Blog

The combination of lenvatinib (Lenvima) and pembrolizumab (Keytruda) plus chemotherapy failed to improve overall survival (OS) vs pembrolizumab plus chemotherapy alone as frontline treatment in patients with metastatic esophageal squamous cell carcinoma (ESCC), according to data from the phase 3 LEAP-014 trial (NCT04949256) presented at the 2025 ESMO Congress.¹

The median OS in the lenvatinib arm was 17.6 months (95% CI, 15.5-19.1) vs 15.5 months (95% CI, 13.5-17.2) in the chemoimmunotherapy arm (HR, 0.92; 95% CI, 0.77-1.10; P = .1852). The 12- and 24-month OS rates were 65% and 34% in the investigational arm vs 61% and 32% in the control arm, respectively.

In patients with a PD-L1 combined positive score (CPS) of at least 10, the median OS was 18.0 months (95% CI, 16.4-20.4) in the lenvatinib arm vs 15.8 months (95% CI, 13.1-17.4) in the control arm (HR, 0.89; 95% CI, 0.71-1.11). The 12- and 24-month OS rates were 67% and 36% in the investigational arm vs 63% and 31% in the control arm, respectively.

“Lenvatinib plus pembrolizumab and chemotherapy did not significantly improve OS as first-line treatment for [patients with] metastatic ESCC vs pembrolizumab plus chemotherapy,” lead study author Jong-Mu Sun, MD, PhD, of the Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine in Seoul, Republic of Korea, stated during the presentation.

What Are the Key Components of the LEAP-014 Trial?

The LEAP-014 trial enrolled patients who were at least 18 years of age with histologically and/or cytologically confirmed locally advanced, unresectable or metastatic ESCC.² Prior treatment, an ECOG performance status of 2 or higher, and lack of measurable disease per RECIST 1.1 criteria excluded patients from enrollment.

A total of approximately 850 patients were randomly assigned 1:1 to treatment.¹ In the investigational arm, patients received 8 mg of oral lenvatinib once daily, plus 400 mg of intravenous (IV) pembrolizumab every 6 weeks, and either cisplatin plus 5-flourouracil (FP) or paclitaxel plus cisplatin (TP) every 3 weeks, or modified FOLFOX6 (mFOLFOX6) every 2 weeks. This was followed by 20 mg of lenvatinib plus 400 mg of pembrolizumab every 6 weeks until progression, unacceptable toxicity, or withdrawal. In the control arm, patients received 400 mg of IV pembrolizumab every 6 weeks plus FP or TP every 3 weeks or mFOLFOX6 every 2 weeks in accordance with local standards.

OS served as the primary end point. Secondary end points included progression-free survival (PFS), objective response rate (ORR), duration of response (DOR) by blinded independent central review, and safety.

A total of 850 patients were randomly assigned to the investigational (n = 423) and control (n = 427) arms: 421 and 426 of whom were treated in the respective arms. In the investigational arm, 2 patients completed treatment and 93 patients remained on therapy. Reasons for discontinuation included progressive disease (n = 206), adverse effects (AEs; n = 72), clinical progression (n = 25), patient decision (n = 18), physician decision (n = 3), and protocol violation (n = 2). In the control arm, 23 patients completed treatment and 70 patients remained on therapy. Reasons for discontinuation included progressive disease (n = 202), AEs (n = 73), clinical progression (n = 22), patient decision (n = 25), physician decision (n = 10), and follow-up loss (n = 1).

Baseline characteristics were well balanced between the treatment groups, Sun said. Within the investigational arm, the median age was 64 (range, 27-86); 49% were above the age of 65. Most patients were male (78%), Asian (66%), and from East Asia (66%). ECOG performance status was 1 in 64% of cases, and most patients had a PD-L1 CPS of 10 or greater (66%). Ninety-nine percent of patients did not have brain metastases and had stage IVB disease. The preferred choice of chemotherapy was mFOLFOX6 (65%), followed by FP (24%) and TP (10%).

Per the study design, efficacy was evaluated in all randomized patients, whereas safety was limited to all randomized patients who received at least 1 dose of study therapy. Of note, PFS and ORR were not tested for statistically significance since OS was not positive. The data cutoff for the present analysis was May 8, 2025, at which point approximately 482 OS events had occurred, or 42 months had passed since the first participant had been randomized.

What Additional Efficacy Data Were Presented?

Sun stated that the OS results were consistent across the prespecified subgroups, including PD-L1 status (CPS ≥10 vs CPS <10), Region (East Asia vs North America and Western Europe vs rest of the world), and chemotherapy (FP vs TP vs mFOLFOX).

Additional results indicated that the median PFS in all randomized patients was 7.2 months (95% CI, 6.8-8.4) in the lenvatinib arm vs 6.9 months (95% CI, 5.8-7.0) in the control arm (HR, 0.89; 95% CI, 0.75-1.04). The 12- and 24-month PFS rates were 31% and 12% in the investigational arm vs 23% and 15% in the control arm, respectively.

The median PFS in patients with a PD-L1 CPS of at least 10 was 8.1 months (95% CI, 6.9-9.5) in the lenvatinib arm vs 6.9 months (95% CI, 6.3-7.2) in the control arm (HR, 0.85; 95% CI, 0.69-1.04). The 12- and 24-month PFS rates were 36% and 16% in the investigational arm vs 25% and 14% in the control arm, respectively.

With respect to response, the ORR was 62.2% (95% CI, 57.4%-66.8%) in the lenvatinib arm vs 54.8% (95% CI, 49.9%-59.6%) in the chemoimmunotherapy arm (delta, 7.3%; 95% CI, 0.7%-13.8%). In the lenvatinib arm, best responses included complete response (CR; 16.3%), stable disease (SD; 22.0%), and progressive disease (PD; 8.0%); 7.3% of patients were not evaluable (NE). In the control arm, best responses included CR (19.2%), PR (35.6%), SD (29.7%), and PD (8.9%); 5.4% of patients were NE.

The median DOR was 8.1 months (range, 1.2+ to 38.7+) in the lenvatinib arm and 17% of patients experienced a response lasting 24 months or greater. In the control arm, the median DOR was 6.8 months (range, 1.2+ to 36.4+) and 21% of patients experienced a response lasting at least 24 months.

What Were the Key Safety Data That Were Presented?

The median duration of treatment exposure was 7.1 months (range, 0.03-41.9) in the lenvatinib arm vs 5.6 months (range, 0.03-28.8) in the control arm. In the investigational arm, any-grade AEs occurred in 99.5% of patients (grade ≥3, 81.2%; grade 5, 9.7%). AEs leading to drug discontinuation occurred in 33.3% of patients. Treatment-related AEs (TRAEs) occurred in 97.1% of patients.

In the control arm, any-grade AEs occurred in 99.3% of patients (grade ≥3, 79.1%; grade 5, 11.5%). AEs leading to drug discontinuation occurred in 39.0% of patients. TRAEs occurred in 96.5% of patients.

AEs that occurred in at least 15% of patients in either arm in order of frequency were decreased neutrophil count, nausea, hypertension, diarrhea, decreased platelet count, anemia, hypothyroidism, stomatitis, decreased white blood cell count, fatigue, decreased appetite, palmar plantar erythrodysesthesia syndrome, proteinuria, aspartate aminotransferase increase, vomiting, rash, weight decrease, constipation, alanine aminotransferase increase, pyrexia, increased lipase, increased amylase, peripheral neuropathy, hypoalbuminemia, and pneumonia.

Immune-mediated AEs and infusion reactions in the experimental arm included adrenal insufficiency (any grade, 4.8%; grade ≥3, 1.4%), colitis (any grade, 1.7%; grade ≥3, 0.9%), encephalitis (any grade, 0.2%; grade ≥3, 0.2%), gastritis (any grade, 2.9%; grade ≥3, 0.2%), hepatitis (any grade, 1.2%; grade ≥3, 0.7%), hyperthyroidism (any grade, 8.1%; grade ≥3, 0%), hypophysitis (any grade, 2.1%; grade ≥3, 0.7%), hypothyroidism (any grade, 33.5%; grade ≥3, 0%), infusion reactions (any grade, 2.1%; grade ≥3, 0.5%), myocarditis (any grade, 0%; grade ≥3, 0%), myositis (any grade, 0.2%; grade ≥3, 0%), nephritis (any grade, 0%; grade ≥3, 0%), pancreatitis (any grade, 0.5%; grade ≥3, 0.5%), pneumonitis (any grade, 7.6%; grade ≥3, 2.6%), severe skin reactions (any grade, 1.9%; grade ≥3, 1.4%), thyroiditis (any grade, 0.7%; grade ≥3, 0%), type 1 diabetes mellitus (any grade, 0.5%; grade ≥3, 0.2%), and vasculitis (any grade, 0.7%; grade ≥3, 0.2%).

“Safety profiles were generally consistent with the known safety profiles of lenvatinib in combination with pembrolizumab and chemotherapy or the pembrolizumab plus chemotherapy regimen,” Sun concluded.

Disclosures: Sun reported financial interests, institutional, and research funding from MSD, Yuhan, AstraZeneca, Ono, and Pfizer.

References

1. Sun JM, Kim SB, Ogata T, et al. Lenvatinib plus pembrolizumab and chemotherapy vs pembrolizumab and chemotherapy in untreated metastatic esophageal squamous cell carcinoma: the randomized phase III LEAP-014 study. Presented at: 2025 ESMO Congress; October 17-21, 2025; Berlin, Germany. Abstract LBA79.
2. Efficacy and safety of pembrolizumab (MK-3475) plus lenvatinib (E7080/​MK-7902) plus chemotherapy in participants with metastatic esophageal carcinoma (MK-7902-014/​E7080-G000-320/​LEAP-014). Clinicaltrials.gov. Updated August 22, 2025. Accessed October 17, 2025. https://clinicaltrials.gov/study/NCT04949256

Sevabertinib monotherapy elicited robust and durable responses as therapy for patients with HER2-mutant advanced non–small cell lung cancer (NSCLC) in both front-line and pretreated settings, according to results from the phase 1/2 SOHO-01 trial (NCT05099172) presented at the European Society for Medical Oncology (ESMO) Congress 2025.¹ Results were simultaneously published in the New England Journal of Medicine.²

In cohort D (n = 81), which included patients who were previously treated, the objective response rate (ORR) by blinded independent central review (BICR) was 64% (95% CI, 53%-75%), with 2% achieving complete responses (CRs), 62% achieving partial responses (PRs), 25% having stable disease, 7% having progressive disease, and 4% not being evaluable. The disease control rate (DCR) was 81% (95% CI, 71%-89%).

The median duration of response (DOR) was 9.2 months (95% CI, 6.3-13.5), with a 12-month DOR rate of 42%; the median progression-free survival (PFS) was 8.3 months (95% CI, 6.9-12.3), with a 12-month PFS rate of 44%.

With a median follow-up of 13.8 months (range, 1-32), the median duration of treatment (DOT) was 9.9 months (range, 0-32). A total of 30% of patients had ongoing treatment, 20% had a treatment duration of more than 15 months, and responses were maintained after dose reductions or interruptions.

Subgroup analyses revealed that ORR was consistent across all subgroups, such as age, sex, race, number of previous systemic treatments, presence of baseline brain metastases, and receipt of prior platinum chemotherapies.

In patients with non-squamous NSCLC and HER2 tyrosine kinase domain (TKD) mutations (n = 70), the ORR by BICR was 71%, and the median PFS was 9.6 months (95% CI, 6.9-14.7). In patients with HER2 Y772_A775dupYVMA alterations (n = 49), the ORR by BICR was 78%; in those with YVMA, the median PFS was 12.2 months (95% CI, 6.9-16.4), and in those with other TKD mutations, the median PFS was 7.0 months (95% CI, 4.0-not estimable [NE]).

Additionally, in cohort E, which included patients who received prior treatment with HER2 antibody drug conjugates (n = 55), the ORR by BICR was 38% (95% CI, 25%-52%), with 5% achieving CRs, 33% achieving PRs, 42% having stable disease, 13% having progressive disease, and 7% not being evaluable; the DCR was 71% (95% CI, 57%-82%).

The median DOR was 8.5 months (95% CI, 5.6-16.4), with a 12-month DOR rate of 29%, and the median PFS was 5.5 months (95% CI, 4.3-8.3), with a 12-month PFS rate of 28%. The median follow-up was 11.6 months (range, 2-22).

In cohort F, which enrolled patients who were treatment-naïve (n = 73), the ORR by BICR was 71% (95% CI, 59%-81%), with 4% achieving CRs, 67% achieving PRs, 22% having stable disease, 3% having progressive disease, and 1% not being evaluable. The DCR was 89% (95% CI, 80%-95%).

The median DOR was 11.0 months (95% CI, 8.1-NE), and the median PFS was NE (95% CI, 9.6-NE). The median follow-up was 9.9 months (range, less than 1-15).

Further, brain metastases were present in 22% of patients in cohort D, 27% of cohort E, and 12% of cohort F. An analysis demonstrated systemic responses were similar in patients with and without brain metastases. In cohort D, patients with and without brain metastases had ORRs of 61% and 65%, respectively; in cohort E, the ORRs were 27% and 43%; and in cohort F, the ORRs were 78% and 70%.

In patients who had brain metastases at baseline in cohorts D, E, and F, post-baseline progression in the brain was observed in 22%, 20%, and 0%, respectively. In those with no brain metastases at baseline in cohorts D, E, and F, post-baseline progression in the brain was observed in 6%, 5%, and 3%, respectively.

“Sevabertinib demonstrated robust and durable responses in patients with HER2-mutant NSCLC in both pretreated and first-line settings. The most common [adverse] effect was diarrhea, which was manageable; there were no reported cases of interstitial lung disease (ILD) or pneumonitis,” said presenting study author Xiuning Le, MD, PHD, associate professor in the Department of Thoracic/Head and Neck Medical Oncology of the Division of Internal Medicine at the University of Texas MD Anderson Cancer Center, and coauthors, in the presentation.¹ “These data support sevabertinib as a potential new targeted therapy for patients with HER2-mutant NSCLC.”

SOHO-01 enrolled patients with advanced NSCLC with HER2 or EGFR mutations; this analysis included data from the cohorts of patients with HER2 mutations. Cohort D enrolled patients who were previously treated, but naïve to HER2-targeted therapies; cohort E enrolled patients who were previously treated with HER2-targeted ADCs; and cohort F enrolled patients who were naïve to systemic therapy for advanced disease.

Across cohorts D, E, and F, the median age of patients was 60 years, 65 years, and 65 years, respectively, and 62%, 64%, and 78% had never smoked. Activating Y772_A775dupYVMA HER2 mutations were observed in 60%, 73%, and 79%, respectively, and 90%, 95%, and 97% had HER2 TKD mutations. Fam-trastuzumab deruxtecan-nxki (Enhertu) was received as previous anti-cancer therapy by 2%, 75%, and 0%, and chemotherapy was received by 96%, 80%, and 8%.

Treatment consisted of increasing oral doses of sevabertinib from 10 mg once daily to 40 mg twice daily, to identify the recommended dose for expansion.

The primary end point of the extension phase was ORR per RECIST v1.1 by BICR. Secondary end points included DOR, DCR, PFS, and safety and tolerability.

Regarding safety, drug-related adverse events (AEs) occurred in 69% of cohort D, 100% of cohort E, and 97% of cohort F; grade 3 AEs were reported in 36%, 31%, and 21%. Notably, grade 3 diarrhea events occurred in 23% of cohort D, 11% of cohort E, and 5% of cohort F. ILD or pneumonitis was not observed in any patients.

Biomarker data from SOHO-01 are being shared in a poster at ESMO on October 18, 2025.³

Sevabertinib is also being evaluated as a first-line therapy for patients with locally advanced or metastatic NSCLC with HER2 mutations in the ongoing phase 3 SOHO-02 trial (NCT06452277).

Disclosures: Dr. Le cited consultant roles for AbbVie, AstraZeneca, BlossomHill, Blueprint Medicines Corporation, Boehringer Ingelheim, Daiichi Sankyo Company, Eli Lilly and Company, Janssen Biotech, Merck KGaA, Novartis, Regeneron Pharmaceuticals, and Spectrum Pharmaceuticals.

References

1. Le X, Kim TM, Dong X, et al. Sevabertinib (BAY 2927088) in advanced HER2-mutant non-small cell lung cancer (NSCLC): results from the SOHO-01 study. Presented at the 2025 European Society for Medical Oncology Congress. Presented at the 2025 ESMO Congress; October 17–20, 2025; Berlin, Germany. Abstract LBA75.
2. Le X, Kim TM, Loong HH, et al. Sevabertinib in advanced HER2-mutant non–small-cell lung cancer. N Engl J Med. Published online October 17, 2025. doi:10.1056/NEJMoa2511065
3. Girard N, Li L, Kim KM, et al. Molecular factors and ctDNA dynamics associated with clinical outcomes in patients with HER2-mutant NSCLC treated with sevabertinib (BAY 2927088): Exploratory analysis of the SOHO-01 study. Will be presented at the 2025 ESMO Congress; October 17–20, 2025; Berlin, Germany. Abstract 2001P.