Introduction
Lung cancer is still the leading cause of cancer-related death worldwide [
1]. Non-small cell lung cancer (NSCLC) accounts for ~85% of all lung cancers. Up to 50% of NSCLCs in Eastern Asia and 8-16% of NSCLCs in the West have epidermal growth factor receptor (
EGFR) mutations [
2,
3,
4].
EGFR L858R missense mutation and in-frame exon 19 deletions (19del) are the most common mutations found in
EGFR-mutant NSCLCs, accounting for ~85% of all
EGFR-activating mutations [
3,
4,
5,
6]. NSCLCs with
EGFR L858R or 19del are sensitive to initial EGFR tyrosine kinase inhibitor (TKI) therapy [
7,
8]. Relative to traditional chemotherapy, EGFR-TKI therapy significantly extends progression-free survival (PFS), overall survival (OS), and improved quality of life due to relatively low toxicity for NSCLC patients harboring sensitive
EGFR mutations [
9,
10,
11,
12,
13,
14,
15,
16,
17,
18]. EGFR-TKI therapy is the first-line treatment for NSCLC patients with sensitive
EGFR mutations in the National Comprehensive Cancer Network (NCCN) guidelines [
19]. However, the remission is still transient for only nine months to around two years [
9,
10,
11,
12,
13,
14,
15,
16,
20]. With the universal application of EGFR TKIs to treat the
EGFR-mutant NSCLCs, EGFR-TKI resistance represents an unmet clinical problem. So far, no further effective treatment options have been approved after progression to EGFR TKIs in
EGFR-mutant NSCLCs, most of whom need chemotherapy-consisting therapy. However, the clinical benefit of the chemotherapy-consisting therapy is modest, with a high potential for toxicity.
Other than targeted therapies, immunotherapy with immune checkpoint inhibitors (ICIs), represented by anti-programmed cell death-(ligand) 1 [anti-PD-(L)1] therapy, has demonstrated improved survival rates, long-term survival benefits, and favorable safety profile over chemotherapy in a subset of advanced NSCLCs [
21,
22,
23,
24,
25,
26,
27], leading 6 ICIs (ipilimumab, nivolumab, pembrolizumab, atezolizumab, avelumab, and durvalumab) approved by FDA as first- or second-line treatments for advanced NSCLCs but not suggested for patients with sensitive
EGFR mutations [
28]. Relative to NSCLCs with
EGFR wild-type, ICI therapy fails to improve survival benefits in most
EGFR-mutant NSCLCs [
29,
30,
31,
32,
33,
34]. Clinical benefit from ICI therapy is associated with inflamed tumors with infiltrated immune cells, high tumor mutation burden (TMBs), and high expression of PD-L1 on tumor cells [
35]. However, the oncogenic EGFR signaling in lung cancer with
EGFR mutations contributes to a non-inflamed tumor immune microenvironment (TIME) characterized by limited infiltration of CD8
+ T cells [
36,
37,
38], a high number of infiltrations of regulatory CD4
+ T cells [
38], and a high number of inactivated infiltrated T cells [
39]. Besides,
EGFR-mutant NSCLCs are generally non-smokers [
34] with low PD-L1 expression [
36,
37,
39] and low tumor mutation burden [
36,
37]. The mechanisms mediating durable effects in a small proportion of
EGFR-mutant NSCLCs are worth further investigation [
40].
Hope:
Although osimertinib treatment decreased PD-L1 expression on
EGFR-mutant NSCLC cell lines by reducing PD-L1 at mRNA level and promoting PD-L1 degradation by proteasomes [
41], the median PFS (mPFS) of osimertinib was similar in PD-L1 negative patients (TC<1%) (18.9 months) and in PD-L1 positive patients (TC≥1%) (18.4 months) [
42], suggesting the osimertinib efficacy was independent of PD-L1 status in
EGFR-mutant NSCLCs.
- 2.
Subtypes of EGFR Alterations May be Sensitive to ICI Therapy
ICI efficacy varies in subtypes of
EGFR alterations in
EGFR-mutant NSCLCs.
EGFR wild-type and L858R NSCLCs had a similar outcome, while
EGFR 19del NSCLCs had a worse outcome when treated with ICI therapy [
43]. One of the possible explanations was that
EGFR L858R-mutant tumors were more likely to have significantly increased CD8
+ T cell infiltration than
EGFR 19del tumors [
39]. The findings were also supported by other groups that ICI therapy alone or in combination exhibited improved survival outcomes in the
EGFR L858R group compared with that in the
EGFR 19del group [
44,
45,
46,
47]. Besides,
EGFR T790M-negative patients were more likely to benefit from ICI therapy than
EGFR T790M-positive patients [
48,
49], as the
EGFR T790M-negative tumors had higher PD-L1 expression than the
EGFR T790M-positive tumors [
48,
50]. Other than the common mutations, 10-20 % of
EGFR-mutant NSCLCs harbor non-
EGFR L858R or 19del mutations, known as uncommon mutations, such as S768I and G719X [
5,
37,
51,
52]. Relative to common
EGFR mutations, uncommon
EGFR-mutant tumors were likely to express high PD-L1 (≥50%), high TMB, and high infiltration of CD8
+ T cells [
37,
49,
52,
53]. Further, 36.7% uncommon
EGFR-mutant NSCLCs with dual high PD-L1 expression and abundant CD8
+ T cell infiltration showed a favorable response to ICI therapy [
49,
53].
- 3.
Heavily Pre-Treated EGFR-Mutant NSCLCs are Likely to Respond to ICI Therapy Relative to Treatment-Naïve EGFR-Mutant NSCLCs
EGFR-TKI treatment results in dynamic changes in host immunity. EGFR-TKI-advanced
EGFR-mutant NSCLC patients trended to have concurrent high levels of PD-L1 expression and high number of CD8
+ T cell infiltration [
34]. It was further found that the proportion of
EGFR-mutant NSCLCs with high PD-L1 (≥50%) was increased from 14% to 28% after EGFR-TKI treatment [
54]. Further, heavily pre-treated EGFR-TKI-advanced
EGFR-mutant NSCLCs with high expression of PD-L1 (≥25% of tumor cells) had a better median OS (mOS) (13.3 vs. 9.9 months) than that with low expression of PD-L1 (<25% of tumor cells) after ICI therapy in a Phase II ATLANTIC study [
29,
55]. It was also found that 2 of 5 EGFR-TKI-advanced
EGFR-mutant NSCLCs with increased PD-L1 after EGFR blockade showed a durable response to subsequent ICI therapy [
54]. Overall, EGFR-TKI-advanced
EGFR-mutant NSCLCs may be the groups responding to ICI therapy. However, ICI therapy immediately followed by EGFR TKI therapy may cause immune-related adverse effects (irAEs), as discussed below.
- 4.
IrAEs Observed During ICI Treatment Are Associated with the Efficacy
Even though irAEs are not necessary for treatment benefit, a growing body of literature supports that the occurrence of irAEs during ICI therapy is associated with improved treatment efficacy. A retrospective analysis proved that the 270 metastatic NSCLCs who experienced irAEs after ICI therapy had improved PFS, overall response rate (ORR), disease control rate (DCR), and OS relative to those who did not experience irAEs (PFS: 5.2 vs. 1.97 months; ORR: 22.9% vs. 5.7%; DCR: 76% vs. 58%; OS: not reached vs. 8.21 months) [
56]. Similarly, secondary analysis of the Phase I CA209-003 clinical trial conducted at 13 US medical centers led to findings that the mOS was significantly longer among patients with irAEs of any grade (19.8 months; 95% CI, 13.8-26.9 months) or grade 3 or more (20.3 months; 95% CI, 12.5-44.9 months) compared with those without treatment-related AEs (5.8 months; 95% CI, 4.6-7.8 months) (
p<0.001 for both comparisons based on hazard ratios) [
21]. Another retrospective study found that around 43.6% of 195 advanced NSCLCs treated with ICI therapy experienced irAEs. Significantly longer mPFS (5.7 s. 2.0 months), PFS (8.5 vs. 4.6 months), mOS (17.8 vs. 4.0 months), and OS (26.8 vs. 11.9 months) relative to those who did not experience irAEs was observed [
57]. Besides the retrospective studies, forty-three advanced NSCLCs were enrolled in a prospective study to evaluate the association between clinical outcomes and irAEs after ICI therapy. Consistently, earlier irAEs were associated with better objective response and disease control rates than those without irAEs after ICI therapy (37% vs. 17% and 74% vs. 29%, respectively) [
58]. Another prospective study with 76 advanced NSCLCs also supported that the mPFS was significantly longer when the NSCLCs experienced irAEs within two weeks of beginning ICI than those who did not (5.0 vs. 2.0 months) [
59]. However, it is difficult to predict the treatment-related toxicity.
- 5.
Positive Clinical Studies Support the Use of ICI Therapy in EGFR-Mutant NSCLCs
Even though the ORR for
KRAS-mutant NSCLCs to ICI therapy is around 20%, 7% of
EGFR-mutant NSCLCs respond to single-agent ICI therapy [
40]. A female NSCLC patient harboring
EGFR L858R mutation treated with ICI therapy resulted in a prolonged PFS (~23 months) [
60]. To achieve further improved outcomes, anti-PD-1/PD-L1 antibody-based ICI therapy, combined with other checkpoint inhibitors or platinum-based chemotherapy, has been widely applied in advanced lung cancers [
28,
61]. Yang et al. observed that chemo-immunotherapy was better than immunotherapy alone for EGFR-TKI advanced
EGFR-mutant NSCLCs (mPFS: 3.42 vs. 1.58 months,
p=0.027) [
62]. A retrospective study of 122 EGFR-TKI-advanced
EGFR-mutant NSCLC patients, especially with
EGFR L858R mutation, revealed better mPFS (5.0 vs. 2.2 months) and mOS (14.4 vs. 7 months) in ICI-based combination therapy than that in ICI alone [
45]. Further, front-line ICI therapy reached better survival benefits than later-line ICI therapy [
45]. In another retrospective study, chemo-immunotherapy also achieved better outcomes than ICI therapy alone of mPFS (4.3 vs. 1.5 months), mOS (14.92 vs. 7.41 months) and ORR (23.1% vs. 3.1%) in the EGFR-TKI-advanced
EGFR-mutant NSCLCs especially with
EGFR L858R mutation, respectively [
63]. Consistent with the findings, patients with
EGFR L858R exhibited a trend of longer mPFS (7.6 vs. 5.4 months) and mOS (23.5 vs. 18.0 months) versus patients with
EGFR 19del receiving chemo-immunotherapy [
46].
VEGF is not only associated with NSCLC progression, recurrence, and metastasis; VEGF, complimentarily with EGFR pathways, promotes an immunosuppressive TIME [
64], and anti-angiogenic therapy can reprogram the TIME from immunosuppressive to immune-supportive [
65,
66]. Currently, anti-angiogenic drugs plus chemotherapy are the most common regimen for EGFR-TKI-advanced NSCLCs [
67]. Chemo-immunotherapy combinations achieved a higher objective response rate (ORR) relative to the chemo-anti-angiogenesis combination (29.5% vs. 13%) [
67]. Longer PFS was also associated with previous anti-angiogenic drug applications in patients receiving chemo-immunotherapy [
62]. Chemo-immuno-anti-angiogenesis combination was superior to either chemo-immunotherapy or chemo-anti-angiogenesis combination on mPFS (10.2, 6.9, 6.9 months), and the mOS (29.4, 19.0, 18.1 months) in the Phase III IMpower150 study in a subgroup of EGFR-TKI-advanced
EGFR-mutant NSCLCs [
68], suggesting that the combination of ICI, anti-angiogenesis, and chemotherapy may be a novel therapy for EGFR-TKI-advanced
EGFR-mutant NSCLCs. Chemo-immuno-anti-angiogenesis combination therapy correlated with the highest rate of grade 3/4 treatment-related AEs compared to chemo-immunotherapy and chemo-anti-angiogenesis (66.7%, 42.4%, and 13.6%) [
68]. Like the IMpower150 clinical trial, 40 EGFR-TKI-advanced
EGFR-mutant NSCLC patients received chemo-immuno-anti-angiogenesis therapy enrolled in a Phase II clinical trial, and the patients achieved impressive mPFS (9.4 months), one-year OS (72.5%), and ORR (62.5%) with only 37.5% reported irAEs [
69]. These positive clinical trials may identify novel biomarkers that predict responders to ICI therapy alone or in combination for EGFR-TKI-advanced
EGFR-mutant NSCLCs.
Challenges:
Expression of PD-1 on activated T cells, B cells, and natural killer (NK) cells blunts the immune response through interaction with its major ligand PD-L1, expressed on tumor cells and infiltrating immune cells [
70,
71,
72]. Disruption of the PD-1/PD-L1 interaction reactivates the anti-tumor T cell-mediated cell cytotoxicity. However, researchers do not consistently agree on PD-L1 expression levels in
EGFR-mutant NSCLCs. In earlier studies, activation of the PI3K-AKT, MAPK-ERK, or JAK-STAT3 pathway by aberrant oncogenic EGFR signaling upregulated PD-L1 expression in
EGFR-mutant NSCLC cell lines [
73,
74,
75,
76,
77], and EGFR TKI treatment decreased PD-L1 expression [
73,
74]. Some studies reported no correlation between PD-L1 expression and
EGFR mutation status [
78,
79]. In more recent studies with clinical samples, the
EGFR-mutant NSCLC group expressed significantly lower PD-L1 than the
EGFR wild-type NSCLC group [
80,
81,
82,
83,
84]{Ji, 2016 #18530}{Dong, 2017 #17138}{Li, 2017 #18540}{Takada, 2018 #18462}{Lee, 2019 #18499}. A pooled analysis of 15 public studies further suggested that
EGFR-mutant NSCLCs have a decreased PD-L1 expression [
36]. PD-L1 expression was more accentuated at portions with higher PD-L1 expression in
EGFR-mutant versus
EGFR wild-type group (51% vs. 68% at TC≥1%, 8% vs. 35% at TC≥25% and 5% vs. 28% at TC≥50%) [
42]. Furthermore, EGFR blockade upregulated PD-L1 expression in
EGFR-mutant NSCLCs [
34,
54,
85], supporting that EGFR-TKI-advanced EGFR-mutant NSCLC patients probably benefit from ICI therapy as discussed above.
PD-L1 was a predictive and prognostic biomarker for response to ICI therapy in NSCLCs [
25]. Several clinical trials reported the association between PD-L1 expression and clinical outcomes in NSCLC patients [
25,
33,
86]. However, PD-L1 levels do not consistently indicate ICI response and correlate with prognosis, especially for
EGFR-mutant NSCLCs [
39]. ICI therapy did not show clinical benefit in TKI naïve
EGFR-mutant NSCLCs, even with high PD-L1 expression. A Phase II trial, despite 73% of enrolled NSCLC patients with more than 50% PD-L1 expression, mostly treatment-naïve with sensitive
EGFR mutations, was discontinued after 11 patients due to lack of efficacy PD-L1 [
33,
86]. Patients with low or even undetectable PD-L1 expression also had improved survival with ICI therapy versus chemotherapy [
25]. Cross-comparison is sometimes challenging due to various immunohistochemical (IHC) assays with different scoring systems and cutoff values [
87,
88], making a universal assay for assessing PD-L1 expression with appropriate cutoff points important.
Besides, the disconnection might be because PD-L1 expression could not reflect the underlying T cell activity in
EGFR-mutant NSCLCs in earlier studies [
39]. PD-L1 is also highly expressed in circulating immune cells, such as dendritic cells and myeloid-derived immune suppressor cells [
89,
90]. In contrast, most clinical trials only assess tumoral PD-L1 expression when investigating ICI efficacy. Assessing PD-L1 expression in tumor cells and tumor-infiltrating immune cells in both POPLAR and OAK trials led to the finding that higher PD-L1 levels in both tumor cells and tumor-infiltrating immune cells were associated with improved patient survival after ICI therapy [
25,
91]. Patients with more than 30% of PD-L1
+CD11b
+ myeloid cells before ICI therapy exhibited a 50% superior response rate [
92]. Furthermore, PD-L2, another ligand identified for PD-1 T cell receptor (TCR) [
93], could compete with PD-L1 with 2-6 fold higher affinity to PD-1 [
94]. EGFR signaling could also regulate PD-L2 expression in NSCLCs [
95].
- 2.
TMB is Low in EGFR-Mutant NSCLCs
TMB is the total number of gene alterations, including substitutions, insertions, and deletions, correlated with higher levels of neoantigens. NSCLC patients with high TMB respond better to ICI therapy [
96]. However,
EGFR mutations are associated with decreased tumor mutation burden compared with tumors with
EGFR wild-type [
36,
37,
97]. Lung tumors with
EGFR 19del harbored an even lower tumor mutation burden than
EGFR L858R lung tumors [
43,
97]. Older people have increased TMB [
98].
EGFR 19del was commonly found in the young, while
EGFR L858R predominated in the elderly [
5]. In contrast to NSCLCs harboring common
EGFR mutations, patients with uncommon
EGFR mutations, especially the G719X mutation, showed the highest TMB (7.5 mutations/Mb), followed by
EGFR exon 20 ins (4.6 mutations/Mb),
EGFR T790M (4.05 mutations/Mb),
EGFR L858R (3.4 mutations/Mb), and
EGFR 19del (3.1 mutations/Mb) (p<0.05) [
37]. Similar to PD-L1, high TMB has been detected in responders and non-responders receiving ICI therapy [
99], indicating TMB is not the only determinant for ICI response.
- 3.
EGFR-Mutant NSCLCs are Non-Smokers, Generally
Tobacco is a known risk factor for lung cancer. However, Smokers are more likely to benefit from ICI therapy [
100,
101,
102]. Although
EGFR mutations are more enriched in never-smokers,
EGFR mutations in NSCLCs are also found in ever-smokers [
6,
103,
104,
105]. Compared with common
EGFR mutations, uncommon
EGFR mutations were significantly associated with smoking [
106]. PD-L1 positivity was also associated with smoking history [
80,
81,
82]. These findings further support the above discussed findings that the NSCLCs with uncommon
EGFR mutations are more likely to have a favorable response to ICI therapy than NSCLCs with common
EGFR mutations [
49,
53].
- 4.
EGFR-Mutant NSCLCs Have a Lymphocyte-Depleted TIME
The crosstalk between cancer cells and TIME has become a hot research topic with the rapid development of immunotherapy in cancer, including lung cancer. Based on the presence or absence of tumor-infiltrating lymphocytes (TILs), there are four different types of TIME in tumors: type I: TIL
+PD-L1
+; type II: TIL
-PD-L1
-; type III: TIL
-PD-L1
+; type IV: TIL
+PD-L1
-, and only type I tumors with lymphocyte infiltration and PD-L1 expression can respond to ICI therapy [
107]. To better reflect the complex relationship of the tumor, host, and environmental factors, tumors have also been classified into the following types: the immune-dessert tumor, the immune-excluded tumor, and the inflamed tumor [
108]. The immune-dessert and immune-excluded tumors are naturally resistant to ICI therapy.
EGFR-mutant tumors have a “lymphocyte depletion” phenotype [
34,
36,
38,
109,
110], characterized by a pronounced lack of the infiltration of CD8
+ T cells [
34,
36,
38,
111], suggesting an immunosuppressive TIME. Single-cell RNA sequencing supported the findings that fewer CD8
+ T cells and more macrophages were in the TIME when tumors were resistant to EGFR-TKI treatment than when EGFR TKIs were effective [
112]. Additionally,
EGFR-mutant NSCLC tumors had markedly less crosstalk between T cells and other cell types via the PD-1/PD-L1 pathway than
EGFR-negative NSCLCs [
113].
Regulatory T cells (Tregs), especially the Forkhead box P3 (Foxp3)
+, play essential roles in immune suppression [
114].
EGFR-mutant NSCLC tumors showed high infiltration of Foxp3
+CD4
+ regulatory T cells [
38,
115,
116]. Retrospective immunohistochemistry analysis of 164
EGFR-mutant and 159
EGFR wild-type tumors revealed that the expression of CD3, CD4, and Foxp3 was significantly higher in
EGFR-mutant NSCLC tumors than that in the
EGFR wild-type tumors [
117]. EGFR blockade increased intratumor CD8
+ T cells and decreased Tregs infiltration in the TIME [
38,
118,
119,
120,
121]. Similar as Tregs, myeloid-derived suppressor cells (MDSCs), known to suppress immune response [
122,
123,
124], were also found to be elevated in
EGFR-mutant NSCLC tumors [
121].
Tumors without detectable EGFR expression responded to EGFR inhibition [
125], implicating that EGFR blocking blockade potentially influenced the tumor-specific immune responses. Deficient
Egfr in murine myeloid cells decreased carcinogenesis, suggesting a tumor-promoting function by myeloid cell-intrinsic EGFR signaling [
126]. Macrophages in the TIME express EGFR [
126,
127]. EGF secreted by tumor cells promoted M2 polarization of tumor-associated macrophages (TAMs), associated with suppression of cytotoxic T cell function [
128]. In contrast to M2-type TAMs, higher infiltrated M1-type TAMs found in NSCLC with uncommon
EGFR mutations (G719X and exon 20s) correlated with longer PFS than common
EGFR mutations [
37].
EGFR-mutant NSCLC cell lines significantly downregulated MHC class I molecule expression compared with the
EGFR wild-type NSCLC cell lines in response to IFNγ [
129]. PI3K-AKT and MAPK pathways, the primary downstream signaling pathways of EGFR {Yan, 2018 #7245}{Yan, 2022 #15218}, also suppressed MHC class I molecule expression [
129,
130,
131,
132,
133], suggesting the downregulation of MHC class I molecules in
EGFR-mutant NSCLCs is through abnormal EGFR signaling. Indeed, EGFR inhibition augmented the expression of MHC class I molecules [
134,
135,
136].
- 5.
EGFR-Mutant NSCLCs Respond Poorly to ICI Therapy Alone or in Combination
Unlike those without driver mutations,
EGFR-mutant NSCLCs generally respond poorly to ICI therapy [
30,
34]. Subgroup analysis of the patients with activating
EGFR mutations in the first Phase III clinical trial of CheckMate-057 revealed no PFS and OS benefit from ICI treatment [
23]. A retrospective analysis revealed that
EGFR mutations were associated with low clinical response to ICI blockade in NSCLCs [
34]. Subgroup analysis of the data from the Phase III clinical trial of KEYNOTE-010 also indicated no improved OS benefit from ICI treatment in
EGFR-mutant NSCLCs [
27]. Similarly,
EGFR-mutant NSCLCs failed to achieve prolonged OS from ICI vs. chemotherapy in Phase III clinical trial OAK [
25]. It was further confirmed in a pooled analysis from 3 clinical trials (CheckMate-057, KEYNOTE-010, and POPLAR) that the
EGFR wild-type but not the
EGFR-mutant NSCLCs had prolonged OS [
31]. Combining data from 5 trials (CheckMate-017, CheckMate-057, KEYNOTE-010, OAK, and POPLAR), Lee et al. additionally confirmed no prolonged OS in
EGFR-mutant NSCLCs receiving ICIs relative to chemotherapy [
32]. In a Phase II trial, treatment of EGFR-TKI naïve
EGFR-mutant patients with ICI was halted due to lack of efficacy and two deaths within six months of enrollment, 1 of which was from pneumonitis [
33]. In the study of WJOG8515L, nivolumab was inferior to the chemotherapy treatment in EGFR-TKI-advanced
EGFR-mutant NSCLCs without T790M mutation with worse mPFS (1.7 vs. 5.6 months) and ORR (9.6% vs. 36%) [
137]. In a retrospective study, 58
EGFR-mutant NSCLC patients who responded to prior EGFR TKIs for more than ten months displayed significantly shorter PFS of ICIs compared to those responding to prior EGFR TKIs for less than ten months (1.6 vs. 1.9 months,
p=0.009) [
138]. More recently, ICI was unfavorable versus chemotherapy in EGFR-TKI-advanced NSCLCs harboring a secondary T790M mutation with worse mPFS (1.7 vs. 5.6 months) and response rate (9.6% vs. 36.0%) and similar OS (20.7 vs. 19.9 months) in Phase II WJOG8515L clinical trial [
137]. Adding ICIs to chemotherapy was associated with worse survival than platinum doublet chemotherapy alone in osimertinib-advanced
EGFR-mutant NSCLCs [
139]. Phase III clinical trial IMpower130 also found no benefit in the
EGFR-mutant subgroup treated with ICIs and chemotherapy combined versus chemotherapy alone [
140]. Based on all these negative findings, the National Comprehensive Cancer Network (NCCN) clinical practice guidelines of NSCLC (version 4, 2021) did not recommend immunotherapy for treating
EGFR-mutant NSCLCs.
- 6.
Safety Concerns and Lower Clinical Outcomes Regarding EGFR TKI and ICI Combined for the Treatment of EGFR-Mutant NSCLCs
Compared to chemotherapy, ICI therapy correlates fewer adverse reactions. However, ICI therapy was associated with immune-related adverse events (irAEs) caused probably due to the disruption of immunologic homeostasis [
141]. There is growing concern that a combination of EGFR TKIs, especially osimertinib and gefitinib, and ICI therapy may be associated with an increased risk of toxicity. Five of seven untreated stage IIIB/IV
EGFR-mutant NSCLC patients (71.4%) treated with ICI plus gefitinib had grade 3/4 liver toxicity, leading to permanent treatment discontinuation in four patients in the Phase 1/2 KEYNOTE-021 clinical trial [
142]. A lung adenocarcinoma patient bearing
EGFR 19del treated with osimertinib following ICI therapy had Stevens–Johnson syndrome and hepatotoxicity [
143]. In a Phase Ib clinical trial of TATTON, 38.5-60% of EGFR-TKI-advanced
EGFR-mutant NSCLCs experienced grade 3/4 irAEs, and 30-40% discontinued to osimertinib and ICI combined therapy due to irAEs [
144]. Concurrent osimertinib and ICI therapy were associated with an increased incidence of irAEs, leading to early termination of a Phase III CAURAL recruitment [
145]. Interstitial lung disease (ILD) is a severe adverse response to EGFR TKIs [
146]. Osimertinib treatment immediately followed prior ICI therapy, which also caused a high incidence of ILD [
147,
148]. A meta-analysis of eight studies further concluded that a higher chance of irAEs was observed in EGFR-TKI plus ICI therapy in EGFR-TKI-advanced
EGFR-mutant NSCLCs [
149].
Other than toxicity concerns, lower clinical outcomes have also been observed in
EGFR-mutant NSCLCs treated with EGFR TKI and ICI combined than those treated with EGFR TKI alone. The Phase Ib clinical trial of TATTON showed that osimertinib and ICI combination therapy only achieved 43% of ORR in the EGFR-TKI-advanced
EGFR-mutant NSCLCs [
144]. Besides, the ORR in gefitinib plus ICI therapy (14.3%) was much worse than the 55% of RR in gefitinib alone for
EGFR-mutant NSCLC patients treated with gefitinib alone [
150]. In contrast to gefitinib plus ICI therapy, the untreated stage IIIB/IV
EGFR-mutant NSCLCs tolerated erlotinib plus ICI therapy and better ORR (41.7%) in the KEYNOTE-021 [
142]. The RR was much lower than ~70% when using erlotinib alone as a first-line treatment in clinical trials [
151,
152]. Decreased ORR was also observed in the osimertinib plus ICI group versus osimertinib alone in the Phase III CAURAL recruitment [
145].
- 7.
Hyper-Progressive Disease (HPD)
Hyper-progressive disease (HPD), characterized by unexpected fast tumor enlargement both at rate and volume and early fatality of patients [
153,
154], was observed in 13.8% of NSCLC patients during treatment with ICIs in a retrospective study [
155]. In contrast, around 20% of
EGFR-mutant NSCLCs showed risk of HPD after ICI therapy [
153].