Immune Checkpoint Inhibitors in Epidermal Growth Factor Receptor-Tyrosine Kinase Inhibitor–Resistant Chemotherapy-Naïve Advanced Non–Small Cell Lung Cancer: A Meta-Analysis Based on Eight Randomized Trials

Abstract

PURPOSE

The efficacy and safety of combination strategies involving immune checkpoint inhibitors (ICIs) in patients with advanced epidermal growth factor receptor (EGFR)–mutant non–small cell lung cancer (NSCLC) who have developed resistance to EGFR-tyrosine kinase inhibitors (TKIs) remains uncertain.

METHODS

We conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) comparing ICIs combined with chemotherapy with or without antiangiogenic therapy (C/A) versus C/A alone in the treatment of advanced NSCLC after resistance to EGFR-TKIs. We searched databases, including PubMed, Cochrane Library, Embase, Web of Science, and meeting abstracts. Hazard ratios (HRs) and 95% CI for median overall survival (OS) and median progression-free survival (PFS) were calculated. Risk ratios (RRs) and 95% CI were used as indicators of objective response rate (ORR) and adverse events (AEs).

RESULTS

Eight RCTs involving 10 cohorts and 2,269 patients were included. Adding ICIs to C/A significantly improved PFS (HR, 0.67 [95% CI, 0.57 to 0.80]; P < .001), OS (HR, 0.89 [95% CI, 0.79 to 0.99]; P = .031), and ORR (RR, 0.80 [95% CI, 0.74 to 0.88]; P < .001) comparedwith C/A alone. Subgroup analyses showed that the benefits were more pronounced in patients with PD-L1 expression ≥50%, specific EGFR mutations (Leu858Arg), absence of Thr790Met mutation, and treatment with pemetrexed-platinum. No significant increase in grade 3 or higher AEs was observed, but rates of discontinuation and specific AEs (rash, hypothyroidism, and hypertension) were significantly higher in the ICI+C/A group.

CONCLUSION

This meta-analysis suggests that the addition of ICIs to C/A may improve survival outcomes in patients with advanced NSCLC after resistance to EGFR-TKIs, particularly in selected subpopulations such as those with high PD-L1 expression or specific EGFR mutations. However, careful monitoring for specific AEs is warranted.

INTRODUCTION

Lung cancer is the leading cause of cancer-related mortality worldwide.^1,2 Non–small cell lung cancer (NSCLC) accounts for approximately 85% of all lung cancer cases, with a 5-year survival rate of <10% in advanced stages.^3,4 Targeted therapy with epidermal growth factor receptor (EGFR)-tyrosine kinase inhibitors (TKIs) has become a cornerstone treatment for NSCLC patients with EGFR-activating mutations. EGFR mutation is present in approximately 60%-80% of actionable genetic alterations (AGAs) found in patients with NSCLC.^5-7 However, resistance to EGFR-TKIs typically develops within 1-2 years. At this stage, the standard of care (SOC)—platinum-based doublet chemotherapy with or without antivascular endothelial growth factor (VEGF) agents—shows limited efficacy, thereby underscoring the urgent need for more effective therapeutic strategies.^8,9

In recent years, immune checkpoint inhibitors (ICIs) targeting PD-1 or its PD-L1, and cytotoxic T-lymphocyte antigen 4 have become the SOC for patients with advanced NSCLC, particularly for those without AGAs. These therapies have significantly prolonged the survival of this patient population.^10,11 However, the efficacy of ICIs in patients with AGAs is generally limited, especially in patients with EGFR-mutant NSCLC during initial treatment.¹² Notably, resistance to TKIs can alter the tumor immune microenvironment (TIME),¹³ potentially enhancing the efficacy of ICIs by increasing PD-L1 expression and tumor-infiltrating lymphocyte density.^14-16 However, evidences regarding the impact of TKI resistance on ICI efficacy remains inconclusive.

Some randomized controlled trials (RCTs) have not shown a clear advantage of ICI monotherapy over chemotherapy in patients who have developed resistance to TKIs.^14,17-19 However, the Checkmate 722 and KEYNOTE-789 studies suggested a trend toward improved progression-free survival (PFS) with ICI plus chemotherapy (chemoimmunotherapy) compared with chemotherapy alone in patients with EGFR-TKI–resistant NSCLC.^20,21 The IMpower150 trial, which combined ICI with chemotherapy and anti-VEGF agents (ICI+CA), showed promising results for PFS and overall survival (OS) in patients with EGFR-TKI–resistant NSCLC.^22,23 Similar benefits were observed in the ORIENT-31 trial.²⁴ Recent clinical trials, including IMpower151, ATTLAS, and HARMONi-A, have shown mixed results regarding the efficacy of ICIs in combination regimens.^25-27 ATTLAS and HARMONi-A demonstrated the beneficial effects of simultaneous blockade of immune checkpoints and VEGF pathways, whether used in combination or as two-in-one bispecific antibodies. However, IMpower151 was unable to replicate these findings.

Given these discrepancies, a meta-analysis focusing on prospective RCTs is warranted to assess the benefits of adding ICIs or ICIs plus anti-VEGF to the initial treatment after resistance to TKIs. This analysis aimed to identify subgroups of patients most likely to benefit, refine combination therapy strategies, and evaluate safety profiles to optimize treatment approaches for patients with EGFR-TKI–resistant NSCLC.

METHODS

Search Strategy

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidance (Data Supplement, Table S1). We conducted a comprehensive electronic search of the PubMed, Cochrane Library, Embase, and Web of Science databases to identify trials published on or before June 30, 2024. In addition, we reviewed conference abstracts from the World Conference on Lung Cancer, American Society of Clinical Oncology, and European Society for Medical Oncology. The search strategy used the following terms: “non-small cell lung cancer,” “NSCLC,” “epidermal growth factor receptor,” “EGFR,” “immune checkpoint inhibitors,” “ICI,” “PD-1,” “PD-L1,” “tyrosine kinase inhibitor,” “TKI,” “randomized controlled trial,” and “RCT.” Two authors independently screened titles, abstracts, and full-text articles.

Eligibility Criteria

The inclusion criteria for eligible trials were as follows: (1) RCT design; (2) enrollment of advanced NSCLC patients with positive EGFR mutation status who had previously received EGFR-TKI treatment and experienced disease progression; (3) treatment of the intervention arm with any PD-1/PD-L1 inhibitor in combination with other antitumor agents; (4) no treatment with PD-1/PD-L1 inhibitor in the control arm; (5) chemotherapy-naïve status and no administration of other antitumor therapies after disease progression on EGFR-TKI; and (6) inclusion of clinical outcomes such as PFS, OS, objective response rate (ORR), or adverse events (AEs). Exclusion criteria included review articles, case reports, opinions, editorials, duplicate publications, studies not related to the research topic, and studies with insufficient data for meta-analysis.

Data Collection and Quality Assessment

The study characteristics extracted included authors, year of publication, country, patient demographics, sample size, and end points. The end points included PFS, OS, ORR, and safety. The quality of the studies was assessed using the Cochrane Collaboration's tool to evaluate the risk of bias.²⁸ Quality assessment criteria included the generation of the allocation sequence, concealment of allocation, blinding, use of intention-to-treat analysis, and the proportion of patients lost to follow-up. Two independent reviewers screened the studies, extracted the data, and performed the quality evaluations. Disagreements were resolved by discussion and consensus among the participating researchers.

Statistical Analysis

PFS and OS are expressed as hazard ratios (HRs) with 95% CIs. ORR and AEs were expressed as risk ratios (RRs) with 95% CIs. Publication bias was assessed using funnel plots, Begg's test, and Egger's test. All statistical tests were two-sided, and statistical significance was defined as P < .05. Heterogeneity between studies was evaluated using the Q test and the I² statistic. An I² statistic >50% or a P value <.10 indicated significant heterogeneity. A fixed-effects model was used when there was no significant heterogeneity between studies. In cases where significant heterogeneity was found, a random-effects model was used. The meta-analysis was performed using STATA software, version 12.0 (StataCorp, College Station, TX).

RESULTS

Study Selection and Characteristics

A total of 1,059 records were screened from databases and conference abstracts. After titles and abstracts review, 32 articles were identified for further evaluation. Finally, eight studies, comprising 10 cohorts and a total of 2,269 patients, met the eligibility criteria and were included in the meta-analysis (Fig 1). Among them, six cohorts compared the addition of ICI to SOC (platinum-based doublet chemotherapy ± anti-VEGF agents; ICI+SOC) versus SOC alone. Three cohorts compared the combination of ICI, anti-VEGF agents, and SOC (ICI + anti-VEGF + SOC) versus SOC alone. Anti-VEGF therapies were administered in five trials, including one trial that used ivonescimab, a bispecific antibody targeting both PD-1 and VEGF. Chemotherapy regimens administered included pemetrexed in five trials and paclitaxel in three trials. Detailed characteristics of the included trials are summarized in Table 1.

FIG 1.

PRISMA flowchart outlining the process of study selection. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

TABLE 1.

Characteristics of Included Studies

Study	Authors	Recruitment Period	Sample Size	Intervention Arm (n)	Control Arm (n)	East Asian (%)	Age, Years (median, range)	Female (%)	Smoker (%)	Median PFS (months)	Median OS (months)	ORR (%)
Checkmate 72220	Mok et al	2017-2020	294	Nivolumab + pemetrexed + platinum (144)	Placebo + pemetrexed + platinum (150)	94.40	64 (29-80)	60.20	38.10	5.6 5.4	19.4 15.9	31% 27%
KEYNOTE-78921	Yang et al	2018-2023	492	Pembrolizumab + pemetrexed + platinum (245)	Placebo + pemetrexed + platinum (247)	60.90	63 (34-87)	61.60	33.90	5.6 5.5	15.9 14.7	29% 27.1%
ORIENT-3124	Lu et al	2019-2022	476	(A) Sintilimab + IBI305 + pemetrexed + cisplatin (158)	(C) Placebo + pemetrexed + cisplatin (160)	100	57 (51-65)	59.24	29.80	(A) 7.2 (B) 5.5 (C) 4.3	(A) 21.1 (B) 20.5 (C) 19.2	(A) 48.1% (B) 34.8% (C) 29.4%
				(B) Sintilimab + pemetrexed + cisplatin (158)
IMpower15023	Reck et al	2015-2016	78	(A) Atezolizumab + bevacizumab + paclitaxel + carboplatin (22)	(C) Placebo + bevacizumab + paclitaxel + carboplatin (28)	40.30	63 (31-82)	54.83	44.35	(A) 9.7 (B) 5.7 (C) 6.1	(A) 27.8 (B) 14.9 (C) 18.1	(A) 71% (B) 36% (C) 42%
				(B) Atezolizumab + paclitaxel + carboplatin (28)
IMpower15125	Zhou et al	2020-2023	163	Atezolizumab + bevacizumab + pemetrexed (98%)/paclitaxel (2%) + carboplatin (81)	Placebo + bevacizumab + pemetrexed (98%)/paclitaxel + carboplatin (82)	100	NA	NA	NA	8.5 8.3	NA	NA
ATTLAS26	Park et al	2019-2022	212	Atezolizumab + bevacizumab + paclitaxel + carboplatin (144)	Pemetrexed + platinum (68)	100	62 (31-90)	56.16	37.28	8.7 5.6	20.6 20.3	69.5% 41.9%
IMpower13029	West et al	2015-2017	44	Atezolizumab + nab-paclitaxel + carboplatin (32)	Nab-paclitaxel + carboplatin (12)	NA	NA	NA	NA	7.0 6.0	14.4 10.0	NA
HARMONi-A27	Zhang et al	2022-2022	322	Ivonescimab + pemetrexed + carboplatin (161)	Placebo + pemetrexed + carboplatin (161)	100	59 (32-75)	51.55	29.50	7.0 4.8	17.1 14.5	50.6% 35.4%

Abbreviations: NA, not available; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.

PFS Comparison Between ICI+C/A and C/A

Ten cohorts from eight trials, comprising 2,269 patients, were analyzed to compare PFS between ICI+C/A and C/A. The ICI+C/A group showed a significant PFS benefit over the C/A group, with a pooled HR of 0.67 (95% CI, 0.57 to 0.80, P < .001; Figs 2A and 2B). Significant heterogeneity was observed among the studies (I² = 61.2%, P = .006; Figs 2A and 2B).

FIG 2.

Forest plots of HR for PFS between ICI+C/A and C/A: (A) stratified by the addition of ICI or ICI+anti-VEGF; and (B) stratified by chemotherapeutic agents. Forest plots of HR for OS between ICI+C/A and C/A: (C) stratified by the addition of ICI or ICI+anti-VEGF; and (D) stratified by chemotherapeutic agents. Forest plots of RR for ORR between ICI+C/A and C/A: (E) stratified by the addition of ICI or ICI+anti-VEGF; and (F) stratified by chemotherapeutic agents. C/A, chemotherapy and/or antiangiogenic therapy; DL, DerSimonian-Laird; HR, hazard ratio; ICIs, immune checkpoint inhibitors; IV, Inverse-Variance; MH, Mantel-Haenszel; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; RR, risk ratio; SOC, standard of care; VEGF, vascular endothelial growth factor.

PFS Benefit of ICI+SOC Subgroup and ICI+Anti-VEGF+SOC Subgroup

In subgroup analyses, the ICI+SOC group demonstrated a significant PFS benefit over the SOC group, with a pooled HR of 0.76 (95% CI, 0.67 to 0.86, P < .001; Fig 2A), and the ICI+anti-VEGF+SOC group also demonstrated a significant PFS benefit over the SOC group, with a pooled HR of 0.51 (95% CI, 0.43 to 0.61, P < .001; Fig 2A).

PFS Benefit of ICI With Different Chemotherapeutic Agents

In the pemetrexed-platinum subgroup, the ICI+C/A group showed a significant PFS benefit compared with the C/A group, with a pooled HR of 0.67 (95% CI, 0.55 to 0.81, P < .001; Fig 2B). By contrast, no significant differences were observed between the groups in the paclitaxel-platinum subgroup, with a pooled HR of 0.70 (95% CI, 0.46 to 1.06, P = .094; Fig 2B). Notably, the point estimates for both chemotherapy types were comparable, and the interaction test for subgroup differences was nonsignificant (P = .73). Additionally, heterogeneity across studies was moderate (I² = 61.2%).

PFS Benefit by Patient Characteristics

On the basis of different clinical characteristics, we performed additional subgroup analyses, including age, sex, Eastern Cooperative Oncology Group performance status (ECOG PS), smoking history, presence of brain or liver metastases, type of EGFR mutation, presence of the T790M mutation, PD-L1 expression levels, and previous lines of EGFR-TKI therapy. We found that ICI combination therapy provided a significant PFS benefit over non-ICI treatment, regardless of age, sex, ECOG PS, smoking status, or brain metastases. However, in subgroups with PD-L1 expression <1%, liver metastases, EGFR exon 19 deletion, presence of T790M mutation, and two previous lines of EGFR-TKIs, the addition of ICI did not confer a PFS benefit. By contrast, in subgroups with PD-L1 expression ≥50%, EGFR Leu858Arg mutation, absence of T790M mutation, and one previous line of EGFR-TKIs, a significant PFS benefit was observed with ICI+C/A compared with C/A (Fig 3).

FIG 3.

Subgroup forest plots of HR for PFS between ICI+C/A and C/A groups. C/A, chemotherapy and/or anti-angiogenic therapy; ECOG PS, Eastern Cooperative Oncology Group performance status; EGFR, epidermal growth factor receptor; HR, hazard ratio; ICI, immune checkpoint inhibitors; PFS, progression-free survival; TKI, tyrosine kinase inhibitor.

OS Comparison Between ICI+C/A and C/A

Nine cohorts from seven trials, comprising 2,106 patients, were analyzed to compare OS between ICI+C/A and C/A. The ICI+C/A group showed a significant OS benefit over the C/A group, with a pooled HR of 0.89 (95% CI, 0.79 to 0.99, P = .031; Figs 2C and 2D). No significant heterogeneity was observed between the studies (I² = 0.00%, P = .900; Figs 2C and 2D).

OS Benefit of ICI+SOC Subgroup and ICI+Anti-VEGF+SOC Subgroup

In subgroup analyses, the ICI+SOC group demonstrated a significant OS benefit over the SOC alone group, with a pooled HR of 0.86 (95% CI, 0.75 to 0.99, P = .032; Fig 2C). By contrast, the ICI+anti-VEGF+SOC group showed a trend toward improved OS over the SOC group, but the difference was not statistically significant (HR, 0.91 [95% CI, 0.75 to 1.10]; P = .315; Fig 2C).

OS Benefit of ICI With Different Chemotherapeutic Agents

In the pemetrexed subgroup, the ICI+C/A group showed significantly better OS than the C/A group, with a pooled HR of 0.87 (95% CI, 0.77 to 0.98, P = .021; Fig 2D). By contrast, there was no significant difference between the two groups in the paclitaxel subgroup, with a pooled HR of 0.99 (95% CI, 0.75 to 1.30, P = .917; Fig 2D).

ORR Comparison Between ICI+C/A and C/A

Eight cohorts from six trials, comprising 2,062 patients, were analyzed to compare the ORR between ICI+C/A and C/A. The ICI+C/A group showed significantly better ORR than the C/A group, with a pooled risk ratio (RR) of 0.80 (95% CI, 0.74 to 0.88, P < .001; Figs 2E and 2F). No heterogeneity was observed between the studies (I² = 47.6%, P = .064; Figs 2E and 2F).

ORR Benefit of ICI+SOC Subgroup and ICI+Anti-VEGF+SOC Subgroup

In subgroup analyses, the ICI+SOC group demonstrated no statistically significant ORR benefit over the SOC group, with a pooled HR of 0.89 (95% CI, 0.79 to 1.01, P = .075; Fig 2E), while the ICI+anti-VEGF+SOC group demonstrated a significant ORR benefit over the SOC group, with a pooled HR of 0.70 (95% CI, 0.62 to 0.79, P < .001; Fig 2E).

ORR Benefit of ICI With Different Chemotherapeutic Agents

In the pemetrexed subgroup, the ICI+C/A group showed a significantly higher ORR than the C/A group, with a pooled RR of 0.82 (95% CI, 0.72 to 0.94, P = .004; Fig 2F). By contrast, in the paclitaxel subgroup, there was no statistically significant difference in ORR between the two groups, with a pooled RR of 0.73 (95% CI, 0.49 to 1.11, P = .140; Fig 2F).

Safety

To evaluate safety profiles, treatment-emergent AEs (TEAEs), treatment-related AEs (TRAEs), grade ≥3 TEAEs or TRAEs, death-related TEAEs or TRAEs, TEAEs or TRAEs leading to discontinuation, and specific AEs (anemia, rash, hypothyroidism, and hypertension) were included. The risks of specific AEs were significantly higher in the ICI+C/A group compared with the C/A group, including rash (RR, 1.41 [95% CI, 1.23 to 1.62]; P < .001), hypothyroidism (RR, 1.62 [95% CI, 1.46 to 1.80]; P < .001), hypertension (RR, 1.57 [95% CI, 1.30 to 1.90]; P < .001), TEAEs leading to discontinuation (RR, 1.42 [95% CI, 1.20 to 1.69]; P < .001), and TRAEs leading to discontinuation (RR, 1.45 [95% CI, 1.31 to 1.61]; P < .001). No significant differences were observed between the two groups for all TEAEs (P = .763), grade ≥3 TEAEs (P = .484), death-related TEAEs (P = .843), all TRAEs (P = .174), grade ≥3 TRAEs (P = .089), death-related TRAEs (P = .559), or anemia (P = .463; Table 2).

TABLE 2.

Meta-Analysis of Adverse Events

Adverse Events	ICI+C/A v C/A
	No. of Studies	No. of Patients		Main Effect		Heterogeneity
		ICI+C/A	C/A	RR (95% CI)	P	I2 (%)	P	Model
All TEAEs	5	874	801	1.10 (0.60 to 2.02)	.763	69.10	.012	Random
TEAEs ≥grade 3	5	874	801	1.06 (0.89 to 1.27)	.484	73.00	.005	Random
TEAEs leading to discontinuation	3	475	481	1.42 (1.20 to 1.69)	<.001	0.00	.540	Fixed
TEAEs related to death	4	720	727	1.04 (0.73 to 1.46)	.843	39.60	.174	Random
All TRAEs	5	854	783	1.24 (0.91 to 1.71)	.174	50.60	.088	Random
TRAEs ≥grade 3	5	854	783	1.15 (0.98 to 1.36)	.089	70.20	.009	Random
TRAEs leading to discontinuation	5	854	783	1.45 (1.31 to 1.61)	<.001	49.20	.069	Fixed
TRAEs related to death	5	854	783	1.13 (0.75 to 1.69)	.559	0.00	.648	Fixed
Rash	3	468	394	1.41 (1.23 to 1.62)	<.001	0.00	.948	Fixed
Hypothyroidism	4	713	640	1.62 (1.46 to 1.80)	<.001	0.00	.571	Fixed
Anemia	5	854	783	1.04 (0.94 to 1.16)	.408	0.00	.591	Fixed
Hypertension	3	468	394	1.57 (1.30 to 1.90)	<.001	62.30	.07	Random

Abbreviations: C/A, chemotherapy with or without antiangiogenic therapy; ICIs, immune checkpoint inhibitors; RR, risk ratio; TEAEs, treatment-emergent adverse events; TRAEs, treatment-related adverse events.

Study Quality and Publication Bias

The risk of bias assessment for the included RCTs was low, indicating that all studies were of high quality (Data Supplement, Table S2). Sensitivity analyses for PFS and OS were performed to assess heterogeneity, and no significant heterogeneity was detected in PFS or OS across the studies (Data Supplement, Figs S1 and S2). To minimize publication bias, strict inclusion criteria were applied to the selected studies, and several methods were used to detect publication bias. Visual inspection of funnel plots did not reveal any significant asymmetry. Additionally, both the Egger linear regression test and Begg rank-correlation test showed no evidence of publication bias (Data Supplement, Figs S3 and S4).

DISCUSSION

To our knowledge, this comprehensive meta-analysis provides the first large-scale evidence demonstrating potential survival benefits of ICI-based regimens in patients with EGFR-mutant NSCLC who have developed resistance to EGFR-TKIs. Unlike previous meta-analyses,^30,31 which included a broader range of study designs, our analysis strictly focused on RCTs that specifically evaluated initial treatment strategies after resistance to EGFR-TKIs. The majority of trials included in this analysis provided updated data from the past 2 years, ensuring the relevance and timeliness of our findings. Our results demonstrate that the addition of ICIs to chemotherapy (with or without antiangiogenic agents) significantly improves PFS and OS in EGFR-TKI–resistant NSCLC, despite inconsistent OS benefits in individual RCTs.

The optimal combination strategy for ICIs was investigated in this meta-analysis. Typically, ICI monotherapy was associated with inferior PFS, OS, and ORR compared with chemotherapy in EGFR-mutant NSCLC, regardless of previous EGFR-TKI exposure.^30,32,33 Therefore, ICI monotherapy is not recommended for these patients. Notably, even in EGFR-mutant patients with high PD-L1 expression, the efficacy of ICI monotherapy is often suboptimal, likely because of poor infiltration of CD8+ T cells and low tumor mutation burden,³⁴ highlighting the limited predictive value of PD-L1 in this population. Chemoimmunotherapy appears effective in overcoming this limitation, with our analysis demonstrating superior efficacy over chemotherapy alone. Preclinical studies suggest this synergy arises from mechanisms such as TIME remodeling, PD-L1 upregulation, immunogenic cell death, neoantigen load increase, and CD8+ T-cell functional restoration.³⁵

In trials like ORIENT-31 and IMpower151, adding antiangiogenic agents to chemoimmunotherapy prolonged PFS. In the subgroup analysis of our meta-analysis, ICI + antiVEGF + SOC demonstrated superior efficacy compared with SOC alone, with significantly improved PFS (HR, 0.51 v HR, 0.76 for ICI + SOC) and ORR benefits. By contrast, ICI + SOC showed no significant ORR advantage over SOC. Our findings suggest that the combination of anti-VEGF agents with ICIs exerts a more pronounced effect on tumor response. Notably, the phase III APPLE trial observed a PFS benefit (9.7 v 5.8 months, HR, 0.67) and a trend toward OS improvement with adding bevacizumab to chemoimmunotherapy in EGFR-TKI-resistant NSCLC.³⁶ Mechanistically, antiangiogenic agents facilitate the enhancement of antigen-specific T-cell trafficking into the tumor microenvironment through the inhibition of VEGF. This modulation not only reduces the presence of immunosuppressive macrophages and monocytes but also potentiates T-cell–mediated cytotoxicity, thereby enhancing the ability of the immune system to target and eliminate cancer cells.³⁷ These findings suggest that, for patients with EGFR-TKI–resistant NSCLC, the synergistic use of ICIs in combination with chemotherapy and antiangiogenic therapies is a particularly promising strategy. This approach appears to be particularly beneficial in scenarios where the tumor burden is significantly elevated or requires rapid reduction.

The choice of chemotherapy backbone (pemetrexed v paclitaxel) in combination with ICIs remains a topic of clinical interest. Although pemetrexed-based regimens showed statistically significant improvements in PFS, OS, and ORR compared with C/A alone, the apparent advantage over paclitaxel should be interpreted cautiously. The HR point estimates for both chemotherapy types were similar (pemetrexed HR, 0.67 v paclitaxel HR, 0.70), and the overlapping confidence intervals, coupled with a nonsignificant interaction test (P = .73), indicate no conclusive evidence of superiority for either regimen. The nonsignificant result for paclitaxel may reflect limited statistical power because of smaller sample sizes in this subgroup (three trials, n = 228 v five trials, n = 1,252 for pemetrexed) rather than a true lack of efficacy.

These findings align with the established role of pemetrexed in lung adenocarcinoma—the predominant histology for EGFR-mutant NSCLC—where its mechanism of action and toxicity profile may synergize better with immunotherapy.^38,39 However, no definitive conclusions can be drawn regarding preferential use of pemetrexed over paclitaxel without direct comparative trials. Future studies should prioritize head-to-head comparisons of chemotherapy backbones in ICI-based regimens, with standardized stratification for PD-L1 expression, EGFR mutation subtypes, and previous therapy lines to minimize confounding.

The efficacy of ICIs is modulated by several factors, including PD-L1 expression level, age, ECOG performance status, metastatic sites, and EGFR mutation status.⁴⁰ In our meta-analysis, stratified assessments of baseline characteristics revealed a notable PFS benefit associated with ICI supplementation, independent of age, sex, ECOG performance status, smoking history, or the presence of brain metastases. However, subgroups that did not show additional benefits from ICIs included those with PD-L1 expression below 1%, those with EGFR exon 19 deletions, T790M mutations, or those who had received two previous EGFR-TKI therapies. Interestingly, patients harboring EGFR L858R mutations refractory to EGFR-TKIs benefited from the ICI+C/A regimen, in contrast to the lack of benefit observed in patients with EGFR exon 19 deletions, suggesting potential differences in the TIME influenced by different EGFR mutations. In addition, as the number of EGFR-TKI lines increases, the subsequent efficacy of ICIs may decrease, potentially explaining the divergent results in relevant RCTs, such as ATTLAS, ORIENT-31, and IMpower151, all of which evaluated the efficacy of ICI+CA. Although ATTLAS and ORIENT-31 showed significant improvement in PFS, IMpower151 did not show comparable results. In particular, none of the three studies showed a PFS benefit in the subgroup of patients who received sequential first- or second-generation EGFR-TKIs, followed by third-generation EGFR-TKI therapy. Conversely, patients treated with single-agent EGFR-TKIs showed a PFS benefit in ATTLAS and ORIENT-31. The difference in the proportion of patients receiving single-line EGFR-TKI therapy (46.1% in IMpower151 v 65.3% in ATTLAS and 73.8% in ORIENT-31) may have contributed to the divergent results of IMpower151.^23,25,26 However, the underlying mechanisms remain unclear and require further investigation.

In terms of safety profiles, the integration of ICIs did not significantly increase the overall TEAEs/TRAEs incidence, grade ≥3 TEAEs/TRAEs, or mortality. However, the ICI+C/A cohort showed higher rates of treatment discontinuation, hypertension, immune-mediated dermatologic reactions, and hypothyroidism, aligning with previous ICI RCTs. Hypertension may be closely associated with concomitant use of antiangiogenic agents. Therefore, it is imperative to evaluate the safety profile of ICIs in combination with antiangiogenic therapies, considering the unique characteristics of patients with resistance to EGFR-TKIs. Previous studies have found that the combination of EGFR-TKIs and ICIs increases the incidence of certain AEs, such as liver injury and interstitial lung disease (ILD).⁴¹ Although sequential EGFR-TKI and ICI use in included studies did not elevate ILD incidence (eg, ORIENT-31: 1%-3% across arms), limited reporting precluded pooled ILD analysis.

Despite the focus of our study on evaluating the efficacy of ICI-based therapeutic approaches in patients with EGFR-TKI–resistant NSCLC, management of these patients remains a major clinical challenge. The emergence of novel agents and evolving treatment paradigms in this setting is encouraging. A recent clinical trial evaluating the combination of amivantamab with chemotherapy, either alone or in combination with lazertinib, showed promising results, demonstrating a significant improvement in PFS compared with chemotherapy alone.⁴² However, this regimen was associated with a significantly increased incidence of TRAEs, particularly grade ≥3 TRAEs.⁴² In addition, antibody-drug conjugates targeting TROP-2 and HER-3 have shown encouraging potential in early clinical trials.^43,44 The resulting challenge for health care providers is to refine and tailor treatment strategies with these innovative agents and ICIs to individual patient characteristics to maximize therapeutic benefits while mitigating AEs.

This study has several limitations. Primarily, efficacy within the PD-L1 expression range of 1%-50% remains equivocal because of insufficient stratified RCT data. Further investigation of factors, such as tumor mutational burden, microsatellite instability, TP53 comutations, and KRAS mutation status, is required. Second, our meta-analysis had a low representation of patients with NSCLC treated with third-generation EGFR-TKIs as the first-line therapy, a scenario that has become increasingly common in recent years. Among the RCTs reviewed, the HARMONi-A trial uniquely evaluated osimertinib and showed tentative PFS benefits and a suggestive OS improvement with the addition of ICI after osimertinib resistance. However, the OS data from this trial remain immature, and updated RCTs are needed to confirm the benefits of ICIs in this subgroup. Finally, different trials used different ICIs, including multiple PD-L1 and PD-1 inhibitors, increasing the possibility of data bias and heterogeneity in the results.

In conclusion, this meta-analysis suggests that the addition of ICIs to C/A may improve survival outcomes in patients with advanced EGFR-mutant NSCLC who have developed resistance to EGFR-TKIs. The observed benefits were particularly pronounced in specific subgroups, including patients with PD-L1 expression ≥50%, EGFR L858R mutation, absence of T790M mutation, and limited previous EGFR-TKI exposure (≤1 line). Regarding the choice of chemotherapy regimens, pemetrexed-based treatments appeared to show improvements compared with C/A alone, but there is no conclusive evidence of superiority over paclitaxel-based regimens because of similar overall effects and potential limitations in sample size. Despite these advances, limitations such as patient selection bias, tumor heterogeneity, and variability in PD-1/PD-L1 inhibitors may influence outcomes. Future studies must validate these findings in diverse populations. Additionally, mechanistic investigations are needed to elucidate how EGFR mutations modulate immunotherapy responsiveness.

Notably, the decision to use ICI + C/A as the next line of therapy should be carefully considered. Subsequent rebiopsy and sequencing after EGFR-TKI therapy often reveals a heterogeneous molecular phenotype, including MET amplification, new KRAS mutations, and SCLC transformation, which may respond better to specific targeted therapies or other treatment strategies. Therefore, it is essential to perform comprehensive genomic profiling after progression on EGFR-TKIs to identify potential targetable mutations or transformations.

In summary, although this meta-analysis supports the potential benefit of adding ICIs to C/A particularly in certain subpopulations of EGFR-TKI–resistant patients, it does not provide sufficient evidence to recommend this combination as a preferred or priority strategy for all patients after EGFR-TKI resistance. Personalized treatment strategies on the basis of detailed molecular characterization and biomarker assessment are crucial for optimizing outcomes in these patients.

DATA SHARING STATEMENT

A data sharing statement provided by the authors is available with this article at DOI https://doi.org/10.1200/PO-24-00907

All data analyzed in this study are available in this published article and supplementary material.

AUTHOR CONTRIBUTIONS

Conception and design: Letian Huang, Jietao Ma, Chengbo Han

Financial support: Letian Huang

Collection and assembly of data: Letian Huang, Shuling Zhang, Li Sun, Chengbo Han

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).

No potential conflicts of interest were reported.