Immunotherapy in Non-Small Cell Lung Cancer With Actionable Mutations Other Than EGFR

While first line targeted therapies are the current standard of care treatment for non-small cell lung cancer (NSCLC) with actionable mutations, the cancer cells inevitably acquire resistance to these agents over time. Immune check-point inhibitors (ICIs) have improved the outcomes of metastatic NSCLC, however, its efficacy in those with targetable drivers is largely unknown. In this manuscript, we reviewed the published data on ICI therapies in NSCLC with ALK, ROS1, BRAF, c-MET, RET, NTRK, KRAS, and HER2 (ERBB2) alterations. We found that the objective response rates (ORRs) associated with ICI treatments in lung cancers harboring the BRAF (0–54%), c-MET (12–49%), and KRAS (18.7-66.7%) alterations were comparable to non-mutant NSCLC, whereas the ORRs in RET fusion NSCLC (less than10% in all studies but one) and ALK fusion NSCLC (0%) were relatively low. The ORRs reported in small numbers of patients and studies of ROS1 fusion, NTRK fusion, and HER 2 mutant NSCLC were 0–17%, 50% and 7–23%, respectively, making the efficacy of ICIs in these groups of patients less clear. In most studies, no significant correlation between treatment outcome and PD-L1 expression or tumor mutation burden (TMB) was identified, and how to select patients with NSCLC harboring actionable mutations who will likely benefit from ICI treatment remains unknown.


INTRODUCTION
NSCLC accounts for 85% of all lung cancers, with lung adenocarcinoma being the major subtype (1). Platinum-based combination chemotherapy is the historical first-line standard of care for patients with advanced NSCLC who have no actionable mutations (2). The introduction of ICIs, such as anti-programmed cell death protein ligand 1 (anti-PD-L1) and anti-programmed cell death protein 1 (anti-PD-1) antibodies, as well as the anti-cytotoxic T-lymphocyte-associated protein 4 (anti-CTLA-4) antibody, have revolutionized the treatment of NSCLC, and is typically offered with or without chemotherapy in the front-line setting for incurable NSCLC that does not have any actionable mutations (2). A number of actionable genetic alterations have been identified in NCSLC, including ALK, ROS1, c-MET, RET, NTRK, BRAF V600E, KRAS, and ERBB2 (HER2) (3)(4)(5)(6)(7)(8)(9)(10)(11). MET, RET, HER2, ALK, NTRK, and ROS-1 are receptor tyrosine kinases; BRAF is a serine/threonine kinase mediating cellular signal from RAS to MEK1/2; KRAS is a RAS protein which functions as a GDP-GTP-regulated binary on-off switch. While c-MET, BRAF and KRAS altered NSCLC may develop in both smokers and non-smokers, ALK, ROS1, RET, NTRK, and HER2 alerted NSCLC tend to occur in non-smokers. In patients with actionable driver mutations, namely, EGFR, ALK, BRAFV600E, RET, c-MET, NTRK or ROS1 alterations, the standard of care is to treat with a Food and Drug Administration (FDA) approved targeted agent, which typically can achieve ORRs of 60-80% in treatment naive patients (2,12). After targeted therapies are exhausted in these patients, systemic therapy with chemotherapy is typically available for them. While incorporating immunotherapy in the regimen is a standard of care option for them, the efficacy of immunotherapy in those with actionable mutations remains poorly defined due to the limited numbers of these patients included in the randomized prospective trials. In addition to the genetic alterations for which targeted therapies have been approved by FDA in the first line setting in NSCLC, KRAS G12C has a targeted agent that was approved recently in the beyond first-line setting. Moreover, HER2 mutations have emerged as new therapeutic targets with promising therapeutic agents in development. The efficacy of ICI in the KRAS G12C or HER2 mutant NSCLC is also of great clinical interest.
In this modern era with a booming number of treatment options for NSCLC and continued improvement in survival, further guidance is needed on what to expect from the use of immunotherapy in those with these genetic abnormalities. The goal of this review is to add valuable information on the use of immunotherapy in NSCLC with actionable alterations in genes including ALK, ROS1, BRAF, c-MET, RET, NTRK, KRAS, and HER2. Epidermal growth factor receptor (EGFR) mutations are not included in this review as they are included in another manuscript by our group which was submitted separately. In this review, we showed that the sensitivity to ICIs can be heterogenous and differs according to the driver alteration considered. ALK and RET fusions were found to be associated with low responses to ICI while BRAF, KRAS, and c-MET alterations were associated with responses that were comparable to non-mutant NSCLC, and PD-L1 positive KRAS mutant NSCLC may be associated with better outcome when treated with ICI monotherapy as suggested by two retrospective studies. The responses to ICIs are less clear in HER2, ROS1 or NTRK altered NSCLCs due to low patient numbers. While an association between PD-L1 expression level or TMB and the responses to ICI has not been consistently observed across all driver alterations, the overall lack of response to ICI treatment appeared to be more common among NSCLC with driver alterations that are typically associated with non-smokers, raising the question whether the absence of tobacco exposure may predict the lack of benefit from ICI treatment. Moreover, the emerging data in the role of co-mutations in response to ICI had also shed a light in the potential underlining mechanism of resistance to ICI, and particularly in the presence of KRAS mutation, co-mutations in TP53, STK11, and KEAP1 have been found to modulate the response to ICIs in several studies (13)(14)(15).

ALK
Anaplastic lymphoma kinase (ALK), a member of the insulin receptor tyrosine kinase family, has been identified as a fusion partner of nearly 30 different proteins in oncogenic signaling in many different cancer types (3). While there are now over 20 ALK fusion partners identified in NSCLC, EML4 represents the most common fusion partner with 29-33% of gene fusions identified to date (16,17). The fusion of the 5′ end partner EML4 to the coding region of the intracellular tyrosine kinase domain of ALK leads to aberrant expression of the ALK fusions in the cytoplasm. The domains in the partner proteins also promote dimerization and oligomerization of the fusion proteins, leading to constitutive activation of ALK kinase and its downstream signaling pathways including RAS-mitogenactivated protein kinase, phosphoinositide 3-kinase-AKT, and JAK-STAT pathways. This subsequently results in uncontrolled cellular proliferation and promotes survival (3,18). ALK fusions are seen in 3-5% of NSCLC patients and are more common among the following groups: no prior smoking history, adenocarcinoma histology, younger age, female gender, and tumors with wild type EGFR and KRAS (16,(19)(20)(21). Several ALK inhibitors have been approved by the FDA for metastatic NSCLC, including crizotinib, brigatinib, alectinib, lorlatinib and ceritinib (22)(23)(24)(25)(26)(27)(28)(29). The data on the efficacy of ICIs in the ALK fusion positive NSCLC has been scarce. It has been postulated that EML4-ALK oncoprotein can upregulate the PD-L1 expression in lung cancer cells. In one report of 100 patients, fifty patients (50.0%) were PD-L1 negative, 34 patients (34.0%) were PD-L1 low expression (tumor proportion score [TPS] 1-50%), and 16 patients (16.0%) had a strong PD-L1 expression (TPS ≥ 50%) (30). Despite the expression of PDL1 in these tumors, the overall response to ICIs in the ALK fusion positive population has been disappointing except in one study ( Table 1).
Although small numbers of patients with ALK fusion NSCLC were included in the randomized phase 3 CheckMate 057 and KEYNOTE-010 studies comparing ICI versus docetaxel in previously treated NSCLC patient population, the outcomes in this specific population were not reported (35,36). In a retrospective study using the IMMUNOTARGET registry which included 551 patients receiving ICI monotherapy for advanced NSCLC with at least one oncogenic driver alteration, 23 patients with ALK fusion NSCLC were identified (31). The objective response rate to ICI treatment was 0%. The Median PFS was 2.5 (1.5; 3.7) months. The median OS from start of ICI therapy was 17.0 (3.6; NR) months. Among the 10 patients with available PD-L1 status, the median percentage of cells expressing PD-L1 was 7.5% ( Table 1).
In a retrospective study conducted at the Massachusetts General Hospital, the ORR to ICI treatment among patients with EGFR mutations or ALK rearrangements was only 1/28 (3.6%) while the ORR among EGFR WT/ALK-negative patients was 7/30 (23.3%) (P = 0.053) (32). Since the lone partial response was seen in an EGFR-mutant patient, it appeared that none of the six ALK fusion NSCLC patients had a response ( Table 1).
In the randomized Impower130 study, atezolizumab plus chemotherapy (Nab-Paclitaxel and Carboplatin) did not show improved overall survival versus chemotherapy alone in the subset of 44 patients with EGFR or ALK genomic alterations in the first line setting (37). However, in the Impower150 study, the addition of Atezolizumab to Bevacizumab, Carboplatin, and Paclitaxel improved the median PFS for patients with EGFR or ALK genomic alteration whose diseases had progressed on TKI or who were unable to tolerate TKI (median, 8.3 months vs. 6.8 months; stratified hazard ratio, 0.61; 95% CI, 0.52 to 0.72). Of note, 34 patients with ALK fusion and 80 patients with EGFR mutant nonsquamous metastatic lung cancer were included in this study, and information on the benefit of atezolizumab in ALK fusion NSCLC was not reported separately (38). In another report of 83 patients with ALK mutation treated with ICI, a mPFS of 2.34 months was reported (33).
A recent prospective multicenter trial presented at the World Conference on Lung Cancer evaluated pembrolizumab and chemotherapy in the setting of recurrent EGFR/ALK-positive NSCLC. The study enrolled a total of 33 patients, including 26 EGFR mutant NSCLC and seven ALK fusion positive NSCLC patients. Most of the patients had one prior targeted therapy. No more than one prior line of platinum-based chemotherapy for advanced NSCLC was allowed. In those with ALK-positive tumors, the ORR was seen in 2/7 (28.6%), and the mPFS and mOS were both 2.9 months, suggesting lack of benefit of ICI in this group of patients (34).

BRAF
BRAF is a serine/threonine kinase mediating cellular signal from RAS to MEK1/2, and BRAF activation can result in phosphorylation and activation of extracellular signal-regulated kinase (ERK)1/2, leading to cell survival and proliferation (4). BRAF mutations are found in 1.5-3.5% of NSCLC with V600E accounting for approximately half of those mutations (39). Besides adenocarcinoma, BRAF mutations have been reported in sarcomatoid carcinomas, large-cell neuroendocrine carcinomas, and squamous cell lung cancer (40,41). BRAF mutations can occur in both smokers and non-smokers (42). Selective kinase inhibitors have been recommended for the firstline and second-line treatments of BRAF V600E mutant advanced NSCLC with a reported ORR as high as 64% in this group of patients (39). The outcomes associated with ICIs in this population have been studies in multiple retrospective analyses ( Table 2). Although the data vary significantly among different studies, responses to ICI have been seen in most of the studies.
In a retrospective study including seven participating Israeli cancer centers reported by Dudnik et al., PD-L1 expression level, tumor mutational burden (TMB), and microsatellite instability status were assessed in both BRAF V600E and non-V600E BRAF mutation positive NSCLC, and the outcome with ICI treatment was reported (43). High (≥50%) PD-L1 expression was found to be more common in the non-BRAF V600E mutant group than the V600E BRAF mutant group (50% vs 42%, p = 0.05). No MSI-H was found in both groups, and the median TMB was 5 (1-42) muts/Mb and 11 (7)(8)(9)(10)(11)(12)(13)(14) muts/Mb in the BRAF V600E and the non-V600E BRAF mutant groups, respectively. ICI therapy was associated with ORRs of 25 and 33% in the BRAF V600E and the non-V600E BRAF mutant positive groups, respectively (p = 1.0) ( Table 2). Among the six patients with high PD-L1 and BRAF V600E mutant NSCLC, two patients had major tumor shrinkage while two other patients had hyperprogression (43). Among the 1,588 advanced non-squamous NSCLC patients enrolled in the Italian Expanded Access Program of second line nivolumab, 210 patients were assessed for BRAF mutations, and 11 patients (5%) were found to be positive. Median OS was comparable among different groups, and was found to be 11.0 months (range: 9.8 to 12.2 months), 11.2 months (range: 9.2 to 13.2 months) and 10.3 months (range: 2.1 to 18.5 months) in the population with unknown BRAF status, BRAF wild-type subgroup, and BRAF mutated subgroup, respectively (44) ( Table 2).
A retrospective study was conducted to evaluate the clinical response to immunotherapy and chemotherapy among 31 patients with BRAF mutant metastatic NSCLC treated at the Duke University Hospital (45). PD-L1 expression information was only available for 11 patients. PD-L1 expression levels ranged from 0 to 90%, with six patients with PD-L1 expression levels greater than 50%. TMB was only available on five patients, ranging from 3 to 18 mutations/Mb. The median PFS in patients who received first-line chemotherapy was 6.4 months (95% CI, 2.3 to 13.0) while the PFS of each of the three patients who received first-line immunotherapy was 0.17, 1.4, and 4.4 months. The median OS in patients who received first-line chemotherapy was 18.4 months (95% CI, 7.4 to 28.6), and the OS of each of the three patients who received first-line immunotherapy was 0.17, 6.8, and 7.5 months ( Table 2).
In the retrospective study using the IMMUNOTARGET registry, among the 43 patients with BRAF mutations, PFS was significantly higher in smokers than never smokers (4.1 versus 1.9 months, P = 0.03), however shorter in the V600E subgroup (1.8 months) compared with other BRAF mutations (4.1 months, P = 0.20) (31) ( Table 2). The ORR was 24.3%. Among the nine patients with available PD-L1 status, the median percentage of cells expressing PD-L1 was 50%.

MET
MET is a proto-oncogene receptor tyrosine kinase that mediates cell proliferation, survival, and metastasis (5). Recurrent somatic splice site alterations at MET exon 14 (METex14) can result in exon skipping, decreased MET degradation, and MET activation. METex14 is involved in cancer through promoting angiogenesis, cell migration, and invasion (49,50). METex14 occurs in 3-4% of lung cancers and 8-30% of sarcomatoid lung cancers (51,52). The occurrence of METex14 appears to be independent of smoking status (53). FDA has granted accelerated approval to capmatinib and tepotinib for adult patients with metastatic NSCLC whose tumors have a mutation that leads to METex14 alterations (54,55).
Among the 551 patients in the IMMUNOTARGET registry, 13 patients with MET amplification and 23 patients with METex14 were identified (31). Median OS from ICI initiation of this 36-paitent cohort was 18.4 months (7.0; NR) (31). Progressive disease (PD) was found to be the best response to ICI among 50% of patients, and median PFS was found to be 3.4 months (1.7; 6.2). Long-term responders were seen in 23.4% of patients ( Table 3). Among the 15 patients with available PD-L1 status, the median percentage of cells expressing PD-L1 was 30%.
In a retrospective study that included eight cases of NSCLC with METex14 and four cases of NSCLC with MET amplification treated with ICI, median PFS with ICI was 4.0 months (95% CI, 2.4-NR) in patients with METex14 and 4.9 months (95% CI, 2.4-NR) in patients with MET amplification (46). ORR comprised 12% (1/8) and 25% (1/4) in patients with METex14 and MET amplification respectively. Median OS with ICI was not reached in patients with METex14 (95% CI, 4.1-NR) or in patients with MET amplification (95% CI, 3.5-NR) ( Table 3). Among patients with known PD-L1 TPS, TPS high (≥50%) was seen in 67% of the cases. TMB high (≥10 mut/Mb) or MSI-H/I was not seen.
In a case series, among 13 patients with METex14 NSCLCs treated with ICI, 46.2% (6/13) patients responded to immunotherapy. Six patients had prolonged duration of responses ranging from 18 months (still ongoing) to 49 months (56).

RET
RET is a proto-oncogene receptor tyrosine kinase that binds with the ligand-co-receptor complex of glial cell line-derived neurotrophic factor (GDNF) family ligands (GFLs) and subsequently activates signaling pathways such as RAS/ mitogen activated protein kinase (MAPK), RAS/ERK, phosphatidylinositol 3-kinase (PI3K)/AKT, and c-Jun N-terminal kinase (JNK). Aberrant activation of the RET receptor have been associated with multiple endocrine neoplasia 2 (MEN2), sporadic medullary thyroid cancer, papillary thyroid carcinoma (PTC), and non-small cell lung cancer (NSCLC) (6,7). RET rearrangements have been identified in 1-3% of NSCLC and were found to have significantly higher frequencies in younger (<60 years of age), female, non-smokers, and adenocarcinoma histology (57)(58)(59)(60). RET fusion positive NSCLC is usually associated with low PD-L 1 expression (61). Two potent selective RET inhibitors, selpercatinib and pralsetinib, have been approved by the FDA for RET fusion-positive NSCLC (62,63). The activity of ICI in RET altered NSCLC has been evaluated in multiple studies, and the benefit of ICI was found to be low in most of the studies ( Table 4).
In a single center retrospective study conducted in Korea, the median progression-free survival for ICI among 13 patients with RET fusion-positive NSCLC treated with ICI was 2.1 (95% CI: 1.6-2.6) months, and the ORR was 7.7% (64). The median PFS and OS were 2.1 (95% CI: 1.6-2.6) and 12.4 (95% CI: 2.9-21.8) months, respectively. Among patients with PD-L1 expression 25% and above, 2/5 patients demonstrated stable disease, while the best response in the other three patients was disease progression ( Table 4). In contrast, the ORR and DCR among 46 patients treated with pemetrexed-based regimens in this study was 63.0 and 91.3%, respectively, and the median PFS was 9.0 (95% CI: 6.9-11.2) months.
Among the 16 patients with RET fusion-positive NSCLC in the IMMUNOTARGET registry, the median OS from the start of   (31). The rate of any partial or complete response was very low and was 6.3% (1/16) ( Table 4). Among the six patients with available PD-L1 status, the median percentage of cells expressing PD-L1 was 26%.
In a retrospective study conducted at the Memorial Sloan Kettering Cancer Center, 13 patients with RET-rearranged NSCLC treated with ICI were assessed for clinical and/or radiologic response (30). No response to immunotherapy was observed. The median PFS was 3.4 months (95% CI, 2.1 to 5.6 months). No difference in OS between patients with advanced RET-rearranged lung cancers who received immunotherapy (n = 16) and those who did not receive immunotherapy (n = 46), (hazard ratio, 1.4 [95% CI, 0.7 to 2.9]; log-rank P = .35) ( Table 4). Only one patient was found to have PD-L1 expression ≥50%, and the disease of this patient did not respond to ICI. No patient had TMB >10 mut/Mb.
In the single institution retrospective study published by Dudnik et al., four patients with RET fusion NSCLC and one patient with RET mutant NSCLC were treated with ICI (46). No objective response was observed. Median PFS was 3.0 months (95% CI, 1.9-3.1) in patients with RET fusion and 6.9 months in patient with RET mutation. Median OS since start of ICP were 14.9 months (95% CI, 7.2-19.7) in patients with RET fusion and 15.3 months in patient with RET mutation ( Table 4). Among patients with known PD-L1 TPS, TPS high (≥50%) was seen in 13 and 0% of the RET fusion and the RET mutant NSCLC cases, respectively. TMB high or MSI-H/I was not seen.

ROS1
ROS proto-oncogene 1 (ROS1) belongs to the subfamily of tyrosine kinase insulin receptors (65). ROS1 fusion can lead to constitutive activation of kinase activity, resulting in increased cell proliferation, survival, and migration due to the upregulation of JAK/STAT, PI3K/AKT, and MAPK/ERK signaling pathways (8). ROS1 rearrangements account for 1-2% of NSCLC patients (66,67). This alteration more frequently occurs in adenocarcinoma and in younger patients with no or light smoking history (68,69).
Seven patients with ROS1 fusion NSCLC treated with ICI were identified in the IMMUNOTARGET registry (31). The objective response rate to ICI treatment was 17% ( Table 5).
In the single institution retrospective study published by Dudnik et al., only one patient with ROS1 fusion NSCLC treated with ICI was identified, and the reported PFS and OS were both 0.1 month (46) ( Table 5). Among the five patients with available PD-L1 status, the median percentage of cells expressing PD-L1 was 90%.
In the Japanese retrospective study, 15 ROS1 altered NSCLC cases were identified. High expression of PD-L1 (>50% of tumor cells by 22C3) were observed in 53% cases, however, no response to immunotherapy was observed (70).

KRAS G12C
KRAS is one of the RAS proteins (KRAS4A, KRAS4B, NRAS, and HRAS) which function as GDP-GTP-regulated binary onoff switches and regulate cell survival, cell cycle progression, cell polarity, movement, and nuclear transport by transducing signals from transmembrane receptors to cytoplasmic signaling pathways such as the MAPK pathway (10,11). It is the most common proto-oncogene identified in NSCLC. KRAS mutations occur in 15-25% of lung adenocarcinomas and are more prevalent in smokers than nonsmokers (73,74). Majority of the KRAS mutations in NSCLC occur on exon 2 or 3 (G12, G13, and Q61), with the most frequent being the G12C followed by G12V and G12D (75,76). Sotorasib has been approved by the FDA for patients with KRAS G12C mutant locally advanced or metastatic NSCLC in the beyond the first line setting (72). It is associated with an objective response rate of 37.1% in this group of patients (77). The efficacy of ICIs in KRAS mutant NSCLC has been studied in several retrospective studies, and most of the data support the benefit on ICIs in KRAS mutant NSCLC ( Table 7). In a retrospective analysis in patients enrolled in the KEYNOTE-042 evaluating pembrolizumab monotherapy vs platinum-based chemotherapy as the first-line therapy among patients with PD-L1positive (TPS ≥1%) advanced non-squamous histology NSCLC, 301 patients were evaluable by whole-exome sequencing (WES). KRAS mutations were found in 69 (23%) patients, among which, 29 (10%) patients were found to have KRAS G12C (78). PD-L1 TPS and TMB were found to be higher in patients with KRAS mutations than without KRAS mutations, although the differences were not significant. The OS associated with pembrolizumab was better than chemotherapy in both the KRAS mutant group and KRAS G12C subgroup, with the HRs being 0.42 (0.22-0.81) and 0.28 (0.09-0.86), respectively. Conversely, there was no significant OS difference seen between pembrolizumab and chemotherapy in the KRAS wild-type patients, and HR was 0.86 (0.63-1.18). A superior PFS was also observed when pembrolizumab was compared with chemotherapy in the KRAS mutant patients. The data supported the benefit of single agent pembrolizumab in the PD-1 TPS >1% KRAS mutant (including KRAS G12C) NSCLC patients, underlining the important role of ICI in the treatment of this group of patients.
The efficacy of ICIs in the first line setting in PD-L1 TPS ≥50% advanced NSCLC was also investigated in a retrospective analysis using the Flatiron Health database (79). Among the1,127 patients with PD-L1 expression of 50% or greater who were treated with either ICI monotherapy or chemoimmunotherapy, 573 (50.8%) had KRAS alterations and 554 (49.2%) had wild type KRAS. Among the patients treated with ICI monotherapy, a better mOS was seen in the KRAS mutant group when compared with the wild-type group (mOS, 21.1 vs 13.6 months; P = .03). Interestingly, this OS advantage was not observed among patients treated with chemoimmunotherapy, and the mOS was 20.0 vs 19.3 months; P = .93 in the KRAS mutant and wild type patients. Furthermore, no mOS difference was seen between ICI monotherapy and chemoimmunotherapy in the KRAS mutant NSCLC patients (mOS, 21.1 vs 20.0 months; P = .78), suggesting that the use of ICI monotherapy in the PD-L1 TPS ≥50% is an acceptable option in the KRAS mutant advanced NSCLC.
The efficacy of chemoimmunotherapy in KRAS mutant NSCLC was also analyzed retrospectively in the participants of another randomized trial, the KEYNOTE-189 study of pembrolizumab plus pemetrexed and platinum chemotherapy vs placebo plus chemotherapy as first-line therapy for metastatic non-squamous NSCLC (80). Among the 289 patients who had evaluable WES data, 89 (31%) patients were found to have KRAS mutations including KRAS G12C, which was found in 37 (13%) patients. As observed in the KEYNOTE-042 study, the higher PD-L1 TPS and TMB tended to be seen with KRAS mutant patients. Although unlike the observation in the KEYNOTE-042, the OS benefit associated with the addition of ICI was only detected in the KRAS wild-type patients. PFS improvement associated with the additional of ICI was seen in both the KRAS mutant and wild type group but not in the KRAS G12C subgroup, which could be related to the small sample number.
In addition to ICI monotherapy and chemoimmunotherapy, the combination of VEGF receptor targeted agent and chemoimmunotherapy represents another first-line treatment option for advanced NSCLC based on the IMpower150 study (84). A post hoc analysis evaluated the efficacy outcomes in   Table 7), supporting the use of this regimen in KRAS mutant NSCLC. The correlation between STK11/LKB1 genomic alterations and the efficacy of ICI treatment in KRAS mutant NSCLC was also evaluated using the Stand Up To Cancer (SU2C) dataset (13). Unlike the post hoc analysis of the IMpower150 study, this study showed that the concurrent STK11/LKB1 mutation in KRAS mutant NSCLC was associated with an inferior ORR to PD-1 blockade when compared with KRAS mutation without STK11/LKB1 mutation and KRAS mutation with P53 mutations groups (7.4, 28.6 and 35.7% (P <0.001)). The details of the ICI therapy in this dataset were not available, and it is unclear whether this group of patients also received angiogenesis targeted agent treatment.
The efficacy of ICI in KRAS mutant non-squamous NSCLC in the beyond first-line setting was also investigated in patients who received nivolumab in an Italian expanded access program (EAP) study (82). Among the 530 patients evaluated, 206 (39%) had KRAS mutations. No significant differences in OS, PFS or ORR were seen between KRAS mutant and KRAS wildtype patients in this study, supporting that nivolumab should be considered for patients regardless of KRAS mutation status. Interestingly, any significantly higher grade and grade 3-4 treatment related adverse events were seen in the KRAS mutant group than the wild-type group, although the underlining mechanism for the finding is unknown.
KRAS mutant NSCLC was also evaluated in the IMMUNOTARGET study. Two hundred and seventy-one patients treated with ICIs were found to have KRAS mutations. An encouraging ORR of 26% was found, and the mPFS and mOS were 3.2 and 13.5months, respectively (31).
In a single instituation retrospective study conducted in France, a total of 162 KRAS-mutant advanced NSCLC were identified among the 282 subjects analyzed. No significant difference was seen in ORR, mPFS or mOS between the KRAS mutant and the KRAS wild-type groups. The ORR, mPFS, and mOS associated with ICI of KRAS mutant NSCLC were 18.7%, 3.09 months and 14.29 months. No significant difference in treamtent outcomes was seen among the KRAS mutation subtypes including G12A, G12C, G12D, G12V, and G13C (83).

HER 2
Human epidermal growth factor 2 (HER2 erbB-2/neu) is one of the four receptor tyrosine kinase members of the human epidermal growth factor receptor family. Upon forming homoor hetero-dimers with other family members, HER2 becomes activated and signal through the PI3K-AKT and MEK-ERK downstream pathways to activate proliferation (86). In NSCLC, activating HER2 mutations occur in 2-4% of cases, most commonly in adenocarcinoma histology and never smokers (87). Patients with HER2 mutant NSCLC have worse OS if treated without HER2 targeted therapy (88). Although there has not been any HER 2 targeted agent approved by NSCLC by the FDA, several agents have showed promising activity. Adotrastuzumab emtansine, a HER2-targeted antibody-drug conjugate was found to be associated with an ORR of 44% in NSCLC with HER2 exon 20 insertions and point mutations (89), and another HER2-targeted antibody-drug conjugate, trastuzumab deruxtecan, also showed an encouraging ORR of 55% in patients with metastatic HER2-overexpressing or HER2mutant NSCLC whose disease had relapsed during standard treatment or was refractory to standard treatment (90). Both agents are included as novel therapeutic options for HER2 mutant NSCLC in the current NCCN guidelines (2). Poziotinib, a tyrosine kinase inhibitor targeting EGFR/HER2 exon 20 insertion mutation, was found to have an ORR of 27% in HER2 exon 20 mutant NSCLC, gaining fast track designation by FDA (91,92).
The efficacy of immunotherapy in patients whose cancer harbors HER2 mutation is largely unknown. The ORR associated with ICI among the 29 patients with exon 20 activating mutations in the IMMUNOTARGET study was only 7%. PFS was 2.5 months, and the 12-month PFS was 13.6 months. The OS was 20.3 months (31). The ORR among the 23 patients with exon 20 insertions included in the IMAD2 study by the French Lung Cancer Group was 27.3%. PFS was similar to the findings in the IMMUNOTARGET study and was 2.2 months, and the 12-month PFS was 22.9%. The mOS was an encouraging 20.4 months (47) ( Table 8).

DISCUSSION
To delineate the benefit of ICI treatment in NSCLC harboring actional mutations other than EGFR alterations, we reviewed the current available data in this area. We found that the ORR, median PFS, and OS with ICPi varied significantly across genetic alteration subgroups. While the ORR observed in the BRAF, c-MET, and KRAS altered NSCLC appeared to be similar to what had been observed in the non-selected NSCLC groups, the ORRs in the ALK and RET altered NSCLC groups were much lower (2). Unlike ALK and RET fusions, BRAF, MET, and KRAS mutations can be seen in both smokers and non-smokers. The higher prevalence of smoking history in these patients could be a potential reason of the higher response rates since smoking has been found to be associated with the benefit derived from ICI treatment in some of the literatures (93,94), although not confirmed by other studies (95). Other known predictive biomarkers for ICI treatment include PD-L1 expression level, microsatellite instability-high (MSI-H) or mismatch repair deficient (dMMR), and TMB (96)(97)(98). A higher percentage of PD-L1 TPS high (67%) was reported in BRAF non V600E mutant and MET mutant NSCLC in some of the reports (46), and a relatively higher response was seen in patients with PD-L1 TPS high BRAF mutant NSCLC (43), albeit the sample numbers was too small to draw any firm conclusion.
The current NCCN guidelines support the use of targeted therapy in the first line setting for advanced NSCLC with actionable genomic alterations involving EGFR, ALK, ROS1, BRAF, NTRK1/2/3, METex14 skipping, and RET. After disease progression, chemoimmunotherapy is recommended for this population based on the guidelines (2). Although most of the chemoimmunotherapy trials excluded the EGFR and ALK altered NSCLC patients, these groups of patients were evaluated in the IMpower150 study if they have had progression with or unacceptable side effects from treatment with at least one approved tyrosine kinase inhibitor. The data showed improved PFS associated with the additional of atezolizumab to the bevacizumab, carboplatin, and paclitaxel combination in EGFR and ALK altered NSCLC (9.7 months vs. 6.1 months), providing direct evidence supporting the use of the regimen in this population (38). In our review, while not robust, we see that ICI's do have activity in patients with NSCLC harbouring actionable mutations, and that a response can be seen after progression on targeted therapy, supporting offering chemoimmunotherapy in the post-targeted therapy setting. For KRAS G12C NSCLC, ICI monotherapy or chemoimmunotherapy is the current standard of care first line treatment. The data included in this review did confirm the benefits of ICI in the KRAS mutant NSCLC, supporting the current treatment approach. The benefit of ICI monotherapy in PD-L1 positive KRAS mutant NSLCL was suggested in retrospective studies, warranting further investigation on the selection of ICI monotherapy vs chemoimmunotherapy in this population (78,79). Furthermore, data from prospective studies will be helpful to identify the best treatment sequence among targeted therapy, ICI, and chemotherapy.
Even in the subgroups where the benefits of ICIs were observed, the ORRs tended to be low. Therefore, developing predictive biomarkers for ICI therapy would be of great importance.
Co-occurring genomic alterations have been reported to be related to responses to immunotherapy through altering the microenvironment. For example, LKB1/STK11 genomic alterations, a frequent co-occurring mutation in of KRAS mutant NSCLC, have been found to be associates with "immune-inert" state (99). This was supported by several studies including a retrospective study conducted in 103 NSCLC patients receiving ICIs. In this study, among the patient with KRAS mutations, the presence of concurrent STK11 mutation or STK11/TP53 mutations were associated with worse survival with ICI therapy. This association was not observed with chemotherapy, supporting the predictive roles of these co-mutations for ICI therapy in KRAS mutant NSCLC (15). The data from a retrospective analysis suggested that cooccurring LKB1/STK11 mutations in KRAS mutant NSCLC may predict lower ORR, while data from another group showed no PFS or OS differences with or without concurrent mutant LKB1/STK11 and/or mutant KEPA1 in patients receiving combined chemoimmunotherapy and angiogenesis targeted agent (13,81), raising the question whether angiogenesis targeted agent may help to overcome the challenge of the immune-inert state. Other cooccurring genomic alterations such as P53, KEAP1, ATM, PTEN, CDKN2A are common among KRAS-mutant NSCLC, and may play a role in determining response to ICI (100). Furthermore, a recent study showed that co-occurring mutations such as NOTCH and HR pathways were also found to be associated with increased efficacy of immunotherapy in advanced NSCLC (101). Therefore, identifying co-occurring mutations that are responsible for ICI response or resistance could potentially help to identify the candidate for ICI treatments and warrants further investigation in this group of patients.
How to overcome the resistance to ICI therapy is another great challenge. The mechanism of resistance is complex and is a combination of tumor-intrinsic and extrinsic factors. Many factors such as immune contexture and tumor microenvironment, expression of PD-L1 and LAG3, TMB, genetic and epigenetic alterations, antigen-presenting molecules (MHC, HLA) and microbiota may all contribute to the resistance to immunotherapy (102). The tumors with higher initial mutational burdens have been found to be associated with higher sensitivity to ICIs in some studies, although this association may be negated by other factors such as intratumoral heterogeneity and mutations (103). RET fusion positive NSCLC was found to have poor response to ICIs, and the alterations appears to be associated with lower TMB. In the analysis by Offin M. et al., the median TMB of RET altered NSCLC was significantly lower than that of the RET wild-type NSCLCs (1.75 versus 5.27 mutations/Mb, P <.0001) (61). The best outcome in patients in this study was stable disease which only lasted 5.6 months. In the report by Dudnik E. et al, the TBM was low in all 13 patients except one patient who had intermediate TMB, and the ORR in this report was also 0% (46). Nevertheless, an ORR of 37.5% was found among the nine evaluable patients reported by Guisier F. et al. Unfortunately, the TMB information was not available in this study, and it was unclear if the treatments were ICI monotherapy or chemoimmunotherapy. A prospective study to allow uniform treatment and collection of information on biomarkers such as TMB, PD-L1, MSI/MMR, tumor-infiltrating lymphocytes, wholeexome sequencing analysis on tumor samples and intestinal microbiome composition may be helpful to identify the resistance mechanisms.
ALK fusion positive NSCLC showed poor response to ICI in retrospective studies. However, this group of patients did benefit from ICI in the IMpower 150 trial, raising the question if the inhibition of angiogenesis could sensitize cancer cells to ICI therapy. Tumor angiogenesis can lead to immunosuppression through various mechanisms including maintaining an acidic/ hypoxic and immunosuppressive environment, development of dysfunctional blood vessels which limits T cell trafficking, and suppression of dendritic cell maturation. Moreover, the angiogenic factors such as VEGF are also immunosuppressive (104). Therefore, further investigation is warranted in coinhibition of angiogenic factors in NSCLC harboring actionable driver mutations undergoing ICI treatment.
With regard to the combination of ICIs and targeted therapies, a number of studies had evaluated the combiantion of ALK TKIs and different ICIs in NSCLC, including the combination of nivolumab with ceritinib or crizotinib and the combiantion of alectinib plus atezolizumab (105)(106)(107). However, significant toxicities were observed without survival benefit. In addition, there has been some concerning safety signals where ICI treatment is followed with targeted therapy (108). Reports showed risk of hepatotoxicity in a series of patients with ALK, ROS1, or MET exon 14 alterations who received ICI before crizotinib. Among the eleven patients treated with crizotinib following ICI, five patients (45.5%) developed grade 3 or 4 hepatotoxicity, whereas only 8% of those patients who received crizotinib alone experienced hepatotoxicity. The increased hepatotoxicity in sequentially treated patients led to permanent discontinuation of crizotinib in four of the five patients (108), highlighting the importance of establishing the presence of actionable mutations prior to initiating ICI therapy in patients with advanced NSCLC. The frequency and severity of toxicities associates with sequential use of ICI followed by targeted therapy may vary among different therapeutic agents. In the CodeBreaK100 phase II study evaluating Sotorasib in the beyond first-line setting, even though 91.3% patients had received ICI treatment prior to Sotorasib, the tolerability remained acceptable. Ongoing clinical trials DESTINY-Lung03 (NCT04686305) and the HUDSON trial (NCT03334617) are investigating the combination of T-DXd with immunotherapy, chemotherapy, novel anticancer agents and will hence shed more light on the approach in HER2 mutant subgroup NSCLC patients.
Our review certainly has limitations. We were unable to comment on the response of HER2, ROS1 and NTRK altered NSCLC to ICI as there were few reports in the literature, and the patient numbers in these reports were often very small. The challenges are obviously associated with the low incidences of these alterations. A recent report from Negrao et al. showed that RET, ROS1 and ALK alterations were associated with low sensitivity to ICIs. However, there were only three ROS1 fusion NSCLC patients included in the study, and the outcome of all three alterations were reported collectively (15). Furthermore, we were also unable to compare the responses to ICIs among different alterations which can be better investigated in prospective studies. Moreover, many studies included in this reivew did not have the biomarker information on all the evaluable patients. The ICI treatments and the number of lines of treament received previously by the patients also varied significantly. Additionally, it was not always clear whether the ICI treament was given as a monotherapy or in combination with cytotoxic chemotherapy. Randomized prospective studies would undoubtly provide more definitive information on this topic.
In conclusion, we see low responses to ICI in ALK and RET altered NSCLCs whereas BRAF, KRAS and c-MET alterations were associated with benefit from ICIs, and PD-L1 positive KRAS mutant NSCLCs may be more responsive to ICI monotherapy. Furthermore, the response to ICIs in KRAS mutant NSCLCs may vary depending on co-existing mutations, and responses to ICIs in HER2, ROS1 and NTRK altered NSCLCs are less clear and varies significantly across a small number of studies. Ultimately, immunotherapy in the second line after progression on targeted agents can be considered as a treatment option at the discretion of treating physicians, following a mutual discussion with patients about the pros and cons of this approach.