P1.16-25 A Propensity Score Model for Appropriate Treatment Selection (Sublobar Resection vs. SBRT) In Patients With cStage I NSCLC

Abstract

Stereotactic body radiotherapy (SBRT) and sublobar resection (SLR) are current treatment options for patients with cStage I non-small-cell lung cancer (NSCLC) who are operable, but at high risk for lobectomy. However, optimal selection of the two treatments remains controversial. Purposes of this study are to identify pre-treatment factors affecting treatment decision and to evaluate their impact on outcomes. We retrospectively reviewed patients who underwent SBRT or SLR for cStage I NSCLC because of medical comorbidity between 2003 and 2009. Patients who were with performance status of 2 or worse, those who had no data on pre-treatment pulmonary function test, or those who had no histological confirmation of NSCLC were excluded. A propensity score (PS) model of treatment decision (0 toward SBRT, and 1 toward SLR) was generated using stepwise logistic regression incorporating pre-treatment factors. Ninety-two and 65 patients who underwent SBRT and SLR, respectively, were enrolled into this analysis. Median potential follow-up period was 8.6 years. The following factors remained in the PS model after stepwise selection: age, sex, Charlson comorbidity index (CCI), body mass index (BMI), forced expiratory volume in 1 second (FEV1) and tumor diameter. Old age, male, CCI of 1 or more, underweight BMI and large tumor had coefficients toward SBRT (Table). In a cohort with PS of 0.5 or more, overall survival was significantly better in SLR patients than in SBRT patients (79.5% and 47.5% at 5 years, respectively; P = 0.004). In a cohort with PS<0.5, whereas, overall survival was similar between SLR and SBRT (43.3% and 39.7% at 5 years, respectively; P = 0.805). The PS model would help appropriate treatment selection for high-risk operable patients. Although patients with PS of 0.5 or more benefit from SLR, SBRT provides comparable outcomes for patients with PS<0.5.