First-in-human trial of multikinase VEGF inhibitor regorafenib and anti-EGFR antibody cetuximab in advanced cancer patients

Angiogenesis and EGFR signaling have now well-established roles in cancer biology. VEGF plays a pivotal role in tumor angiogenesis, while activation of the EGFR has been linked to many processes crucial to tumor progression (1, 2). Close relationships exist between these 2 pathways. Preclinical studies suggest that the EGFR may have a role in angiogenesis, and also that inhibition of the EGFR downregulates VEGF (3-6). Conversely, VEGF upregulation independent of EGFR signaling seems to contribute to resistance to EGFR inhibition (7). Moreover, VEGF inhibition may also block EGFR autocrine signaling and thereby inhibit cancer cell growth (8).

Given that the EGFR and VEGF share common downstream signaling pathways, combined inhibition of these 2 targets may enhance efficacy. In vivo preclinical data have demonstrated decreased angiogenesis as well as increased tumor and endothelial cell apoptosis with combined inhibition of VEGF and EGFR (9).

Inhibitors of VEGF and EGFR have become key therapies in several tumor types. Regorafenib is a multikinase inhibitor, with targets including VEGF receptors 1–3, KIT, and PDGFR-α and -β. It is approved for use in patients with refractory colon cancer as well was gastrointestinal stromal tumors. Cetuximab is one of the earliest employed monoclonal antibodies targeting the EGFR, and is approved for use in metastatic KRAS wild-type colorectal cancer (CRC) and surgically unresectable squamous cell carcinoma of head and neck. All patients receiving regorafenib or cetuximab eventually progress, and a search for more effective treatments continues.

Recently, Napolitano et al. studied the in vitro effect of the combination of regorafenib plus cetuximab in KRAS- and BRAF-mutated human colorectal cell lines as a model of primary resistance to cetuximab as well as colorectal cell line models of acquired resistance to cetuximab (10). The study demonstrated a synergistic antiproliferative and apoptotic effect of this combination by blocking MAPK and AKT pathways, compared with either drug alone or control. The authors then injected nude mice with colorectal cancer cell lines with primary and secondary resistance to cetuximab, and demonstrated significant tumor inhibition with the combination of regorafenib and cetuximab compared with either drug alone or control. Similar results were also observed in an orthotopic KRAS-mutated human CRC xenograft.

We hypothesized that the combination of regorafenib and cetuximab may enhance antitumor efficacy. Therefore, this phase I trial was conducted to determine maximum tolerated dose (MTD) and dose-limiting toxicities (DLTs) with this regimen, and to assess the antitumor efficacy of this combination. We also report on comprehensive genomic profiling (CGP) in a patient with CRC and mismatch repair (MMR) deficiency who had an exceptional response of stable disease beyond 20 cycles of therapy. Recently, MMR status was shown to predict response to checkpoint immunotherapy in CRC and non-CRC patients (11). More patients with CRC are being tested for microsatellite instability (MSI) status for checkpoint therapy and treatment options are needed for checkpoint progressors.

All 27 patients enrolled in the trial were evaluable for toxicity and 24 (89%) (13 male [54%]) were evaluable for response, as they completed a full cycle with restaging and were included in the analysis. The median age of patients was 54 (range 28–68) years. Sixty-three percent of the patients had CRC (17 patients). Patients had received a median of 4 (range 0–7) lines of systemic therapies. Twenty patients had received local treatments including surgery, radiation, chemo-embolization and/or radiofrequency ablation. Twenty-two patients had an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1. Patient characteristics are summarized in Table 1.

Table 1

Patient Characteristics on the phase 1 trial with cetuximab and regorafenib

Toxicities. None of the 19 patients treated at dose level 1 experienced any DLTs. Two of 5 patients treated at dose level 2 experienced DLT. Therefore, MTD was considered to have been exceeded at dose level 2, and dose level 1 was determined to be the recommended phase II dose. DLTs included grade 3 thrombocytopenia in 1 patient with a diagnosis of glioblastoma, and grade 3 intra-abdominal bleed in 1 patient with a diagnosis of CRC. Other treatment-related adverse events are summarized in Table 2. The most common toxicities included rash (20 patients), fatigue (7 patients), hand-foot syndrome (4 patients), myalgia (4 patients), and nausea (4 patients); all these common toxicities were grade 1 or 2.

Table 2

Treatment-related adverse events in the phase 1 trial with cetuximab and regorafenib.

Patients were able to continue treatment for a median of 5 (range 1–22) cycles; 1 patient is continuing treatment after 27 cycles without any evidence of progression. Thirteen patients received treatment for fewer than 6 cycles, 9 received treatment for 6 to 10 cycles, and 2 for 20 or more cycles.

Responses. Twenty-four patients were evaluable for response. One patient achieved partial response, and 16 had stable disease as best response. Eleven (46%) of patients had clinical benefit from treatment and included 8 patients with CRC and 1 patient each with squamous cell cancer of head and neck, cancer of unknown primary, and glioblastoma (with EGFR G598V mutation).

Among 17 patients with CRC, all had previously received anti-VEGF as well as anti-EGFR therapy, and 5 had received regorafenib (Table 3). Among these, 7 patients received treatment for fewer than 6 cycles, 6 received treatment for 6 to 10 cycles, and 2 patients for 20 or more cycles. Best response among 14 response-evaluable patients included partial response in 1 patient and stable disease in 10 patients. Eight (53%) patients with CRC had clinical benefit from treatment. Six had up to a 50% decrease and 5 had more than a 50% decrease in serum carcinoembryonic antigen (CEA). Three patients had an increase in CEA. Although we selected for KRAS wild-type CRC, there were no NRAS or atypical KRAS mutations included.

Table 3

Mutational analysis of patients enrolled in the trial

We identified 2 patients who had an exceptional benefit from this treatment for further study. The first patient had CRC and had previously progressed on cetuximab, bevacizumab, and regorafenib as individual monotherapies. The patient achieved a partial radiographic response with a 46% decrease in tumor measurements per RECIST v1.1 that lasted for approximately 15 months. A 50-gene next-generation sequencing analysis of the tumor specimen revealed a mutation in TP53.

The second patient had CRC with Lynch syndrome (MSI high) and had also progressed on cetuximab, bevacizumab, and regorafenib. The 50-gene analysis of the tumor specimen revealed mutations in the CTNNB1 and FGFR1 genes. Subsequent comprehensive genomic profiling identified alterations in the AKT2, BRCA1, BRCA2, CTNNB1, DNMT3A, FGFR3, MSH2, MSH6, PIK3CA, CDKN1B, BCORL, CIC, CREBBP, SETD2, SMARCA4, SOX9, and STAT4 genes. These alterations were frameshift mutations and are summarized in Tables 4–6. The mutational burden in this tumor was 99 mutations/megabase, which exceeds 99.3% of other tumors (Frampton et al., manuscript in preparation, personal communication). This patient harbors a MSH2 R389* nonsense mutation, and there is an additional MSH2 splice site mutation and an MSH6 frameshift mutation (Tables 4–6). This patient has ongoing stable disease after 20 cycles of treatment (Figure 1).

Figure 1

Dose escalation schema showing the number of dose levels, doses, number of patients enrolled, and dose-limiting toxicities (DLTs).

Table 4

Comprehensive genomic profile of patient that had prolonged stable disease and clinical benefit on protocol

Table 6

Comprehensive genomic profile of patient that had prolonged stable disease and clinical benefit on protocol

This open-label phase I trial studied the safety and tolerability of the regorafenib plus cetuximab combination among patients with advanced cancer refractory to several lines of therapy. Dose level 1 was determined to be the MTD, and no patients experienced any DLT at this dose level. Toxicities observed were consistent with previously published literature about these drugs (12–14).

To our knowledge, this is the first trial to study the combination of regorafenib and cetuximab. Previous trials have studied combinations of other anti-VEGF and anti-EGFR agents. Combination of bevacizumab, an anti-VEGF monoclonal antibody, in combination with erlotinib, a small-molecule EGFR inhibitor, in patients with non–small cell lung cancer demonstrated encouraging efficacy and safety profile (15, 16).

Although based on a phase II BOND-2 trial, there was initial enthusiasm for combining anti-VEGF plus anti-EGFR therapy with chemotherapy among patients with CRC, which showed encouraging efficacy of a combination of cetuximab, bevacizumab, and irinotecan among bevacizumab-naive irinotecan-refractory patients (17). However, subsequent trials demonstrated less favorable results. The phase II BOND-2.5 trial studied the same combination among bevacizumab-refractory patients and reported modest efficacy (18). Similarly, the CAIRO-2 trial evaluated the addition of cetuximab to chemotherapy plus bevacizumab and reported worse outcomes with dual antibody therapy; no benefit of dual antibody therapy was observed even among patients with wild-type KRAS colorectal cancer (19). The PACCE trial evaluated the addition of panitumumab, an anti-EGFR antibody, to chemotherapy and bevacizumab, and reported increased toxicity and progression, irrespective of KRAS status (20). Another trial of a phase III randomized, placebo-controlled study of cetuximab plus brivanib alaninate versus cetuximab plus placebo in patients with metastatic, chemotherapy-refractory, wild-type K-RAS CRC was also negative despite positive effects on progression-free survival and objective response (21).

More recently, results were reported for a phase I study of cetuximab with erlotinib and bevacizumab in heavily pretreated cancer patients, most of whom had previously received prior cetuximab as well as bevacizumab. The combination was well tolerated and demonstrated clinical benefit in 34% patients with CRC (22). A retrospective review of patients treated with EGFR therapy who progressed and rechallenged supported the notion of anti-EGFR retreatment in metastatic CRC (23).

The current study demonstrates that the combination of regorafenib (80 mg daily for 5 days every week for 3 weeks, followed by a week off) and cetuximab (200 mg/m² loading dose, followed by 150 mg/m² every week) is well tolerated. The combination, although at a lower dose level of the drug, showed a clinical benefit of 46% and 53% among all patients and CRC patients, respectively. Among patients with CRC, 79% had a decrease in CEA and 36% had a more than 50% reduction in CEA. These are noteworthy results for patients with advanced CRC, given that all the patients had previously received anti-EGFR as well as anti-VEGF therapy, and several patients had previously received regorafenib. While this trial selected for KRAS wild-type CRC, there were no NRAS or atypical KRAS mutations included in the group, as this is the emerging definition of cetuximab-appropriate patients.

We present a few putative mechanisms underlying the efficacy we observed with dual anti-EGFR and anti-VEGF inhibition in CRC patients. First, unlike earlier studies (17–20, 22), we selected only KRAS wild-type CRC patients for this study. Second, the combination of regorafenib and cetuximab may overcome the resistance to cetuximab, as suggested by preclinical data reported by Napolitano et al. (10). Third, regorafenib is a multikinase inhibitor of VEGF receptors 1–3 that also targets KIT, PDGFR-α and -β, RET, FGFR 1 and 2, TIE2, DDR2, TrkA, Eph2A, RAF-1, BRAF, SAPK-2, PTK5, and ABL. It is plausible that inhibition of one of the other targets of regorafenib contribute to overcoming resistance to previous anti-VEGF or anti-EGFR therapy.

Notably, the exceptional responder with Lynch syndrome may represent a distinct mode of benefit from the combination of cetuximab and regorafenib. Genomic profiling demonstrated an exceptionally high mutational burden even in the context of MSI. The benefit here may be linked to either the very high mutational burden or to the specific mutations in this tumor. Although the other patients in this study were microsatellite stable, CGP was not performed to assess tumor mutational burden (TMB) in these patients to further support this hypothesis, including a lack of characterization of the partial responder patient as to whether this patient had high TMB despite displaying MSI.

In this vein, an important followup study would be to apply this treatment regimen to additional MSI-H, high mutational burden CRC patients to assess replication of the response. The clinical need is clear, as although MSI-H CRC patients are known to have a good response rate to programmed cell death protein 1 (PD-1) inhibitors, optimal treatment for such patients is unclear after progression on immunotherapy. The combined cetuximab/regorafenib regimen could also benefit those CRC patients who are de novo refractory to PD-1 treatment or contraindicated from receiving such treatment due to comorbidities such as autoimmune disease. In conclusion, the combination of regorafenib (80 mg daily for 3 weeks, followed by a week off) and cetuximab (200 mg/m² loading dose, followed by 150 mg/m² every week) is reasonably well tolerated and demonstrated early signals of activity. We recommend a larger phase II study with this combination among patients with KRAS wild-type colorectal cancer who have previously received anti-EGFR and anti-VEGF therapy, and an explicit effort to include MSI and/or high TMB patients to replicate the exceptional responder outcome seen here.

Patients were eligible if they were aged 12 years or older, had evaluable or measurable advanced or metastatic cancer refractory to standard therapies, and had significant organ function reserve, defined as leukocyte count (WBC) greater than or equal to 3,000/ml, absolute neutrophil count (ANC) greater than or equal to 1,000/ml, platelets greater than or equal to 100,000/ml, creatinine less than or equal to 2 times the upper limit of normal (ULN), total bilirubin less than or equal to 2 times ULN, alanine aminotransferase (ALT, formerly serum glutamic pyruvic transaminase [SGPT]) less than or equal to 3 times ULN, and the ECOG performance status less than or equal to 2. For patients with liver metastases, total bilirubin less than or equal to 3 times the ULN, ALT or aspartate aminotransferase (AST, formerly serum glutamic oxaloacetic transaminase [SGOT]) less than or equal to 5 times the ULN were allowed. Patients with KRAS-mutated CRC, recent history of major surgery, hemoptysis or uncontrolled bleed, uncontrolled hypertension, clinically significant cardiovascular disease, or untreated or uncontrolled brain metastases were excluded.

Using the standard 3+3 trial design, patients received an escalating dose of regorafenib and cetuximab (Table 1) in 28-day cycles (Figure 2 and Table 7). During each cycle, regorafenib was administered orally daily for 21 days; patients did not receive any regorafenib during week 4 of each cycle. Cetuximab was administered i.v. every week following an initial loading dose on day 1 of cycle 1.

Figure 2

Representative examples of imaging studies of patients with response in the phase 1 trial. Top 2 rows: A 54-year-old woman with metastatic colorectal cancer and partial response. (A–D) A left lower lobe nodule decreases in size over the course of the first 3 followup scans, obtained approximately every 6 weeks. (E–H) A right hepatic lobe lesion decreases in size and enhancement over the same time period. Bottom 2 rows: A 35-year-old woman with metastatic carcinoma of unknown primary and stable disease. (I–L) Maximum intensity projections from ¹⁸F-FDG PET/CTs obtained approximately every 6 weeks show a left breast mass, left axillary adenopathy, and a right lower lobe metastasis. (M–P) Axial fused images at the level of the left breast mass show stable maximum standard uptake value (SUVmax) when accounting for scan-to-scan variation. The postobstructive consolidation in the right lower lobe related to a right perihilar lesion can also be seen.

Table 7

Planned dose escalation schema, number of patients enrolled in each cohort, and dose-limiting toxicities (DLTs)

Three patients were enrolled at the same dose level; if none of them experienced DLT, the next cohort of 3 patients was treated at the next higher dose level. If 1 of 3 patients treated at a dose level experienced DLT, 3 more patients were enrolled at the same dose level. If no other patient experienced DLT, the next cohort of 3 patients was enrolled at the next higher dose level. If 2 or more patients treated at a dose level experienced DLT, the MTD was considered to have been exceeded. In that case, 3 more patients (for a total of 6) were enrolled at the next lower dose level unless 6 patients had already been treated at that dose. In summary, the MTD was defined as the highest dose studied in which the incidence of DLT was less than 33%.

No patients were enrolled in the next dose level until 3 patients enrolled at the previous dose level had completed at least 1 cycle of therapy. If a DLT was observed in a patient, dose escalation did not proceed until a total of 6 patients in the cohort had been assessed for toxicity after 1 cycle. DLTs were defined as adverse events related to the study agents during the first cycle of treatment and included any grade 3 or 4 nonhematologic toxicity (except nausea and vomiting responsive to appropriate regimens, alopecia, or correctable electrolyte imbalances), any grade 4 hematologic toxicity lasting 2 weeks or longer despite supportive care, or any severe or life-threatening complication or abnormality. All toxicities were graded according to Common Terminology Criteria for Adverse Events, version 4.0 (CTCAE v4.0).

Once the MTD was determined, additional patients were enrolled; if 33% or more of patients at that dose level experienced DLT, then that dose level was to be considered above the MTD. Patients had radiologic restaging after every 2 cycles and were allowed to continue treatment in the absence of disease progression or significant toxicity. Radiographic response or progression was evaluated on the basis of Response Evaluation Criteria in Solid Tumors, version 1.1 (RECIST v1.1) (24).

Molecular profiling. All histology was centrally reviewed and molecularly tested at MD Anderson as part of standard of care as previously published (23). Paraffin-embedded tumor sections that were macrodissected were used for DNA extraction. KRAS exon 2 (codons 12 and 13) mutations, whenever done were expanded to also report KRAS, NRAS, BRAF V600E, and PIK3CA mutations. This was based on the standard of care testing results done in accordance with the Clinical Laboratory Improvement Amendment (CLIA)-certified lab (23).

One patient with durable response had biopsy samples submitted as formalin-fixed, paraffin-embedded blocks for CGP. Using FoundationOne, DNA of 315 cancer-related genes plus select introns from 31 genes involved in rearrangements was sequenced to a median depth of coverage of 481×.

Statistics. The primary objective of this study was to assess safety and tolerability as well as to define the MTD of combination treatment with regorafenib and cetuximab in patients with advanced cancer and CRC. Secondary objectives included a preliminary assessment of antitumor efficacy of the combination (objective response by RECIST) and a preliminary assessment of the exceptional responder using clinical next-generation sequencing. Descriptive statistics were performed to report patient characteristics, adverse events, and responses observed. Clinical benefit was defined as objective response or no evidence of progression for 6 or more cycles (partial response or stable disease ≥ 6 cycles).

Study approval. This was an investigator-initiated, non–company sponsored, open-label phase I dose escalation trial conducted at The University of Texas MD Anderson Cancer Center. Informed written consent was obtained from all patients, and the study was reviewed and approved by the IRB of MD Anderson Cancer Center.