, non-responding patients showed significant upregulation of 3 important immune suppressive biomarkers; T-cell immunoglobulin and mucin domain containing 4 (TIMD-4) (p<0.0361*), Nectin 2 (p<0.0310*) and the carcinoembryonic antigen (CEA) tumor marker (p<0.0484*).

Conclusions

The study shows that soluble immune suppressive/stimulatory biomarkers can serve as plausible early biomarkers of response to ICB treatment.