A multicenter randomized feasibility study of paclitaxel plus carboplatin vs. S-1 in patients with locally advanced completely resected NSCLC was conducted. A total of 40 patients underwent complete resection and were diagnosed with pathological stage II or IIIA NSCLC according to the 7th edition of the Tumor-Node-Metastasis classification (25) at the Nagoya City University Hospital (Nagoya, Japan) and its affiliated hospitals, between January 2008 and December 2013. Written informed consent was obtained from all the patients, and the study protocol was approved by the Institutional Review Board of each participating institution. This study was registered on the UMIN Clinical Trial database (ID:000001510).

The eligibility criteria were as follows: Histologically confirmed NSCLC, completely resected, pathological stage II or IIIA disease, no previous chemotherapy or radiotherapy, age 20–75 years, Eastern Cooperative Oncology Group performance status of 0 or 1, white blood cell count 3,500–12,000/mm³ (normal, 3,000–8,500/mm³), absolute neutrophil count ≥2,000/mm³ (normal, 38.3–74.7%), platelet count ≥100,000/mm³ (normal, 150,000–361,000/mm³), hemoglobin level ≥10 g/dl (normal, 10.8–14.9 g/dl), aspartate aminotransferase and creatinine level <upper limit of normal (ULN), creatinine clearance rate >60 ml/min, percutaneous oxygen saturation concentration by room air ≥95%, and aspartate aminotransferase, alanine aminotransferase and total bilirubin levels <2 times the ULN; the patients also had to have started chemotherapy within 8 weeks following surgery and been able to receive oral intake. The exclusion criteria were patients with previous chemotherapy or radiotherapy, concomitant malignancy within 5 years, interstitial pneumonia with clinical symptoms, and significant cardiac arrhythmia or heart failure.

The randomization was performed centrally at the Department of Oncology, Immunology and Surgery of the Nagoya City University Graduate School of Medical Sciences (Nagoya, Japan). The patients were randomly assigned either to arm A (18 cases) receiving paclitaxel plus carboplatin bi-weekly, or to arm B (19 cases) receiving S-1. The treatments performed in the present study are schematically summarized in Fig. 1. Randomized allocation factors included facility, age, histological type and stage.

The infusing dosage of paclitaxel was 120 mg/m² on days 1 and 15. Carboplatin at an area under the curve (AUC) dose of 3 was also administered on days 1 and 15. The patients received adjuvant chemotherapy with paclitaxel plus carboplatin every 4 weeks for up to 4 cycles. The Calvert's formula was used to calculate the dose of the AUC for carboplatin (26), while the creatinine clearance was determined with the Jelliffe formula (27).

The dosage of S-1 was established as follows: Patients with a body surface area (BSA) <1.25 m² received 40 mg twice daily (80 mg/day); those with BSA ≥1.25 m² but <1.5 m² received 50 mg twice daily (100 mg/day); and those with a BSA ≥1.5 m² received 60 mg twice daily (120 mg/day). S-1 was administered for 2 weeks followed by a 1-week rest period for up to 1 year. Both arms A and B continued on the above prescription unless there was any evidence of relapse, other malignancies, or severe adverse events.

Throughout the study, the dosage of paclitaxel plus carboplatin was adjusted according to the presence and severity of hematological and non-hematological toxicities. For patients exhibiting evidence of hematological or non-hematological toxicity, the treatment on day 15 was omitted and the dosage for the next course was reduced by one level (from paclitaxel 90 mg m² and carboplatin AUC 2, to paclitaxel 60 mg m² and carboplatin AUC 2). The dosage of S-1 was also planned to be reduced by 1 level (15 mg/m²) up to 2 times for patients exhibiting evidence of grade ≥3 hematological or non-hematological toxicities. All dose reductions were limited to two levels.

Recurrence was diagnosed on the basis of imaging study findings. Chest and abdominal computed tomography and positron emission tomography plus head magnetic resonance imaging were performed at 6- and 12-month intervals, respectively. In addition, when the patients complained of any symptoms or exhibited elevated tumor markers on blood tests, imaging studies were performed.

All the eligible patients who had received any definitive treatment were considered as assessable for feasibility and toxicity. The feasibility was evaluated based on the rate of treatment completion (4 cycles completed for carboplatin plus paclitaxel, and 1 year completed for S-1) and safety (rate of grade ≥3 toxicities). Adverse events were graded according to the National Cancer Institute Common Toxicity Criteria (Common Terminology Criteria for Adverse Events) version 3.0 (28).

The sample size was determined based on a phase II study reported by Kawamura et al applying docetaxel plus gemcitabine as an adjuvant chemotherapy in 35 patients (29). This previous study reported a 2-year DFS rate of ~52%, with a 95% confidence interval (CI) of 35–69%. Based on this result, the expected and threshold values of the 2-year DFS were 40 and 65%, respectively. The number of patients required was determined with an α risk of 0.05 and a β risk of 0.1. The number of patients in each arm was calculated using the Fleming method and found to be 32 per arm. Sufficient data for patients in the present study could not be gathered within the study period. The primary endpoint was the 2-year DFS rate, and the secondary endpoints were the feasibility and toxicity. The characteristics, feasibility, adverse events, DFS and overall survival of 37 patients were analyzed. The cumulative total administration cycles, 2 and 5-year DFS and overall survival (OS) were examined by the Kaplan-Meier method and the difference between the two arms was calculated by the log-rank test. The differences in the rate of adverse events were evaluated by the χ² test. All the data were analyzed with EZR software (30). P≤0.05 was considered to indicate a statistically significant difference.

A total of 40 patients with stage II or IIIA NSCLC who had received surgically complete resection were enrolled. Of the 40 patients, 3 were excluded in accordance with the exclusion criteria: 1 patient refused to continue participating in this study after registering, 1 patient used another chemotherapy regimen during the follow-up period (under recurrence-free conditions) and 1 patient had a history of multiple cancers. The remaining 37 patients were randomized to either arm (18 cases in arm A and 19 in arm B). The patient characteristics are summarized in Table I. Briefly, the patients included 6 women and 31 men, with a mean age of 62.8 years (range, 39–75 years). In arm A, 2 patients with pneumonectomy (2/37, 5%) were included. These 2 patients completed four cycles of carboplatin plus paclitaxel, but were confirmed to have recurrence within 1 year following surgery. When interpreting the results of the present study, this point should be kept in mind.

In total, 50% of the patients in arm A received paclitaxel plus carboplatin and 52.6% of the patients in arm B received S-1, along with the planned schedule and at the planned dose (Table II). In arm B, 1 of the 2 patients refused to continue the treatment due to financial difficulties, and the other interrupted the treatment due to continued grade 1 anorexia.

The drug-related adverse events are listed in Table III. The main adverse events in arm A were anaphylaxis, hematological toxicity, neuropathy and alopecia. Two patients developed a grade 4 allergic reaction (anaphylactic shock); however, immediately after cessation of the infusion of paclitaxel, and following treatment with steroid therapy, the patients recovered without sequelae. Both patients discontinued adjuvant chemotherapy with paclitaxel plus carboplatin: 1 patient went on to receive 4 cycles of gemcitabine plus carboplatin as adjuvant therapy, and the other patient received no further adjuvant chemotherapy. There were no grade 3 or 4 adverse events in arm B. Adverse events occurred in 15 patients (83.3%) in arm A, and in 11 patients (57.9%) in arm B; the difference was non-significant (P=0.151). In total, 3 patients (16.7%) in arm A and 6 patients (31.6%) in arm B discontinued drug administration due to adverse events caused by the agent and the patients' wishes. In addition, 6 patients (33.3%) in arm A and 3 (15.8%) in arm B required a dose reduction due to adverse events. No treatment-associated deaths occurred.

Ultimately, 9 patients (50%) in arm A and 8 (42.1%) in arm B relapsed. In arm A, the relapse sites were the brain (4 cases), the brain and mediastinal lymph nodes (1 case), the bone (1 case), the adrenal glands (1 case), the mediastinum (1 case), and the mediastinum and supraclavicular lymph nodes (1 case). In arm B, the relapse sites were the brain and mediastinal lymph nodes (1 case), the trachea (1 case), the supraclavicular lymph nodes (1 case), the mediastinal lymph nodes (2 cases), and the intrathoracic cavity (3 cases). Two patients were excluded from the analysis of DFS and OS as they were unable to continue paclitaxel plus carboplatin chemotherapy due to an anaphylactic reaction following infusion of paclitaxel. The median follow-up time was 47 months (range, 13–79 months). The 2-year DFS rates were 54.2% (95% CI: 27.1–75.0%) in arm A and 84.2% (95% CI: 58.7–94.6%) in arm B (Fig. 2A). No statistically significant difference in the 2-year DFS was noted between the two arms, although there was a weak tendency toward an improved rate in arm B. For further detailed breakdowns of the effects on survival, the 5-year DFS (Fig. 2A) and OS (Fig. 2B) was also investigated.