Prognostic factors (PF) have a pivotal role in Clinical Oncology. They are helpful in the selection of treatment, provide insights into the disease process and the therapeutic response, and are fundamental in the design of clinical trials or in the interpretation of data from the literature. The number of possibly useful PFs in lung cancer is large (certainly more than one hundred). This paper attempts to provide a comprehensive listing of PFs and other variables potentially associated with outcome of lung cancer. This is achieved using tables, where studies relevant to each PF are referenced in relationship to their results, statistical power, type of analysis, number of variables incorporated in multivariate tests, and cell type. Tables include the outcome of an extensive retrieval of the literature and indicate visually where much of the evidence resides for the contribution of a variable to prognosis. Each table is briefly discussed and systematically comprises one group of PFs. Among the many pinpointed, the best predictive models still belong to studies based on clinical and routine laboratory data. Recent researches have clarified the role of new PFs (such as the biological factors); others (eg. quality of life measures, the serum content of acute phase reaction proteins or the pathological evidence of tumour neoangiogenesis) might be recognized as important in the future. Like infinity, the fate of the individual patient will never become a completely measurable entity. However, as the discovery of new PFs proceeds, the assessment of the future prospects for patients is becoming more reliable.
