Primary bronchial cancer is responsible for at least 20,000 deaths per year in France. The treatment of cancer in the clinical phase remains disappointing. Numerous trials, which are reviewed here chronologically, have looked for a way to improve the prognosis by an earlier detection of this tumour. Unfortunately numerous methodological approaches have not been able to avoid the fact that the real value of such an early diagnosis is not always known. The only clear conclusion of these studies is that cytological examination of the expectorate does not lead to an improved survival in patients when compared to that which is obtained with the single radiographic film. However, taking account of the natural history of the disease, it is probably necessary to abandon the concept of an early diagnosis of a tumour which is already too late; it is towards a search for pre-cancerous lesions which are still reversible that the research effort should be turned.