Carlos Jordá Aragón1, Juan Carlos Peñalver Cuesta2, Nuria Mancheño Franch3, Karol de Aguiar Quevedo2, Francisco Vera Sempere3, José Padilla Alarcón4. 1. Servicio de Cirugía Torácica, Hospital Universitario y Politécnico La Fe, Valencia, España. 2. Servicio de Cirugía Torácica, Instituto Valenciano de Oncología, Valencia, España. 3. Servicio de Anatomía Patológica, Hospital Universitario y Politécnico La Fe, Departamento de Patología, Universidad de Valencia, Valencia, España. 4. Servicio de Cirugía Torácica, Instituto Valenciano de Oncología, Valencia, España. Electronic address: jpadilla@comv.es.
Abstract
BACKGROUND AND OBJECTIVE: Survival studies of non-small cell lung cancer (NSCLC) are usually based on the Kaplan-Meier method. However, other factors not covered by this method may modify the observation of the event of interest. There are models of cumulative incidence (CI), that take into account these competing risks, enabling more accurate survival estimates and evaluation of the risk of death from other causes. We aimed to evaluate these models in resected early-stage NSCLC patients. PATIENTS AND METHOD: This study included 263 patients with resected NSCLC whose diameter was ≤ 3 cm without node involvement (N0). Demographic, clinical, morphopathological and surgical variables, TNM classification and long-term evolution were analysed. To analyse CI, death by another cause was considered to be competitive event. For the univariate analysis, Gray's method was used, while Fine and Gray's method was employed for the multivariate analysis. RESULTS: Mortality by NSCLC was 19.4% at 5 years and 14.3% by another cause. Both curves crossed at 6.3 years, and probability of death by another cause became greater from this point. In multivariate analysis, cancer mortality was conditioned by visceral pleural invasion (VPI) (P=.001) and vascular invasion (P=.020), with age>50 years (P=.034), smoking (P=.009) and the Charlson index ≥ 2 (P=.000) being by no cancer. CONCLUSIONS: By the method of CI, VPI and vascular invasion conditioned cancer death in NSCLC >3 cm, while non-tumor causes of long-term death were determined.
BACKGROUND AND OBJECTIVE: Survival studies of non-small cell lung cancer (NSCLC) are usually based on the Kaplan-Meier method. However, other factors not covered by this method may modify the observation of the event of interest. There are models of cumulative incidence (CI), that take into account these competing risks, enabling more accurate survival estimates and evaluation of the risk of death from other causes. We aimed to evaluate these models in resected early-stage NSCLCpatients. PATIENTS AND METHOD: This study included 263 patients with resected NSCLC whose diameter was ≤ 3 cm without node involvement (N0). Demographic, clinical, morphopathological and surgical variables, TNM classification and long-term evolution were analysed. To analyse CI, death by another cause was considered to be competitive event. For the univariate analysis, Gray's method was used, while Fine and Gray's method was employed for the multivariate analysis. RESULTS: Mortality by NSCLC was 19.4% at 5 years and 14.3% by another cause. Both curves crossed at 6.3 years, and probability of death by another cause became greater from this point. In multivariate analysis, cancer mortality was conditioned by visceral pleural invasion (VPI) (P=.001) and vascular invasion (P=.020), with age>50 years (P=.034), smoking (P=.009) and the Charlson index ≥ 2 (P=.000) being by no cancer. CONCLUSIONS: By the method of CI, VPI and vascular invasion conditioned cancer death in NSCLC >3 cm, while non-tumor causes of long-term death were determined.