Introduction

Lung cancer is the second most frequent cancer and the cancer with the highest mortality worldwide [1]. Based on morphological features, lung cancer has been historically classified as small cell lung cancer (SCLC: 15%) and non-small cell lung cancer (NSCLC: 75%) [2]. NSCLC patients have often an advanced, unresectable disease stage not amenable to curative treatment or a metastatic disease stage at their initial diagnosis [3]. Historically, chemotherapy has long been the standard of care for the treatment of advanced and metastatic lung cancer. Compared with patients who receive best supportive care only, patients undergoing platinum-based combination chemotherapy have an improved 1-year survival rate from 20 to 29% (hazard ratio [HR]: 0.77, 95% CI: 0.71–0.83) [4].

Until the advent of immune checkpoint inhibitors (CPIs) in the recent treatment of the aforementioned group of NSCLC patients without oncogenic driver mutations systemic therapy was limited to chemotherapy associated with poor survival times and an unfavorable toxicity profile [5]. The availability of CPIs, which are monoclonal antibodies targeting immune systems T cells or ligands on the tumor cells [6], caused a paradigm shift in the treatment landscape of NSCLC opening new perspectives for a relevant number of patients, particularly of those with comorbidities and a reduced performance status [7].

Results from the pivotal clinical studies CheckMate-017, CheckMate-057, KEYNOTE-010 and OAK revealed marked improvements in survival with monoclonal antibodies targeting programmed cell death receptor 1 (PD-1) or programmed cell death receptor-ligand 1 (PD-L1) as compared with standard chemotherapies and led to the approval of nivolumab, pembrolizumab and atezolizumab after platinum-based chemotherapy for the treatment of metastatic NSCLC [8].

Due to the success of the CPIs in previously treated NSCLC patients, their effectiveness unter real life condition of a lung cancer as single agent therapy in the frontline setting was evaluated. Currently, efficacious first-line treatment options with ICI monotherapy can be offered to patients with an expression of PD-L1 on ≥ 50% of the tumor cells or on ≥ 10% of the immune cells according to phase 3 clinical trials with pembrolizumab [911], atezolizumab [12, 13], and cemiplimab [14, 15].

Regarding the approval status, there are some differences between atezolizumab, cemiplimab, and pembrolizumab. Pembrolizumab (a PD-1- antibody) was approved on October 24, 2016 by the U.S. Food and Drug Administration (FDA) and on January 31, 2017 by the European Medicines Agency (EMA) as first-line therapy for metastatic NSCLC patients with a PD-L1-Tumor Proportion Score (TPS) of ≥ 50%, after the amazing results of the KEYNOTE-024 trial were presented [9]. In contrast to the EMA, the FDA extended approval of pembrolizumab for every therapy-naïve, metastatic NSCLC patient with any PD-L1 expression in the tumor cells after the KEYNOTE-042 trial met its primary endpoints [16, 17].

The positive results of the IMpower-110 trial led to approval of atezolizumab (a PD-L1 antibody) as first-therapy in metastatic NSCLC patients with a Tumor Cell Score (TCS) of ≥ 50% or an Immune Cell Score (ICS) of ≥10% on May 18, 2020 by the FDA and on May 5, 2021 by the EMA [12, 13].

Only cemiplimab got approval on February 22, 2021 by the FDA and on June 21, 2021 by the EMA for locally advanced stage IIIB/ IIIC NSCLC patients with a TPS ≥ 50% and with unresectable tumor disease and who are unfit undergoing definitive chemo-/ radiotherapy or for metastatic patients according to the results of the EMPOWER-Lung1 trial [11].

However, from 2016/2017 until 2020/2021 pembrolizumab was the one and only first-line therapy option for metastatic NSCLC patients with a TPS = 50%. After the approval of atezolizumab in a rather similar setting, some treating physicians changed to the administration of atezolizumab or even cemiplimab due to a slightly broader approval status. Furthermore, the lower incidence of pneumonitis in NSCLC patients treated with PD-L1 inhibitors [18] versus those receiving PD-1 inhibitors might have prompted oncologists to prefer atezolizumab. But data about the comparison of these antibodies are very limited. Therefore, we aimed to compare mainly atezolizumab with pembrolizumab as single-agent firstline therapy for patients with metastatic NSCLC and a high PD-L1 expression not only on the tumor cells, but also on the immune cells. Particularly, data about the efficacy of pembrolizumab in NSCLC patients with a high CPS or ICS and in patients with a poor performance status (Eastern Cooperative Oncology Group (ECOG): 2) caused by the underlying malignant disease or serious comorbidities (e.g., chronic obstructive pulmonary disease (COPD), cardiac or renal failure or polyneuropathy) are lacking. The last group of patients tend to be treated with CPIs more often, even if they are more suffering from tumor related symptoms, usually requiring combined chemo-/immunotherapy.   

Materials and Methods

Study Design and Patients

Data of a total of 161 locally advanced stage IIIB/IIIC NSCLC patients with unresectable tumor disease  who were unfit to undergo definitive chemo-/radiotherapy or metastatic patients, who started their first systemic palliative therapy from January 31, 2017 until April 30, 2024 at the Lung Cancer Center EssenMitte, which is certified by the German Cancer Society (Deutsche Krebsgesellschaft: DKG), were retrospectively analyzed. The clinical data collected for this analysis were age, gender, Eastern Cooperative Oncology Group performance status (ECOGPS), the pretreatment absolute counts of leukocytes, neutrophil granulocytes and lymphocytes, as well as the serum level of lactate dehydrogenase (LDH), the smoking status and TNM or IASLC/ Union for International Cancer Control (UICC) stages. Some of the values were used to calculate either the Lung Immune Prognostic Index LIPI or the modified LIPI (mLIPI).

As previously defined the LIPI score was calculated [19], using LDH > upper limit of normal and derived neutrophil-to-lymphocyte ratio (dNLR) > 3 as cutoff-points. dNLR is calculated as follows: absolute neutrophil count/(total leukocyte count-absolute neutrophil count). Therefore patients could be assigned a score of 0, 1 or 2 based on their LDH and dNLR values, corresponding to good, intermediate and poor LIPI, respectively. Furthermore, patients were additionally scored on the basis of the mLIPI [20], specifically, NLR =3, LDH >1.5× the upper limit of the normal value, and ECOG-PS =2 were assigned 1 point each. According to the total score, the patients were divided into good (0 points), intermediate (1 point), and poor/very poor (=2 points) groups.

Regarding the retrospective analysis of adverse events (AE), only serious AEs leading to treatment interruption or discontinuation with CPIs were extracted from the patients` records.

All tumor sample evaluations were conducted by board-certified pathologists. All tissues were stained with hematoxylin and eosin (H&E), rabbit monoclonal immunoglobulin as a negative reagent control, and with the VENTANA PD-L1 (SP263) Assay, which is one of the recommended assays for PD-L1 diagnostics [21]. The H&E staining was performed to determine the adequacy of tumor. A tissue sample was adequate for the assay interpretation if it contained at least 100 viable tumor cells. Tumor-associated stroma was not required for tumor cell scoring, but was essential for scoring of immune cells [22]. For each staining run, prequalified human benign tonsil tissue was used as positive and negative tissue control. Tonsil tissue stained with PD-L1 was assessed for staining in lymphocytes and macrophages in germinal centers, and scattered PD-L1 staining cells among PD-L1-negative cells in paracortical regions. Tonsil tissue was also assessed for the presence of diffuse staining observed in the reticulated crypt epithelial cells with the absence of staining of superficial squamous epithelial cells. Patient-matched tissue stained for negative reagent control was evaluated for the presence and acceptability of nonspecific background staining. Once the H&E and the control slides were deemed acceptable, the PD-L1 stained slide was assessed.

In case of the suspicion of lung cancer as the underlying malignant disease leading to probe sampling reflex testing for PD-L1 usually occurs, if specimens contain a sufficient number of vital tumor cells. The PD-L1 Tumor Proportion Score (TPS) was calculated as the percentage of at least 100 viable tumor cells with complete or partial membrane staining and was separated into the following three groups: <1% (no expression: category 0), 1%–49% (low expression: category 1) and =50% (high expression: category 2).

The expression on immune cells was assessed as the proportion of tumor area occupied by PD-L1-positive immune cells of any intensity. Only immune cell staining in the tumor microenvironment was evaluated, including different patterns of staining (aggregates and single cells dispersed among tumor cells) and staining in different immune cell types (lymphocytes, macrophages, dendritic cells, and granulocytes). The coverage of the tumor area by PDL1 expressing tumor infiltrating immune cells was graded into the following four groups: <1% (ICS 0), 1%–4% (ICS 1), 5%–9% (ICS 2) and =10% (ICS 3) [12, 13, 22].

Initially, the mononuclear immune cell density score (MIDS), defined as the ratio of the number of PD-L1-expressing immune cells to that of all tumor cells, was used in an attempt to capture immune cell expression. However, MIDS demonstrated poor reproducibility in preliminary studies. Given that the TPS is the ratio of the number of PD-L1-expressing tumor cells to that of all tumor cells, it is mathematically feasible to combine it with the MIDS (both are fractions with a common denominator). The result of this combination is the ratio of the number of all PDL1–expressing cells (tumor cells, lymphocytes, macrophages) to the number of all tumor cells, which is termed the Combined Proportion Score (CPS). As the name implies, CPS considers PDL1 expression on tumor cells and immune cells combined [23].

From January 31, 2017 (date of first approval of pembrolizumab as first-line therapy for metastatic NSCLC by the EMA) until May, 5 of 2021 (date of first approval of atezolizumab as first-line therapy for metastatic NSCLC by the EMA) only TPS was assessed by the pathologists. Therefore, CPS and ICS had to be retrospectively analyzed for patients who started CPS first-line treatment during this time period.

Immuncyto- and -histochemistry were performed at the Practice for Pathology Essen-Mitte (Zentrum fuer Pathologie Essen-Mitte), Essen, Germany, which is accredited by the German Accreditation Body (Deutsche Akkreditierungsstelle: DAkkS). Furthermore, PD-L1-testing is certified by the Initiative for Quality Assurance in Pathology (Qualitätssicherungsinitiative in der Pathologie: QuIP) of the German Society of Pathology (Deutsche Gesellschaft für Pathologie: DGP).

Next Generation Sequencing (NGS) for DNA and RNA alterations was performed at GENOPATH GbR, Bonn, Germany using the ARCHER panel, if molecular diagnostics were requested by the treating physician. In general, DNA mutations of the BRAF, cMET (exon 14 skipping mutation), EGFR, Her2 and KRAS gene as well as RNA fusions of the ALK, NTRK, RET, and ROS1 genes were investigated.

Treatment and Assessment

Atezolizumab (1.200 mg, d1, q3w), Cemiplimab (350 mg, d1, q3w) or Pembrolizumab (200 mg, d1, q3w) were intravenously administered according to local standard. Radiological evaluation of response to treatment was carried out using RECIST 1.1 [24]. For progression-free survival (PFS) and overall survival (OS) analysis, patients were followed up until April, the 30th of 2024.

Statistical Analysis

The retrospective study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and approved by the institutional review board of the DKG-certified Lung Cancer Center (approval on March 11, 2024: ITT-EP-03-2024). Individual consent for this retrospective analysis was waived.

The statistical tests were performed using IBM SPSS Statistics (version 29), the associated figures were created using GraphPad PRISM (version 7). The Chi-square test was used to compare categorical variables (when necessary, Fisher’s exact test). Students t-test was used for age, Mann Whitney U test for „pack years“ and PD-L1 expression (%). Multivariate logistic regression was used to estimate associations (wald test). Models were adjusted for all variables. Kaplan-Meier methodology was used for estimation of medians, survival curves and their comparisons. Cox proportional hazard model was performed to identify prognostic factors in both the univariate analysis and the multivariate analysis (wald test).

Results

Clinicopathological Features

Detailed clinicopathological information is shown in table 1. A total of 161 patients with locally advanced NSCLC with unresectable tumor disease being unfit to undergo definitive chemo-/radiotherapy or metastatic NSCLC were identified. All patients received atezolizumab, cemiplimab or pembrolizumab as single-agent first-line palliative therapy. The majority of the patients (70.2%) had metastatic disease at the initial diagnosis (table 2). LIPI and mLIPI was calculated in 150 patients (93.2%).

Table 1: Comparison of clinicopathological factors between locally advanced NSCLC patients with unresectable tumor disease being unfit to undergo definitive chemo-/radiotherapy and metastatic NSCLC patients, who received atezolizumab, cemiplimab or pembrolizumab as single-agent first-line therapy for palliative treatment.

PD-L1 TPS was assessed in each patient (table 1). Median PD-L1 TPS was 80%. PD-L1 CPS was analyzed in 138 patients (85.7%) with a median of 90. PD-L1 ICS was determined in 139 patients (86.3%), with a median of 3 %. A high ICS (=10%) was noted in 33 patients (23.8%). Only 21 patients (13.0 %) patients had a high TPS and ICS, 106 patients (65.8%) had a high TPS, but low ICS and only 12 patients (7.5%) had a low TPS, but high ICS.

In the univariate analysis of the clinicopathological features (table 1), significant differences between patients receiving atezolizumab, cemiplimab or pembrolizumab regarding age, synchronous development of metastases, PD-L1 expression (TPS, CPS or ICS) or pathogenic gene mutations (e.g., KEAP1 or TP53) were noted. In contrast, LIPI and mLIPI factor distribution were quite similar. In the subgroup of patients with metastatic disease at primary diagnosis, significant differences between atezolizumab and cemiplimab regarding PD-L1 expression or pathogenic mutations were found (table 2).

Table 2: Comparison of clinicopathological factors between metastatic NSCLC patients, who received atezolizumab or pembrolizumab as single-agent first-line therapy for palliative treatment.

Gene Mutations

In 119 patients of the entire cohort (74.5 %), DNA & RNA mutational diagnostics were performed. The majority of these patients suffered from an adenocarcinoma of the lung (90 resp. 75%) or had metastatic disease at the initial diagnosis (84 resp. 70%). Any type of an activating or a pathogenic mutation was observed in 85 patients (70.8%). Of these, activating KRAS gene mutations were found in 42 patients (49.4%), pathogenic TP53 gene mutations in 56 patients (65.9%) or CDKN2A gene mutations in 12 patients (14.1%). All other gene mutations had a frequency lower than 10% (Suppl. table 1).

Suppl. Table 1: Genetic mutations (activating or pathogenic mutations) found in patients with DNA- and RNA mutational diagnostics. Of note, in 35 patients only wild type genes were detected. No gene transfusions or translocations were observed. In some patients, several mutations were found (comutations).

Treatment Effectiveness

ORR was assessable in 127 patients (78.9%), but 15 patients (23.1%), 2 patients (22.2%), and 17 patients (19.5%) discontinued therapy with atezolizumab, cemiplimab, or pembrolizumab before radiographic assessment of response, mainly due to clinical deterioration. Partial response (PR) occurred in 54.4%, stable disease (SD) in 21.6% and progressive disease (PD) in 24.0% of the assessable patients (table 1). There were significant differences between the treatment groups (p=0.034, table 1). In the subgroup of stage IV patients ORRs between atezolizumab and pembrolizumab did not differ significantly (table 2).

Median PFS of the entire cohort was 10.5 months and median OS was 15.0 months (figure 1A/B). There were no significant differences between the treatment groups. However, patients receiving pembrolizumab compared to atezolizumab tended to receive longer treatment without progression (13.0 vs. 9.1 months, p=0.063) in the univariate analysis.

Figure 1A: Progression-free survival (PFS) of locally advanced NSCLC patients with unresectable tumor disease being unfit to undergo definitive chemo-/radiotherapy or metastatic NSCLC patients, who received atezolizumab, cemiplimab or pembrolizumab as single-agent first-line therapy for palliative treatment.

Figure 1B: Overall survival (OS) of locally advanced NSCLC patients with unresectable tumor disease being unfit to undergo definitive chemo-/radiotherapy or metastatic NSCLC patients, who received atezolizumab, cemiplimab or pembrolizumab as single-agent first-line therapy for palliative treatment.

Non-smoking status (p=0.026), a high ICS and low TPS (p=0.002), a lower number of treatment cycles (p<0.001) or progressive disease (p<0.001) as response to therapy resulted in a worse PFS (table 3) in the multivariate analysis. An ECOG PS of 2, detection of at least three genetic mutations (compared to wild type of all genes tested), stage IV at initial diagnosis, 2 factors in the LIPI, 2 or 3 factors in mLIPI, a lower number of treatment cycles or PD as response to CPIs were associated with a poorer OS (table 4) in the multivariate analysis. In contrast, the kind of the used CPI had no influence on OS (p=0.509 for atezolizumab vs. cemiplimab or p=0.0326 for pembrolizumab vs. atezolizumab).

Table 3: Multivariate logistic regression analysis by choosing PFS as the terminal point variable.

Table 4: Multivariate logistic regression analysis by choosing OS as the terminal point variable.

In the subgroup of patients with stage IV at initial diagnosis, median PFS was 10.5 months and median OS was 12.0 months (figure 2A/B). Patients receiving pembrolizumab compared to atezolizumab tended to be longer without progression (11.0 vs. 9.1 months, p=0.058) and to survive longer, although there was no significant difference in OS between pembrolizumab or atezolizumab treatment (14.2 vs. 8.7 months, p=0.065).

Figure 2A: Progression-free survival (PFS) of metastatic NSCLC patients, who received atezolizumab or pembrolizumab as single-agent first-line therapy for palliative treatment.

Figure 2B: Overall survival (OS) of metastatic NSCLC patients, who received atezolizumab or pembrolizumab as single-agent first-line therapy for palliative treatment.

Adverse events leading to treatment interruption or permanent discontinuation

In 20 patients (12.4%) treatment was either temporarily or permanently discontinued due to the advent of serious adverse events (Suppl. table 2). The main causes were colitis (6 patients), dermatitis, hepatitis or pneumonitis (each 3 patients). In patients treated with cemiplimab, no serious adverse were noted. Of note, one case of hyperprogression during atezolizumab therapy (proven by CT, the course of serum levels of LDH and tumor markers) was observed.

Suppl. Table 2: Adverse events leading to temporary or permanent treatment interruption.