Perioperative immunotherapy for stage II-III non-small cell lung cancer: a meta-analysis base on randomized controlled trials

Background In recent years, we have observed the pivotal role of immunotherapy in improving survival for patients with non-small cell lung cancer (NSCLC). However, the effectiveness of immunotherapy in the perioperative (neoadjuvant + adjuvant) treatment of resectable NSCLC remains uncertain. We conducted a comprehensive analysis of its antitumor efficacy and adverse effects (AEs) by pooling data from the KEYNOTE-671, NADIM II, and AEGEAN clinical trials. Methods For eligible studies, we searched seven databases. The randomized controlled trials (RCTs) pertaining to the comparative analysis of combination neoadjuvant platinum-based chemotherapy plus perioperative immunotherapy (PIO) versus perioperative placebo (PP) were included. Primary endpoints were overall survival (OS) and event-free survival (EFS). Secondary endpoints encompassed drug responses, AEs, and surgical outcomes. Results Three RCTs (KEYNOTE-671, NADIM II, and AEGEAN) were included in the final analysis. PIO group (neoadjuvant platinum-based chemotherapy plus perioperative immunotherapy) exhibited superior efficacy in OS (hazard ratio [HR]: 0.63 [0.49-0.81]), EFS (HR: 0.61 [0.52, 0.72]), objective response rate (risk ratio [RR]: 2.21 [1.91, 2.54]), pathological complete response (RR: 4.36 [3.04, 6.25]), major pathological response (RR: 2.79 [2.25, 3.46]), R0 resection rate (RR: 1.13 [1.00, 1.26]) and rate of adjuvant treatment (RR: 1.08 [1.01, 1.15]) compared with PP group (neoadjuvant platinum-based chemotherapy plus perioperative placebo). In the subgroup analysis, EFS tended to favor the PIO group in almost all subgroups. BMI (>25), T stage (IV), N stage (N1-N2) and pathological response (with pathological complete response) were favorable factors in the PIO group. In the safety assessment, the PIO group exhibited higher rates of serious AEs (28.96% vs. 23.51%) and AEs leading to treatment discontinuation (12.84% vs. 5.81%). Meanwhile, although total adverse events, grade 3-5 adverse events, and fatal adverse events tended to favor the PP group, the differences were not statistically significant. Conclusion PIO appears to be superior to PP for resectable stage II-III NSCLC, demonstrating enhanced survival and pathological responses. However, its elevated adverse event (AE) rate warrants careful consideration. Systematic review registration https://www.crd.york.ac.uk/PROSPERO/#recordDetails, identifier CRD42023487475.


Introduction
For decades, lung cancer (LC) has been the leading global cause of cancer-related deaths, with over 80% attributed to non-small cell lung cancer (NSCLC) (1,2).Comprehensive treatment based on surgery is the standard of care for selected resectable stages II-III NSCLC (3).In previous approaches to neoadjuvant and adjuvant treatment for stage II-III NSCLC, chemotherapy played a vital role, but its solitary use yielded unsatisfactory results (4).In recent years, immunotherapy has gained widespread acceptance in solid tumor treatment, demonstrating superior efficacy in both neoadjuvant and adjuvant treatment for resectable NSCLC (5)(6)(7).Nevertheless, controversy persists in clinical settings regarding whether perioperative immunotherapy (neoadjuvant +adjuvant) can yield superior results (8).
The use of immunotherapy in the perioperative period of resectable lung cancer has been a hot topic in recent years.In neoadjuvant therapy, the CheckMate 816 study demonstrated that the addition of nivolumab to platinum-based chemotherapy (PBC) could significantly increase event-free survival (EFS) and drug responses (9).Similar results were also validated in the TD-FOREKNOW study (Camrelizumab) (10).In adjuvant therapy, the KEYNOTE-091 study showed that the addition of pembrolizumab to PBC could significantly increase disease-free survival (DFS) (11).The IMpower010 study also confirmed that adding atezolizumab to PBC could improve DFS and overall survival (OS), especially in patients with programmed cell death 1 ligand 1 (PD-L1)-positive NSCLC (12).Regarding the use of immunotherapy in combination of neoadjuvant and adjuvant therapy, both the KEYNOTE-671 study (pembrolizumab) and the AEGEAN study (durvalumab) found that perioperative immunotherapy could significantly improve OS and EFS, and similar results were also validated in the NADIM II study (nivolumab) (13)(14)(15).
This study conducted a meta-analysis based on randomized controlled trials (RCTs) to evaluate the impact of perioperative immunotherapy with neoadjuvant PBC on survival, pathological responses, and adverse reactions.

Materials and methods
This study was conducted in accordance with PRISMA guidelines and registered in PROSPERO (ID: CRD42023487475) (Supplementary Table S1).
Articles lacking initial data, as well as meta-analyses, conference articles, and case reports, were not considered for inclusion.Distinct articles covering the same trial with diverse outcomes were included, but for identical outcomes, only the most recent data were utilized in the analysis.

Data extraction
Two investigators independently extracted data, including study characteristics (publication date, first author, etc.), participant details (sex, age, etc.), cancer specifics (histopathology, stage, etc.), antitumor effectiveness (OS, EFS, pathological responses, etc.), and counts of adverse events (total AEs, serious AEs, etc.).Disagreements were resolved through a process of re-evaluation and discussion.

Quality assessment
We assessed the quality of RCTs using the Jadad scale, a 5-point system reflecting randomization, blinding, and patient inclusion.A score of ≥3 points was considered indicative of high quality (17).Additionally, the Cochrane Risk Assessment Tool was employed, which evaluates bias related to selection, performance, detection, attrition, and reporting and categorizes risk as low, unclear, or high (18).The results are presented in a bias graph.
We assessed the quality of the results using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) method, which primarily encompasses bias, indirectness, inaccuracy, and publication bias.The outcomes are classified into four levels: very low, low, medium, and high (19).

Statistical analysis
The pooled data were assessed using Review Manager 5.3.Hazard ratios (HR) were employed for the analysis of survival data, favoring the PIO group when HR < 1.For dichotomous variables, we used the risk ratio (RR), with results favoring the PP group when RR > 1, particularly in the AE analysis.Conversely, support for the PIO group emerged in the analysis of OSR, EFSR, and drug responses.Heterogeneity was assessed using the I 2 statistic and c2 test.In cases where I 2 was less than 50% or p was greater than 0.1, indicating the absence of significant heterogeneity, we employed a fixed-effects model; otherwise, a random-effects model was utilized.Statistical significance was defined by P values less than 0.05, and we assessed publication bias by visually inspecting funnel plots.

Search results
Three high-quality RCTs (KEYNOTE-671, NADIM II, and AEGEAN) were included in the analysis.The PIO group included 820 patients, and the PP group included 803 patients (Figure 1, Supplementary Figure S1, Supplementary Table S3) (13)(14)(15).These comprised two global multicenter studies (KEYNOTE-671 and AEGEAN) and one study conducted in Spain (NADIM II) (13)(14)(15).As per the GRADE method, the quality of all results was categorized within the medium-high range (Supplementary Table S4).Table 1 provided a summary of the baseline information for the included studies.
In the neoadjuvant treatment phase, total AEs, grade 3-5 AEs, serious AEs, and fatal AEs tended to favor the PP group without a significant difference (Table 2, Supplementary Figure S7).More cases of rash, pruritus, increased alanine aminotransferase, hypothyroidism, and pneumonitis were found in the PIO group   Forest plots of overall survival and event-free survival associated with perioperative immunotherapy versus perioperative placebo.

A B D C
Comparisons of overall survival rate (6-48 months, A: trend of overall survival rate; C: trend of risk ratios) and event-free survival rate (6-48 months, B: trend of event-free survival rate; D: trend of risk ratios) associated with perioperative immunotherapy versus perioperative placebo according to survival time.(Supplementary Table S5).There was no significant difference in the incidence of all grade 3-5 adverse events between the two groups in the neoadjuvant treatment phase (Supplementary Table S6).
In the surgical treatment phase, total AEs, grade 3-5 AEs, serious AEs, and fatal AEs tended to favor the PP group without a significant difference.PIO treatment was associated with more ALTD (4.79% vs. 1.75%,RR: 2.73 [1.16, 6.43]) (Table 2, Supplementary Figure S8).More diarrhea of any grade was found in the PIO group (Supplementary Table S7).There was no significant difference in the incidence of all grade 3-5 adverse events between the two groups in the surgical treatment phase (Supplementary Table S8).
In the adjuvant treatment phase, PIO treatment resulted in a greater incidence of total AEs (40.09% vs. 20.51%,RR: 1.97 [1.58, 2.46]) and grade 3-5 AEs (7.30% vs. 3.75%, RR: 1.95 [1.06, 3.58]).Serious AEs and fatal AEs tended to favor the PP group, but the difference was not significant (Table 2, Supplementary Figure S9).More grade pruritus, rash, and hypothyroidism were found in the PIO group (Supplementary Table S9).There was no significant difference in the incidence of all grade 3-5 adverse events between the two groups in the adjuvant treatment phase (Supplementary Table S10).

Sensitivity analysis
Analysis of ORR, surgery rate, and R0 resection rate revealed significant heterogeneity.Excluding any study did not affect the Subgroup analysis of event-free survival.Forest plots of pathological responses (objective response rate, pathological complete response, and major pathological response) associated with perioperative immunotherapy versus perioperative placebo according to survival time.stability or reliability of the results, as indicated by the sensitivity analysis (Supplementary Figure S10).

Discussion
Resectable stage II-III NSCLC cases can have improved outcomes if neoadjuvant and/or adjuvant treatment is given in addition to surgery (20)(21)(22).However, although traditional PBC can improve patient survival, it is very limited (23,24).In recent years, the introduction of immunotherapy in neoadjuvant therapy and adjuvant therapy for resectable NSCLC has brought new hope to the long-term survival of these patients (9)(10)(11)(12)(13)(14)(15).This study represents the first meta-analysis analyzing the perioperative use (neoadjuvant +adjuvant) of immunotherapy for stage II-III NSCLC based on RCTs.The results suggested that PIO exhibited superior efficacy in OS, EFS, ORR, PCR, MPR, R0 resection rate, and rate of adjuvant treatment compared with PP.In safety assessment, more serious AEs and ALTD were found in the PIO group.
Neoadjuvant immunotherapy may have improved survival benefits, although a direct comparative randomized trial would need to be conducted to determine this (30,31).Therefore, the pathological response and its impact on surgical treatment are crucial indicators for evaluating drug efficacy.In summary, the ORR, PCR and MPR were 51.46%, 19.02% and 32.44% in the PIO Funnel plots of survival summary (A), pathological responses (B), adverse events' summary during all treatment phase (C), adverse events' summary during the neoadjuvant treatment phase (D), adverse events' summary during the surgical treatment phase (E), adverse events' summary during the adjuvant treatment phase (F) associated with perioperative immunotherapy versus perioperative placebo according to survival time.
group, which was similar to the results of NADIM study and SAKK 16/14 study (32,33) Safety is another concern in the perioperative and long-term use of immunotherapy after surgery.The IMpower010 trial reported that Atezolizumab-related adverse events leading to hospitalization occurred in 7% of the surgery groups (34).In clinical practice, although the incidence of AEs in immunotherapy is often much lower than that in chemotherapy, immune related AEs (such as pneumonitis, myocarditis, etc.) are often challenging to manage and can substantially impact the quality of life (35).At different periods of this study, it was observed that the incidence of total AEs, grade 3-5 AEs, serious AEs, and fatal AEs was higher in the PIO group than in the PP group in varying degrees, especially during the neoadjuvant treatment phase.In this phase, the top 5 AEs in the PIO group were nausea (41.15%), anemia (36.17%), neutrophil count decreased (30.28%), constipation (26.87%), and fatigue (23.17%), similar to those in the PP group.These common AEs are often associated with chemotherapy (13).The incidences of rash, pruritus, alanine aminotransferase increased, hypothyroidism, and pneumonitis were significant higher in the PIO group.These significantly increased AEs are often associated with immunotherapy (36).Therefore, although PIO can substantially improve survival, the monitoring and treatment of AEs at different phases still requires close attention.
This meta-analysis has limitations.Firstly, the inclusion of only English articles may introduce language bias.Secondly, including only 3 RCTs may reduce the overall clinical value.Thirdly, all the data analyzed were extracted from previously published articles, leading to increased data heterogeneity.Fourthly, the absence of individual patient data prevented a meta-analysis at the patient level, potentially decreasing the clinical value.Fifthly, variations in median follow-up times across studies might contribute to increased data heterogeneity.

Conclusion
PIO appears superior to PP for resectable stage II-III NSCLC, exhibiting better survival (OS and EFS) and improved pathological responses.Survival tended to favor the PIO group across almost all subgroups.Additionally, PIO demonstrated an increased advantage in survival compared to PP with longer follow up and increased PD-L1 expression.However, the higher rate of AEs in the PIO group warrants serious consideration.organizations, or those of the publisher, the editors and the reviewers.Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

TABLE 1
Characteristics of the three randomized controlled trials (KEYNOTE-671, NADIM II and AEGEAN).

TABLE 2
Summary of adverse events.