First-generation EGFR tyrosine kinase inhibitor therapy in 106 patients with compound EGFR-mutated lung cancer: a single institution’s clinical practice experience

Lung cancer continues to be the primary cause of cancer death both in China and worldwide. The 5-year survival rate for lung cancer has increased from 12% in the 1970s to 17.7% in 2016 [1], and these sufferers primarily include numerous patients with advanced lung cancer. Many advances have been made in the treatment of advanced lung cancer, especially with regard to targeted therapies, and these treatment strategies have resulted in considerable improvements in survival. Among them, the receptor tyrosine kinases (RTKs) super-family of cell surface receptors serve as mediators of cell signaling by extra-cellular growth factors. Members of the ErbB family of RTKs, such as ErbB1 (also known as EGFR), ErbB2, ErbB3 and ErbB4, have received much attention, given their strong association with malignant proliferation [2, 3].

Over the past decade, three small-molecule ErbB tyrosine kinases inhibitors (TKIs) have been shown to efficiently target tumour cell survival pathways in advanced non-small cell lung cancers (NSCLC) expressing the epidermal growth factor receptor (EGFR): gefitinib (approved by the US Food and Drug Administration in May 2003), erlotinib (approved by the US Food and Drug Administration in November 2004) and icotinib (approved by China’s State Food and Drug Administration in June 2011). The use of these agents has resulted in higher overall response rates (ORRs, up to 60%–70%) and longer progression-free survival (PFS; 9–16 months) and overall survival (OS; exceeding 20 months) than current first-line platinum-based chemotherapies [1–5].

These patients with classical EGFR genetic mutations that are sensitive to TKIs involve in-frame deletions of exon 19 and L858R substitutions of exon 21 and occur in approximately 85%–90% of all EGFR-mutated patients [1, 4–6]. However, compound EGFR mutations with or without classical mutations have been detected within the same tumour tissues in some patients [7–14]. Previous studies reported that only 2%–15% of the population with EGFR mutations exhibits these rare compound mutations [7–14]. The characteristics of this rare population, the efficacy of EGFR TKIs, and the prognostic value of the compound mutations have not been clarified because of the very low rate of these mutations. Furthermore, the lack of adequate evidence-based medical research hinders treatment decisions when these co-existing double-site mutants are detected.

The present cohort analysis examined two major questions: first, we investigated the incidence of different compound EGFR mutation subtypes in a single institution; and second, we investigated this population’s characteristics and the efficacy (ORRs and PFS) of first-generation small-molecule TKI treatment and prognosis compared with patients bearing the classical mutation alone.

Our single institution study identified 114 patients with compound EGFR mutations among 3925 patients with EGFR mutations (114/3925, 2.9%). The RR of this rare population to first-generation TKI therapy was 50.9%, which was lower than that of patients with a single common mutation, but the difference was not significant (P = 0.088). Patients with compound mutations exhibited a shorter duration of first-generation TKI therapy in multivariate analysis than patients with a single common mutation (HR, 1.981; 95% confidence interval (CI) 1.466–2.676; P < 0.001). Exclusion of patients with a co-occurrence of mutations in exon 20 (excluding T790M) revealed that the duration of targeted TKI therapy was even shorter for other types of compound mutations (P = 0.002).

19Del and L858R mutations are classical sensitizing mutations, and the strong response of these two mutations to TKIs has been demonstrated in many prospective studies [2–5, 9]. However, these two common mutations are frequently detected concomitant with other mutations in the compound EGFR-mutated population. A compound mutation co-existing with a 19Del or L858R mutation was the most common combination in previous reports (203/278, 73.0%) (Table 4) and in our cohort (85/106, 80.2%) (Table 3). Xu et al. [16] reported that tumours with double common EGFR mutations (19Del + L858R, n = 18) exhibited similar antitumour responses to small-molecule TKIs as tumours with single common mutations, and the median PFS and ORR rates were 9.53 months and 71.4% (10/14), respectively, which is consistent with our results (10.1 months and 60%, respectively) and those of Hata Akito (16.5 months and 86%, respectively) [11]. Some case reports also found that first-generation EGFR TKIs may be a desirable therapeutic strategy for patients with advanced lung cancer with synchronous 19Del and L858R mutations [8, 9, 35].

In patients harbouring common plus rare mutation, the L858R mutation was more frequently observed than the 19Del mutation. For example, approximately 10% and 17.3% of NSCLC patients harboured the L858R mutation concomitantly with rare mutations in the two cohorts reported by Wu et al. [17] and Kobayashi et al. [8], respectively. Similarly, in our common plus rare mutation subgroup, L858R, was identified in the majority of the cases (9/11, 81.8%).

The response to TKIs in patients with common plus rare mutations and whether TKI therapy prolonged PFS remains controversial because of the relatively large heterogeneity. Keam [14] reported an RR of 68.8% and median PFS time of 8.1 months in 16 patients, which are similar to our observed RR (45.5%) and median PFS time (10.5 months) in 11 patients. Notably, this above finding was also reported previously [16, 18]. As a whole, this population may benefit from TKIs, but to lesser extent than the population harbouring a single common mutation. We found that the patients with L858R + K757R mutations (exon 21 + exon 19) and L858R + I744M mutations (exon 21 + exon 19) exhibited a partial response to gefitinib and obtained PFS of 9.0- and 9.6-month, respectively. Klughammer et al. [19] and Kempf et al. [20] reported that a single I744M mutation or a single K757R mutation in exon 19 may be TKI-sensitizing mutations, and these mutations were also observed to have PR to oral TKI therapy. Therefore, the above double or single mutation(s) patterns may be candidates for TKI therapy. Patients with exon 20 mutations are considered resistant to TKIs (discussed below), but our study included a patient with L858R plus the R776H mutation (exon 21 + exon 20) showing PR to TKIs for 10.5 months, which is also highly consistent with prior reports [8, 21]. Other patients with the L858R mutation associated with S720P, S768I, L858Q or L833V were classified in the insensitive to TKIs group, and most (5/6, 83.3%) exhibited PD or SD to TKI therapy. To the best of our knowledge, our study is the first to report the combinations of L858R + S720P mutations (exon 21 + exon 18) and L858R + L833V mutations (exon 21 + exon 21). No patient with L833V + H835L mutations (exon 21 + exon 21) was detected in our cohort; however, patients with this combination have been reported to have a good response to gefitinib [22, 23]. One case is especially notable. Leventakos et al. [24] demonstrated that patients with L858R + S768I mutations may be sensitive, or at least not resistant, to TKI therapy, which is in contrast to our results. Our patients with a single L861Q mutation or compound mutations with L861Q exhibited a high RR (66%) and non-inferior PFS (median, 6 months) to TKI therapy. We also detected two cases with 19Del + L861Q mutations, and a good response to TKIs (one PR and one SD) and prolonged PFS (11.4 and 11.9 months) were observed. However, the patient with L858R + L861Q exhibited PD after only 1 month of TKI therapy.

Notably, common mutations concomitant with an initial T790M mutation accounted for 65.1% of all compound mutations in our cohort, which was higher than in previous reports [13]. However, Su et al. [13] verified that pre-treatment of a co-existing EGFR T790M mutation was not a rare event (23/73, 31.5%), and the PFS was significantly shorter than that in patients without T790M (median, 6.7 months vs. 10.2 months, respectively; P = 0.035) [13]. T790M status also affected PFS in our cohort compared with a single-sensitizing EGFR mutation (9.1 months vs. 13.0 months, respectively; P < 0.001). We hypothesize that the scarcity of reports may be due to the bias of excluding patients with T790M [21]. Previous research on compound mutations may have overlooked the fact that—T790M may occur in patients before receiving EGFR TKI treatment [1, 21]. In addition, direct sequencing may be used to detect the classical sensitizing mutations in exons 19 and 21, which leads to a missed opportunity to discover patients with co-occurring T790M. The impact of EGFR TKIs in these patients with 19Del or L858R plus T790M was not clarified because of the scarcity of patients and the varying durations of PFS in the published literature. However, approximately one-third and one-half of patients with concomitant initial T790M as one of the compound mutations in previous studies [13] and our cohort obtained more than 8 months of PFS with the aid of TKI therapy. Therefore, small molecule TKIs may be an optional therapeutic strategy to identify potential beneficiaries after explaining the bias of the therapy to patients in detail to ensure patient understanding and informed consent.

Patients who harbour a single exon 20 mutation in EGFR are reportedly insensitive to small-molecule TKIs [14, 19, 25–27]. However, whether patients with an EGFR exon 20 mutation accompanied by another mutation are candidates for TKI therapy remains unanswered. Marius Lund-Iversen and his colleagues reported seven exon 20-positive patients who received oral TKI, including five patients with single exon 20 mutation and two patients with double mutations. The five patients with single exon 20 mutation were found to have progressive disease at the first post-treatment follow-up, but the two patients with double mutations obtained 11 and 14 months of an ongoing response [28]. Chen et al. [18] also concluded that patients with compound mutations involving mutation in exon 20 benefited from TKIs more than single exon 20 mutations [18]. The duration of response to TKIs in compound EGFR-mutated patients with concomitant exon 20 mutation (excluding T790M) (6.5 months) was still shorter than compound mutated patients without exon 20 mutation (9.1 months) and patients with single common mutations (13.0 months), which is consistent with Keam et al. [14] (< 5 months). The analytical results of Kancha et al. [29] and Wu et al. [17] also support this finding. Together this suggests that first-generation EGFR TKIs may not be suitable for patients with an exon 20 mutation regardless of the presence of other mutations.

Overall, patients with double rare mutations or a rare mutation plus T790M exhibited a lower RR (38.5% and 37.5%, respectively) and worse PFS to TKI therapy (median, 6.5 and 5.4 months, respectively) in our cohort, which is consistent with a previous publication [14]. A patient with a single L861Q point mutation at exon 21 and a G719X point mutation at exon 18 may be classified into the TKI-sensitive mutation group [30]. However, patients with a L861Q or G719X mutation co-occurring with a rare mutation or T790M affected the effectiveness and sensitivity to TKI therapy in our clinical practice (RR, 28.6%; median PFS, 5.1 months). We found five patients with a rare mutation plus T790M, which is more than the overall number of reported cases. One patient with G719X + T790M mutations and one patient with L858Q + T790M mutations exhibited PR to TKI therapy and obtained more than 10 months PFS, which was similar to a case report from Chen et al. [18]. Balak et al. [31] found that the D761Y mutation in exon 19 was a novel secondary resistance mutation to EGFR TKIs [31]. Therefore, the two patients with two resistance mutations (D761Y + T790M) in our study exhibited disease progression very soon after initiating TKI therapy, which was not surprising.

The population in our study was a fairly large cohort to investigate the effectiveness of TKI therapy in patients with compound EGFR-mutated lung cancer. However, this study was a retrospective analysis, which may limit the reliability of the results. Potential selective bias may be unavoidable because of the low incidence of occurrence of these types of mutations. In addition, the RR and PFS of patients with compound EGFR mutations were compared only between patients who received EGFR TKI therapy without inclusion of patients who received chemotherapy. Our data were collected from a single institution, and patients from other areas of China should be examined. All reported cases between 2004 and 2017 were enrolled, but the literature from Asia still accounts for the majority of available data.

Although NSCLC patients with compound mutations exhibited a shorter RFS and lower RR in response to TKI therapy than those with a single common mutation, TKI therapy may still benefit patients with compound mutations. Therefore, after explaining the biases of TKI therapy to patients in detail to ensure their understanding and informed consent, a trial of first-generation small molecule TKIs may be an optional therapeutic strategy to identify potential beneficiaries.