Stereotactic body radiotherapy versus lobectomy for operable clinical stage IA lung adenocarcinoma: comparison of survival outcomes in two clinical trials with propensity score analysis (JCOG1313-A).

Objective: No randomized controlled trials comparing stereotactic body radiotherapy and lobectomy for operable early-stage non-small-cell lung cancer have been successfully conducted. This study compared survival outcomes in two multi-institutional clinical trials for stereotactic body radiotherapy (Japan Clinical Oncology Group JCOG0403) and lobectomy (Japan Clinical Oncology Group JCOG0201) with propensity score analysis.
Methods: Inclusion criteria were operable, cT1N0M0 and adenocarcinoma diagnosed prior to registration of each trial. Forty of 169 patients from JCOG0403 and 219 of 811 patients from JCOG0201 were included. The primary endpoint was overall survival adjusted with propensity score analysis. The patient selection factors included in the logistic model to estimate the propensity score were age, sex, tumor diameter and consolidation/tumor ratio.
Results: Among patient selection factors, age distribution was quite different with little overlap: the median was 79 (interquartile range: 74.5-83.5) in stereotactic body radiotherapy and 62 (interquartile range: 55-68) in lobectomy. In propensity score analysis, 21 patients from each group were matched and the hazard ratio for stereotactic body radiotherapy over lobectomy was 9.00 (95% confidence interval: 1.14-71.04). In the post hoc subgroup analysis with propensity score analysis of inverse probability of treatment weighting, patients were limited to be aged 75 or younger because JCOG0201 only included them when aged 75 or younger. Thirteen patients for stereotactic body radiotherapy and 219 for lobectomy were compared, and the hazard ratio for stereotactic body radiotherapy over lobectomy was 1.19 (95% confidence interval: 0.38-3.73).
Conclusions: The point estimates of hazard ratio favored lobectomy over stereotactic body radiotherapy in the limited number of patients. A randomized controlled study is needed for valid comparison.
(© The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.)