Lung cancer ranks first among newly diagnosed cancers and remains the leading cause of cancer-related death worldwide.1 In Spain, it is the third most common cancer after colorectal and breast cancer and remains the leading cause of cancer-related death.2

Histological characterisation of the tumour is essential because it determines prognosis and guides treatment. Approximately 85% of new lung cancer diagnoses are non-small cell lung cancer (NSCLC), and most patients (70%) present with advanced or metastatic disease at diagnosis.

The aim of treatment for metastatic lung cancer is to prolong survival, improve quality of life, and control disease-related symptoms.

Before the advent of immunotherapy, platinum-based combination therapy using either cisplatin or carboplatin was the standard first-line treatment for patients with advanced or metastatic NSCLC, unless specific molecular characteristics, such as EGFR, ALK, or ROS alterations, made them candidates for targeted therapies.

The arrival of immunotherapy in oncology has brought about a paradigm shift in the treatment of certain cancers by changing the focus from directly targeting tumours to strengthening patients' immune systems.

The incorporation of immunotherapy with immune checkpoint inhibitors (ICIs) into first-line treatment for patients with NSCLC has improved survival rates in both squamous and non-squamous histologies.

First-line treatment options with ICIs for advanced or metastatic NSCLC are guided by the tumour proportion score (TPS) for PD-L1. The KEYNOTE-024 clinical trial3,4 established pembrolizumab as the standard first-line treatment for NSCLC in patients whose tumours expressed PD-L1 with a TPS ≥50% and with no EGFR or ALK alterations. After a median follow-up of 5 years, median overall survival was 12.9 months longer with pembrolizumab than with chemotherapy (hazard ratio [HR] = 0.6; 95% CI: 0.5–0.9; p = 0.002). The clinical practice guidelines from ESMO,5 ASCO,6 NCCN,7 and SEOM8 reflect this recommendation.

Furthermore, the use of ICIs has been associated with an adverse-effect profile related to their mechanism of action, which differs substantially from that of conventional anticancer therapies and may lead to so-called immune-related adverse events (irAEs).9 The KEYNOTE-024 clinical trial reported an incidence of immune-mediated adverse effects associated with pembrolizumab of 29.2%. Although most of these adverse effects are mild, they can be severe in some cases, requiring rapid recognition and management, particularly in the context of advanced disease. Various studies have shown that the development of irAEs in patients treated with ICIs may correlate with clinical response. However, further investigation is warranted given the inconsistency in study findings regarding the association between the occurrence of irAEs and response rates, progression-free survival (PFS), and OS in NSCLC treated with ICIs.10–14

The conditions under which drugs are used in routine clinical practice are often far less restrictive than those in the clinical trials that supported their approval. This means that efficacy and safety outcomes may differ from those observed in the clinical trials. In fact, many of the situations that met the exclusion criteria in the clinical trial are relatively common in clinical practice and partly account for the observed differences in efficacy and safety outcomes. Real-world studies are therefore crucial to complement evidence from clinical trials and to provide a more comprehensive understanding of treatment performance in more diverse patient populations and in real-world clinical settings.

For these reasons, the present study aimed to evaluate the efficacy and safety of pembrolizumab monotherapy as first-line treatment for patients with advanced or metastatic NSCLC in routine clinical practice in Spain.

After a median follow-up of 17.9 months, our study showed a median PFS of 8.7 months (95% CI: 7.2–10.1) and a median OS of 18.0 months (95% CI: 16.2–21.3), whereas the KEYNOTE-024 trial reported a median PFS of 7.7 months (95% CI: 6.1–10.2), similar to ours, and a median OS of 26.3 months (95% CI: 18.3–40.4), which was higher than ours. The difference in OS may be due to the fact that 11% of patients in our sample had an ECOG PS ≥2, which was an exclusion criterion in the KEYNOTE-024 trial. It is also considered to be a predictor of poor treatment response, as demonstrated by our results and those of other published real-world studies.18–22 Recording performance status in routine clinical practice shows considerable variability. Although PS ECOG categories are well defined, studies have demonstrated substantial interobserver heterogeneity in their assignment. In this context, some patients classified as ECOG PS 1 in clinical records may have been considered ECOG PS 2 in a more controlled setting, such as a clinical trial, potentially leading to an underestimation of the true proportion of patients with ECOG PS 2 in our cohort.

Although there were no differences in the proportion of patients with squamous-cell NSCLC—which is associated with poorer survival outcomes—between the KEYNOTE-024 trial and our study, the proportion of non-smokers differed, at 3.2% and 7.7%, respectively. This may represent another explanation for the lower OS observed in the present study, as non-smoking status is associated with poorer survival outcomes.20,21,23

Another distinguishing feature of our cohort is the use of concomitant radiotherapy, a practice not included in the KEYNOTE-024 trial, in a considerable proportion of patients (32%). Although we did not observe a statistically significant association between OS and radiotherapy, its use may serve as an indirect marker of higher symptom burden or poorer baseline clinical status that might not be fully captured by ECOG PS scores recorded in medical records.

Furthermore, the bivariate analysis showed that a BMI ≥25 kg/m2 was associated with improved PFS and OS, which is consistent with the findings of recently published studies.24 Our results support the value of LIPI as a predictor of response to ICIs, which has also been demonstrated in previous studies.25 In our study, prior exposure to corticosteroids, antibiotics and, in particular, PPIs—which together with PS ECOG constituted one of the predictors of poorer overall survival in the multivariate analysis—was significantly associated with worse survival outcomes. Although this relationship is well documented in the literature,21,22,26,27 the heterogeneity of the studies, particularly with regard to the description of drug exposure, prevents clear conclusions from being drawn.

In terms of safety, the incidence of adverse events of any grade was somewhat lower than that reported in the pivotal clinical trial (61.7% vs 76.6%), and a similar pattern was observed for grade 3 or higher adverse events (15.3% vs 31.2%). The proportion of patients who discontinued pembrolizumab due to toxicity was also slightly lower (12.7% vs 13.6%).

In our cohort, 4 patients (0.4%) died as a result of adverse events, including 2 cases of pulmonary toxicity (pneumonitis) and 2 cases of cardiac toxicity (acute myocardial infarction and pericardial effusion). In the KEYNOTE-024 clinical trial, 2 patients (1.3%) died, one suddenly from an unknown cause at treatment initiation and the ot due to pneumonitis.

Our results confirm that the development of toxicity during pembrolizumab treatment may be associated with significantly improved survival rates. We also observed higher OS in patients who developed 2 or more adverse events, indicating a relationship between the number of adverse events experienced and clinical benefit.13

Finally, the main limitations of this study stem from its retrospective design, which is associated with a risk of information loss due to under-reporting and with reduced accuracy in data collection. Its strengths lie in its multicentre design and the large number of patients included, making it one of the largest recently published real-world studies in this context.19–21 Furthermore, this study was conducted across 42 hospitals in Spain, reflecting real-world clinical practice in this setting.

Overall, in terms of both efficacy and safety, the results of this study are consistent with those reported in the KEYNOTE-024 clinical trial and in previously reviewed real-world studies. The main differences, particularly in OS, may be attributed to the inclusion in these real-world studies of patient groups systematically excluded from clinical trials, such as those with an ECOG PS ≥2.

Our results confirm that ECOG PS ≥2 and prior use of PPIs are predictive factors for lower OS, and that the development of toxicity during treatment is statistically associated with improved survival rates.

The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Ethics and Drug Research Committee of Alicante General Hospital on 1 April, 2021, registration number 2021–002. Exemption from obtaining informed consent was authorised in the case of deceased patients.

This project received financial support through the SEFH 2020–2021 call for project grants to working groups.

Amparo Burgos: Writing – review & editing, Writing – original draft, Validation, Supervision, Project administration, Methodology, Data curation, Conceptualization. Teresa Calleja: Writing – review & editing, Writing – original draft, Validation, Supervision, Project administration, Methodology, Data curation, Conceptualization. María Sacramento Díaz: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. Garbiñe Lizeaga: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. Cristina Ibáñez: Writing – review & editing, Writing – original draft, Validation, Data curation. Carmen María Valencia: Writing – review & editing, Writing – original draft, Validation, Data curation. Mónica Carbajales: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. David Conde: Writing – review & editing, Writing – original draft, Validation, Data curation. Juan Francisco Marín: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. Margarita Garrido: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. Raúl Díez: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. María José Martínez: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. María Larrosa: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. Estela Moreno: Writing – review & editing, Validation, Methodology, Conceptualization.