European Journal of Epidemiology

, Volume 20, Issue 11, pp 907–914 | Cite as

Heterogeneity of Prognostic Profiles in Non-small Cell Lung Cancer: Too Many Variables but a Few Relevant

  • Agustín Gomez de la Cámara
  • Angel López-Encuentra
  • Paloma Ferrando
  • Bronchogenic Carcinoma Cooperative Group of the Spanish Society of Pneumolody and Thoracic Surgery (GCCB-S)*,**


Objective:Many prognostic factors, exceeding 150, for non-small cell lung cancer (NSCLC) are mentioned in the literature. The different statistical weight of the some variables at issue, their heterogeneity and their clinical uselessness is reviewed. Study design and setting: Survival analysis of a cohort of NSCLC operated (n = 1730, 1993–1997) was carried out utilizing different statistical approaches: Cox proportional hazard analysis (CPHA), logistic regression (LRA), and recursive partitioning (CART). Results:CPHA identified 13 prognostic variables and 11 LRA. Of the 17 possible variables, 10 are coincident. CART provided five different diagnostic groups but only three differentiated survival levels. Parsimonious models were constructed including only T and N cancer staging variables. Areas under the ROC curve of 0.68 and 0.68 were found for CPHA and LGA parsimonious models respectively, and 0.72 and 0.71 for complete models. Conclusion: Variables with a minimal impact on the respective models and thus with little or scarce predictive clinical repercussion were identified. Differences in the prognostic profile of survival can be caused by the different methodological approaches used. No relevant differences were found between the parsimonious and complete models. Although the amount of information managed is considerable, there continues to be a large predictive gap yet to be explained.


Follow up studies Lung cancer Non-small cell lung cancer Prognostic research Risk factor Survival 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Boyle, P, d’Onofrio, A, Maisonneuve, P,  et al. 2003Measuring progress against cancer in Europe: Has the 15% decline targeted for 2000 come about?Ann Oncol1413121325PubMedGoogle Scholar
  2. 2.
    Skuladottir, H, Olsen, JH 2001Epidemiology of lung cancerEur Respir Mon17112Google Scholar
  3. 3.
    Sobin, LH, Wittekind, Ch 1997TNM Classification of Malignant Tumors. UICC International Union Against Cancer5Wiley-LissNew York, NYGoogle Scholar
  4. 4.
    López-Encuentra, A, Bülzebruck, H, Feinstein, AR,  et al. 2000Tumor staging and classificatión in lung cancerLung Cancer297983PubMedGoogle Scholar
  5. 5.
    López-Encuentra A, Duque-Medina JL, Rami-Porta R, Gómez de la Cámara A, Ferrando P for the Bronchogenic Carcinoma Co-operative Group of the Spanish Society of Pneumology and Thoracic Surgery. Staging in lung cancer: Is 3 cm a prognostic threshold in pathologic stage I non-small cell lung cancer? A multicenter Study of 1,020 patients. Chest 2002; 121: 1515–1520Google Scholar
  6. 6.
    Gospodarowicz, MK, Henson, DE, Hutter, RVP,  et al. 2001Prognostic Factors in Cancer2Wiley-LissNew York, NYGoogle Scholar
  7. 7.
    Brundage, MD, Davies, D, Mackillop, WJ 2002Prognostic factors in non-small cell lung cancer. A decade of progressChest12210371057CrossRefPubMedGoogle Scholar
  8. 8.
    Grupo Cooperativo de Carcinoma Broncogénico de SEPAR (GCCB-S)1995Cirugía del carcinoma broncogénico en España. Estudio descriptivoArch Bronconeumol31303309Google Scholar
  9. 9.
    López Encuentra A, the Bronchogenic Carcinoma Cooperative Group of the Spanish Society of Pneumology and Thoracic Surgery (GCCB-S), 1998; Criteria of functional and oncological operability in surgery for lung cancer. A multicenter study. Lung Cancer 1995; 20: 161–168Google Scholar
  10. 10.
    López Encuentra A, Gómez de la Cámara A, for the Bronchogenic Carcinoma Cooperative Group of the Spanish Society of Pneumology, Thoracic Surger (GCCB-S). Validation of a central review board of staging prior to surgery for non-small-cell lung cancer – impact on prognosis: a multicenter study. Respiration 2002; 69: 16–24Google Scholar
  11. 11.
    López Encuentra A, Gómez de la Cámara A, Varela de Ugarte A, Mañes N, Llobregat N y GCCB-S. El “fenómerno Will-Rogers”. Migración de estadios en carcinoma broncogénico, tras aplicar criterios de certeza clasificatoria. Arch Bronconeumol 2002; 38: 166–171Google Scholar
  12. 12.
    American Thoracic Society1983Clinical staging primary lung cancerAm Rev Respir Dis12716Google Scholar
  13. 13.
    Izbicki, JR, Passlick, B, Karg, O,  et al. 1995Impact of radical systematic mediastinal lymphadenectomy on tumor staging in lung cancerAnn Thorac Surg59209214CrossRefPubMedGoogle Scholar
  14. 14.
    Grupo Cooperativo de Carcinoma Broncogénico de SEPAR (GCCB-S)1996Control de calidad en un registro multiinstitucional de carcinoma broncogénicoArch Bronconeumol3270Google Scholar
  15. 15.
    Feinstein, AR, Wells, CK, Walter, SD 1990A Comparison of multivariable mathematical methods for predicting survival-I. Introduction, rationale, and general strategyJ Clin Epidemiol43339347PubMedGoogle Scholar
  16. 16.
    Wasson, J, Sox, HC, Neff, RK, Goldman, L 1985Clinical prediction rules. Applications and methodological standardsN Engl J Med313793799PubMedCrossRefGoogle Scholar
  17. 17.
    Cook, F, Goldman, L 1984Empiric comparison of multivariate analytic techniques: Advantages and disadvantages of recursive partitioning analysisJ Chron Dis37721731CrossRefPubMedGoogle Scholar
  18. 18.
    Walter, SD, Feinstein, AR, Wells, CK 1990A comparison of multivariable mathematical methods for predicting survival-II. Statistical selection of prognostic variablesJ Clin Epidemiol43349359PubMedGoogle Scholar
  19. 19.
    Mitchell, H, Katz,  2003Multivariable análisis: A primer for readers of medical researchAnn Intern Med138644650Google Scholar
  20. 20.
    Long, WJ, Griffith Selker, JL HP, D’Agostino, RB 1993A comparison of logistic regression to decision-tree induction in a medical domainComput. Biomed Res.267497CrossRefPubMedGoogle Scholar
  21. 21.
    Stokes, ME, Davis, ChS, Koch, GG 1995Categorical Data Analysis Using the SAS® SystemSAS institute IncCary, NC499Google Scholar
  22. 22.
    SAS Institute Inc.,1996SAS/STAT® Sofware: Changes and Enhancements Through Release 6.11SAS Institute incCary, NC1104 Google Scholar
  23. 23.
    Laupacis, A, Wells, G, Richardson, WS, Tugwell, P 1994Users’ guides to the medical literature. V. How to use an article about prognosis. Evidence-Based Medicine Working GroupJAMA272234237CrossRefPubMedGoogle Scholar
  24. 24.
    Mountain. Regional lymph node classification for lung cancer staging. Chest 1997; 111: 1718–1723.Google Scholar

Copyright information

© Springer 2005

Authors and Affiliations

  • Agustín Gomez de la Cámara
    • 1
  • Angel López-Encuentra
    • 1
  • Paloma Ferrando
    • 1
  • Bronchogenic Carcinoma Cooperative Group of the Spanish Society of Pneumolody and Thoracic Surgery (GCCB-S)*,**
  1. 1.Unidad de Investigación-Epidemiologia Clínica MadridSpain

Personalised recommendations