LOCALLY ADVANCED NON-SMALL-CELL LUNG-CANCER - DO WE KNOW THE QUESTIONS - A SURVEY OF RANDOMIZED TRIALS FROM 1966-1993

Background. Substantial lack of consensus exists regarding the appropr iate management of patients with locally advanced non small cell lung cancer (NSCLC). The purpose of the present study was to investigate wh y published clinical trials have not resolved this uncertainty, and to examine the potential of current randomized studies to resolve the ma jor controversies regarding the treatment of locally advanced NSCLC. M ethods. A literature search identified papers addressing the therapy o f locally advanced NSCLC published in the English language from Januar y 1966 through March 1993. The treatment modalities studied in these t rials were recorded. The CD-ROM Physician Data Query database was used to identify ongoing studies in NSCLC. For phase III trials in stage I II NSCLC, the treatment modalities, eligibility criteria, outcome meas ures, and statistical considerations were recorded. Results. A total o f 164 reports of phase III trials were identified, representing 11% of the 1516 publications meeting search criteria, A wide range of compar isons have been reported; the number of study arms, the number of diff erent modalities employed as control arms, and the number of modalitie s employed as investigational arms increased over time. Eighteen activ e phase III protocols open to patients with stage III NSCLC were ident ified. In trials which enrolled patients with stage IIIB disease, ther apy in control arms employed six different strategies of surgery, radi ation, or chemotherapy, alone or in combination, and investigational a rms were equally heterogeneous. Variation was also present in the spec trum of disease stages studied, in patient eligibility criteria, and i n the clinical outcome measures investigated. The magnitude of improve ment in survival sought was varied in its absolute magnitude, in the s election of survival probability for the control arm, and in the time point of its evaluation. Implications. We demonstrated diversity in re search practice reflected in five major types of variation: (i) select ion of control arms, (ii) selection of study investigational arms, (ii i) choice of eligibility criteria, (iv) outcome measures selected for study, and (v) type and magnitude of benefit sought in the primary out come measure. This variation has important implications regarding the inability of these studies to address some fundamental management cont roversies, and the ability to generalize the results of these trials t o the general population of NSCLC patients. This diversity reflects a poorly defined process for setting the goals of clinical research. The generation of future trials may be improved by strategies that identi fy the most important controversies, identify important outcome measur es, improve consensus among physicians, and provide the opportunity to incorporate patients' preferences in this clinical situation.