Pronóstico individualizado en el carcinoma de pulmón no microcítico. Una aproximación multivariada.
Por:
Soler JJ, Perpiñá M, Padilla J, Greses JV, Sales G and Paris F
Publicada:
1 ago 1995
Resumen:
With the aim of providing a system to give individualized initial prognosis for patients with non-small cell lung cancer (NSLC), we performed a multivariate discriminant analysis with combinations of various prognostic factors, studying 93 patients with diagnoses of stage II NSLC and complete followup information. All had undergone surgery at our hospital. Survival longer than or less than one year was defined as the dependent variable: independent variables more clinical, analytical, lung function, histological, anatomical and surgical data. Among the 31.2% of patients who died within one year, tumor size was noticeably greater (6.0 +/- 2.1 cm versus 4.8 +/- 2.0 cm; p = 0.11) and serum albumin was lower (3.7 +/- 0.7 g/dl versus 3.9 +/- 0.6 g/dl; p = 0.039); we found no significant differences among the remaining independent variables in the preliminary univariate analysis. The linear function obtained with discriminant analysis allowed us to classify the patients correctly in 87.1% of cases, with a diagnostic sensitivity of 79.3%, specificity of 90.6%, positive predictive value of 79.3% and negative predictive value of 90.6% for patients who died during the first year. The independent variables that were associated with poor prognosis were large tumor size, long duration of symptoms, low albumin level, high alkaline phosphate level, presence of 2 or more N1 affected nodes, pneumonectomy and presence of perioperative complications. Our findings allow us to conclude that the simultaneous analysis of a variety of prognostic factors can help to give an accurate prognosis for individual patients with the same anatomical stage classification.
Filiaciones:
:
Servicio de Neumología, Hospital Universitario La Fe, Valencia
|