Pietro Marino

Single-agent chemotherapy versus combination chemotherapy in advanced non-small cell lung cancer: a quality and meta-analysis study

STUDY OBJECTIVE To estimate the quality of the studies and to compare single-agent with combination chemotherapy in advanced non-small cell lung cancer. DESIGN Identification of published randomized trials and extraction of essential results directly from the published reports. MEASUREMENTS AND RESULTS Survival probability at 1 year, as estimated from the published survival curves, has been considered as the end-point of interest. Quality scoring of the studies has also been performed. Arithmetical calculation, concerning the estimation of quantities necessary for the meta-analysis of the literature, has been addressed. The estimated pooled Odds Ratio of death was 0.8, with 95% confidence interval of 0.6-1.0, thus favoring combination chemotherapy. CONCLUSIONS The results of our meta-analysis favor combination chemotherapy. They must, however, be considered in the light of their clinical relevance and of the balance between quality of life, toxicity and costs of chemotherapy.

In survival curves is the P-value enough?

Considering the ‘results’ in papers, we view with particular attention the famous ‘P’. Almost all publications require the presentation of a P-value of a hypothesis test before a statement can be made concerning the superiority of one treatment modality to another. As it is known, only when the P-value is small are we willing to believe that the observed difference in treatment is very likely real and not simply due to chance. Usually if the P-value is >0.05 we consider the results as ‘not significant’. Often such ‘not significant’ results are due to the following causes: (1) sample size is small because of expense and time constraints and/or (2) variation (standard deviation) is large because of difficulties in obtaining sufficient, reliable, and repeatable data, or because of instrumentation and technique errors, or an inherently large variation between subjects, etc. [l]. Recently, in a meta-analysis study, we reviewed a lot of lung cancer trials conceming the treatment of non-small cell lung cancer [2], and other trials concerning single versus combination-chemotherapy in the treatment of NSCLC (data not yet published). The majority of those trials did not present the confidence interval (CI) in the statistical analysis concerning the survival curves. Looking at those survival curves, we observed that the majority of the authors had presented their results as ‘not significant’ although they showed two or more different curves. In this common situation, the reader is often unable to evaluate the result. In such cases, the authors should show the confidence interval for median difference, that gives most helpful information when P > 0.05. Here, the chance of error in claiming a difference is large and greater than 5%, thus leading us to the conclusion of ‘no significance’. The 95% confidence interval calculated around the median difference in our data will in-

Increased levels of soluble interleukin-2 receptors in supernatants from peripheral blood mononuclear cells of patients with lung cancer

INTERLEUKIN-2 (IL-i?) stimulates immunity by acting on specific IL-2 cell surface receptors, expressed by activated but not resting immune cells [ 11. It has been demonstrated that lymphocytes may produce a soluble form of IL-2 receptors(sIL-2r) [2]. We have reported increased sIL-2r serum levels in lung cancer Table 1. Median values of sIL-2r in healthy subjects and in lung cancer patients