TKI beyond progression
In clinical practice, clinicians may prescribe EGFR TKI therapy beyond progression, especially when patients suffer from asymptomatic progression. Nishie et al. retrospectively analyzed Japanese patients with
EGFR mutations. Continuous use of EGFR TKIs beyond progression in patients with activating
EGFR mutations may prolong OS compared with switching to cytotoxic chemotherapy [
74]. In addition, the phase II ASPIRATION study demonstrated that continued erlotinib therapy following progression is feasible in selected patients [
75]. The NCCN Panel recommended continuing EGFR TKIs, whether erlotinib, gefitinib, or afatinib, and considering local therapy in patients with asymptomatic progression [
48].
A flare-up phenomenon (rapid disease progression) occasionally is noted after discontinuation of EGFR TKIs. Intratumor heterogeneity is the possible mechanism of the phenomenon. Compared to the resistant clone with indolent behavior, rapid regrowth of TKI-sensitive clones causes rapid clinical deterioration when EGFR TKIs are discontinued [
76]. One retrospective study also showed that 14 of 61 (23%) patients suffered from disease flare after stopping EGFR TKIs [
77]. Therefore, some patients were prescribed EGFR TKIs after acquired resistance to EGFR TKIs.
The phase III IMPRESS trial aimed to evaluate the efficacy and safety of continuing gefitinib combined with chemotherapy versus chemotherapy alone in patients with
EGFR-mutation-positive advanced NSCLC with acquired resistance to first-line gefitinib. A total of 265 patients were enrolled. However, continuation of gefitinib after disease progression on first-line gefitinib did not prolong PFS in patients treated with platinum-based doublet chemotherapy as subsequent treatment. A long-term follow-up found that median OS was 13.4 months in the combination arm and 19.5 months in the control arm (HR 1.44;
p = 0.016) [
78]. Besides, the gefitinib group had more side effects and grade 3 or worse AEs. According to the results of the IMPRESS trial, continuation of chemotherapy with first-generation EGFR TKIs after acquired resistance to EGFR TKIs is not considered as standard treatment.
Next-generation (third-generation) epidermal growth factor receptor tyrosine kinases inhibitors (EGFR TKIs)
The third-generation EGFR TKIs can form an irreversible covalent binding to EGFR. They are pyrimidine-based compounds, and differ from quinazolines-based first-and second-generation EGFR TKIs (Table
2) [
86]. Third-generation EGFR TKIs can attenuate
EGFR T790M activity and have less epithelial toxicity due to less wild-type
EGFR activity [
86,
87]. Among them, osimertinib (AZD9291) received FDA and European Medicines Agency (EMA) approval in November 2015 and February 2016, respectively, for treatment of patients with T790M mutation-positive NSCLC after acquired resistance to first-line EGFR TKIs treatment. Table
3 shows the available efficacy data of different third-generation EGFR TKIs in clinical trials.
Table 2
Different generations of EGFR TKIs
1st-generation | Reversible; | Gefitinib |
EGFR del19, L858R
| Skin rash/acne, abnormal LFT | FDA approved |
competitive | Erlotinib |
EGFR del19, L858R
| | FDA approved |
2nd-generation | Irreversible; covalent | Afatinib | EGFR del19, L858R, uncommon mutations, HER2, HER4 | Diarrhea, paronychia. Skin rash | FDA approved |
| Dacomitinib | EGFR del19, L858R, HER2, HER4 | Diarrhea, skin rash/acne | Phase III |
| Neratinib | EGFR G719X, HER2, HER4 | Diarrhea, dyspnea, N/V | Phase II |
3rd-generation | Irreversible; | Osimertinib | EGFR mutations and T790M | Diarrhea, skin rash | FDA approved |
covalent | Rociletinib | EGFR T790M mutation, IGF-1R | Hyperglycemia, QTc prolong | Withdrawn |
| Olmutinib | EGFR T790M mutation | Diarrhea, skin exfoliation, nausea | Approved in South Korea |
| ASP8273 | EGFR L858R, del19, T790M, | Diarrhea, N/V, thrombocytopenia | Phase III Discontinued |
| Nazartinib | EGFR L858R, del19, T790M, | Rash, diarrhea, pruritus | Phase I/II |
| Avitinib (AC0010) | EGFR L858R, del19, T790M, | Diarrhea, skin rash, abnormal LFT | Phase I/II |
| HS-10296 | EGFR sensitive mutations (G719X, del19, L858R, L861Q) +/− T790M | None reported | Phase I/II |
| PF-06747775 | EGFR L858R, del19, T790M, | None reported | Phase I/II |
Table 3
Efficacy of third-generation EGFR TKIs in EGFR T790M-positive NSCLC patients
Osimertinib | | Total: 253 T790M(+): 138 | 20-240 mg QD | T790M(+): 61% T790M(−): 21% | T790M(+): 9.6 T790M(−): 2.8 |
AURA phase I T790M(+) | 63 | 80 mg QD | 71% | 9.7 |
AURA phase II | 210 | 80 mg QD | 70% | 9.9 |
AURA phase II extension [ 132] | 411 | 80 mg QD | 62% | 12.3 |
| 416 -Osimertinib arm: 279 -Chemotherapy arm: 140 | | 71% 31% Odds ratio:5.39 (95% CI: 3.47–8.48) | 10.1 4.4 HR: 0.30 (95% CI: 0.23–0.41) |
Rociletinib | | Total: 69 T790M(+): 51 | 500, 625 or 750 mg bid | 45% | T790M(+): 9.6 T790M(−): 2.8 |
Olmutinib | HM-EMSI-101 phase I/II T790M(+) [ 133] | 76 | 800 mg QD | 62% | 6.9 |
ASP8273 | NCT02113813 phase I/II [ 134] | Total: 63 T790M(+): 58 | 300 mg QD | 29% | 6.8 |
Nazartinib | NCT02108964 phase I/II [ 105] | 152 | 75-350 mg QD | 46.9% | 9.7 |
Avitinib (AC0010) | NCT02330367 phase I/II [ 106] | 136 | 50-350 mg QD | 44% | |
Osimertinib (AstraZeneca, Macclesfield, UK) is an irreversible mono-anilino-pyrimidine EGFR TKI that covalently binds to the ATP-binding site, CYS797, of the EGFR tyrosine kinase domain. In EGFR recombinant enzyme assays, osimertinib showed potent activity against diverse activating
EGFR mutations with/without T790M. According to the preclinical data, osimertinib has 200 times greater potency against L858R/T790M than wild-type
EGFR [
88]. Two circulating metabolites of osimertinib, AZ5104 and AZ7550, were detected, and both had comparable potency to sensitizing
EGFR mutation and T790M [
89]. There was no significant difference in pharmacokinetic exposure between Asian and non-Asian patients, showing a minimal food effect [
90]. In addition, unlike first- and second-generation EGFR TKIs, osimertinib exposure was not affected by concurrent administration of omeprazole [
91].
AURA (NCT01802632) is a phase I/II dose-escalation clinical trial of osimertinib, which enrolled 253 Asian and western NSCLC patients with acquired resistance to first- or second-generation EGFR TKIs, as defined by Jackman criteria [
22,
92]. Patients were not preselected according to T790M status [
92]. Thirty-one patients were treated across five dose-escalation cohorts (20, 40, 80, 160 and 240 mg oral, daily) and 222 were treated in the dose-expansion cohort.
In the dose-escalation cohort, there was no dose-limiting toxicity (DLT) and the maximum tolerated dose (MTD) has not been reached. Of the 239 evaluable patients, the objective response rate (ORR) was 51% and the disease control rate (DCR) was 84%. Patients with
EGFR-T790M mutation had a better ORR (61% vs. 21%), DCR (95% vs. 61%), and longer median PFS (9.6 months vs. 2.8 months) than patients without an
EGFR-T790M mutation. The drug is relatively safe, and most of the AEs were grade 1 and 2. The most common AEs were diarrhea (47%), skin toxicity (40%), nausea (22%), and anorexia (21%). When patients took higher dose levels (160 and 240 mg), there was an increasing incidence and severity of AEs (rash, dry skin, and diarrhea). Based on efficacy and safety, 80 mg daily was selected as the recommended dose for further clinical trials [
92].
Then, a phase II “AURA2” study (NCT02094261) was initiated to enroll NSCLC patients with an
EGFR-T790M mutation and acquired resistance to approved EGFR TKIs; the enrollment criteria were similar to those of the AURA study extension cohort. A preplanned pooled analysis was performed, including 201 patients from the 80 mg osimertinib expansion cohort of AURA and 210 patients from AURA2; ORR was 66%, DCR was 91%, and median PFS was 11.0 months [
93].
In the phase III AURA3 study, 419 patients were randomized into osimertinib or platinum-pemetrexed chemotherapy (maintenance pemetrexed was allowed) groups after they had acquired resistance to first-line EGFR TKI therapy. The investigator-assessed PFS (primary endpoint) was significantly longer in the osimertinib arm than in the chemotherapy arm (median 10.1 vs. 4.4 months; HR 0.30; p < 0.001). The FDA has granted regular approval to the third-generation EGFR TKI, osimertinib, for the treatment of patients with metastatic EGFR T790M mutation-positive NSCLC.
In the preclinical study, osimertinib demonstrated greater penetration of the mouse blood-brain barrier than gefitinib, rociletinib, or afatinib [
94]. There were several reports of dramatic intracranial response to osimertinib in patients with
EGFR T790M lung cancer [
94,
95]. A phase I study (BLOOM, NCT02228369), which has enrolled pretreated
EGFR-mutant NSCLC patients with leptomeningeal metastasis treated with 160 mg osimertinib once daily, is ongoing. The preliminary data is promising [
96].
Rociletinib, a 2,4-disubstituted pyrimidine compound, is an oral, irreversible, mutant-selective inhibitor of activating
EGFR mutations, including T790M, and spares wild-type
EGFR [
97]. TIGER-X (NCT01526928A), a phase I/II trial of rociletinib, enrolled 130
EGFR-mutant NSCLC patients with acquired resistance to first- or second-generation EGFR TKIs [
83]. The ORR was 59% for the 46 evaluable T790M mutation-positive patients and 29% for the 17 T790M mutation-negative patients [
83]. Because of targeting of IGF-1R, hyperglycemia (22%) was detected as the most common grade 3 AE. An independent updated analysis of the TIGER-X trial showed that the T790M mutation-positive patients had an ORR of 45% [
98]. In addition, a series of cases with response to osimertinib after resistance to rociletinib were reported [
99]. Clovis Oncology, Inc. decided to stop enrollment in all ongoing rociletinib studies and terminate the future development program in May 2016.
A phase I/II dose escalation clinical trial, HM-EMSI-101 (NCT01588145), was initiated in South Korea [
100]. Patients took olmutinib in doses ranging from 75 to 1200 mg/day. Among the 34 patients with NSCLC harboring T790M detected by a central laboratory, the ORR was 58.8%. The DCR was 97.1% for patients treated with olmutinib in doses greater than 650 mg. The most common DLTs involved gastrointestinal symptoms, abnormal liver function (AST/ALT), and increasing amylase/lipase levels. Therefore, 800 mg/day was selected as the recommended phase II dose. Seventy-six patients with centrally confirmed T790M mutation-positive NSCLC were enrolled in part II of the study, and 70 were evaluable for response. The ORR was 61% and median PFS was 6.9 months [
101]. Based on the aforementioned result, olmutinib was first approved in South Korea in 2016. However, Boehringer Ingelheim decided to stop the co-development of this drug because of an unexpected grade 3/4 skin toxicity (including palmoplantar keratoderma) [
102].
Preclinical data showed ASP8273 had antitumor activity against EGFR TKI-resistant cells, including those with resistance to osimertinib and rociletinib [
103]. A multi-cohort, phase 1 study (NCT02113813) was initiated to evaluate the safety and efficacy of ASP8273 in NSCLC patients with disease progression after EGFR TKI treatment. The most common AEs included diarrhea (47%), nausea (42%), and fatigue (32%). The most common grade 3/4 AE was hyponatremia (17%). Across all doses, the ORR was 30.7%, and median PFS was 6.8 months in patients with
EGFR T790M [
104]. A phase III randomized clinical trial (SOLAR) was conducted to compare the efficacy and safety of ASP8273 with that of erlotinib or gefitinib as first-line treatment for advanced
EGFR-mutant NSCLC (NCT02588261). However, Astellas Pharma (OTCPK: ALPMY) terminated the phase III SOLAR study in May 2017 because the treatment advantage apparently was not adequate enough to justify continuation.
A phase I/II first-in-human study, NCT02108964 (EGF816X2101), investigated nazartinib in
EGFR-mutant patients. A total of 152 patients were treated across seven cohorts using doses ranging from 75 to 350 mg [
105]. Among the 147 evaluable patients, the ORR and DCR were 46.9% and 87.1%, respectively. The median PFS across all dose cohorts was 9.7 months. Skin rash (54%), diarrhea (37%), and pruritus (34%) were the most common AEs. The skin rashes related to nazartinib were different from those caused by other EGFR TKIs in pattern, location, and histology. The most common grade 3/4 AE was diarrhea (16%) [
105]. A phase II clinical trial with six cohorts is ongoing. In addition, a phase Ib/II trial (NCT02335944 and NCT02323126) is ongoing to investigate the efficacy of combined treatments with INC280, a specific MET inhibitor, and with nivolumab, an anti-PD-1 monoclonal antibody in patients with
EGFR-T790M mutation after acquired resistance to first-line EGFR TKI.
A phase I/II, first-in-human dose-escalation and expansion phase clinical trial (NCT02330367) was carried out with advanced NSCLC patients with acquired T790M mutation after first-generation EGFR TKIs treatment [
106]. In all, 136 patients have been treated across seven cohorts (50, 100, 150, 200, 250, 300, and 350 mg BID), and MTD has not been reached. The most common drug-related AEs were diarrhea (38%), rash (26%) and ALT/AST elevation. Grade 3/4 AEs of diarrhea (2%), rash (2%) and ALT/AST elevation (4%, 2%) were recorded. The 124 evaluable patients had ORR and DCR of 44% and 85%, respectively. Because of the drug safety profile and activity against NSCLC with acquired T790M mutation, a phase II, AEGIS-1 study is ongoing to evaluate treatment efficacy for patients with T790M mutation-positive NSCLC with acquired resistance to first-generation EGFR TKIs. An open label, randomized phase III trial (NCT03058094) also is ongoing to compare AC0010 (300 mg, BID) with pemetrexed/cisplatin (4–6 cycles) in patients with advanced NSCLC who have progressed following prior therapy with EGFR TKI. T790M in biopsy samples was confirmed by a central laboratory.
An open-label, multicenter, phase I/II dose escalation and expansion trial (NCT02981108) is currently recruiting patients with locally advanced or metastatic NSCLC after acquired resistance to first- and/or second-generation EGFR TKIs.
PF-06747775 has potent antitumor efficacy against NSCLC harboring a classical mutation with/without T790M. It significantly attenuates T790M activity and has less toxicity because of the reduction of proteome reactivity relative to earlier EGFR TKIs [
107,
108]. A phase I/II clinical trial (NCT02349633) involving patients with advanced NSCLC harboring
EGFR mutations (Del19 or L858R with/without T790M) is ongoing.
Combination therapy
Cetuximab is a recombinant human/mouse chimeric EGFR IgG1 monoclonal antibody. Combining afatinib and cetuximab may be useful for patients who have progressed after receiving EGFR TKI therapy and chemotherapy [
109]. Among 126 patients, the response rate of patients with T790M-positive and T790M-negative tumors was comparable (32% vs. 25%;
p = .341). The two groups showed no statistical difference in PFS. The NCCN Panel recommends considering an afatinib/cetuximab regimen for patients who have progressed after receiving EGFR TKIs and chemotherapy [
48]. However, skin rash (90% all grades) and diarrhea (71% all grades) were the two most common adverse effects. Grades 3 and 4 adverse effects were 44% and 2%, respectively. Because of the high rate of AEs with this combination therapy, it is no longer a preferred treatment for patients with tumor harboring
EGFR T790M mutations [
110].
Since bypass signaling pathway activation is an important acquired resistance mechanism of EGFR TKIs, it is reasonable to combine inhibition of EGFR pathway signaling and inhibitors for the bypass signaling pathway to overcome resistance. Different horizontal combination strategies are being investigated, but results are preliminary and immature (Table
4).
Table 4
Main mechanisms involved in acquired resistance to EGF receptor-tyrosine kinase inhibitors and the associated targetable drugs
HER2 amplification | | Afatinib, Trastuzumab, ado-trastuzumab emtansine (TDM1) |
MET overexpression/genetic alteration | | ● Anti-HGF antibody: Rilotumumab, Ficlatuzumab ● Anti-c-MET antibody: MET Mab, Emibetuzumab (LY2875358) ● Selective c-MET inhibitor: Tivantinib (ARQ197), Capmatinib (INC280), Savolitinib (AZD6094), Tepotinib (EMD 1214063), SGX523, SAR125844, ● Multikinase inhibitors: Crizotinib, Cabozantinib (XL184), Glesatinib (MGCD265), Merestinib (LY2801653), S49076 |
PIK3CA | PI3K-AKT-mTOR | ● PI3K inhibitor: Pilaralisib (XL147), Dactolisib (BEZ235) and Pictilisib (GDC-0941), Buparlisib (BKM120) ● AKT inhibitor: MK-2206 ● mTOR inhibitor: Everolimus, Temsirolimus, Ridaforolimus |
BRAF | Ras-Raf-MEK-ERK | Vemurafenib (PLX4032), Dabrafenib (GSK2118436), Selumetinib, LY3009120 |
AXL overexpression | GAS6-AXL | ● Tyrosine kinase inhibitor: Cabozantinib (XL 184) ● AXL antibody: E8, D9, Mab173 ● AXL decoy receptor: AXL-Fc, MYDI |
MET amplification is an important mechanism of acquired resistance to EGFR TKI therapy [
31,
111]. A randomized, open-label, phase 2 study enrolled patients with advanced NSCLC (enriched for
EGFR-mutant disease) who developed acquired resistance to erlotinib to receive emibetuzumab (LY2875358), a humanized IgG4 monoclonal bivalent MET antibody, with or without erlotinib therapy. The ORR of patients whose re-biopsy samples harbored MET overexpression (≥60%) was 3.8% in the combination arm and 4.8% in the monotherapy arm [
112]. In Japan, another phase II clinical trial enrolled 45 patients with advanced
EGFR-mutant NSCLC who developed acquired resistance to first-generation EGFR TKIs to receive tivantinib (ARQ197) and erlotinib combination therapy. The response rate was 6.7%. High MET expression (≥ 50%) was detected by immunohistochemical stain in 48.9% of the patients, including all three partial responders [
113]. In addition, a combination of capmatinib (INC280) and gefitinib was tested in a phase 2 study (NCT01610336) in
EGFR-mutant NSCLC patients after acquired resistance to gefitinib.
EGFR T790M NSCLCs were excluded and high cMET expression was required. Of the 65 evaluable patients, the ORR was 18% and DCR was 80%. More responses were seen in tumors with
MET amplifications [
114].
In addition to
MET amplification, different medications are being investigated to inhibit other bypass signaling pathways, including a heat shock protein 90 inhibitor, AUY922 (
ClinicalTrials.gov: NCT01259089 and NCT01646125); a JAK inhibitor, ruxolitinib (
ClinicalTrials.gov: NCT02155465 and NCT02145637); a MET/AXL/FGFR inhibitor S- 49076 (
EU Clinical Trials Register: EudraCT Number: 2015–002646-31) and a PI3K inhibitor, buparlisib (BKM120) (
ClinicalTrials.gov: NCT01570296 and NCT01487265).
Furthermore, combination therapy with osimertinib has been investigated. The TATTON study (
ClinicalTrials.gov: NCT02143466) enrolled patients who received osimertinib-based combination therapy with either a MET inhibitor (savolitinib), MEK inhibitor (selumetinib), or anti-PD-L1 monoclonal antibody (durvalumab) [
115]. However, the rate of drug-related interstitial disease was high in the osimertinib plus durvalumab arm, so the development of this combination therapy was discontinued [
116]. Other clinical trials, including osimertinib in combination with ramucirumab, necitumumab, bevacizumab, or navitoclax (
ClinicalTrials.gov, NCT02789345, 02496663, 02803203 and 02520778), are ongoing.
Combination therapies have higher rates of toxicities and side effects than a single agent does. Although the aforementioned medications have been evaluated in clinical trials, clinicians should keep in mind the possibility of AEs when prescribing combination therapy.
Immunotherapy
For subsequent therapy, or immunotherapy, nivolumab and pembrolizumab have been approved as standard treatment, and high-level PD-L1 expression in tumors can predict a higher response rate. Phase III trials assessing pembrolizumab, nivolumab, or atezolizumab compared to docetaxel as subsequent therapy for patients with metastatic NSCLC found there were no survival benefits for
EGFR-mutant lung cancer patients. Also, there were not enough patients with these mutations to determine whether there were statistically significant differences. However, immunotherapy was comparable to chemotherapy and was better tolerated. [
117‐
119]. Until now, there is not enough evidence to recommend pembrolizumab, nivolumab, or atezolizumab as subsequent therapy for
EGFR-mutant patients.
In vitro,
EGFR-mutant lung cancer cells inhibited antitumor immunity by activating the PD-1/PD-L1 pathway to suppress T-cell function [
120]. This finding indicates that EGFR functions as an oncogene through cell-autonomous mechanisms and raises the possibility that other oncogenes may drive immune escape [
120]. However, retrospective studies showed that NSCLCs harboring
EGFR mutations were associated with low response rates to PD-1/PD-L1 inhibitors, which may have resulted from low rates of concurrent PD-L1 expression and CD8(+) TILs within the tumor microenvironment [
119]. A retrospective study on the efficacy of nivolumab in patients with
EGFR mutation-positive NSCLC after EGFR TKI failure found that T790M-negative patients were more likely than T790M-positive patients to benefit from nivolumab [
121].