Lobectomy is the standard of care for early-stage non-small-cell lung cancer (NSCLC). The survival and clinical benefits of segmentectomy have not been investigated in a randomised trial setting. We aimed to investigate if segmentectomy was non-inferior to lobectomy in patients with small-sized peripheral NSCLC.

We conducted this randomised, controlled, non-inferiority trial at 70 institutions in Japan. Patients with clinical stage IA NSCLC (tumour diameter ≤2 cm; consolidation-to-tumour ratio >0·5) were randomly assigned 1:1 to receive either lobectomy or segmentectomy. Randomisation was done via the minimisation method, with balancing for the institution, histological type, sex, age, and thin-section CT findings. Treatment allocation was not concealed from investigators and patients. The primary endpoint was overall survival for all randomly assigned patients. The secondary endpoints were postoperative respiratory function (6 months and 12 months), relapse-free survival, proportion of local relapse, adverse events, proportion of segmentectomy completion, duration of hospital stay, duration of chest tube placement, duration of surgery, amount of blood loss, and the number of automatic surgical staples used. Overall survival was analysed on an intention-to-treat basis with a non-inferiority margin of 1·54 for the upper limit of the 95% CI of the hazard ratio (HR) and estimated using a stratified Cox regression model. This study is registered with UMIN Clinical Trials Registry, UMIN000002317.

Between Aug, 10, 2009, and Oct 21, 2014, 1106 patients (intention-to-treat population) were enrolled to receive lobectomy (n=554) or segmentectomy (n=552). Patient baseline clinicopathological factors were well balanced between the groups. In the segmentectomy group, 22 patients were switched to lobectomies and one patient received wide wedge resection. At a median follow-up of 7·3 years (range 0·0-10·9), the 5-year overall survival was 94·3% (92·1-96·0) for segmentectomy and 91·1% for lobectomy (95% CI 88·4-93·2); superiority and non-inferiority in overall survival were confirmed using a stratified Cox regression model (HR 0·663; 95% CI 0·474-0·927; one-sided p<0·0001 for non-inferiority; p=0·0082 for superiority). Improved overall survival was observed consistently across all predefined subgroups in the segmentectomy group. At 1 year follow-up, the significant difference in the reduction of median forced expiratory volume in 1 sec between the two groups was 3·5% (p<0·0001), which did not reach the predefined threshold for clinical significance of 10%. The 5-year relapse-free survival was 88·0% (95% CI 85·0-90·4) for segmentectomy and 87·9% (84·8-90·3) for lobectomy (HR 0·998; 95% CI 0·753-1·323; p=0·9889). The proportions of patients with local relapse were 10·5% for segmentectomy and 5·4% for lobectomy (p=0·0018). 52 (63%) of 83 patients and 27 (47%) of 58 patients died of other diseases after lobectomy and segmentectomy, respectively. No 30-day or 90-day mortality was observed. One or more postoperative complications of grade 2 or worse occurred at similar frequencies in both groups (142 [26%] patients who received lobectomy, 148 [27%] who received segmentectomy).

To our knowledge, this study was the first phase 3 trial to show the benefits of segmentectomy versus lobectomy in overall survival of patients with small-peripheral NSCLC. The findings suggest that segmentectomy should be the standard surgical procedure for this population of patients.

National Cancer Center Research and the Ministry of Health, Labour, and Welfare of Japan.