A Phase II Study of Lapatinib in Recurrent/Metastatic Squamous Cell Carcinoma of the Head and Neck

Head and neck cancer is the fifth most common cancer in the world. In the United States, it accounts for 3% to 5% of all malignancies annually, representing approximately 2% of all cancer deaths in 2010 (1). Therapeutic options are limited for patients who have recurrent/metastatic (R/M) squamous cell carcinoma of the head and neck (SCCHN), and most patients will die within 1 year of recurrence (2). Few patients with recurrent disease are suitable for potentially curative salvage surgery, and some patients can benefit from radiation with or without chemotherapy, with a dismal overall response to second-line therapy. In addition, radiation, salvage surgery, and chemotherapy have high-toxicity profiles and should be carefully planned in the palliative setting (3).

In this context, novel agents, particularly those targeting epidermal growth factor receptor (EGFR) have been extensively studied. Overexpression of EGFR occurs in almost all SCCHN (4–6). EGFR is a member of the ErbB family of receptors, composed of EGFR (HER-1 or ERBB1), ERBB2 (HER-2/neu), ERBB3 (or HER-3), and ERBB4 (or HER-4). Upon ligand binding, EGFR forms a homodimer or heterodimer with other members of the family, resulting in dimerization of the receptors, autophosphorylation, and activation of downstream signaling pathways. ERBB2 has no known ligand, but it is the preferred heterodimerization partner for EGFR. The presence of ERBB2 has been established in SCCHN, and it is hypothesized that the EGFR/ERBB2 heterodimer cross-talk may have a role in tumor progression (7). ERBB3 lacks tyrosine kinase activity and thus participates in signal propagation through dimerization with other family members. Also, the mechanisms of resistance to EGFR inhibitors have not been well described, and cross-activation of the downstream EGFR signaling molecular pathways by other receptors, such as ERBB2 or ERBB3, may play a role (8). We thus rationalized that dual inhibition of EGFR and ERBB2 was promising.

Lapatinib is a competitive reversible inhibitor of EGFR and ERBB2 that has been previously reported to inhibit growth in SCCHN and other tumor xenografts expressing EGFR and ERBB2 (9, 10). Prior phase I clinical trials have shown the tolerability of lapatinib in the locally advanced and R/M settings (11–15). We conducted a multicenter trial of lapatinib in 2 groups of patients with R/M disease: those who have never received EGFR target therapy (EGFR naive) and those who had already been exposed to an EGFR inhibitor. Exploratory correlative analyses included pretreatment immunohistochemistry (IHC) staining [EGFR, ERBB2, ERBB3, PTEN, and insulin growth receptor-1 (IGF-1R)] and effect of lapatinib on tumor growth and survival pathways [EGFR, ERBB2, extracellular signal–regulated kinase (ERK), and AKT] in paired tumor biopsies obtained before and after 14 days of therapy.

This was an open-label, uncontrolled, multicenter phase II study, conducted by the University of Chicago Phase II Consortium. Patients were required to have histologically or cytologically confirmed R/M SCCHN with measurable disease, 18 years or older with Eastern Cooperative Oncology Group performance status less than 2 and intact organ function, as previously described (16). Patients were enrolled into 2 cohorts based on prior exposure to EGFR inhibitor: those without prior exposure to an EGFR inhibitor (arm A) and those with prior exposure to an EGFR inhibitor (arm B). Patients were allowed to have no more than 2 prior regimens for R/M SCCHN, and prior irradiation was allowed. Institutional Review Board approval was obtained, and all patients provided informed consent.

All subjects were treated with lapatinib 1,500 mg daily orally or via a gastrostomy tube, and advised to take it on an empty stomach (either 1 hour before or 1 hour after meals). Patients continued on lapatinib single-agent therapy until progressive disease, clinical deterioration, or unacceptable adverse events were observed. Adverse events were summarized by type and grade.

The trial design was specific to each patient cohort and decisions about early stopping and treatment efficacy were made independently. For patients who had not been previously treated with an EGFR inhibitor (naive cohort A), the primary outcome was the objective response rate, defined as complete response (CR) and partial response (PR). A Simon (17) optimal 2-stage design was used to test the null hypothesis that the response rate (CR + PR) was 10% against the alternative that it was 25%. We would enroll 21 evaluable patients in the first stage and if 2 or less responses were observed, further study of this cohort would be terminated due to lack of treatment efficacy. During the second stage, an additional 29 patients would be enrolled for a total of 50. This design had an α level of 10%, a power of 90%, and a probability of early stopping if the true response rate was only 10%.

In the patients who had been previously treated with an EGFR inhibitor (pretreated cohort B), the primary endpoint was the progression-free survival time (PFS). We would test the null hypothesis that the median PFS time was 2 months against the alternative that it was 4 months. Using the method described in Brookmeyer and Crowley (18), we calculated the lower, 1-sided, 90% confidence limit for the median PFS time, and if this value exceeded 2 months we would consider the treatment regimen worthy of further study. Assuming exponential survival and no losses to follow-up before 2 months, a sample size of n = 38 patients would provide 88% power for this evaluation, based on a binomial test at the (1-sided) 10% significance level. As an early stopping rule, we would examine the data after 19 patients had been followed for 2 months and if the observed PFS rate was no greater than 0.5, we would consider terminating this cohort. This simple rule provides a 50% probability of stopping early under the null hypothesis and the results of Wieand and colleagues (19) indicate that the power loss is minimal (<2%).

Archival paraffin-embedded, formalin-fixed tissue was requested from every patient enrolled on study. IHC was carried out with commercially available antibodies against EGFR, ERBB2, ERBB3, IGF-1R, and PTEN, as previously described (20). Samples that were adequate were evaluated further using a 4-point scoring system on the basis of the number of cells that stained positively (0, no staining; 1, ≤10%; 2, 10% to 50%; 3, 50%). Histologic examination was carried out on all samples by a single pathologist, who was blinded to timing of biopsy and response data.

Tumor core biopsies were carried out and tissue collected on all consenting patients within 1 week before starting therapy (between day −7 and day 0) and after 2 weeks of therapy (day 14 or 15). Tumor biopsies were carried out with a 14-guage biopsy needle, instantly frozen, and stored at −80°C, as previously described (20). To measure potential target inhibition by lapatinib and downstream signaling, antibodies against human phosphorylated EGFR (pEGFR; Y1173; Santa Cruz Biotechnology; 1:100), EGFR (1005; Santa Cruz; 1:100), pERK (Calbiochem; 1:100), ERK (Calbiochem, 1:100), pAKT (Calbiochem catalog# 1240011:100), and AKT (Cell Signaling catalog# 2966;1:100) were incubated with the rehydrated slides as previously described (21). Quantification of protein expression was determined by laser scanning cytometry (LSC) analysis which obtains 2- and 3-color mean fluorescence intensity (MFI) information as previously described (21, 22). Cell nuclei were contoured by the red fluorescence (propidium iodide) detector. Biomarker positive events, that is, phosphorylated and total EGFR, ERBB2, ERK, and AKT, were detected with the green detector. The presence of known ERBB2-activating mutations was also investigated as previously described (23).

Paired t tests or Wilcoxon signed-rank tests were carried out to examine the magnitude and significance of pre-post treatment changes. To determine whether these markers were correlated with tumor response, both the baseline levels and the pre-post changes would be compared between responders and nonresponders using the nonparametric, Wilcoxon rank-sum test.

From November 2004 to February 2006, 45 patients were enrolled (27 in the EGFR-naive arm A; 18 in the pretreated arm B). Most patients were male (84%), and the most common primaries tumors were in the oropharynx (24%), hypopharynx (24%), and oral cavity (20%). Fourteen patients in arm A and 13 in arm B had been treated with prior chemotherapy for R/M disease. Three patients with no chemotherapy for R/M disease in arm A, and 1 patient in arm B, relapsed within 6 months of potentially curative chemoradiotherapy. For a median follow-up of 163 days (range, 2–2223 days), the median number of cycles of therapy delivered was 2 (range, 1–9), with 10 patients (22%) receiving 1 or less cycle. Baseline characteristics are provided in Table 1.

Twenty-seven EGFR-naive patients were enrolled in arm A. Six patients were not evaluable for response due to clinical deterioration before response assessment. There was no complete or PR reported among these 27 patients. Eleven patients (41%) achieved a best response of stable disease, with a median duration of response of 50 days (range, 34–159 days). The overall stable disease rate was 37% (95% CI, 19.4–57.6) by intention to treat analysis, whereas 4 patients died on treatment. The 6-month PFS rate based on the Kaplan–Meier curve for these patients was 7% (95% CI, 1–21) and the median PFS was 52 days (95% CI, 34–109 days). The median and 6-month overall survival (OS) rate were 288 days (95% CI, 62–374) and 52% (95% CI, 32–69), respectively. At the time of data analysis, all 27 patients had progressed, a total of 26 patients had died, and 1 was alive at day 2063.

Similar to arm A, there were no objective responses in any of the 18 patients in arm B. At the time of analyses all patients had died and the median and 6-month PFS were 52 days (95% CI, 39–64) and 6% (95% CI, 0.4–22) respectively. The median OS was 155 days (95% CI, 75–242) and the 6-month OS rate, 39% (95% CI, 17–60). Three patients (17%) had stable disease as their best response, with median duration of response of 163 days (range, 135–195 days). Kaplan–Meier OS and PFS curves are shown in Fig. 1.

The most common toxicities were diarrhea (49%) and rash (33%), although the severity was grade 1 or 2 in most patients. Grade 3 events occurred in 7 patients: fatigue in 3, hyponatremia in 2, diarrhea in 1, and vomiting in 1. One patient had a reduction in left ventricular ejection fraction (60%–40%) which was asymptomatic and recovered to baseline upon discontinuation of lapatinib. There was no treatment-related death. Seven patients died while on treatment, and in 6, the cause of death was reported to be disease progression, and 1 patient with coronary artery disease had a sudden cardiac death, considered not related to the drug. Table 2 summarizes grade 1 or higher toxicities considered to be related to lapatinib in all patients.

Pretreatment IHC staining for EGFR, ERBB2, ERBB3, PTEN, and IGF-1R was carried out in 23 baseline specimens, as shown in Table 3. Most patients expressed EGFR and ERBB2. However, PTEN expression was noted in only 1 patient, whereas IGF-1R was not noted in this sample of patients. Figure 2 shows the rate of MFI of the biomarkers prior and after treatment as represented in box plots. Posttreatment biopsies showed a significant decreased pERBB2/ERBB2 ratio (P = 0.048) when compared with pretreatment biopsies. Reduction in phosphorylation of EGFR, ERK, and AKT was not statistically significantly different after therapy compared with baseline. Further, in an exploratory analyses, 3 available samples were analyzed for the presence of ERBB2-activating mutations and only a nondamaging variant was found.

The most active cytotoxic regimens for R/M SCCHN include a platinum agent in combination with fluorouracil or a taxane and have been associated with a 30% response rate, a 3- to 4-month median PFS, and a median OS of 6 to 8 months (2, 24). Patients progressing on a platinum-based therapy have limited treatment options and poor prognosis. Building on the results of the Erbitux in First-Line Treatment of Recurrent or Metastatic Head and Neck Cancer (EXTREME) study, that confirmed the benefit of adding cetuximab to chemotherapy as a first-line treatment in the R/M setting (25), with no decrease in quality of life, the interest in EGFR inhibitors has been intense. The response rate of SCCHN to EGFR tyrosine kinase inhibitors when used as single agents ranges between 4% and 11% (20, 26). In an effort to increase efficacy of targeted therapy, we pursued dual inhibition of EGFR and ERBB2 administering lapatinib, an orally reversible inhibitor of EGFR and ERBB2 tyrosine kinases that has been previously shown to be well tolerated in phase I clinical trials. Although some toxicity concerns have arisen when lapatinib was combined with induction chemotherapy in the locally advanced setting (27), this trial further confirmed that lapatinib monotherapy is well tolerated in heavily pretreated patients. However, monotherapy with daily 1,500 mg of lapatinib was inactive with no clear benefit in this study of R/M SCCHN patients, regardless of prior exposure to an EGFR inhibitor. Stable disease was the best achieved response in a minority of patients in both groups, and the PFS for both groups was only 52 days. Both arms were closed at interim analysis due to lack of efficacy.

These findings were unexpected given that prior studies assessing lapatinib in SCCHN showed modest but promising data. In a phase 1 trial of lapatinib in patients with metastatic carcinomas, Burris and colleagues (13) reported 3 patients with SCCHN who achieved stable disease for greater than 6 months, including a case of stable disease lasting 21 months. A further phase 1 trial from the same group showed a CR in a patient with locally recurrent SCCHN who received 1,250 mg of lapatinib daily (12). In a phase II study, 107 therapy-naive patients with locally advanced squamous cell carcinoma of the head and neck (LAHNC) were randomized (2:1) to receive 1,500 mg of oral lapatinib or placebo (28). In this “window of opportunity study,” tumor biopsies were taken at time of study enrollment (day 0) and after 2 weeks of study participation (day 14) for biomarker analysis. During the short duration (14 days) of lapatinib treatment, there was a statistically significant reduction in mean tumor cell proliferation index in patients who had received lapatinib compared with the placebo arm (−6% vs. −1.7%, respectively, P = 0.030). In 24 patients who had radiological scans following a short duration of lapatinib monotherapy, 4 patients (17%) had a CR or PR, compared with no responders in the placebo arm. Following standard treatment with concurrent platinum-based chemotherapy and radiation therapy, there also an improvement in objective response rate (CR and PR) for patients who had received lapatinib compared with the placebo arm (86 vs. 63%, respectively). Most recently, another randomized phase II trial comparing lapatinib (1,500 mg per day) or placebo with concurrent chemoradiotherapy followed by maintenance therapy with lapatinib or placebo in unresected LAHNC was reported (29). The CR rate at 6 months postchemoradiotherapy was 53% with lapatinib versus 36% with placebo. Given the overlapping mechanisms of radiotherapy and EGFR inhibition on cell proliferation, survival, angiogenesis, and DNA repair, these prior trials may suggest a radiosensitizer role of lapatinib, despite its lack of activity when used as single agent in the current trial.

The totality of our data strongly suggests that lapatinib is not an effective EGFR inhibitor, as may have been evident by the relatively low incidence of the typical EGFR inhibitor induced rash. Both the IHC staining and the MIF analyses are hypothesis generating, and the small sample size and lack of objective response precluded further correlation analyses between these data and clinical outcomes. In this regard, we hypothesize that the lack of EGFR inhibition may explain the lack of activity of lapatinib in this study. Although there was evidence of significant ERBB2 inhibition, reduction in phosphorylation of EGFR was not statistically significantly different after therapy compared with baseline. Because EGFR plays a critical role in the biology and pathogenesis of SCCHN, we hypothesize that ERBB2 inhibition alone is insufficient for appreciable efficacy. This hypothesis is further supported by a multiinstitutional randomized study comparing carboplatin and paclitaxel with or without trastuzumab, where the addition of the ERBB2 monoclonal antibody did not improve activity compared with control (30).

Notwithstanding, there may be SCCHN patients that do benefit from ERBB2 inhibition. We previously reported that response to tyrosine kinase inhibitors in a subset of SCCHN patients may in fact be due to mutations in ERBB2 rather than EGFR (23). This hypothesis may explain the CR previously shown in a patient treated with lapatinib but these events seem to be rare (12). In fact, recent data from SCCHN sequencing studies show that ERBB2 can be mutated or amplified in a minority of tumors, suggesting that there may be some patients that could benefit from lapatinib (31, 32). We can also speculate that the lack of clinical activity may be related to the absent baseline expression of IGF-1R, as it has been previously shown that, in absence of a functional IGF-1R, HER-2 does not become phosphorylated, and thus, may not be a pathway with clinical significance (33).

In conclusion, administration of lapatinib as monotherapy in subjects with R/M SCCHN did not produce any objective responses nor durable PFS in a significant number of subjects. Our data would suggest that lapatinib is a weak EGFR inhibitor and this underlies the lack of observed efficacy. Conversely, lapatinib seems to be a potent ERBB2 inhibitor and has been approved in breast cancers with ERBB2 amplification. Moreover, although lapatinib has not shown efficacy as either a single agent or with chemotherapy in SCCHN, its use as a radiosensitizer is currently under investigation. A phase III study (NCT00424255) is in progress to assess the benefit of postoperative lapatinib in high-risk SCCHN patients. Approximately 680 patients will be randomized to receive lapatinib (1,500 mg) or placebo once-daily with radiotherapy and cisplatin for 7 weeks. Patients will then receive lapatinib or placebo maintenance therapy for 1 year. Given the lack of efficacy of lapatinib as a single agent and the biologic significance of EGFR as a target in SCCHN, caution should be exercised before embarking on further trials assessing this agent.

No potential conflicts of interest were disclosed.

This study was supported by NIH grant no. N01-CM-17102, the University of Chicago Cancer Research Center (grant no. P30 CA14599), and Glaxo-Smith Kline.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.