Pneumonitis Incidence in Patients With Metastatic Non-small Cell Lung Cancer on Immunotherapy: A Systematic Review and Meta-Analysis

Non-small cell lung cancer (NSCLC) is the most common type of lung cancer, often diagnosed at the advanced stage (metastatic). Treatment options for metastatic NSCLC include radiotherapy, chemotherapy, target drug therapy, and immunotherapy. Immunotherapy (utilization of checkpoint inhibitors) boosts the immune system to recognize and destroy cancer cells. However, it is often associated with immune-related complications such as pneumonitis. This review aims to determine the incidence of pneumonitis in metastatic NSCLC patients treated with different immunotherapy drugs. PubMed, Cochrane Library, and Embase databases were scoured for randomized controlled trials (RCTs) until October 2023. Published RCTs with similar research objectives were included, while non-English articles, reviews, case reports, ongoing trials, non-randomized studies, conference abstracts, and studies on small cell lung cancer (SCLC) were excluded. The Cochrane risk-of-bias tool for randomized trials (RoB 2) was used to assess the risk of bias among the included studies. The statistical analyses were performed with the Comprehensive Meta-Analysis software. The subgroup analysis of the 16 included RCTs showed that metastatic NSCLC patients treated with nivolumab and pembrolizumab had a higher incidence of any grade pneumonitis than those treated with atezolizumab (4.5% and 5.1% vs. 1.6%, respectively). Similarly, the incidence of grade ≥3 pneumonitis was higher among patients receiving nivolumab (1.3%) and pembrolizumab (2.4%) than those receiving atezolizumab (0.7%). Furthermore, the subgroup analysis showed that patients with naive-treated NSCLC on immunotherapy had a higher incidence of any grade pneumonitis than those with previously treated NSCLC (6.5% vs. 3.9%). Treatment-naive patients recorded higher grade ≥3 pneumonitis incidences than those previously treated (3.1% vs. 1.3%). Programmed death 1 (PD-1) inhibitors (i.e., pembrolizumab and nivolumab) have higher incidences of pneumonitis than programmed death-ligand 1 inhibitors (atezolizumab).


Introduction And Background
Lung cancer is among the deadliest malignancies worldwide and is the leading contributor to cancer-related deaths among women and men [1].Non-small cell lung cancer (NSCLC), a subdivision of lung cancer, equals the majority of incidences, with surgery being the leading management for fully developed NSCLC [2].It is the second leading cancer type in the USA in terms of incidence and the leading in terms of attributable deaths [3].Study evidence indicates that fully developed NSCLC has a poor prognosis with low one-and fiveyear survival rates [4].Despite the utilization of paclitaxel-and platinum-based chemotherapeutic agents and recent addition of biologics, such as vascular endothelial growth factor inhibitors (ramucirumab and bevacizumab), there are still poor survival outcomes [5].Lately, the utilization of immunotherapy-based intervention regimens has indicated better outcomes in enhancing progression-free and overall survival in NSCLC.
With the development and evolution in the treatment and management of NSCLC in recent decades, the approval of targeted therapies, including immune checkpoint inhibitors (ICIs) and tyrosine kinase inhibitors (TKIs), has led to promising long-term survival, reducing the role of traditional chemotherapy [3,6,7].Despite their promising survival outcomes, these therapies are linked to possibly fatal treatment-induced adverse events, such as pneumonitis.Immunotherapy-induced pneumonitis was characterized by the emergence of unusual imaging results, with or without respiratory symptoms, after starting immunotherapy and is classified according to the Common Terminology Criteria for Adverse Events (CTCAE) [7].
Adverse events, including interstitial lung abnormalities in treatment-naïve advanced NSCLC patients, are associated with smoking, a well-established risk factor leading to a shorter survival rate among lung cancer patients [8].Immunotherapy causes life-threatening adverse events, such as pneumonitis, complicating its use among lung cancer patients [9].Pneumonitis, also referred to as checkpoint inhibitor pneumonitis (CIP), occurs if the foreign irritant substances inflame the minute air sacs in an individual's alveoli (lungs) [10].
The severity of pneumonitis symptoms ranges from mild to life-threatening depending on its grade.According to the National Cancer Institute of the National Institutes of Health, pneumonitis can be asymptomatic (grade 1), symptomatic (grade 2), severe symptoms (grade 3), life-threatening compromise (grade 4), or death related to adverse outcomes (grade 5) [11,12].Research evidence shows it is a problematic condition among immune-related adverse events (irAEs) [5,11].CIP mechanisms remain indistinguishable; however, they are believed to be associated with immune dysregulation caused by ICIs [12].Higgins and Thompson [13] highlighted four possible mechanisms underlying CIP as an irAE.Foremost, adverse events can be associated with escalated T-cell activity against cross-antigens in normal and tumor tissues.Increased activated alveolar T-cell percentage and attenuation of the anti-inflammatory Treg phenotype might contribute to T-cell activity dysregulation, triggering irAE [14].Second, increased preexisting autoantibody levels might contribute to irAEs.Fehrenbacher et al. indicated that preexisting antinuclear antibodies, anti-rheumatoid factor antibodies, anti-thyroid peroxidase antibodies, and anti-thyroglobulin are possibly associated with the advancement of irAEs in NSCLC patients [15].Third, increased inflammatory cytokine levels are related to the appearance of irAEs.A patient diagnosed with NSCLC with advanced CIP after atezolizumab treatment is noted to have escalated interleukin-6 (IL-6) and C-reactive protein levels compared to baseline levels [16].Similarly, cytokines can be biomarkers for adverse measures, and their escalated expression is related to acute ICI toxicity [17,18].Another possible mechanism is the direct binding of anti-CTLA-4 (cytotoxic T lymphocyte-associated antigen) antibodies with CTLA-4 on normal tissues, for instance, the pituitary gland.This method might also explain why pituitary inflammation is a specific adverse event of anti-CTLA-4 antibody [19].
Pneumonitis has been recognized among patients treated with programmed death ligand 1 (PD-L1)/programmed death 1 (PD-1) ICIs [7].Although most patients respond to oral corticosteroids, some might develop substantial dyspnea and hypoxia, necessitating NSCLC therapy discontinuation, supplemental oxygen, or intravenous corticosteroids and additional immunosuppressive interventions, such as mycophenolate mofetil, cyclophosphamide, or infliximab [7].Even though it barely occurs, high-grade (grade 3/4) pneumonitis is linked to significant morbidity and mortality in a small proportion (1.0%) of the affected patients [8].
Pneumonitis incidence has been documented to be between 3.0% and 5.0% in clinical trial evaluations [8,9], but pneumonitis incidence in NSCLC is not yet known, especially across patients treated with different immunotherapy drugs.Only a few meta-analyses and systematic reviews have been performed to evaluate the safety and efficacy of palliative radiotherapy (pRT) and ICIs in metastatic NSCLC patients [10] and risk of pneumonitis among patients receiving durvalumab treatment and radiotherapy [11].Consequently, there is a need to evaluate the available evidence regarding the differences in incidences among NSCLC patients treated with various immunotherapy drugs.This systematic review and meta-analysis was conducted to understand the difference in pneumonitis incidences in metastatic NSCLC patients treated with different immunotherapy drugs.

Study Design
This study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Eligibility Criteria
Predetermined eligibility criteria were developed to assess the search records.Included and excluded studies were based on the following conditions (Table 1).Secondary studies: systematic reviews, meta-analyses, scoping reviews, and narrative reviews.Systematic reviews and meta-analyses were excluded since they do not have the aggregated data about patients to perform the analysis.Similarly, the meta-analyses and systematic reviews can compare different interventions, making it impossible to plot pooled outcomes in a forest plot.

TABLE 1: Inclusion and exclusion criteria developed to screen the search records.
SCLC: small cell lung cancer; NSCLC: non-small cell lung cancer; RCTs: randomized controlled trials.

Information Sources and Search Strategy
The reviewers performed an electronic search to locate relevant studies in PubMed, Embase, and the Cochrane Library.An extensive search was performed on these databases for RCTs published up to November 2023.The following search terms were used in Embase, PubMed, and Cochrane Library to identify relevant articles: ((non-small cell lung cancer) AND (pneumonitis)) AND (immunotherapy).Detailed search terms for the three databases are shown in Appendix 1.There was no date (year) restriction in search terms for all the databases.

Data Extraction
The eligible studies were analyzed, and data were abstracted for reviewing and analysis.The extracted data from the eligible studies were study ID (author(s) and publication year), patient characteristics (sample size (patients with metastatic NSCLC treated with immunotherapy only) and median age), follow-up period, immunotherapy agents utilized, comparator/control intervention, number of patients with pneumonitis of any grade, and number of patients with pneumonitis of grade ≥3.

Risk of Bias Assessment
The risk of bias assessment was performed via the Cochrane RoB 2 tool for RCTs, according to the Cochrane Handbook for Systematic Reviews of Interventions [12].RoB 2 instrument has five aspects utilized to generate the overall risk of bias: (1) randomization process, (2) deviations from the intruded intervention, (3) missing outcome data, (4) measurement of the outcome, and ( 5) selection of the reported result.Each aspect was assessed with either of the following options: "high risk of bias," "low risk of bias," or "some concerns."Following each aspect evaluation, studies were categorized with a "low risk of bias" judgment if they had one "some concerns" in five of the domains, "Some concerns" judgment if they had two or more "some concerns" in five of the domains, or a "high risk of bias" judgment if they had one "high risk of bias" in one domain.

Statistical Analysis
The random effects model was used to sum up the effect size in various studies.The effect size was reported regarding event rate with a 95% CI.The heterogeneity in studies was evaluated using the I 2 statistics, with a significance level of <0.05.Substantial heterogeneity was ascertained if the value of I 2 statistics was more than 75%, moderate if the value was between 50% and 75%, and low heterogeneity if less than 50% [13].All the analyses were performed using the Comprehensive Meta-Analysis (CMA) software version 3 (Englewood, USA).

Results
An initial search of the three electronic databases yielded a total of 474 records.The initial scanning led to the exclusion of 308 articles due to duplication.The remaining 164 articles were screened based on titles and abstracts, leading to the exclusion of 91 articles.The remaining 73 articles were sought for retrieval, and all were retrieved.They were subjected to screening based on the eligibility criteria, and 57 articles were excluded.The reasons for exclusion were non-English (n=10), SCLC (n=6), non-RCTs (n=16), and secondary studies, including systematic reviews, meta-analyses, and narrative reviews (n=25).A PRISMA flow chart of the study selection is shown in Figure 1.

Study Characteristics
The baseline and relevant characteristics of each study that met the inclusion criteria are summarized in Table 2.The 16 studies included in the meta-analysis were conducted between 2015 and 2021, with sample sizes ranging between 52 and 811 patients.Cumulatively, a total sample of 5,662 patients was included in this meta-analysis.The immunotherapy agents evaluated in the studies included nivolumab, atezolizumab, and pembrolizumab.NSCLC: non-small cell lung cancer; NR: not reported.

Discussion
This study aimed to assess the incidence rate of pneumonitis among metastatic NSCLC patients treated with various immunotherapy drugs.The analyses show that PD-1 inhibitors (pembrolizumab and nivolumab) have significantly higher incidences of pneumonitis than PD-L1 inhibitors (atezolizumab).Similarly, pneumonitis incidences among patients on immunotherapy were significantly higher for patients with naive-treated metastatic NSCLC.
Checkpoint inhibitor immunotherapy is gaining popularity and utilization in cancer care [30].Pneumonitis is among the adverse events in utilizing NSCLC immunotherapy drugs that determine their safety.Such an adverse event can be unpredictable in relation to severity and timing, resulting in significant mortality (occasionally) and morbidity.
Therefore, early detection and understanding of pneumonitis incidence rates can be vital for clinicians and medical practitioners to reduce irreversible lung damage among patients caused by such adverse events [31].
Additionally, evidence has indicated that physicians' knowledge regarding pneumonitis incidence rates among NSCLC patients can assist in devising mechanisms for reducing pneumonitis' impact, improving treatment adherence, and enhancing outpatient outcomes [31].
Various studies, including systematic reviews, have evaluated the incidences of several serious adverse events without solely focusing on pneumonitis [8,31,32].Zhou et al.'s [32] meta-analysis showed that incidences of serious adverse event (rash, hepatobiliary disorders, pneumonitis, and colitis) were 37% in the chemotherapy arm, 37% in PD-L1, and 33% in the PD-1 arm, while 47% in combined groups of PD-L1 plus chemotherapy and 43% in the PD-L1 plus chemotherapy arm [32].These results are similar to the outcomes of the present analysis, where PD-1 inhibitors had significantly higher pneumonitis incidences than PD-L1 inhibitors.Despite PD-1 inhibitors' high pneumonitis incidence rate, evidence has shown that pembrolizumab significantly prolongs overall and progression-free survival with less high-grade toxic adverse events in advanced melanoma patients compared to ipilimumab standard care [33].Also, nivolumab monotherapy offers better patient progress and clinical safety mechanisms (meaningful pneumonitis incidences) in advanced refractory squamous NSCLC [24].Reck et al. also noted that in comparison with chemotherapy treatment, pembrolizumab experienced more immune-related adverse events (pneumonitis) but had a better overall safety profile [23].
Shankar and Naidoo reported similar findings, showing that patients with PD-L1 monotherapy had a lower incidence of pneumonitis compared to those treated with PD-1 monotherapy [34].Atezolizumab offers better survival benefits in previously treated NSCLC patients [19].Its expression on tumor-infiltrating immune and tumor cells is predictive of such an advantage.For an extended follow-up period, atezolizumab shows a favorable safety profile than docetaxel as compared to PD-1 inhibitors [19].The nature and incidence of treatment-related adverse events of any grade and grade 3 were low even after longer exposure to treatment [19].Lower pneumonitis incidence among patients treated with PD-L1 may be due to the action mechanism of an anti-PD-1 agent blocking the PD-L1 and PD-L2 interactions, while anti-PD-L1 agents allow interactions between PD-L2 and PD-1 [35].
With any presentation, immune-related adverse events should be contemplated in multiple diagnoses and necessitate emphasis from the clinicians.Pneumonitis is a toxicity of specific concern for doctors managing NSCLC, with a higher incidence of pneumonitis induced by immunotherapy.Risk factors that contribute to higher incidences of pneumonitis may include patients with a smoking history or underlying lung pathology who can develop grade 3 pneumonitis [30].Similarly, pneumonitis can occur more frequently in patients with renal cell carcinoma and NSCLC than in metastatic melanoma patients [36].Wu et al. [37] noted that the lowest pneumonitis rates are displayed in patients treated for advanced melanoma (0.720%) and the highest in advanced NSCLC-treated patients (4.70%) [37].Studies in this meta-analysis and previous research have also shown that incidence varies with various tumor types, with pembrolizumab-induced pneumonitis being less frequent in melanoma (1%) than in NSCLC patients (5%) [20,33,36].The results of this meta-analysis align with this study's outcomes that pneumonitis incidences among patients on immunotherapy were significantly higher for patients with naive-treated metastatic NSCLC.NSCLC patients are more likely to have lung diseases, for instance, chronic obstructive pulmonary disease, likely increasing their vulnerability to developing higher-grade pneumonitis [38].
Furthermore, pneumonitis incidences are prominent among NSCLC patients, and it can be attributed to poor pulmonary functioning with the burden of tumors among such patients [36].Patients with PD-L1 and PD-1 inhibitor-induced pneumonitis show lymphocyte infiltration in lung biopsies and predominant lymphocytes in bronchoalveolar lavage [39].Since pneumonitis is reported as the most common contributor to fatal immune-related adverse events in both PD-L1 and PD-1 inhibitors [40], medical practitioners should enhance pneumonitis management in NSCLC.For instance, healthcare practitioners should decide the treatment duration and the utilization of immune-related drugs based on the severity of the adverse event.
There should also be consideration in the re-introduction of ICIs after discontinuation and emphasis on the treatment and diagnosis of each patient [36,41].

Limitations of the Study
Some limitations should be noted when interpreting the results of this study.There is no consensus about the diagnostic criteria for pneumonitis.Hence, determining pneumonitis was based on clinicians' experience, which can contribute to bias.Two studies by Hui et al. [28] and Leigh et al. [29] had a high risk of bias due to missing outcome data, which were not reported.This study has provided specific safety insights into the presentation and incidence of pneumonitis in metastatic NSCLC patients treated with different immunotherapy drugs.The meta-analysis illustrates that PD-1 inhibitors, pembrolizumab and nivolumab, have higher incidences of pneumonitis than PD-L1 inhibitors (atezolizumab).Incidences of pneumonitis among patients on immunotherapy were higher for patients with naive-treated metastatic NSCLC.The reporting of these findings might be essential for clinical decisions, including advising on risk factors like smoking habits, among metastatic NSCLC patients.Early detection of pneumonitis is crucial as it allows for timely intervention, which can prevent further morbidity or mortality.Early diagnosis and management of pneumonitis can significantly reduce the severity of symptoms, improve patient outcomes, and potentially prolong survival.By identifying high-risk patients and implementing preventive measures, healthcare providers can minimize the incidence and impact of pneumonitis, ensuring better overall treatment efficacy and enhancing patients' quality of life.

FIGURE 1 :
FIGURE 1: PRISMA flow diagram of search strategy.PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

FIGURE 2 :
FIGURE 2: Summary of the risk of bias for each domain among the included studies.

FIGURE 4 :
FIGURE 4: Pneumonitis incidence rate of any grade according to the immunotherapy drug.

FIGURE 6 :
FIGURE 6: Incidence of pneumonitis of any grade according to the NSCLC treatment (overall p-value, p < 0.05).