Epidermal growth factor receptor (EGFR) mutations are major genetic variants reported in lung adenocarcinomas, with reported incidences of approximately 50% in Asians and 10–15% in Caucasians [1, 2]. Lung cancer patients with EGFR mutations tend to have little or no smoking history. EGFR-tyrosine kinase inhibitors (TKIs) have been successfully developed for EGFR-mutated non-small cell lung cancer (NSCLC) and continue to stand as robust first-line treatments for advanced NSCLC harboring EGFR mutations [3–6]. Despite their efficacy, resistance to EGFR-TKIs occurs in almost all patients [7]. However, optimal treatment strategies have not yet been established.

For patients who have developed resistance to EGFR-TKIs, treatment strategies based on resistance mechanisms are currently under investigation. However, at present, for patients with any EGFR mutation who have progressed on EGFR-TKIs with no T790M mutation or whose disease has progressed on osimertinib, treatment based on the nondriver mutation guideline may be offered [8].

Unlike cytotoxic anticancer drugs, immune checkpoint inhibitor (ICI) therapy is an attractive option due to its potential for a durable response. However, EGFR-mutated NSCLC may be at a disadvantage for immunotherapy because of the lack of smoking history and low tumor mutation burden (TMB). Nevertheless, programmed cell death ligand 1 (PD-L1) expression has been reported to be increased in EGFR-mutant NSCLC [9–11]. Furthermore, in preclinical models, programmed death 1 (PD-1) antibody blockade improved survival in mice with EGFR-driven adenocarcinoma by enhancing effector T cell function and reducing the levels of tumor-promoting cytokines [12], suggesting that ICI treatment for EGFR-mutated cases may be beneficial. Therefore, many clinical trials have been conducted to investigate the efficacy of ICI therapy in patients with NSCLC with EGFR mutations.

However, several clinical trials have shown that the efficacy of ICI therapy in patients with EGFR mutations is limited, and there are concerns regarding its toxicity. Pneumonitis has been reported to be enhanced by the concomitant or sequential use of ICIs [13–15]. Therefore, ICI therapy for patients with EGFR mutations is challenging. Recent data show a trend toward the addition of vascular endothelial growth factor (VEGF) being effective in EGFR-mutated cases; however, this is not conclusive.

In addition, ICIs have recently been introduced for locally advanced NSCLC and advanced stages [16–18]. However, whether immune-oncology (IO) should be administered to patients with EGFR mutations at this stage needs to be considered. Thus, summarizing and discussing the findings regarding ICI treatment for patients with EGFR mutations may help us to consider whether patients with EGFR-mutated locally advanced NSCLC should be treated with ICI. In this review, we summarize the treatments, including ICI, and consider whether they are necessary for patients or not.

Although EGFR-TKIs are the first-line treatment for EGFR-mutated NSCLC, due to their impressive clinical efficacy, almost all patients develop resistance to EGFR-TKIs [7, 19]. Several mechanisms have been reported for the acquired resistance to EGFR-TKIs. One example is the T790M second mutation, which is a resistance mechanism against first- and second-generation EGFR-TKIs. T790M is reported in approximately 50% of patients who acquire resistance to first-/second-generation EGFR-TKIs [20, 21]. Osimertinib, a third-generation EGFR-TKI developed to overcome this resistance, has shown efficacy against T790M-positive EGFR-mutated NSCLC and has been approved as a second-line treatment [22]. However, osimertinib was superior to first-generation EGFR-TKIs in the FLAURA trial and has been approved as a first-line therapy in many countries (in certain countries, only as second-line therapy). The mechanisms underlying osimertinib resistance are more diverse than those of the first or second generation [23–25]. Typical examples include MET amplification, C797S mutations, and signaling pathways other than EGFR, such as RET [23, 26]. Many attempts have been made to overcome these mutations as a strategy for osimertinib resistance, including the ORCHARD trial, which used an adaptation strategy for each resistance mutation, and patritumab deruxtecan (HER3-Dxd) and amivantamab plus lasertinib combination therapies, which target broad resistance [27–29]. Prolonged overall survival (OS) has been reported in patients who received platinum-doublet chemotherapy and EGFR-TKIs [30, 31]. As resistance to targeted therapy is expected to develop at a certain point, cytotoxic chemotherapy will continue to hold a prominent position in the treatment of EGFR-mutated NSCLC.

Furthermore, it has been reported that cancers generally become more heterogeneous during the disease [32, 33]. Therefore, the need for broader strategies, such as cytotoxic chemotherapy or immunotherapy, may strengthen after the first and second treatments fail.

The KEYNOTE-001 study was the first attempt to investigate the efficacy of ICIs as a first-line treatment for patients with EGFR-mutated NSCLC. The objective response rate (ORR) to pembrolizumab in four EGFR-TKI-naive patients was 50%, with a median progression free survival (mPFS) of 157.5 days and a median OS (mOS) of 559 days. By contrast, the efficacy was limited in 26 patients previously treated with EGFR-TKIs (ORR, 4%; mPFS, 56 days; mOS, 120 days) [34]. These results led to a phase II trial of pembrolizumab in TKI-naive patients expressing PD-L1. However, the interim analysis did not show the efficacy of pembrolizumab in this population; the ORR in the first 10 patients was 0% and the study was terminated early and considered invalid [35]. CheckMate012 reported that nivolumab monotherapy was less effective in patients with EGFR mutations than those with EGFR wild-type (ORR: 14% vs. 30%; mPFS: 1.8 months vs. 8.8 months; mOS: 18.8 months vs. not reached) [36]. Furthermore, in a meta-analysis of phase II/III trials comparing ICIs (nivolumab, pembrolizumab, and, atezolizumab) with docetaxel in the second-line treatment of NSCLC, the OS of ICIs vs. digital therapeutics (DTX) in patients with EGFR mutations was hazard ratio (HR) 1.11 [95% confidence intervals (CI): 0.80–1.53, P = 0.54], indicating that although efficacy has been demonstrated in the overall population, treatment with ICIs is not superior in patients with EGFR mutation-positive [37].

Compared to platinum doublet, in a phase II study (WJOG8515L) comparing carboplatin (CBDCA) plus pemetrexed and nivolumab as second-line therapy after EGFR-TKI failure, nivolumab showed a shorter PFS than CBDCA plus pemetrexed and did not show a survival benefit [38]. In a retrospective analysis (an immunotarget study) investigating the efficacy of ICI monotherapy for each driver gene mutation, the ORR of ICI was 12.0% and the PFS was only 2.1 months (95% CI: 1.8–2.7) in patients with EGFR mutations [39]. The BIRCH and ATRANTIC trials investigated ICI monotherapy in patients with EGFR-mutated PD-L1. In both trials, ICI monotherapy was less effective in EGFR-mutated cases than in wild-type NSCLC [40, 41].

These results indicate that treatment with ICIs is effective in the overall population, but not in patients with EGFR mutations.

The data from these trials (Table 1) suggests that EGFR-mutant NSCLC is less effective than ICI monotherapy. In addition, in a retrospective study, we showed that high PD-L1 expression might not be a predictor of response in patients with EGFR/anaplastic lymphoma kinase (ALK) mutations [42]. A few reports indicate that EGFR is immunologically “cold”, and that the tumor microenvironment (TME) is unfavorable to ICI therapy. Tumors that do not elicit a strong immune response and do not usually respond to immunotherapy are called “cold” tumors. These tumors tend to be surrounded by cells that can suppress the immune response, making it difficult for T cells to attack the tumor cells. Therefore, these TMEs may be responsible for the poor efficacy of ICI monotherapy in EGFR-mutated NSCLC. These trials are summarized in Table 1.

The CheckMate012 trial evaluated the combination of nivolumab and erlotinib in 21 patients with EGFR-mutant NSCLC. The PFS of patients previously treated with EGFR-TKIs (n = 20) was 5.1 months. Responders were PD-L1 positive or PD-L1 status unknown. No grade 4 or 5 treatment related adverse events (TRAEs) were reported, and 2/21 patients discontinued the study due to toxicity [45]. In contrast, the TATTON trial, which evaluated a combination of osimertinib and durvalumab, raised serious safety concerns. In this trial, 48% of the patients developed at least one grade 3 TRAE, 5/23 developed interstitial lung disease, and all patients discontinued the trial [14]. Furthermore, the CAURAL trial comparing osimertinib with durvalumab as second-line treatment was stopped early due to the early discontinuation of the TATTON trial reported at the same time, and one patient developed pneumonitis [46]. In addition, the combination of cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and EGFR-TKI for previously treated EGFR-mutant NSCLC was investigated, but most studies were discontinued in the early phase because of low efficacy and toxicity (e.g., tremelimumab and gefitinib [47], ipilimumab and EGFR-TKI or ALK-TKI [44], erlotinib and atezolizumab [48], and pembrolizumab plus gefitinib or erlotinib [49]). These trials are summarized in Table 2.

The KEYNOTE 021 phase I/II study evaluated pembrolizumab plus ipilimumab as second-line or later therapy. In this study, 27% (12/44) of the patients harbored EGFR or ALK, of whom only one showed an objective response [44]. Although several trials have been conducted [43], severe toxicity concerns and limited efficacy were shown.

Several reports have indicated that the TME has a significant impact on the therapeutic effects of ICIs [70–72].

The most representative biomarker is PD-L1 expression, and it has been reported that PD-L1 expression may be upregulated by multiple pathways in EGFR-mutated NSCLC [9]. However, there are no certain opinions on whether PD-L1 expression is high in EGFR-mutated cases, as a few indicate that PD-L1 expression is high, whereas others indicate that it is low [10, 73]. Real-world studies have reported that PD-L1 expression correlates with the response to first-line osimertinib therapy, and PD-L1 expression is associated with prognosis [74].

No consensus has been observed on the importance of PD-L1 expression in EGFR-mutated NSCLC. However, PD-L1 expression does not appear to be a biomarker for ICI treatment in advanced EGFR cases, and there have been few clinically significant results.

TMB was defined as the number of somatic mutations per megabase in the coding region of a tumor. In advanced NSCLC, there is a significant association between smoking history and genetic alterations and TMB [75], and patients with EGFR-mutated NSCLC are less affected by smoking and therefore have fewer somatic mutations and neoantigens [76]. However, it is reported that TMB is not associated with the therapeutic efficacy of PD-1/L1 blockade in patients with driver mutations [77].

In addition, lymphocyte infiltration into the tumor and surrounding stroma is associated with ICI efficacy, and a higher density of CD8+ TILs is associated with a better ICI response. In contrast, Treg infiltration was associated with poor ICI efficacy. Fewer CD8+ TILs and more Tregs were observed in EGFR-mutated mutations. EGFR-mutated NSCLC is a cold tumor with a non-inflammatory microenvironment. However, the high prevalence of Treg infiltration, which is usually observed in inflammatory microenvironments, is unique. Tregs are induced by EGFR [78].

Various soluble molecules have been reported to interact with EGFR. For example, interleukin-6 (IL6) is reported to be overexpressed in EGFR-mutated mouse models [79], and transforming growth factor-beta (TGF-β) and tumor necrosis factor (TNF) are reported to be increased by EGFR expression [80, 81].

Another potential influence on the TME is the effect of previous EGFR-TKI treatment. Several studies have indicated that EGFR inhibition affects TMEs [73, 78, 79, 82]. Reports have shown that EGFR inhibition improves the TME [78, 82], but the combination of EGFR-TKIs and ICI has shown less clinical benefit in clinical trials.

Current evidence suggests that ICI therapy for EGFR-positive lung cancer remains inadequate, probably due to the EGFR-specific TME. Various attempts have been made to increase the efficacy of ICI in this population. Combination therapy with CTLA-4 has not shown good results in EGFR cases, whereas the combination of VEGF, chemotherapy, and ICI has shown good results. A better understanding of EGFR-specific TME and consideration of suitable combinations is required to establish treatment strategies, including optimal ICI for this population. However, the evidence currently described is still insufficient for ICI to prolong the prognosis of EGFR-mutated NSCLC, and there is hope for the development of new agents such as ADC drugs and dual antibodies. In addition, adjuvant therapy with PD-L1 inhibitors has been introduced for resectable NSCLC and unresectable stage III NSCLC; however, EGFR-positive NSCLC is unlikely to benefit from ICI, even in these patients, and targeted therapy seems to be the most promising.