Potential bias in the clinical trial registration of patients with peripheral small-sized non-small cell lung cancer: Real-world data in the era of segmentectomy

Recent phase III trials demonstrate that survival after segmentectomy is not inferior to that after lobectomy for peripheral small-sized non-small cell lung cancer (NSCLC). However, the influence of patient selection and registration bias remains unclear. This study analyzed real-world data to assess patients who were eligible for the JCOG0802/WJOG4607L trial, but who were not enrolled. We analyzed data on 38 of 61 patients not enrolled in the JCOG0802/WJOG4607L trial, during its enrollment period. Surgical procedures, clinical characteristics, and outcomes of the nonregistered patients were reviewed. Lobectomy was performed in 31 patients and segmentectomy was performed in 7 patients. Radiological features of higher-grade malignancy were seen more frequently in the patients who underwent lobectomy. The most common reason for nonregistration in the JCOG0802/WJOG4607L trial was multiple mild comorbidities (n = 11), followed by tumor aggressiveness (n = 6), and concerns about inadequate margins with segmentectomy (n = 6). Recurrence was found in five patients. The 5-year overall and recurrence-free survival rates were 82.9% and 80.2%, respectively. Our analysis emphasizes the challenges posed by registration bias in clinical trials. Survival outcomes among the study population were worse than those reported in the JCOG0802/WJOG4607L trial, suggesting a high malignant potential on preoperative imaging, and that registration bias occurs even in large-scale randomized trials.