We have systematically reviewed the biomedical literature to try to establish whether laboratory variables might give any additional prognostic information in non-small-cell lung cancer (NSCLC) patients independently of the usual radioclinical parameters. In each study, we acknowledged the independent prognostic value of a biological fluid variable if it had been demonstrated through a multivariate statistical analysis in which at least the following had been included: patient's weight loss, age, gender, performance status, histology, stage and extent of the disease. The clearest conclusion that can be derived from the 42 studies we reviewed is that it remains to be clearly demonstrated whether or not the "new" tests (tumour markers, p53 antibodies, etc.) are superior to the "old" tests (serum LDH, calcium, albumin or other proteins, blood cell counts, etc.), even though a number of studies did suggest that serum cyfra 21-1 has a pre-therapeutic prognostic significance in NSCLC. From the four studies in which the same powerful statistical methodologies were used (i.e. Cox models in association with RECPAM analysis), it could be derived that serum calcium and perhaps the blood neutrophil and lymphocyte counts might have independent pre-therapeutic prognostic significance in advanced NSCLC. Further studies are needed to demonstrate whether repeated measurements during therapeutic follow-up can bring any independent prognostic information. Provided that both laboratory and statistical expertise is clearly guaranteed in future primary studies published in this particular biomedical field, it might perhaps become possible to propose laboratory variables as additional prognostic parameters in NSCLC.