This study recruited advanced or metastatic NSCLC patients with histologic or cytologic-confirmed in Korea. Patients, who had progressive or recurrent disease with 19 to 80 years old following first-line chemotherapy, and who had either a positive EGFR mutation result or met 2 or more of the 3 characteristics: adenocarcinoma, female, and non-smoker, were participated in this study. Patients were also required to have a measurable lesion according to the Response Evaluation Criteria in Solid Tumors (RECIST) criteria, a World Health Organization (WHO) performance status 0~1, and a life expectancy of at least 12 weeks.

Exclusion criteria included patients with any evidence of clinically active interstitial lung disease, malignancies, severe/uncontrolled systemic disease, treatment with a non-approved or investigational drug within 30 days prior to the study treatment, and those who have history of EGFR inhibitor treatment.

Before the start of treatment, all patients underwent the screening procedure, which included a physical examination, medical history, vital signs, WHO performance status, laboratory parameters, tumor assessment (RECIST), Quality of Life (QoL) assessment (by Functional Assessment of Cancer Therapy-Lung [FACT-L] Korean version 4), and EGFR mutation analysis. Determination of EGFR mutation was performed at a central laboratory (Isu Abxis Co., Ltd., 6th Yonsei University Medical Center, Seoul, Korea) or at 2 local laboratories (Konkuk University Hospital and Chonnam National University Hwasun Hospital). DNA was extracted from section of paraffin-embedded tumor (primary or recurred) tissue block, then was analyzed by using standard polymerase chain reaction-techniques and direct sequencing methods on exon 18, 19, and 21 of EGFR gene.

Patients visited every 4 weeks until the confirmation of disease progression and were performed the following assessments/procedures in every visit: physical examination, vital signs, WHO performance status, laboratory test, adverse events (AEs), concomitant medication, chest X-ray, and drug compliance. During the treatment period, tumor assessments by computed tomography scan and QoL by FACT-L were evaluated every 8 weeks. Once disease progression was confirmed, survival rate was follow-up every 12 weeks.

Gefitinib 250 mg (Iressa; AsatraZeneca, London, England) was administered orally once daily until the investigators gave confirmation of disease progression or of unacceptable toxicity or until patient refuse to participate in the study.

Response to treatment was assessed according to the RECIST criteria¹³. According to the RECIST criteria, ORR was defined as the proportion of subjects indicating partial response (PR) or complete response (CR) among the best responses assessed by the investigators. PFS was defined as the interval of the date of starting treatment and the objective disease progression or the absence of objective disease progression or death due to any cause. In addition, OS was defined as the interval of the date of starting treatment and the patient death due to any cause. Toxicity was classified in accordance with guideline of National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) version 3.0.

The planned sample size was 219, the evaluable patients, considering drop-out rate of 10% wherein a response rate (P0) of 21.5% and (P1) of 30%. The establishment was based on Expanded Access Program with α of 0.05 (two-sided) and power of 0.8. Subjects who received at least one dose of gefitinib and had at least one tumor assessment were included in the modified intent-to-treat (mITT) analysis data set and the primary analysis was performed on the mITT set. The primary efficacy endpoint, ORR, was defined as the proportion of subjects showing PR or CR as the best response until final tumor assessment. Sub-analysis was done in subgroups that were defined according to EGFR mutation status, gender, smoking history, and tumor histology. All statistical analyses of ORR were performed using the chi-square test. Median survival time and 95% confidence interval (CI) of PFS or OS was calculated by using the Kaplan-Meier method and survival curve was presented and comparisons of PFS or OS were performed using the log-rank test. The improvement rate for QOL and for symptom improvement was defined as the change of point ≥6 in the summed FACT-L score from baseline and the change of point ≥2 in the summed lung cancer subscale (LCS) of FACT-L from baseline, respectively. Safety summaries were based on the safety set, which included all subjects who had received at least one dose of gefitinib and had at least one safety follow-up. Safety data included any AEs that occurred after the administration of study drug and any abnormalities in laboratory values at each visit. AE coding complied with MedDRA (version 12.0).

From January 2007 to July 2008, 156 subjects were enrolled from 11 centers in Korea and the follow-up was tracked until January 2009.

Among 156 subjects, 154 subjects were included in the evaluable-for-safety population (2 patients did not receive study treatment) and 138 subjects were included in mITT population (8 subjects had no target lesion at baseline; 8 subjects had no tumor assessment after baseline, and hence could not provide RECIST data). Patients' demographics and characteristics at baseline for the mITT population are summarized in Table 1. Breifly, patients had a median age of 64 years old (52.2% were <65 years old) and as anticipated by the entry criteria, the majority were female (87.0%), had adenocarcinoma histology (87.7%) and were non-smokers (84.8%). EGFR mutation was determined to be 109 patients. Among these patients, 59 were found to have EGFR mutation-positive, 50 were found to have EGFR mutation-negative.

The patients were administered gefitinib for approximately 5.8 months and the maximum treatment period was 18.7 months.

Of the 138 patients (mITT population), one subject had a best response of CR and 42 subjects had best responses of PR, resulting in an ORR of 31.2% (95% CI 23.6~39.6) (Table 2). In addition, disease control rate was observed 69.6% (96/138).

The degree of tumor shrinkage per individual patient as well as the number of patients who had tumor shrinkage was presented in Figure 1. This figure displayed the maximum percentage of change in the sum of the longest dimensions from baseline, and included patients who had at least one post-baseline tumor assessment. From this, changes in tumor size did not require confirmation as would necessary be by RECIST for objective response. Subgroup analysis found that ORR showed higher significant in patient groups whose tumors were EGFR mutation-positive compared with EGFR mutation-negative (45.8% vs. 14.0%, p=0.0004). There were no statistically significant differences in ORR between male and female, non-smokers and smokers, and adenocarcinoma and other histological subtypes (Table 3).

In the mITT population, progression of disease or death occurred in 100 cases with the median PFS was 5.7 months (95% CI, 3.9~8.4 months) (Figure 2). As death occurred in only 38 cases, median OS could not be derived (Figure 3). The 6-month PFS and OS were 49.6% and 87.9%, respectively. EGFR mutation-positive status showed a strong relationship with PFS (Figure 4) and OS (Figure 5) among subjects with EGFR mutation compared with others predictors. The 12-month PFS and OS were 23.3% and 69.4%, respectively.

In overall FACT-L score, the rate of improvement with the change of points ≥6 was 43.6% (41/94 patients) (Table 4). The symptom improvement was evaluated by the LCS of FACT-L. Patients who had a change at any time of at least 2 points difference from baseline were counted as having symptom improvement. The symptom improvement rate was 51.6%.

Among 154 evaluable-for-safety patients, 141 (91.6%) reported to have AEs including 105 (68.2%) with AEs considered to be treatment-related. Serious adverse events (SAEs) were reported by 25 subjects (16.2%) and were considered treatment-related in 5 subjects (3.2%). Discontinuation due to AEs occurred in 15 subjects (9.7%) and 5 of those AEs were considered treatment-related (3.2%). SAEs leading to death occurred in 11 subjects (7.1%), 2 were considered treatment-related (1.3%; both due to interstitial lung disease). Grade 3 or 4 AEs were reported in 26 subjects (16.9%) of which 9 subjects (5.8%) were considered treatment-related (Table 5).

The most frequently reported AEs related to the treatment were rash (33.1%), diarrhea (23.4%), and pruritus (15.6%), which were mostly Common Terminology Criteria grade 1 or 2.