Phase I Trial of Erlotinib-Based Multimodality Therapy for Inoperable Stage III Non-small Cell Lung Cancer

Abstract

Introduction

This Phase I trial aimed to determine the maximum-tolerated-dose of erlotinib administered with two standard chemoradiotherapy regimens for non-small cell lung cancer.

Methods

Unresectable stage III non-small cell lung cancer patients were enrolled in this 2-arm dose-escalation study. Erlotinib, given only during chemoradiotherapy, was escalated from 50 to 150 mg/d in 3 to 6 patient cohorts. Arm A: erlotinib with cisplatin (50 mg/m² IV days 1, 8, 29, 36), etoposide (50 mg/m² IV days 1–5, 29–33) and chest radiotherapy (66 Gy, 2 Gy/d) followed by docetaxel (75 mg/m² IV Q21 d) for 3 cycles. Arm B: induction carboplatin (AUC 6) and paclitaxel (200 mg/m²) for two 21-d cycles then radiotherapy with erlotinib, carboplatin (AUC = 2/wk) and paclitaxel (50 mg/m²/wk).

Results

Seventeen patients were treated in each arm. Patient characteristics: performance status 0 to 24 patients, 1 to 10 patients, median age 63 years, adenocarcinoma 21% and female 14 patients. Dose-escalation of erlotinib to 150 mg/d was possible on both chemoradiotherapy regimens. Grade 3/4 leukopenia and neutropenia were predominant toxicities in both arms. Grade 3 chemoradiotherapy toxicities in arm A were esophagitis (3 patients), vomiting (1), ototoxicity (1), diarrhea (2), dehydration (3), pneumonitis (1); and arm B was esophagitis (6). Seven patients (21%) developed rash (all grade 1/2). Median survival times for patients on Arm A and B were 10.2 and 13.7 months, respectively. Three-year overall survival in patients with and without rash were 53% and 10%, respectively (log-rank P = 0.0807). Epidermal growth factor receptor IHC or FISH positive patients showed no significant overall survival difference.

Conclusion

Addition of standard-dose erlotinib to chemoradiotherapy is feasible without evident increase in toxicities. However, the survival data are disappointing in this unselected patient population and does not support further investigation of this approach.

Prognosis of patients with unresectable stage III nonsmall cell lung cancer (NSCLC) has improved over the past two decades with concurrent chemoradiotherapy.^1–4 Attempts to further improve the outcome of unresectable stage III NSCLC with the use of induction or consolidation chemotherapy have not been successful.^5,6 The pressing need for more effective combined modalitiy-therapies in NSCLC lead us to explore targeting the epidermal growth factor receptor (EGFR).

EGFR overexpression occurs in approximately 60% of NSCLC and correlates with higher stage and worse prognosis.^7,8 There is a positive correlation between EGFR expression and tumor radioresistance.^9–12 Furthermore, the degree of radioresistance correlates with the magnitude of EGFR overexpression.⁹ In addition, radiation damage results in activation of EGFR and subsequently augments cell survival and repopulation.¹³ By inhibiting EGFR activation, tumor cells may become more radiosensitive.^14–16 This hypothesis is validated in the treatment of head and neck cancer with concurrent cetuximab and radiotherapy.¹⁷

Based on this rationale, we performed a two-arm phase I study to incorporate erlotinib into two frequently used chemoradiotherapy regimens (SWOG9504¹⁸ and CALGB39801⁵) for unresectable Stage III NSCLC. Erlotinib, an EGFR tyrosine kinase inhibitor (TKI) results in single-agent response rates of approximately 10% in unselected metastatic NSCLC patient populations but demonstrates higher activity in selected clinical or molecular patient subsets.¹⁹ We hypothesized that the addition of erlotinib to concomitant chemoradiotherapy could further increase radio- and/or chemotherapy sensitivity of tumor cells and result in higher locoregional and systemic control rates.

METHODS

This study aimed to determine the maximum tolerated dose (MTD) and dose-limiting toxicity (DLT) of erlotinib administered with chest radiotherapy and concomitant cisplatin-etoposide (Arm A); and carboplatin-paclitaxel (Arm B) for unresectable stage III NSCLC.

Eligibility Criteria

Untreated patients with pathologically confirmed, unresectable, or inoperable Stage III NSCLC were eligible. Patients with malignant pleural effusions were ineligible. All patients were over 18 year of age, had an Eastern Cooperative Oncology Group performance status of 0 to 1 and adequate hematological (white blood cell count ≥3000/µL, absolute neutrophil count ≥1500/µL, platelet count ≥100,000/µL), hepatic (bilirubin within upper limit of normal (ULN), serum transaminases ≤ 1.5 × ULN, alkaline phosphotase ≤ 2.5 × ULN) and renal (creatinine clearance ≥50 mL/min) function. Patients were required to have at least one measurable lesion by chest x-ray, computerized tomography or magnetic resonance imaging.

Treatment Regimen

The treatment schema is shown in Figure 1. Patients received erlotinib only during concurrent chemoradiotherapy. Therapy on arm A, administered as in SWOG9504,¹⁸ consisted of concurrent radiotherapy (66 Gy, 5 fractions/wk over 7 wk) with concurrent cisplatin 50 mg/m² on days 1, 8, 29, and 36 and etoposide 50 mg/m² daily on days 1 to 5 and 29 to 33. After completion of concurrent chemoradiotherapy, consolidation docetaxel 75 mg/m² every 3 weeks for 3 cycles (days 50, 71, and 92) was administered.

FIGURE 1.

Treatment schema of Arms A and Arm B. (C–cisplatin, CB–carboplatin, D–docetaxel, E–Etoposide, P–Paclitaxel).

Therapy on arm B was administered as in CALGB39801.⁵ Treatment consisted of induction chemotherapy with paclitaxel 200 mg/m² and carboplatin AUC 6 on days 1 and 21. Concurrent chemoradiotherapy consisted of weekly paclitaxel 50 mg/m² and carboplatin AUC 2 on days 43, 50, 57, 64, 71, 78, and 85 with concurrent chest radiotherapy beginning on day 43 to a total of 66 Gy. The target volumes of irradiation were defined prior to the administration of induction chemotherapy.

Radiation therapy was given using photon beams with energy between 4 and 25 MV. Clinical target volume included the gross target volume, potential occult disease, the ipsilateral hilum, and mediastinum. Clinical target volume was treated with an initial 44 Gy at 2 Gy/fraction. Boost volume included the gross tumor volume, ipsilateral hilum, and ipsilateral mediastinum; and was treated to 22 Gy at 2 Gy/fraction. No corrections for lung or bone attenuation were made. Maximum dose to any point in the spinal cord was 49 Gy. Two- and three-dimensional treatment planning were allowed.

Supportive Care

Prophylactic use of granulocyte colony stimulating factor was not permitted. Erythropoietin was recommended for hemoglobin below 10 g/dL. Carafate and fluconazole was added during chemoradiotherapy as needed. Topical anti-inflammatory acne therapy and/or oral antibiotics were used to treat rash.²⁰

Toxicity Evaluation and Dose Escalation

Patients were assigned to a dose level, alternating between arm A or arm B (Fig. 2). On both study arms the dose levels of erlotinib started at 50 mg orally daily (level I), and escalated to 100 mg daily (level II) and 150 mg daily (level III).

FIGURE 2.

Dose escalation schema for Arms A and B.

Patients were evaluated weekly for toxicity using the NCI Common Toxicity Criteria 2.0. The trial was initially written to define DLT as the occurrence of grade 4 neutropenia. Since neutropenia is the predominant toxicity encountered during standard concomitant chemoradiotherapy, an amendment to exclude neutropenia as a DLT was approved midway during the accrual for dose level 1. DLT in this trial was defined as: Grade 4 thrombocytopenia, or need for platelet transfusion, grade 4 esophagitis, grade 4 vomiting despite maximal antiemetic support, grade 3 or greater despite antidiarrheal support, and any toxicity grade 3 or greater exceeding 7 days. Anemia, nausea, fatigue and alopecia were not considered DLTs.

A minimum of three assessable patients were entered at each dose level. Erlotinib was started at dose level I (50 mg). If DLT developed, the dose level was expanded to include six patients. Dose escalation continued until greater than one third of patients treated at a given dose level developed DLT. Once the MTD was determined, up to an additional six patients were to be tested at the determined MTD for each sequence.

Response Evaluation and Statistical Considerations

Arm A and B were evaluated separately to determine the MTD of each regimen. Response to treatment was assessed in all patients. Best clinical response to treatment with erlotinib and chemoradiotherapy was determined using the Response Evaluation Criteria in Solid Tumors criteria. Response evaluation for arm A was performed after concurrent chemoradiotherapy and after consolidation chemotherapy. In arm B, response evaluation was performed after induction chemotherapy and after concurrent chemoradiotherapy. Patients who were not evaluable, died early from any cause, or withdrew from the study were considered as failing to respond to the treatment, and were classified as disease progression.

Immunohistochemistry for Epidermal Growth Factor Receptor

Samples of tissue sections from the primary tumors were identified and cut in 5 µm sections onto positively charged slides. The tissue sections were deparaffinized in xylene and hydrated with alcohol before being placed in 3% H₂O₂/methanol blocking solution to quench endogenous peroxidase activity followed by subsequent antigen unmasking. Nonspecific binding was reduced by incubating the slides in a protein blocking solution for 20 min. Incubation with the primary antibodies was done with the monoclonal anti-EGFR antibody clone 31G7 (Invitrogen/Zymed, Carlsbad, CA) with a 1:100 dilution. After washing with TBS, the slides were incubated for 30 min at room temperature with goat anti-rabbit IgG conjugated to a horseradish peroxidase–labeled polymer (Envision System, Carpinteria, CA). Reactions were developed with 3,3′-diaminobenzidine chromogen and counterstained with hematoxylin. Appropriate negative controls for the immunostaining were prepared by omitting the primary antibody step and substituting it with nonimmune rabbit serum.

All of the slides were reviewed by an independent pathologist and standard scoring was performed. For each case, an average number of 1500 cells per section was evaluated utilizing a semi-quantitative grading system based on 4 stages (0, no staining; 1+, staining in 1–10% of considered cells; 2+, staining in 11–25% of considered cells; 3+, staining in >25% of considered cells).²¹ Tumors with 2+ or 3+ staining were considered EGFR positive on IHC.

Fluorescent In Situ Hybridization

Fluorescence in situ hybridization for detection of EGFR gene amplification was performed on formalin-fixed paraffin-embedded tissue sections adjacent to those analyzed by IHC using the Vysis Inc. LSI^® Locus Specific Identifier DNA Probes (Vysis/Abbott Inc., Des Planes, IL), which is a hybridization mixture of a EGFR probe labeled with SpectrumOrange, and a chromosome 7 enumeration probe CEP7, labeled with SpectrumGreen. Slides were pretreated using the Vysis/Abbots Inc. Paraffin Pretreatment Kit and Post-Hybridization Rapid Wash LSI Protocols. Briefly, slides were baked at 70°C overnight, deparaffinized in series of xylenes, dehydrated in 100% ethanol. After incubation in sodium chloride, slides were pretreated in sodium thiocyanate (1 M NaSCN, 80°C for 30 min), washed in standard saline citrate buffer (2 × SSC) and processed for digestion with Pepsin (Sigma, 2300 U/mg, 0.5 mg/mL in 0.9% NaCl, pH = 1.5). Slides were fixed in 10% buffered formalin for 10 min, washed in 2 × SSC, and denatured in 70% formamide at 75°C for 5 min and dehydrated in ethanol series. Probe mixture was diluted in t-DenHyb-2 hybridization buffer (InSitus Biotechnologies, Albuquerque, NM, USA) as described by the manufacturer, denatured and applied to the slide. Hybridization was conducted in a humid chamber at 37°C overnight. Slides were immersed in 2 × SSC/0.1% NP40 at room temperature for 2 minutes followed by wash in 0.4 × SSC/0.3% NP40 at 73°C for 2 min. For visualizing the hybridization, DAPI II counterstain by Vysis was applied.

In each tumor sample an average of 80 (30–200) well-defined malignant nuclei were scored. The absolute number of EGFR signals, the ratio of EGFR signals to CEP7 signals and the percentage of cells with given copy number of each signal per cell were recorded. Tumors with a EGFR:CEP7 signal ratio <2 were considered nonamplified, whereas those with a ratio of 2 or greater (or ≥ 15 copies of EGFR per cells in ≥ 10% of cells) were considered amplified. “Low amplification” was defined as ratio of 2.0 to 3.0 and “high amplification” a ratio of >3. The alterations in EGFR signals due to alterations in chromosome 7 copy number were classified as described previously.^7,22 Briefly, disomy, ≤2 gene copies in more than 90% of cells; trisomy, 3 gene copies in ≥10% of cells; low polysomy, ≥4 gene copies in ≥10% but less than 40%; high polysomy, ≥4 gene copies in ≥40%. Fluorescent in situ hybridization (FISH) positivity was defined by the high number of the copies of EGFR (amplification or high polysomy).^7,22,23

RESULTS

Patient Characteristics

Between July 2002 and March 2006, 34 patients were enrolled (Table 1). There were 14 women and 20 men with a median age of 63 years (39–78 years). All patients were either current or former smokers. Ten patients (24%) had Stage IIIA disease while 24 (71%) had Stage IIIB disease. Histologic subtypes were: poorly differentiated 47%, squamous cell carcinoma 29%, adenocarcinomas 21% and large cell carcinoma 3%.

TABLE 1.

Patient Demographics

n	34
Median age
Sex	63 yr	39.1–78.2 yr
Male	20	59%
Female	14	41%
Race
Caucasian	24	71%
African-American	10	29%
Performance status
0	24	71%
1	10	29%
Smoking status
Current smoker	10	29%
Ex-smoker	24	71%
Stage
IIIA	10	29%
IIIB	24	71%
Histology
Poorly differentiated	16	47%
Squamous cell carcinoma	10	29%
Adenocarcinoma	7	21%
Large cell carcinoma	1	3%

Toxicity

Chemoradiotherapy Toxicity in Arm A

Of four evaluable patients in dose level 1, grade 3/4 neutropenia occurred in three patients (75%), and grade 3 esophagitis in two patients (Table 2). One patient (25%) developed grade 3 pneumonitis 5 weeks into chemoradiotherapy and improved with corticosteroids.

TABLE 2.

Toxicities Observed During Concurrent Chemoradiotherapy in Arm A (Cisplatin-Etoposide-Erlotinib-Radiotherapy) and Arm B (Carboplatin-Paclitaxel-Erlotinib-Radiotherapy)

	Hemoglobin		WBC		ANC		Platelets		Esophagitis		Vomiting		Neuropathy		Diarrhea		Rash		Radiation Dermatitis		Others
Grade	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4
Arm A
Level 1 (n = 4)	4(100%)		1(25%)	3(75%)	1(25%)	3(75%)	1(25%)	1(25%)	1(25%)	2(50%)		1(25%)	1(25%)			1(25%)			1(25%)			1 Pneu
Level 2 (n = 3)	2(67%)		1(33%)	2(67%)		2(67%)		3(100%)	2(67%)		2(67%)				2(67%)				2(67%)
Level 3 (n = 9)	6(67%)		2(22%)	4(44%)	4(44%)	3(33%)	6(67%)	1(11%)	5(56%)	1(11%)	1(11%)		4(44%)		4(44%)	1(11%)	3(33%)		3(33%)			1 Otox
Arm B
Level 1 (n = 5)	2(40%)		2(40%)	1(20%)	1(20%)	1(20%)	1(20%)	1(20%)	5(100%)		3(60%)		2(40%)				2(20%)
Level 2 (n = 3)	2(67%)		2(67%)		1(33%)	1(33%)	2(67%)		1(33%)	2(67%)			2(67%)				1(33%)
Level 3 (n = 7)	7(100%)		2(29%)	4(57%)	4(57%)	1(14%)	3(43%)	1(14%)	3(43%)	4(57%)	3(43%)		2(29%)		5(71%)		1(14%)		2(29%)

Pneu, Radiation pneumonitis; Otox, Otoxicity.

Underlined-italic font indicates dose-limiting toxicity.

Dose level 2 was well tolerated without DLT. One patient in dose level 2 died just prior to starting consolidation chemotherapy from a ruptured abdominal aortic aneurysm that was unrelated to the therapy.

No DLTs were observed in the first three patients in dose level 3. This cohort was subsequently expanded to nine patients. As in previous dose levels, neutropenia was the most common severe hematologic toxicity. Fatigue, esophagitis, mucositis, neuropathy and diarrhea were the most common nonhematologic toxicities. Grade 3 hearing loss was the only DLT observed. It was at this dose level that rash was noted in three patients (33%). Two patients were females and, all three were poorly differentiated carcinomas. Grade 3 dehydration occurred in three patients but improved rapidly with intravenous hydration. Grade 3 diarrhea occurred in one patient and improved with antidiarrheal therapy.

The MTD of treatment arm A was erlotinib 150 mg.With exception of neutropenia and leucopenia noted early in the trial, the only DLTs observed in this arm were hearing loss (1 patient) and pneumonitis (1 patient). Commonly observed grade 3/4 toxicities were esophagitis (3 patients), leucopenia (9 patients), neutropenia (8 patients), and thrombocytopenia (5 patients).

Chemoradiotherapy Toxicity in Arm B

In the five patients treated in dose level 1, grade 4 neutropenia was encountered in one patient (20%). This patient also experienced grade 3 leucopenia and thrombocytopenia. There were no other grade 3/4 toxicities observed in dose level 1. In dose level 2, one patient experienced grade 3 neutropenia; all other hematologic toxicities were mild. Two patients (67%) in this dose level experienced grade 3 esophagitis.

In dose level 3, no DLTs were observed in the first three patients and this cohort was expanded to seven patients. At this dose level, grade 3/4 neutropenia remained the predominant hematologic toxicity. Grade 3 esophagitis occurred in four patients.

The MTD of treatment arm B was erlotinib 150 mg. The only DLT observed was in one patient who developed grade 4 thrombocytopenia. Commonly observed grade 3/4 toxicities were esophagitis (6 patients), and neutropenia (3 patients). Rash was observed in four patients (2 in dose level 1, and 1 each in dose levels 2 and 3). There were two female patients who developed rash and two patients each had poorly differentiated carcinoma and squamous cell carcinoma.

Toxicity from Chemotherapy

The maximum toxicity grades during consolidation (Arm A) and induction (Arm B) chemotherapy are listed in Table 3. Five patients did not receive consolidation docetaxel (died of an aortic aneurysm rupture after concurrent chemoradiotherapy– 1 patient, refusal of further therapy after chemoradiotherapy– 4 patients). Consolidation docetaxel was associated with grade 3/4 leukopenia (3 patients/4 patients). Induction carboplatin-paclitaxel was well tolerated with expected rates of leukopenia, and thrombocytopenia. One patient died of a myocardial infarct and another patient developed grade 4 fatigue after induction chemotherapy.

TABLE 3.

Toxicities Related to Consolidation Docetaxel (Arm A) and Induction Carboplatin-Paclitaxel (Arm B)

	Hemoglobin		Leucopenia		Neutrpenia		Platelets		Anorexia		Fatigue		Nausea		Vomiting		Diarrhea		Neuropathy
Grade	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4	1/2	3/4
Arm A (n = 12)	10(83%)		5(42%)	7(58%)	5(42%)	7(58%)	2(17%)		4(33%)		10(83%)		2(17%)		1(8%)		1(8%)		1(8%)
Arm B (n = 17)	9(53%)		5(29%)	1(6%)	2(12%)	6(35%)	2(12%)	1(6%)	4(24%)		11(65%)	1(6%)	5(29%)		1(6%)		2(12%)		5(29%)

Response Evaluation

Response to therapy was determined on an intention-totreat basis for all patients. In Arm A (n = 17), partial response was achieved in 11 patients for a 65% overall response rate. Stable disease and progressive disease occurred in two (12%) and four patients (24%), respectively. In Arm B (n = 17), the complete and partial response rates were 6% (1 patient) and 53% (9 patients), respectively, for an overall response rate of 59%. Stable disease and progressive disease occurred in two (12%) and five patients (29%), respectively.

Sites of Failure

Relapse occurred in 21 patients. Local relapse occurred in two patients (18%) in Arm A, and four patients (40%) in Arm B. Distant relapse occurred in five patients (45%) in Arm A and one patient (10%) in Arm B. Simultaneous local and distant relapse occurred in four (36%) and three patients (30%) in Arm A and B, respectively. Two patients in Arm B had disease recurrence but their site of relapse was not documented.

Survival

The median follow-up was 11.1 month (1–61 months), seven patients were still alive. The intention-to-treat 12-, 24-, and 36-month overall survivals were 50%, 25%, and 16%, respectively and median survival was 11 months (Fig. 3A). Median progression-free survival for the study was 9 months (Fig. 3B). Intention-to-treat median survival for Arm A was 11 months and Arm B 15 months (Fig. 3C). The overall survival for Arms A and B at 36 months, respectively were 20% and 16% (log-rank P = 0.8979). The progression-free survival at 36 months for Arm A was 13% and Arm B 15% (log-rank P = 0.9168) (Fig. 3D).

FIGURE 3.

Kaplan-Meier curves for A, overall survival in all patients; B, Progression- free survival in all patients according to treatment arm; C, overall survival in all patients according to treatment arm (Arm A = SWOG, Arm B = CALGB); D, Progression-free survival in all patients according to treatment arm; E, Overall survival according to appearance of rash; F, Progression- free survival according to appearance of rash; G, Overall survival according to EGFR immunohistochemistry or EGFR increased gene copy number/amplification status; H, Progression-free survival according to EGFR immunohistochemistry or EGFR increased gene copy number/amplification status.

The median survival for patients who did not develop a rash with erlotinib (n = 24) was 10 months while the median survival for patients who had rash of any grade (n = 7) was not reached (Fig. 3E). The overall survival between the patients with rash and those without at 36 months were 53% and 10%, respectively. This difference had a trend towards improved outcome among patients who had a rash (log-rank P = 0.0807).

EGFR Status and Clinical Outcome Evaluation

Eighteen out of 22 available pathology samples (82%) were successfully tested for EGFR IHC. Ten patients (56%) had high EGFR staining. Twenty out of 22 pathology samples (91%) were available and successfully tested for EGFR FISH. While EGFR amplification was not observed, high polysomy in five tumors (25%) accounted for FISH positivity in this group. Three tumors demonstrated high EGFR staining and FISH positivity.

Both EGFR FISH and IHC results were available in eighteen patients. Three patients who were FISH positive also had high EGFR expression by IHC, while eight patients who were FISH negative had low EGFR expression. The concordance rate between EGFR FISH and IHC was 61%.

Analysis of patients whose tumors either EGFR IHC or FISH positive showed no significant overall survival difference over those who were negative for both (Fig. 3G). In addition there was no significant difference between the two groups in terms of progression-free survival or time to progression (Fig. 3H). In the seven patients who developed acneiform rash, tumor tissue was available in five patients. Three out of the 5 patients (60%) were found to have either EGFR IHC or FISH positive tumors.

DISCUSSION

EGFR inhibitors play a significant role in the management of metastatic NSCLC. EGFR-TKIs and anti-EGFR monoclonal antibodies are active in NSCLC.^19,24 Laboratory and clinical evidence supports the hypothesis that EGFR inhibitors may enhance radiosensitization.^9–13,17 Combining EGFR inhibition and concurrent chemoradiotherapy is of theoretical interest since EGFR inhibition may result in increased tumor radiosensitivity via mechanisms different from those of chemotherapy.

In this study, we sought to investigate the tolerability and safety of erlotinib when combined with concurrent chemoradiotherapy. Erlotinib was well tolerated at 150 mg daily during concurrent chemoradiotherapy with the two commonly used regimens. We did not observe increased rates of in-field dermatitis or radiation pneumonitis.

In this small trial, the response rate in both arms is approximately 60% and appears similar to the response rates achieved by the standard chemoradiotherapy regimens. The median survival in Arm A and Arm B were 10.2 and 13.7 months, respectively. The median survival in Arm B was similar as reported for CALGB 39801 study.⁵ The median survival of Arm A was disappointing when compared with reports on this regimen by SWOG or the Hoosier Oncology Group trials.^6,18

There may be several reasons for this difference. Intertrial comparisons have to be interpreted with caution and, the small number and heterogenous population of our patients must be taken into account. Nevertheless, a positive clinical signal supporting the hypothesis of a benefit from the addition of erlotinib to chemoradiotherapy using either base regimen is not apparent in this trial.

Clinical predictors of response to EGFR TKIs—female, East Asian race, adenocarcinoma histology and nonsmokers, were not well represented in our study. We did, however, observe that patients who developed an acneiform rash had a trend towards improved survival. Skin rash developed in only seven patients (21%) but in these patients, one attained a complete response, four had partial responses, and one had stable disease. This observation suggests an EGFR inhibition effect and is consistent with other erlotinib studies in lung,^19,25 head and neck,²⁶ ovarian²⁷ and pancreatic cancer²⁸ demonstrating an association between response and survival with the development of a rash and its severity. Interestingly, the incidence of rash in our subjects is low compared with other studies where rash occurred between 60 and 75% of subjects.^19,29 The occurrence of erlotinib-associated rash is related to the maximum concentration and AUC of erlotinib attained in patients. There is a wide interindividual pharmacokinetic variability of erlotinib which may in part be related to EGFR intron 1 polymorphism and the drug transporter, ATP-binding cassette G2 (ABCG2).³⁰ The low incidence of rash in our subjects may reflect our limited patient population.

In line with the findings of the BR21 study, we observed that patients who had high EGFR expression or copy numbers tended to also develop a rash and respond to therapy.²³ Furthermore, the BR21 study showed that patients with high EGFR gene copy numbers also had a higher incidence of rash and increased rash severity.³¹ Despite the improved tumor response in our patients who have positive clinical and molecular predictors of response, the improved response rate did not translate into a survival benefit.

EGFR mutation analysis was not performed in our study. When the trial was designed, it was not mandatory for tumor tissue to be banked and sites were given the option to collect three slices of paraffin-embedded tumor tissue. For most of our patients the diagnosis of stage III NSCLC was made on fine-needle aspiration where tumor blocks were not made. This precluded us from obtaining an adequate sample size to perform mutational analyses. However, the presence of EGFR mutations has been predictive of tumor response to erlotinib monotherapy,³² and Ready et al reported a statistical trend between EGFR mutation status and survival when gefitinib was delivered concurrently with radiotherapy.³³

The poorer survival in our study may be attributed to the presence of k-ras mutations which render tumors insensitive to EGFR-TKIs.³⁴ Although this postulation cannot be tested definitively, our study patients are all current or former smokers, which k-ras mutations are commonly associated.³⁵ Furthermore, analysis of the TRIBUTE study showed that patients with k-ras mutations had significantly shorter survival when treated with chemotherapy and erlotinib, suggesting a possible detrimental effect of erlotinib in patients harboring such mutations.³⁶

The CALGB trial by Ready et al may have relevance to our study.³³ The trial administered induction carboplatin-paclitaxel followed by concurrent gefitinib, carboplatin, paclitaxel and radiotherapy to “good-risk” patients (less than 5% weight loss and performance status 0–1) and achieved a median survival of 12 months. However, in “poor-risk” patients, who received induction carboplatin-paclitaxel followed by concurrent radiation therapy and gefitinib daily, an 18-month median survival and 1-year survival of 60% was achieved. This difference in outcome suggests a possible antagonist effect of between chemotherapy and gefitinib during chemoradiotherapy. Such an effect may be responsible for the poor outcome in our trial and in the trials evaluating EGFR-TKIs with chemotherapy.^29,37–39 This finding is also supported by preclinical data demonstrating schedule-dependant interaction between cytotoxic chemotherapy and EGFR-TKIs.^40–43 It is hypothesized that continuous exposure to EGFR-TKIs results in G1 phase arrest, and therefore reducing cell cycle phase-dependent activity of cytotoxic chemotherapy.^42–44 This effect preferentially affects wild-type EGFR.⁴³ Similarly, the SWOG 0023 which randomized patients after concurrent chemoradiotherapy and consolidation docetaxel to either maintenance placebo or gefitinib, showed a worse survival in the gefitinib arm. There have been numerous postulated reasons but they remain unclear pending the molecular analysis of the pathology samples.

EGFR-TKIs are standard second-line treatment for NSCLC. Early trials determined that EGFR-TKIs do not improve outcome when combined with chemotherapy. EGFRTKIs have not been explored with radiotherapy until recently. This is the first study investigating concurrent administration of erlotinib at its recommended single agent dose with chemoradiotherapy. This approach was well tolerated. However, the response rates and overall survivals observed in our small trial do not support further investigation of erlotinib with concurrent chemoradiotherapy in an unselected population despite its tolerability. We did not find molecular parameters associated with a favorable outcome. We did, however, observe that patients who developed a rash appeared to have better outcomes with the addition of erlotinib. This finding suggests that there may be specific patient subsets that could benefit from the addition of erlotinib to chemoradiotherapy and could be targeted in future studies.