Real-world implementation of sequential targeted therapies for EGFR-mutated lung cancer

Abstract

Background: Epidermal growth factor receptor-mutated (EGFR⁺) non-small-cell lung cancer (NSCLC) patients failing tyrosine kinase inhibitors (TKI) can benefit from next-line targeted therapies, but implementation is challenging.

Methods: EGFR⁺ NSCLC patients treated with first/second-generation (1G/2G) TKI at our institution with a last follow-up after osimertinib approval (February 2016), were analyzed retrospectively, and the results compared with published data under osimertinib.

Results: A total of 207 patients received erlotinib (37%), gefitinib (16%) or afatinib (47%). The median age was 66 years, with a predominance of female (70%), never/light-smokers (69%). T790M testing was performed in 174/202 progressive cases (86%), positive in 93/174 (53%), and followed by osimertinib in 87/93 (94%). Among the 135 deceased patients, 94 (70%) received subsequent systemic treatment (43% chemotherapy, 39% osimertinib), while 30% died without, either before (4%) or after progression, due to rapid clinical deterioration (22%), patient refusal of further therapy (2%), or severe competing illness (2%). Lack of subsequent treatment was significantly (4.5x, p < 0.001) associated with lack of T790M testing, whose most frequent cause (in approximately 50% of cases) was also rapid clinical decline. Among the 127 consecutive patients with failure of 1G/2G TKI started after November 2015, 47 (37%) received osimertinib, with a median overall survival of 36 months versus 24 and 21 months for patients with alternative and no subsequent therapies (p = 0.003).

Conclusion: Osimertinib after 1G/2G TKI failure prolongs survival, but approximately 15% and 30% of patients forego molecular retesting and subsequent treatment, respectively, mainly due to rapid clinical deterioration. This is an important remediable obstacle to sequential TKI treatment for EGFR⁺ NSCLC. It pertains also to other actionable resistance mechanisms emerging under 1G/2G inhibitors or osimertinib, whose rate for lack of next-line therapy is similar (approximately 35% in the FLAURA/AURA3 trials), and highlights the need for closer monitoring alongside broader profiling of TKI-treated EGFR⁺ NSCLC in the future.

Keywords: EGFR T790M mutation; EGFR+ NSCLC; overall survival; rebiopsy; second line; tyrosine kinase inhibitor.

Figure 1.

Flow diagram of the entire study population and subsets used to calculate each endpoint. 1G/2G, first/second-generation; EGFR⁺, epidermal growth factor receptor-mutated; FU, follow up; NSCLC, non-small-cell lung cancer; TKI, tyrosine kinase inhibitors.

Figure 2.

Molecular testing, T790M positivity and next-line osimertinib after failure of 1G/2G-generation EGFR inhibitors in the entire study population. (a) Rate of T790M testing, T790M positivity and next-line osimertinib administration among patients with documented radiologic disease progression under 1G/2G EGFR TKI (n = 202, Figure 1). Error bars indicate 95% CI. (b) Reasons for lack of T790M testing in 14% of progressive patients (chi-square p < 0.001). Patient refusal was due to severe side effects from first-line TKI (5/6), and reluctance to undergo bronchoscopy (1/6). Severe competing illness leading to decision against further anticancer therapy was dementia in two cases, and glioblastoma multiforme in the third. Refractory disease was primary progressive disease at the first restaging after start of 1G/2G EGFR TKI in 2/3 cases, and small-cell transformation at the time of TKI failure in 1/3. (c) Reasons for lack of treatment with osimertinib despite T790M positivity in 3% of progressive patients (chi-square p = 0.47). (d) Utilization and positivity rate of liquid rebiopsies, tissue rebiopsies, and their combination for T790M testing in our patients (chi-square p = 0.076 for trend regarding positivity). 1G/2G, first/second-generation; CI, confidence intervals; EGFR, epidermal growth factor receptor; SCLC, small-cell lung cancer; TKI, tyrosine kinase inhibitors.

Figure 3.

Systemic treatment after failure of 1G/2G EGFR inhibitors in the subset of deceased patients. (a) Rate for implementation of any subsequent treatment, subsequent CHT, and subsequent osimertinib after failure of 1G/2G EGFR inhibitors. Error bars indicate 95% CI. (b) Association between administration of any next-line therapy and performance of T790M testing (chi-square p < 0.001). (c) Association between administration of any next-line therapy and results of T790M testing (chi-square p = 0.041). (d) Reasons for lack of T790M testing in 30% of deceased patients (chi-square p < 0.001). Severe competing illness leading to withdrawal of further treatment was dementia in two cases, and glioblastoma multiforme in the third. 1G/2G, first/second-generation; CHT, chemotherapy; CI, confidence intervals; EGFR, epidermal growth factor receptor; TKI, tyrosine kinase inhibitors.

Figure 4.

OS according to next-line treatment in the consecutive subset of EGFR⁺ NSCLC patients. Within the subset of consecutive patients with 1G/2G TKI start after 15 November 2015 (n = 127, Figure 1), median OS from the start of systemic treatment for stage IV disease was significantly longer for patients that received next-line osimertinib, compared with patients that received alternative or no subsequent therapies: 36.0 (95% CI 26.3–45.7) versus 23.5 (18.9–28.2) versus 20.6 (14.4–26.8) months, respectively, logrank test for trend p = 0.0031. 1G/2G, first/second-generation; CI, confidence interval; EGFR⁺, epidermal growth factor receptor-mutated; NSCLC, non-small-cell lung cancer; OS, overall survival; TKI, tyrosine kinase inhibitors.

Figure 5.

Implementation of sequential targeted therapies for metastatic EGFR⁺ NSCLC. Critical parameters, possible improvement strategies and feasibility limit for implementation of sequential targeted therapies in metastatic EGFR⁺ NSCLC. The maximum testing rate of 95% is taken from Supplemental Table S2 and excludes only patients with EGFR TKI discontinuation due to reasons precluding further treatment (i.e. “other reasons” or patient decision); the actual testing rate of 80–85% is taken from Figure 2a, Supplemental Figure S1a and the literature cited in the Discussion; the actual treatment rate for T790M positive patients with osimertinib of 90–95% is taken from Figure 2a and Supplemental Figure S1A; the maximum treatment rate for T790M positive patients of >95% additionally considers that 3/6 T790M positive patients foregoing osimertinib treatment suffered early death before the drug could be started (Figure 2c), which could potentially have been prevented by an earlier change in therapeutic strategy facilitated by closer patient monitoring; the x% rate of marker-positivity is mutation-specific, for example, approximately 55% for EGFR T790M under 1G/2G EGFR TKI (Figure 2a).^, For any next-line targeted therapy, the theoretical upper limit of implementation is the product of these three parameters (0.95*0.95*x), for example, approximately 50% for osimertinib after 1G/2G EGFR TKI. The rate of molecular testing affects the feasibility of implementation for all next-line targeted therapies. This framework appears to be very similar for 1G/2G EGFR inhibitors and osimertinib, as shown in Supplemental Table S2 and explained in the Discussion. 1G/2G TKI, first/second-generation tyrosine kinase inhibitors; EGFR, epidermal growth factor receptor; EGFR⁺, epidermal growth factor receptor-mutated; ePRO, electronic patient-reported outcomes; NGS, next-generation sequencing; NSCLC, non-small-cell lung cancer.