In recent years, the Chinese National Medical Products Administration has approved PD-1 blockades for the treatment of patients with advanced NSCLC. Therefore, a significant number of patients with advanced NSCLC have undergone PD-1 blockade monotherapy in clinical practice. The present study aimed to retrospectively include patients with advanced NSCLC who underwent PD-1 blockade monotherapy (any PD-1 blockade licensed in China) at the Department of Thoracic Surgery of the Affiliated Hospital of Hebei University (Baoding, China) from July 2018 to June 2022. The inclusion criteria were as follows: i) Diagnosis of NSCLC with pathological staging of IIIb or IV confirmed by pathological expert; ii) age ≥18 years; iii) patients with previously-treated advanced NSCLC who received PD-1 blockades monotherapy for at least one cycle in clinical practice; and iv) at least one measurable target lesion. The main exclusion criteria were: i) Patients with a history of autoimmune disease or clinical symptoms unsuitable for PD-1 blockades therapy; ii) patients diagnosed with one or more tumors or serious diseases that might compromise their living status; and iii) insufficient availability of demographic characteristics or efficacy assessment data according to the investigators' judgment. Eventually, a total of 89 patients with advanced NSCLC met the eligibility criteria and were included in the present study. The primary objective of the present study was to identify the association between PD-L1 polymorphisms and clinical outcomes of PD-1 blockades, including ORR, disease contrail rate (DCR), progression-free survival (PFS) and overall survival (OS).

The protocol and additional materials for the present study were approved (approval no. 2022-KY-11053) by the Ethics Committee of the Affiliated hospital of Hebei university (Baoding, China). Each enrolled patient provided written informed consent in accordance to the recommendations of the Declaration of Helsinki (1964).

All 89 patients with advanced NSCLC received PD-1 blockade monotherapy for a minimum of one cycle in clinical practice. The PD-1 blockades utilized in the present study were approved in mainland China and clinically available for Chinese patients. These included nivolumab (Bristol-Myers Squibb Co.), pembrolizumab (Merck KGaA), camrelizumab (Jiangsu Hengrui Pharmaceuticals Co., Ltd.), sintilimab (Innovent) and tislelizumab (BeiGene, Inc.). Camrelizumab, sintilimab, pembrolizumab and tislelizumab were administered intravenously at a dose of 200 mg on day 1, while nivolumab was given intravenously with a dose of 360 mg on day 1. A treatment cycle of 21 days was completed. (20). The administration of these five PD-1 blockades persisted until either disease progression or intolerable adverse reactions in the patients.

Since PD-1 blockades monotherapy was used in the present study, the iRECIST criteria were utilized to evaluate the therapeutic response of the patients (21). As aforementioned, all 89 patients included in the present study had at least one measurable target lesion, assessed through radiological scans such as CT or MRI at baseline and throughout PD-1 blockades treatment. ORR and DCR were calculated according to the best overall response during administration of PD-1 blockades. Specifically, ORR was defined as the proportion of patients with complete response (CR) and partial response (PR) among the 89 patients. DCR was determined as the proportion of patients with CR, PR and stable disease (SD) among the 89 patients.

Furthermore, clinical and demographic characteristics of the 89 patients were retrieved from the electronic medical record system at the Affiliated Hospital of Hebei University (Baoding). Additionally, the disease progression status of each patient was assessed, and follow-up was conducted through phone communication to gather prognostic data. The therapeutic regimens of patients who experienced progression after PD-1 blockades monotherapy were documented and their health status was primarily obtained accordingly. PFS was defined as the duration from the initiation of PD-1 blockade treatment to the date of disease progression or death, whichever occurred first. OS was defined as the initiation from the date of PD-1 blockades treatment to the date of death from any cause.

Concerning the analysis of PD-L1 gene polymorphism, the DNA specimens of each patient were primarily extracted from their respective peripheral blood or cancer tissue biopsies, obtained before the initiation of PD-1 blockade treatment, according to a previous study (22).

Additionally, single nucleotide polymorphisms in PD-L1 gene of the present study were adopted from a previous study (23), including rs2297136, rs17718883, rs822339 and rs1411262. The preliminary analysis comparing the genotype status of these polymorphisms and PFS of the 89 patients is presented in Table I. Notably, only rs2297136 exhibited clinical significance. Therefore, the present study primarily focused on the results of rs2297136.

Genotyping of rs2297136 polymorphism was carried out using PCR-RFLP methods derived from a previous study (23). The forward primer for the PCR products of rs2297136 was 5′-GGAGGAGACGTAATCCAGCA-3′, and the reverse primer was 5′-CCAGGCTCCCTGTTTGACT-3′, resulting in a PCR product of 216 bp. PCR product was disposed using PspOMI restriction enzyme. The genotyping of rs2297136 was determined based on the following criteria: AA (216 bp stripe); AG (216 bp stripe, 104 bp stripe and 112 bp stripe); GG (104 bp stripe and 112 bp stripe).

Aiming to investigate the potential association between PD-L1 polymorphism and PD-L1 gene mRNA expression, available fresh specimens of peripheral blood mononuclear cells (PBMC) were initially collected from 89 patients with advanced NSCLC. Unfortunately, 8 patients failed to obtain the qualified RNA specimens. Ultimately, a total of 81 patients were included in PD-L1 gene mRNA expression analysis. The methodology of PD-L1 mRNA expression analysis was adopted from a previous study (23). Total RNA samples were extracted with TRIzol^® (Thermo Fisher Scientific, Inc.) using RNAiso Plus reagents (Takara Biotechnology Co., Ltd.) as the RNA extraction buffer according to the manufacturer's instructions, and stored at −80°C for mRNA expression analysis. A total of 500 ng RNA extracted from the PBMC was used as the templates for reverse-transcription polymerase chain reaction to prepare the first-stand of cDNA with the PrimeScript RT reagent Kit. Relative quantitative analysis of PD-L1 gene mRNA expression was carried out with Roche LightCycler 480 (Roche Diagnostics, Ltd.) using a SYBR Premix EX Taq system. The forward primer of PD-L1 was 5′-TTCAATGTGACCAGCACACTGAG-3′, the reverse primer was 5′-TTTTCACATCCATCATTCTCCCT-3′. The amplification system was comprised of a 20 µl containing 10 µl SYBR Premix EX Taq, 0.2 µl of each primer (20 µM), 7.6 µl double distilled water (ddH2O) and 2 µl cDNA. PD-L1 mRNA expression level was calculated by comparative Cq (2^−ΔΔCq) method (24), with GAPDH mRNA expression serving as an endogenous control. The forward primer sequences used for GAPDH mRNA expression was 5′-GAAGGTGAAGGTCGGAGTCAAC-3′, the reverse primer was 5′-CAGAGTTAAAAGCAGCCCTGGT-3′ (25). The thermocycling conditions were as follows: 50°C for 2 min and 95°C for 2 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 30 sec.

The statistical analysis presented in the present study was conducted using SPSS software version 25.0 (IBM Corp.). The conformity of the genotype status of the rs2297136 polymorphism with Hardy-Weinberg equilibrium was assessed through the chi-square test (22). Regarding the analysis of baseline characteristics, the distribution between proportion variables and genotype status of rs2297136 was carried out using the Mann-Whitney U non-parametric test, between the two groups. Data in the present study were presented as median (range) and the number of patients in percentages based on corresponding data category. PFS and OS were defined as aforementioned. Survival curves were generated using Stata 14.0 to illustrate survival data according to rs2297136 genotype status, with a log-rank test used to determine significant differences. Cox analysis was used for OS in multivariate analysis. P<0.05 was considered to indicate a statistically significant difference.

Baseline characteristics of the 89 patients with advanced NSCLC were shown in Table II. It was revealed that all the 89 patients included in the present study were individuals commonly encountered in clinical practice with previously treated advanced NSCLC. Notably, 38.2% of the patients had an Eastern Cooperative Oncology Group (ECOG) performance status score of 2–3. Interestingly, 18 cases (20.2%) exhibited positive EGFR mutation, and 4 patients (4.5%) exhibited anaplastic lymphoma kinase positive rearrangement. Among the patients, 19 received PD-1 blockades as second-line therapy, while 70 underwent third-line or subsequent treatments. The present study utilized five PD-1 blockades, including camrelizumab (28.1%), sintilimab (23.6%), tislelizumab (22.5%), pembrolizumab (16.9%) and nivolumab (8.9%).

As outlined in the methods section, only rs2297136 demonstrated clinical significance in the preliminary analysis, as demonstrated in Table I. The prevalence of rs2297136 among the 89 patients with advanced NSCLC is detailed as follows: The AA genotype was observed in 58 cases (65.2%), the AG genotype was found in 28 cases (31.5%), and the GG genotype was noted in 3 cases (3.4%), resulting in a minor allele frequency (MAF) of 0.19, consistent with Hardy-Weinberg Equilibrium (P=0.865). Given the rarity of patients with GG genotype, patients with GG and AG were combined into one group in the subsequent analysis. The association between genotype status of rs2297136 and baseline characteristics is presented in Table II. Evidently, baseline characteristics of patients with AA and AG/GG genotypes were comparable and well-balanced (P>0.05).

Radiological evidence for the target lesions of the 89 patients with advanced NSCLC who received PD-1 blockade treatment was collected and assessed. According to iRECIST criteria, the best overall response during PD-1 blockade treatment indicated a CR in one patient (1.1%), PR in 19 patients (21.3%), SD in 37 patients (41.6%) and progressive disease (PD) in 32 patients (36.0%). This resulted in an ORR of 22.5% (95% CI: 14.3–32.6%) and a DCR of 64.0% (95% CI: 53.2–73.9%; Fig. 1). Specifically, the changes in the target lesions of the 89 patients after PD-1 blockade treatment according to genotype status of rs2297136 are depicted in Fig. 1. Target lesions of some patients shrunk significantly after the PD-1 blockade treatment. It was noteworthy that even patients with AG/GG genotype numerically demonstrated a higher ORR compared with patients with AA genotype [ORR of AG/GG vs. AA: 29.0% (95% CI: 14.2–48.0%) vs. 19.0% (95% CI: 9.9–31.4%)]. However, this difference did not reach statistical significance (χ2=1.18, P=0.278). Additionally, the DCR for patients with AA and AG/GG genotype was 58.6% (95% CI: 44.9–71.4%) and 74.2% (95% CI: 55.3–88.1%), respectively (χ2=2.13, P=0.145).

Regular follow-up was performed for the 89 patients with advanced NSCLC included in the present study, resulting in a mature prognostic data ultimately. The data cut-off date of the present study was November 15, 2022 and the median follow-up duration was 10.2 months (follow-up range: 0.9–32.5 months). Among these patients, 75 were observed to experience progression or death events, providing a maturity of 84.3% for PFS data. The PFS survival curve is presented in Fig. 2, revealing a median PFS of 3.4 months (95% CI: 1.80–5.00) for the 89 patients treated with PD-1 blockade monotherapy. Notably, a total of 15 patients experienced a sustained PFS benefit lasting over 12 months.

Additionally, 61 patients were documented to have succumbed, resulting in a maturity of OS data of 68.5%. The OS survival curve, also depicted in Fig. 2, revealed a median OS of 11.3 months (95% CI: 7.93–14.67) for the 89 patients with advanced NSCLC treated with PD-1 blockades. Interestingly, 10 patients achieved a sustainable OS benefit lasting over 24 months. Furthermore, as shown in Table SI, 49 patients received subsequent treatment upon progression during PD-1 blockade therapy. Among them, 21 received anlotinib regimen, 13 underwent chemotherapy, 9 were administered traditional Chinese medicine and the remaining 6 received PD-1/PD-L1 related therapy.

In exploring the connection between prognosis and the genotype status of rs2297136, additional survival analysis was conducted. As demonstrated in Fig. 3, patients with AG/GG genotype showed a tendency towards improved PFS compared with those with AA genotype [median PFS: 5.30 months (95% CI: 3.42–7.18) vs. 2.95 months (95% CI: 2.58–3.32)], reaching marginal statistical significance (χ2=4.30, P=0.038). Furthermore, the association between OS and genotype status of rs2297136 was separately examined. As illustrated in Fig. 4, patients with AG/GG genotype exhibited a longer OS compared with those with AA genotype [median OS: 18.4 months (95% CI: 8.61–28.19) vs. 8.8 months (95% CI: 4.82–12.78)] and this difference was statistically significant (χ2=6.43, P=0.011).

Additionally, to investigate the independent prognostic implication of rs2297136 for patients with advanced NSCLC, multivariate Cox analysis for OS was adopted subsequently. Initially, association analysis between OS and baseline characteristic subgroups in univariate analysis was carried out separately. The median OS and 95% CI according to baseline characteristic subgroups in univariate analysis were presented in Table III. Notably, it appeared that almost all patients might uniformly benefit from PD-1 blockades monotherapy uniformly. However, ECOG performance status and number of metastatic lesions exhibited a significant association with OS in the univariate analysis, as shown in Table III. This suggested that patients with ECOG performance status 0–1 score had a longer OS than that of patients with 2–3 score (median OS: 15.5 vs. 9.5 months, P=0.008), and patients with number of metastatic lesions ≤3 demonstrated improved OS than those >3 metastatic lesions (median OS: 13.5 vs. 9.5 months, P=0.019). Interestingly, patients with EGFR positive mutation demonstrated a trend towards inferior OS compared with those with EGFR negative mutation, although the difference was not statistically significant (median OS: 9.5 vs. 13.8 months, P=0.131). Subsequently, variables significantly associated with OS were incorporated into multivariate Cox analysis furthermore. As illustrated in Table III, after multivariate adjustment, the Cox multivariate analysis demonstrated that ECOG performance status (HR=0.63, P=0.011), the number of metastatic lesions (HR=0.73, P=0.026) and PD-L1 rs2297136 genotype status (HR=2.01, P=0.018) were all independent risk factors for OS.

Ultimately, mRNA analysis was conducted on a total of 81 patients. The prevalence of rs2297136 polymorphism among these patients was as follows: The AA genotype was observed in 53 cases (65.4%), the AG genotype was noted in 26 cases (32.1%), the GG genotype was found in 2 cases (2.5%). The MAF was 0.19, aligning with Hardy-Weinberg Equilibrium (P=0.567) and demonstrating similarity with the genotype distribution frequency among the 89 patients with advanced NSCLC. The median relative expression level of PD-L1 mRNA was 3.30 (ranging from 1.55 to 5.32) and the mean relative expression level was 3.24±0.835 in 81 PBMC specimens. Subsequently, the association between PD-L1 mRNA expression and genotype status of rs2297136 is illustrated in Fig. 5. In comparison with AA genotype, AG/GG genotypes of rs2297136 exhibited a higher relative expression of PD-L1 mRNA in PBMC specimens (4.09±0.538 vs. 2.59±0.644), demonstrating statistical significance (P<0.001).

PD-L1 immunohistochemical expression results in two patients among the present study and matched with PD-L1 mRNA expression status is shown in Fig. S1. PD-L1 mRNA expression status was divided into PD-L1 high expression (PD-L1 H) and PD-L1 low expression (PD-L1 L) according to the median expression threshold value (3.30). Patients with PD-L1 H and PD-L1 L were observed in 41 and 40 cases, respectively. As exhibited in Fig. S2, patients with PD-L1 H conferred a trend for superior PFS compared with those with PD-L1 L (median PFS: 5.3 vs. 2.8 months), although the difference was not statistically significant (χ2=3.438, P=0.064). Furthermore, as shown in Fig. S3, patients with PD-L1 H conferred a significantly improved OS compared with those with PD-L1 L (median OS: 13.5 vs. 7.8 months), demonstrating statistical significance (χ2=4.559, P=0.033).

Additionally, some patients with advanced NSCLC examined the expression the of immunohistochemistry (IHC) of PD-L1 using biopsy cancer tissue samples in the third-party testing agency to predict the efficacy of PD-1 blockades. These test results were collected and matched with mRNA expression results correspondingly. As shown in Fig. S1, the relative mRNA expression level of PD-L1 gene was correlated with the PD-L1 IHC expression consistently, the relatively high mRNA expression level was associated with high TPS score of PD-L1 IHC expression.