Nomogram for Predicting the Benefit of Surgery for Stage IA-IIB Small Cell Lung Cancer.

Y. Wang,Z. Wang,Q. Zheng
DOI: https://doi.org/10.1093/annonc/mdy298.016
IF: 51.769
2018-01-01
Annals of Oncology
Abstract:Background: The role of surgical resection remains controversial in small cell lung cancer (SCLC) although there are some retrospective and population-based studies indicate that patients with very early stage SCLC has longer survival compared with those not given surgery. The specific aim of this study was to identify the survival benefit of surgery for patients with Stage IA-IIB SCLC and nomogram predictive model was created to select patients who are eligible to surgery. Methods: Patients diagnosed with stage IA-IIB SCLC between 2004 and 2014 were selected from the SEER database. The primary endpoint was overall survival. Multivariate Cox proportional regression and coefficients of the predictors were calculated. A nomogram was constructed for predicting 1- and 3-year overall survival probability. All statistical analysis was performed with R software. Results: 2246 patients with stage I-II were enrolled. 618 (27.5%) received surgery and 1628 (72.5) not. Unadjusted median overall survival (OS) was 23 months (95% CI: 21-24), which was 35 months (95% CI: 31-44) vs. 19 months (95% CI: 18-21) in surgery and non-surgery groups respectively (p <0.0001). We used a propensity score to balance observed covariates. OS benefit was observed in all subgroups between the surgery and non-surgery group except in the non-White race, well or moderately grading, stage IIA or IIB and N1 lymph involvement. Multivariate Cox proportional hazards regression analysis showed a survival benefit in the surgery group compared with non-surgery group no matter balanced by propensity score weighting or not. The competing-risks nomogram was built for predicting 1-year and 3-year survival. The age, tumor size, extent of tumor, N0/1 and surgery with radiation and chemotherapy were introduced as variables. The calibration of internal validation for predicting survival at 1 and 3-year by this nomogram-predicted probability was identical to the actual probability. Conclusions: Surgery was proved to benefit patients with stage IA-IIB SCLC by this relatively large number population-based study and a nomogram built from a parametric survival model from the SEER database can be used to predict which patients with stage IA-IIB SCLC may benefit from surgery. Legal entity responsible for the study: Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department 1 of Thoracic Oncology Medicine, Peking University Cancer Hospital & Institute. Funding: National Natural Science Foundation of China, Award ( 81401914) and Beijing Municipal Administration of Hospitals' Youth Programme (20161112). Disclosure: All authors have declared no conflicts of interest.
What problem does this paper attempt to address?