Efficacy and safety of pyrotinib in advanced lung adenocarcinoma with HER2 mutations: a multicenter, single-arm, phase II trial

Background There is currently a lack of effective treatments for non-small cell lung cancer (NSCLC) patients harboring HER2 mutations. We examined the efficacy and safety of, and potential resistance mechanism to, pyrotinib, a pan-HER inhibitor, in advanced NSCLC carrying HER2 mutations. Methods In this multicenter, single-arm, phase II trial, stage IIIB-IV NSCLC patients harboring HER2 mutations, as determined using next-generation sequencing, were enrolled and treated with pyrotinib at a dose of 400 mg/day. The primary endpoint was 6-month progression-free survival (PFS) rate, and secondary endpoints were objective response rate (ORR), PFS, overall survival (OS), disease control rate (DCR), and safety. The impact of different HER2 mutation types on sensitivity to pyrotinib and the potential of utilizing mutational profile derived from circulating tumor DNA (ctDNA) to predict disease progression were also explored. Results Seventy-eight patients were enrolled for efficacy and safety analysis. The 6-month PFS rate was 49.5% (95% confidence interval [CI], 39.2–60.8). Pyrotinib produced an ORR of 19.2% (95% CI, 11.2–30.0), with median PFS of 5.6 months (95% CI, 2.8–8.4), and median OS of 10.5 months (95% CI, 8.7–12.3). The median duration of response was 9.9 months (95% CI, 6.2–13.6). All treatment-related adverse events (TRAEs) were grade 1–3 (all, 91.0%; grade 3, 20.5%), and the most common TRAE was diarrhea (all, 85.9%; grade 3, 16.7%). Patients with exon 20 and non-exon 20 HER2 mutations had ORRs of 17.7% and 25.0%, respectively. Brain metastases at baseline and prior exposure to afatinib were not associated with ORR, PFS, or OS. Loss of HER2 mutations and appearance of amplification in HER2 and EGFR were detected upon disease progression. Conclusions Pyrotinib exhibited promising efficacy and acceptable safety in NSCLC patients carrying exon 20 and non-exon 20 HER2 mutations and is worth further investigation. Trial registration Chinese Clinical Trial Registry Identifier: ChiCTR1800020262 Supplementary Information The online version contains supplementary material available at 10.1186/s12916-022-02245-z.

Pyrotinib is an oral, irreversible pan-HER TKI, which has been adopted as the combination partner of capecitabine for treating advanced HER2-positive breast cancer in China [9]. In patient-derived lung cancer xenograft mouse models harboring HER2 exon 20 insertions, pyrotinib demonstrated stronger antitumor activities than T-DM1 or afatinib [10]. In a phase II study (N = 60) conducted by Zhou C et al., chemotherapy-treated NSCLC patients with HER2 mutations within exon 20 and 19 achieved an ORR of 30% upon pyrotinib, with mPFS of 6.9 months and median overall survival (mOS) of 14.4 months [11]. Evidence regarding efficacy and safety of pyrotinib remains to be confirmed in larger sample sizes, particularly in patients with HER2 mutations outside of exon 20. Moreover, the underlying mechanism of resistance to pyrotinib and its efficacy in patients who had brain metastases and prior exposure to anti-HER2 therapy has not been well elucidated.
The aim of this study was to evaluate the efficacy and safety of pyrotinib in advanced NSCLC patients harboring HER2 mutations. The impact of different HER2 mutation types on sensitivity to pyrotinib, the association between baseline characteristics and response, and the potential of utilizing mutational profile information derived from circulating tumor DNA (ctDNA) to predict disease progression were also explored.

Patients
Patients were recruited at 11 Chinese hospitals from December, 2018 until April, 2020. Patients were enrolled if they were 18 years or older and had histocytologically confirmed unresectable stage IIIB or IV NSCLC, HER2 mutations as determined using next-generation sequencing (NGS), an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0-2, and at least one radiographically measurable lesion per Response Evaluation Criteria in Solid Tumors (RICIST) version 1.1 [12]. Exclusion criteria included having had undergone surgery, chemotherapy, or radiotherapy for NSCLC within 4 week before the study treatment. Written informed consent was provided by each patient before the onset of any trial-related treatment. The study protocol was approved by each site's institutional review board in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.

Study design and treatment
This is a multi-center, single-arm, phase II trial (Clinical trial registration: ChiCTR1800020262). Pyrotinib was administrated orally at 400 mg/day within 0.5 h after breakfast until intolerable toxicity, disease progression, or discontinuation at the patient's request. In case of intolerable toxicity, the dose of pyrotinib was reduced to 320 mg daily. Depending on sample availability, biopsy tissue sample or blood sample was obtained from each patient at baseline, followed by NGS analysis. Under patients' consents, blood samples were also collected from some patients upon disease progression for NGS analysis.

Outcome assessment
The primary end point was 6-month PFS rate, which was defined as the proportion of PFS at 6 months after the first dose of pyrotinib. Secondary endpoints included safety, ORR (the frequency of patients who have had obtained partial response [PR] or complete response [CR] at two consecutive evaluations at least 4 weeks apart), PFS (the time between the first dose of pyrotinib and disease progression or death due to any reason), OS (the time between the first dose of pyrotinib and death due to any reason), and disease control rate (DCR, the frequency of patients who have had achieved a stable disease or PR or CR for ≥ 6 weeks before disease progression). Radiological assessment was conducted every six weeks in the first year, and every 9 weeks thereafter. Adverse events were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0. Upon disease progression, patients were followed up every 3 months until death. Exploratory endpoints included the association between different HER2 mutation types and ORR, PFS, OS, or DCR and the feasibility of using ctDNA to monitor disease progression.

Next-generation sequencing
Baseline tissue or blood samples were subjected to NGSbased molecular profiling to identify gene aberrations including alterations in the driver genes (EGFR, ALK, ROS1, MET, BRAF, RET, HER2, and KRAS) recommended by NCCN guidelines for NSCLC, while blood samples obtained from patients at disease progression were analyzed using a panel spanning 150 cancer-related genes at 3D Medicines, Inc., a clinical laboratory accredited by the College of American Pathologists (CAP) and certified by the Clinical Laboratory Improvement Amendments (CLIA) laboratory (Additional file 1: Supplementary Method for NG S [13,14], Additional file 2: Table S1).

Statistical analysis
According to previous study [15,16], the 6-month progression-free rate of chemotherapy is hypothesized to be 30%, then 67 patients would provide 80% power to detect a 6-month progression-free rate of 45% at 5% alpha level. A total of 75 patients would need to be enrolled with the consideration of a dropout rate of 10%.
All statistical analyses were performed using the SPSS statistical software (version 20.0) and GraphPad prism (version 7). PFS and OS were estimated using Kaplan-Meier curves, with P value determined by a log-rank test. The difference in ORR and DCR between different groups were analyzed using the Fisher's exact test. Cox regression was applied for calculating hazard ratio (HR) and 95% confidence intervals (CIs). A two-tailed P < 0.05 was defined as statistically significant.

Patients
Between December, 2018 and April, 2020, 80 patients with HER2 mutations were screened for eligibility. Two patients were excluded for withdrawing informed consents before study treatment; hence, a total of 78 patients were enrolled in this study and were included in the efficacy and safety analyses (Fig. 1). As data cut-off (December 30, 2020), the median duration of follow-up time was 10.5 months (range, 1.0-21.4 months). A total of 19 patients were still on treatment and 59 patients discontinued treatment, among which 50 for disease Fig. 1 Study flow. We screened 80 patients and two patients were excluded due to withdrew informed consent before treatment. Hence the intention-totreat population consisted of 78 patients, in which the efficacy and safety analyses were performed. As of December 2020, fifty-nine patients discontinued the study treatment for disease progression and intolerable adverse effect. The remaining 19 patients continued to receive pyrotinib progression, 4 for intolerable adverse events, and the rest for other reasons.
Baseline characteristics were summarized in Table 1. The median age of the 78 patients was 62 years (range, 31-85 years). All patients had stage IV adenocarcinoma and 20 (25.6%) had brain metastases. Seven patients (9.0%) had an ECOG PS of 2 and the rest were 0-1. Most patients were non-smokers (65.4%). Twenty-one patients had a prior exposure to afatinib (first-line, N = 3; second-line or higher, N = 18). The majority of the patients received pyrotinib in the second-line or higher (first-line, 29.5%; second-line or higher, 70.5%). Among the enrolled patients, 62 carried HER2 exon 20 mutations (79.5%) while the other 16 patients (20.5%) harbored mutations outside of exon 20. Of the 62 patients carrying exon 20 mutations, 42 and 11 patients had Y772_A775dup and G776delinsVC, respectively, and 9 carried other types of exon 20 mutations. Among the 78 patients, two patients harbored ≥ two HER2 mutations. A total of 81 HER2 mutations were detected at baseline, 73 fell in the kinase domain, three were in the transmembrane domain (TMD), three in extracellular domain, and the other two in other region of the coding region (Additional file 2: Fig. S1). HER2 mutation types identified at baseline were summarized in Additional file 2: Table S2.
Feasibility of using ctDNA to monitor disease progression upon pyrotinib treatment Of the 78 patients in the analysis cohort, twelve patients who acquired resistance to pyrotinib had blood samples available both at baseline and upon disease progression. These blood samples were subjected to NGS analysis to monitor disease progression. Concurrent HER2 CNA and EGFR CNA, which were not presented at baseline blood samples, were detected from two patients upon PD, suggesting that co-occurrence of HER2 CNA and EGFR CNA may have played a role in resistance to pyrotinib. One of these two patients' representative CT images captured at baseline, best response, and PD are shown in Additional file 2: Fig. S7. Another four patients had a loss of HER2 mutation upon PD, rending it rational to speculate that the loss of HER2 mutations may confer resistance to pyrotinib. In addition, appearance of EGFR (p.E330K), KRAS (p.G12D), MET CNA, and BRAF CNA were also detected in three patients at PD (Additional file 2: Table S4). Since KRAS and BRAF are both downstream of HER2 in the RAS/RAF signaling pathway, our results suggested that gene alterations in the RAS/RAF pathway may serve as a potential mechanism of resistance to pyrotinib.   ALT alanine aminotransferase, AST aspartate aminotransferase, WBC white blood cell *No grade 4 or higher adverse events occurred HER2 mutations in different exons, patients harboring non-exon 20 aberrations achieved comparable ORR than those with exon 20 mutations. Patients who had brain metastases and prior exposure to anti-HER therapy could benefit from pyrotinib. Moreover, loss of HER2 mutations, appearance of HER2 amplification, and aberrations in EGFR, MET, KRAS, and BRAF were detected upon disease progression, suggesting their potential roles in the resistance to pyrotinib. Chemotherapy, the current standard treatment for advanced NSCLC patients with HER2 mutations, typically elicits an ORR of 10% and an mPFS of 4.3 months in a second-line setting (6). TKIs targeting HER2 or pan-HER have been investigated for treating HER2-mutated lung cancer patients. However, afatinib, neratinib, and dacomitinib only elicited ORR of 7.7%, 3.8%, and 12% [1][2][3]. The ORRs upon T-DM1 and T-DXd treatment could reach up to 44% (8/18) and 72.7% (8/11), respectively [4,5]. The mPFS of T-DM1-treated NSCLC patients as previously reported was 5.0 months, which was similar to that observed in the present study (5.0 vs. 5.6 months). Most recently, the results of the phase II study DESTINY-Lung trial were released in which T-DXd showed an ORR of 55% (50/91) and mPFS of 8.2 months in patients with previously treated NSCLC with HER2 mutation [17]. Albeit encouraging anti-tumor activity, grade 4 and 5 TRAEs occurred upon T-DXd, whereas in our study, no grade 4 or 5 TRAEs were observed, suggesting that pyrotinib is safer than T-DXd [5,17]. Poziotinib, another promising anti-HER2 TKI, has exhibited an ORR of 42% in HER2-mutated NSCLC patients (N = 12), causing grade 3 or 4 AEs in 66.7% of the patients [18].

Discussion
Treatment of HER2-mutated NSCLC with pyrotinib has been previously reported. In phase II trials conducted by Wang Y et al. and Zhou C et al., treatment with pyrotinib was associated with ORRs of 53.3% and 30%, and mPFSs of 6.4 months and 6.9 months in cohorts of 15 and 60 HER-mutated advanced NSCLC patients [10,11]. Both studies reported better efficacy than our observations (ORR, 19.2%; PFS, 5.6 months). This could have been explained by the fact that our study enrolled patients with a PS score of 2 (7/78, 9%) whereas Zhou C's study only included patients with a PS score of 0-1. A higher percentage of patients in our cohort had brain metastases at baseline (25.6% vs. 20%) and more patients received pyrotinib in the third line or higher (51.3% vs. 41.6%) than in their study. In addition, patients who had prior exposure to HER2-targeted drugs were also included in our study. Of note, the duration of response in the present study was 9.9 months, which was longer than 6.9 months documented in Zhou C's study.
The sensitivities to anti-HER2 TKIs in patients bearing different HER2 mutations were also distinct. In patients with HER2-mutated NSCLC, the major HER2 mutation type was exon 20 insertions, occurring in 1.5% of NSCL C and accounting for 90% of all NSCLC with HER2 mutations [19][20][21][22]. Previous studies have been mainly focusing on these insertions. Two prospective studies investigating pyrotinib employed the ADx HER2 Mutation Detection Kit for HER2 genotyping, which only allows for detection of exon 20 and 19 mutations [10,11]. In our study, we utilized NGS to detect HER2 mutations, which was capable of identifying mutations outside of exons 20 and 19. Indeed, patients carrying mutations outside of exon 20 were also able to benefit from pyrotinib. A numerically higher ORR was observed among patients carrying non-exon 20 mutations, especially those carrying exon 19 mutations. These observations were consistent with previous findings that HER2 exon 20 insertions are less sensitive to currently available TKIs than mutations in other exons, potentially due to the structural difference of mutant in this exon from in others [19]. HER2 exon 20 insertions primarily affected two structural regions: the αChelix, comprising residues 770-774, and the loop region at residues 775-783 [20,21,23]. Structure-based comparison of behaviors between these variant types needs to be further studied.
Patients with HER2 exon 20 mutation Y772_A775dup, the most common HER2 mutation in NSCLC, failed to respond to afatinib and dacomitinib as reported [1,24,25]. Surprisingly, pyrotinib produced an ORR and a DCR of 23.8% and 78.6%, respectively, in 42 patients harboring Y772_A775dup in our study [24,25]. Consistent with the results of Zhou C's study, although none of the 11 patients carrying G776delinsVC achieved PR in our study, the DCR of this subset reached 63.6%, which was similar to that of the other mutation types [11]. Clinical efficacy regarding anti-HER2 TKIs has been poorly investigated in patients with HER2 TMD mutations [26,27]. In our study, three patients harbored HER2 TMD, including two with V659E and one with I655V. The PFS and OS of the patients with V659E was 2.9-5.6 months and 5.3-5.6 months, respectively. The other patient bearing I655V, however, experienced PD three weeks after initiation of pyrotinib. Collectively, our results revealed variable efficacy of pyrotinib in NSCLC patients with different HER2 mutations and warrant further validation in larger randomized clinical trials.
Another point to be noted was the monitoring of acquired resistance to pyrotinib by using blood sample profiling, highlighting the importance of liquid biopsy in this setting. In this study, we also explored potential resistance mechanisms underlying disease progression upon pyrotinib. HER2 CNA was identified from two patients upon PD, consistent with a previous report that HER2 CNA conferred resistance to anti-HER2 TKIs in HER2-mutated NSCLC patients [28]. Of note, EGFR CNA was also detected from these two patients upon PD, indicating the concurrent HER2 CNA and EGFR CNA may engender resistance to pyrotinib. In another four PD patients, HER2 mutation, which existed at baseline, was not detected from the blood sample at PD, rending it rational to speculate that the loss of HER2 mutations may engender resistance to pyrotinib as well. In addition, MET CNA, KRAS (p.G12D), BRAF CNA, and EGFR (p.E330K) were also detected from patients at PD. MET CNA has been reported to be associated with resistance to anti-HER2 TKIs in EGFR-mutant NSCL C, HER2-amplified breast cancer, and HER2-mutated NSCLC [28][29][30]. Based on these results, we propose that strategies combining pyrotinib and EGFR TKI or other TKIs targeting the above alternations might be a potential treatment option to vanquish resistance or potentiate the antitumor activities in treating this subset of patients.
Indeed, Rolfo C et al. summarized a series of novel agents that has potential against HER2-mutated NSCL C [8]. Interestingly, the combinational treatment of a pan-HER inhibitor (neratinib) and T-DM1 or T-DXd induced a superior activity compared with T-DM1 alone [31]. Similarly, preclinical studies revealed that the novel pan-HER TKI poziotinib could up-regulate HER2 cell-surface expression and increase the activity of T-DM1 in tumors with HER2-mutation [32]. In addition, Bob T. Li et al. reported that the combination of T-DM1 and irreversible pan-HER inhibitors (neratinib or afatinib) could enhance the duration of the responses in HER2-altered lung cancers [31]. Pyrotinib is an irreversible pan-HER inhibitor, also presenting promising activity in HER2-mutated NSCLC as observed in our study. Part of data of this trial (ChiCTR1800020262) was published recently which has shown the efficacy of pyrotinib in NSCLC patients with HER2 amplification (6-month PFS rate: 51.9%, ORR: 22.2%, mPFS: 6.3 months, mOS: 12.5 months) [33]. Therefore, a combination of T-DM1/T-DXd and pyrotinib may become a potentially effective therapy for these HER2-altered patients. These results indicate that combining T-DM1/T-DXd and anti-HER2 TKI might be a potential treatment option to increase antitumor activity or conquer resistance to targeted therapies. The above proposals are a ray of hope shining the future of patients with HER2 alternations.
Despite being the largest prospective study investigating pyrotinib effects in NSCLC, our study is still limited by the small sample size due to the low prevalence of HER2 mutations in NSCLC. Second, comparison with chemotherapy or other targeted therapies was not feasible due to a lack of control arm. The findings of the current study should be examined in larger randomized clinical trials.

Conclusions
Pyrotinib exhibited promising efficacy and acceptable safety in treating NSCLC patients with both exon 20 and non-exon 20 HER2 mutations.  Table S1-S4., Figure S1-S7. Table S1. List of genes in the 3DMed 150-gene panel. Table S2. HER2 mutations identified at baseline. Table S3. Clinical response to pyrotinib according to different HER2 mutation types. Table S4. The detected molecular alterations at baseline and progression. Fig. S1. HER2 mutational map at baseline. Green: receptor L domain; red: furin-like cysteine rich region; blue: growth factor receptor domain IV; yellow: protein tyrosine kinase. Fig. S2. Survival curves of pyrotinib treated HER2-mutated NSCLC patients according to baseline characteristics. (A, B) progression-free survival (PFS) and overall survival (OS) according to the ECOG performance status. PS represents ECOG performance score. (C, D) PFS and OS of pyrotinib treated patients with or without brain metastasis. mPFS, median progression-free survival; mOS, median overall survival; HR, hazard ratio; 95%CI, 95% confidence interval. Fig. S3 Objective response rate in pre-specific subgroups. Fig.  S4 Survival curves of NSCLC patients treated with pyrotinib according to previous treatment. (A, B) progression-free survival (PFS) and overall survival (OS) according to the treatment lines of pyrotinib. (C, D) PFS and OS of patients according to the prior exposures to afatinib. mPFS, median progression-free survival; mOS, median overall survival; HR, hazard ratio; 95%CI, 95% confidence interval. Fig. S5 Survival curves of pyrotinib treated NSCLC patients with different HER2 mutation. mPFS, median progression-free survival; mOS, median overall survival; HR, hazard ratio; 95%CI, 95% confidence interval. Fig. S6 Survival curves of pyrotinib treated HER2-mutated NSCLC patients according to molecular characteristics. (A, B) progression-free survival and overall survival according to HER2 amplification. (C, D) PFS and OS of pyrotinib treated patients according to the occurrence of co-mutations in other driver genes. mPFS, median progression-free survival; mOS, median overall survival; HR, hazard ratio; 95%CI, 95% confidence interval. Fig. S7 Pyrotinib resistance in a patient with HER2 and EGFR amplification. CT scans were performed at baseline (2 weeks before starting pyrotinib), best response (2 months after starting pyrotinib) and disease progression (7 months after starting pyrotinib), respectively. Mm, millimeter. SX, JH, DW, DL, SL, XH, CX, XW, JF, FH, WW, CX, SC2, and SL. Data interpretation: ZS, YL, SC1, SY, and TB. Manuscript drafting: ZS, SC1, TB, and CG. Critical revision of the manuscript: YZ, ZS, YL, SY, SC1, TB, and SC2. Statistical analysis: ZS, and SC1. All authors read and approved the final manuscript.

Funding
This study was granted by Foundation of CSCO-Hengrui (Z. Song, Y-HR2019-0173). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations
Ethics approval and consent to participate The study protocol was approved by each site's institutional review board in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Written informed consent was provided by each patient before the onset of any trial-related treatment.

Consent for publication
Not applicable.