Patient disposition and baseline characteristics
Between September 2019 and June 2022, a total of 717 Chinese patients were enrolled and screened for eligibility. Of them, 318 were excluded for reasons listed in Supplementary Table 1. The remaining 336 patients met the inclusion and exclusion criteria and were randomized in a 2:1 ratio to receive mefatinib or gefitinib. A total of 223 patients received mefatinib, and 113 patients received gefitinib (Fig. 1). As shown in Table 1, the two groups had well-balanced and comparable demographics and baseline clinicopathological features. The mefatinib group had a median age of 63.0 years (range: 55.0–69.0), with 57.8% females. The gefitinib group had a median age of 60.0 years (range: 55.0–66.0), and 57.5% were females. Given that the EGFR genotype was a stratification factor, almost half of the cohort in each treatment arm harbored L858R (51.2%), and the other half had ex19del (48.8%), with their baseline characteristics summarized in Supplementary Table 2.
Fig. 1
Study participant flow diagram
Table 1 Patient demographics and baseline characteristics in the full analysis set
Primary efficacy endpoint
At the data cutoff date (June 30, 2023), the median follow-up was 15.9 months (range: 1.1–36.3) for the mefatinib group and 18.5 months (range: 1.1–36.4) for the gefitinib group. Mefatinib had a significantly longer PFS than gefitinib (13.7 months [95% CI: 11.0–15.2] vs 9.7 months [95% CI: 8.3–12.4 months]; HR = 0.68 [95% CI: 0.53–0.87]; p = 0.002; Fig. 2a). The independent review committee (IRC)-assessed PFS was consistent with the investigator-assessed PFS (Supplementary Fig. 2; Appendix p 9). Moreover, as shown in Fig. 2b, the subgroup analysis demonstrates the PFS advantage with mefatinib over gefitinib regardless of clinical factors, such as sex, age, smoking history, Eastern Cooperative Oncology Group (ECOG) performance status (PS), and EGFR genotype.
Fig. 2
Progression-free survival of the cohort. a Kaplan‒Meier curve comparing progression-free survival based on the independent review committee of the mefatinib and gefitinib groups of the full analysis cohort. The risk table below shows the number of cases included and censored per time point. b Forest plot showing the subgroup analysis of the impact of various clinical factors on progression-free survival
PFS by EGFR mutation subtype
Subgroup analyses of PFS outcomes based on EGFR genotypes were also performed. Patients with EGFR ex19del had a comparable median IRC-assessed PFS with mefatinib and gefitinib (12.6 months [95% confidence intervals (CI): 11.0–15.2] vs 12.5 months [95% CI: 9.6–15.1]) (hazard ratio [HR] = 0.83, 95% CI: 0.58–1.19; p = 0.307; Fig. 3a). IRC- and investigator-assessed clinical outcomes were broadly comparable between treatment arms among patients with EGFR ex19del (p > 0.100; Supplementary Table 2). Furthermore, patients with EGFR L858R had a significantly longer median IRC-assessed PFS when treated with mefatinib than gefitinib (13.7 months [95% CI: 9.7–15.3] vs 8.3 months [95% CI: 7.0–9.7]; HR = 0.55 [95% CI: 0.38–0.78]; p = 0.001; Fig. 3b). Consistently, the median investigator-assessed PFS was significantly longer with mefatinib than with gefitinib among patients with EGFR L858R (13.1 months [95% CI: 10.9–13.9] vs 9.6 months [95% CI: 6.9–11.0]; HR = 0.60 [95% CI: 0.43–0.85]; p = 0.003; Supplementary Table 3).
Fig. 3
Progression-free survival based on EGFR genotype. Kaplan‒Meier curves comparing progression-free survival based on the independent review committee between patients with EGFR exon 19 deletion (ex19del) (a) and patients with EGFR L858R (b) treated with mefatinib and gefitinib. The risk table below shows the number of cases included and censored per time point
Secondary efficacy endpoints
IRC-assessed and investigator-assessed treatment outcomes are summarized in Table 2 and Supplementary Table 3, respectively. No significant difference was observed between the two groups in terms of IRC-assessed objective response rate (ORR; 81.2% vs 77.9%; odds ratio: 1.23 [95% CI: 0.70–2.14]; p = 0.477) and disease control rate (DCR; 89.7% vs 90.3%; p = 0.868) or investigator-assessed ORR and DCR. Among the treatment responders, those treated with mefatinib had a significantly longer median IRC-assessed duration of response (DoR; 12.5 months [95% CI: 11.0–14.0] vs 9.7 months [95% CI: 7.0–11.1]; HR = 0.68 [95% CI: 0.52–0.91]; p = 0.008; Table 2; Supplementary Fig. 3A). Patients with EGFR ex19del similarly benefited from treatment with mefatinib and gefitinib (12.6 months vs 11.1 months; HR = 0.83 [95% CI: 0.58–1.19]; p = 0.284; Table 2). Notably, patients with EGFR L858R benefited more from mefatinib than gefitinib, as shown by the significantly longer IRC-assessed DoR (12.5 months [95% CI: 9.7–15.2] vs 8.2 months [95% CI: 6.7–9.7]; HR = 0.54 [95% CI: 0.36–0.81]; p = 0.003; Table 2).
Table 2 Clinical outcomes of the full analysis set assessed by the independent review committee
IRC-assessed time to progression (TTP) was also significantly longer for mefatinib than for gefitinib (13.8 months [95% CI: 11.2–15.2] vs 9.7 months [95% CI: 9.5–12.4]; HR = 0.70 [95% CI: 0.54–0.90]; p = 0.005; Supplementary Fig. 3C). Specifically, patients with EGFR L858R had a significantly longer IRC-assessed TTP with mefatinib than gefitinib (13.8 months [95% CI: 10.9–16.4] vs 8.3 months [95% CI: 8.2–9.8]; HR = 0.56 [95% CI: 0.39–0.81]; p = 0.002; Table 2). The median time to treatment failure (TTF) was significantly longer for mefatinib than gefitinib in the overall cohort (12.9 months vs 11.5 months; HR = 0.76 [95% CI: 0.60–0.97]; p = 0.026) and in patients with EGFR L858R (12.6 months vs 9.8 months; HR = 0.64 [95% CI: 0.46–0.89]; p = 0.008) (Supplementary Table 3).
Overall survival
The overall survival (OS) data for both groups remain immature. As of the data cutoff date, OS events occurred in 37.2% (83/223) of the mefatinib group and 39.8% (45/113) of the gefitinib group. There was no significant difference in OS between the two treatment groups (35.4 months vs 33.3 months; HR for death: 0.90 [95% CI: 0.63–1.30]; p = 0.577; Supplementary Fig. 4). The 30-month OS rate was 60.2% for mefatinib and 54.3% for gefitinib. Among the patients with EGFR L858R, OS events occurred in 41.2% (47/114) of those who received mefatinib and 50.0% (29/58) of those who received gefitinib. Patients with EGFR L858R had a trend of longer median OS, albeit with no significant difference (35.9 months [95% CI: 29.1–not evaluable (NE)] vs 29.1 months [95% CI: 22.1, NE]; HR = 0.76 [95% CI: 0.48–1.20]; p = 0.236 Supplementary Fig. 5). The 30-month OS rates in patients with EGFR L858R were 56.6% [95% CI: 45.6%, 66.2%] for those treated with mefatinib and 43.7% [95% CI: 26.9%, 59.3%] with gefitinib (Supplementary Table 3). Supplementary Table 4 summarizes the second-line therapy among patients who had experienced disease progression as of the data cutoff. Third-generation EGFR-TKI alone or in combination with either an antiangiogenic inhibitor or MET inhibitor was the most frequently administered agent, received by over half of the patients in both the mefatinib arm (56.0%, 79/141) and gefitinib arm (60.7%, 51/84).
Health-related quality of life
The mefatinib and gefitinib groups had generally comparable health-related quality of life (HRQoL) in terms of functional status, global health status, and symptoms, including fatigue, nausea and vomiting, pain, shortness of breath, insomnia, and loss of appetite. The HRQoL assessments for both groups from baseline and at different time points are summarized in Supplementary Tables 5 and 6. Based on the EORTC QLQ-C30 assessment, treatment with either mefatinib or gefitinib did not result in significant changes in functional scales, symptom scales, or overall health status, except for the significantly higher rates of diarrhea reported in the mefatinib group after receiving 24 weeks of treatment relative to baseline than in the gefitinib group (p < 0.001; Supplementary Table 5). Based on the EORTC QLQ-LC13 assessment, the mefatinib group had significantly higher rates of mouth pain (p = 0.002) and tingling hands or feet/peripheral neuropathy (p = 0.004) after receiving 24 weeks of treatment relative to baseline than the gefitinib group (Supplementary Table 6).
Safety
The mefatinib group had a mean treatment duration of 16.2±8.5 months (95% CI: 1.2–36.4), while the gefitinib group had a mean treatment duration of 17.6±10.1 months (95% CI: 1.1–36.5). AEs were observed in almost all participants, with adverse events (AEs) of any grade reported in 99.6% of the mefatinib group and 100.0% of the gefitinib group (Supplementary Table 7). In the mefatinib group, the most frequently reported treatment-related AEs (TRAEs) were diarrhea (96.0%), rash (83.9%), and stomatitis (62.8%). On the other hand, the most frequently reported TRAEs in the gefitinib group were rash (62.8%), diarrhea (61.9%), and alanine aminotransferase (ALT) increase (57.5%) (Table 3). The incidence rates of TRAEs of special interest are summarized in Supplementary Table 8.
Table 3 Treatment-related adverse events (TRAE) observed in ≥10% of the safety population
The mefatinib group experienced a higher incidence of Grade 3 and higher AEs than the gefitinib group (61.0% vs 45.1%), with Grade 3 and higher TRAEs of 45.7% and 24.8%, respectively. Compared with the gefitinib group, the mefatinib group had higher rates of treatment-related grade 3 or higher diarrhea (20.0% vs 0.9%), stomatitis (6.3% vs 0.9%), rash (11.7% vs 1.8%), weight loss (3.1% vs 1.8%), hypokalemia (4.0% vs 0.0%), and paronychia (2.7% vs 0.9%) (Table 3). The gefitinib group showed higher rates of treatment-related grade 3 or higher ALT increase (7.1% vs 3.1%) and elevated aspartate aminotransferase (AST) (6.2% vs 2.2%) than the mefatinib group (Table 3).
Notably, the mefatinib and gefitinib groups had a comparable rate of serious AEs (31.8% vs 31.0%), with the mefatinib group having a numerically lower rate of serious TRAEs than the gefitinib group (8.5% vs 10.6%). Among the serious TRAEs, the mefatinib group had a higher incidence of gastrointestinal AEs than the gefitinib group. However, the gefitinib group had a higher incidence of AEs classified under investigation, mainly reflected by the increased incidence of ALT and AST elevation, than the mefatinib group. The mefatinib group had ten cases (4.5%) resulting in death due to AEs, with TRAEs observed in a case who died of unknown etiology. In contrast, the gefitinib group had seven (6.2%) deaths due to AEs, with a case that succumbed to treatment-related grade 5 interstitial lung disease.
The mefatinib group had a higher incidence of AEs and TRAEs leading to permanent discontinuation than the gefitinib group (AEs, 5.4% vs 2.7%; TRAEs, 4.0% vs 1.8%). AEs that led to permanent discontinuation summarized according to system organ class were respiratory, thoracic and mediastinal disorders (1.8% vs 1.8%) and skin and subcutaneous tissue disorders (1.3% vs 0%). There was a low incidence of TRAEs that led to permanent discontinuation in both the mefatinib and mefitinib groups, with treatment-related rash observed at an incidence of 1.3% in the mefatinib group and not reported in the gefitinib group.
Compared with the gefitinib group, the mefatinib group had a higher incidence of AEs, leading to dose reduction (5.8% vs 2.7%). AEs that led to dose reduction were categorized under skin and subcutaneous tissue disorders (3.1% vs 0.9%), gastrointestinal disorders (1.3% vs 0%), and infection and infestations (1.3% vs 0%). Rash (3.1% vs 0.9%) and paronychia (1.3% vs 0%) were the most frequent TRAEs that led to dose reduction.
The incidence of AEs and TRAEs leading to treatment interruption was numerically higher in the mefatinib group than in the gefitinib group (AEs, 35.9% vs 27.4%; TRAEs, 30.9% vs 23.0%). The most frequent TRAEs that led to treatment interruption were diarrhea (9.0% vs 0%), stomatitis (4.9% vs 1.8%), rash (13.0% vs 3.5%), increased ALT (3.1% vs 8.8%), increased AST (3.1% vs 11.5%), increased gamma-glutamyl transferase (0.9% vs 1.8%), neutropenia (0% vs 1.8%), and paronychia (2.7% vs 0.9%). The incidence of diarrhea and rash leading to treatment interruption was higher with mefatinib than with gefitinib. In contrast, the incidence of an ALT/AST increase leading to treatment interruption was higher with gefitinib than with mefatinib.
