Background
Gefitinib was the first epidermal growth factor receptor tyrosine kinase inhibitor (EGFR-TKI) to become available for the treatment of non-small cell lung cancer (NSCLC). Several studies have demonstrated that gefitinib is effective for the second-line treatment of NSCLC [
1‐
3]. Although the phase III ISEL trial failed to prove the superiority of gefitinib treatment compared to placebo in previously treated patients, a subgroup analysis demonstrated improved survival in particular populations (Asians and non-smokers) [
4]. Further analyses in other studies have also revealed that clinical factors (Asians, females, non-smokers, and adenocarcinoma histology) are associated with the response to gefitinib treatment [
5]. EGFR mutations, such as the deletion of exon 19 and the single L858R mutation in exon 21, have also been reported to be correlated with a longer survival and were found more frequently in Asian patients [
6‐
8]. Recently, a superior progression-free survival (PFS) with gefitinib compared with the combination of carboplatin and paclitaxel in untreated NSCLC patients with predictors of gefitinib sensitivity was proven in two large phase III studies [
9,
10]. Gefitinib is now recommended for advanced or metastatic NSCLC patients under such circumstances as a first or a second-line treatment.
Despite the high disease control rate (DCR), gefitinib treatment is not curative and eventually there is disease recurrence, even in patients with predictors of sensitivity. For the many NSCLC patients who previously responded to gefitinib but later showed tumor progression, very few treatment options are available.
Some investigators have conducted studies to evaluate the efficacy of EGFR-TKI re-administration [
11‐
14]. In most of those studies, both gefitinib responders and non-responders were retreated with gefitinib or erlotinib, and gefitinib responders tended to benefit from the 2
nd EGFR-TKI.
Here, we retrospectively analyzed the efficacy of the 2nd EGFR-TKI administration after failure of the initial gefitinib treatment in NSCLC patients who had previously achieved disease control with gefitinib. The risks of the 2nd administration of EGFR-TKI, especially the association with adverse events in the initial gefitinib treatment, were also evaluated.
Discussion
To the best of our knowledge, 18 cases of patients who received gefitinib re-administration after failure of the initial gefitinib treatment have been reported to date, including 3 cases reported by our group (Table
6) [
17‐
21]. All 18 patients responded to the initial gefitinib treatment, and most of the cases underwent cytotoxic chemotherapy between the first and second gefitinib therapy. Fourteen patients benefited from the 2
nd gefitinib treatment, and the overall DCR was 78%. In our 3 patients, the toxicity of the 2
nd gefitinib was similar to that observed for the initial gefitinib treatment and was acceptable. Gefitinib retreatment is likely a good option for patients who have demonstrated a response to a previous gefitinib treatment.
Table 6
Patient characteristics of the previous studies of gefitinib readministration
Yokouchi H et al. | 9 | 9 | - | 8 | 1 |
Yoshimoto A et al. | 1 | 1 | - | 1 | 0 |
Yano S et al. | 3 | 3 | - | 2 | 1 |
Hashimoto N et al. | 1 | 1 | - | 0 | 1 |
Kurata T et al. | 1 | 1 | - | 1 | 0 |
Clinical studies have demonstrated that erlotinib is effective even in patients who are not considered to be good responders to gefitinib, such as those with a negative EGFR mutation, squamous cell carcinoma, or a history of smoking [
22]. Because erlotinib is used at its maximum tolerated dose, whereas gefitinib is used at only about one-third of its maximum tolerated dose in daily practice, the biological activity of erlotinib at standard doses may be higher than that of gefitinib [
2,
4,
23‐
25]. These reports suggest that erlotinib may be active even in patients who demonstrated tumor progression during a prior gefitinib treatment. Thus, erlotinib has been selected as a treatment option for use after gefitinib failure (Table
7) [
11‐
14,
26‐
33]. In most studies, including the present investigation, favorable results have been documented, and the authors have concluded that erlotinib appears to be a useful treatment after gefitinib failure.
Table 7
Patient characteristics of the previous studies of erlotinib after gefitinib failure
Lee DH et al. | 23 | 17 | 6 | 2 | 21 | 9 |
Cho BC et al. | 21 | 10 | 11 | 6 | 15 | 29 |
Viswanathan A et al. | 5 | 4 | 1 | 0 | 5 | 0 |
Costa DB et al. | 18 | 16 | 2 | 4 | 14 | 22 |
Sim SH et al. | 16 | 11 | 5 | 4 | 12 | 25 |
Chang JW et al. | 1 | 1 | 0 | 1 | 0 | 100 |
Garfield DH et al. | 1 | 1 | 0 | 1 | 0 | 100 |
Vasile E et al. | 8 | 8 | 0 | 5 | 3 | 63 |
Gridelli C et al. | 3 | 3 | 0 | 3 | 0 | 100 |
Wong AS et al. | 14 | 9 | 5 | 5 | 9 | 36 |
Zhou ZT et al. | 21 | 15 | 6 | 10 | 11 | 48 |
Wong MK et al. | 21 | 18 | 3 | 12 | 9 | 57 |
Although it is difficult to address the precise mechanism underlying these results, several studies have suggested a possible explanation for the clinical benefit of EGFR-TKI retreatment. Some cytotoxic agents have been reported to restore the sensitivity of NSCLC cells to gefitinib in vitro by increasing EGFR phosphorylation [
34,
35]. It is also possible that chemotherapy during the EGFR-TKI-free interval could decrease EGFR-TKI resistant tumor cells. However, no significant differences in PFS or OS were observed between our patients who received chemotherapy before the 2
nd EGFR-TKI and those who received the 2
nd EGFR-TKI immediately after gefitinib failure. In addition, the duration between the initial gefitinib and the 2
nd EGFR-TKI treatments was not associated with the response to 2
nd EGFR-TKI. Similarly to these findings, other researchers have found no evidence that either chemotherapy among the 1
st and 2
nd EGFR-TKIs or the duration of the EGFR-TKI-free period affects either PFS or OS in the 2
nd EGFR-TKI [
31,
33].
Secondary EGFR mutations might be associated with the efficacy of erlotinib after gefitinib failure. MET amplification and secondary EGFR mutations, such as T790 M, L747 S, D761Y, and T854A have been identified in NSCLC patients with an acquired resistance to EGFR-TKI [
36‐
42]. T790 M mutation was found in 50%, MET amplification in 20%, and other secondary mutations in less than 5% of the NSCLC patients carrying EGFR mutations with TKI resistance [
43,
44]. In vitro studies have revealed that tumor cells carrying non-T790 M mutations show a partial resistance to EGFR-TKI, but are much less resistant compared to cells with T790 M. These data suggest that an increased EGFR-TKI dose might circumvent the acquired resistance caused by non-T790 M mutations. Previous studies have indicated that the serum concentration of erlotinib is several-fold higher than that of gefitinib at standard doses [
24,
25]. This difference in biological activities between the TKIs may contribute to the efficacy of erlotinib after gefitinib failure in patients carrying non-T790 M mutations.
In conclusion, our findings suggest that a 2
nd EGFR-TKI can be a treatment option for patients who benefited from a previous gefitinib treatment. However, as shown in Table
7 some studies failed to demonstrate the efficacy of 2
nd EGFR-TKI after gefitinib failure. Cho et al. mentioned that the tumor response to 1
st gefitinib treatment can be a predictive marker [
14]. They described that patients who showed SD during 1
st gefitinib treatment had better survival with 2
nd EGFR-TKI, however those who had PD to 1
st gefitinib rarely responded to 2
nd EGFR-TKI. The difference in the percentage of patients with a good predictor might affect the results of these trials about 2
nd EGFR-TKI. Intense research has been devoted to clarifying the mechanism responsible for acquired resistance, but it is difficult to obtain clinical samples from all patients to confirm MET amplification or secondary mutations. Jackman et al. recently published a clinical definition of acquired resistance to EGFR-TKI [
45]. This consensus definition will facilitate the establishment of standard entry criteria for studies investigating acquired resistance. All of our patients except one met these criteria (no. 8 in Table
2). Despite rapid tumor progression during a previous cytotoxic chemotherapy, this patient obtained SD with an initial gefitinib therapy of 4.3 months, and therefore we considered this patient to have benefited from the gefitinib treatment. Further clinical trials are required to develop a novel treatment for patients with acquired resistance.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SW conducted the study and drafted the manuscript. JT conceived and designed the study and collected the clinical data. TO, RK, HT, HK and TM participated in the patient care, and collected the data. KI, JK and JB analyzed and interpreted the data. IN and HY provided the administrative support. All the authors have read and approved the final version of the manuscript.