Competing events determining relapse-free survival in limited small-cell lung carcinoma. The French Cancer Centers' Lung Group.

PURPOSE We report results in terms of relapse-free survival (RFS), obtained in patients with limited small-cell lung carcinoma (SCLC) treated by four consecutive alternating protocols, using a competing risk approach with local recurrences, distant metastases, and death unrelated to cancer as competing events. PATIENTS AND METHODS Two hundred two patients with limited SCLC were included in four consecutive protocols alternating radiotherapy and chemotherapy (CT). The alternating schedule consisted of six cycles of CT (doxorubicin, etoposide [VP16213], and cyclophosphamide [CAVP16], plus methotrexate in the first protocol; cisplatin replaced methotrexate in the other three protocols) and three courses of thoracic radiotherapy at a total dose of 45, 55, 65, and 61 Gy in the four consecutive protocols, respectively (accelerated hyperfractionation was used in the first course of the fourth protocol). A 1-week rest followed each CT cycle and each course of radiotherapy. Seventy-six percent of patients were in complete remission at the end of the induction treatment. RFS variables were determined according to a model assuming competing risks to define the first cause of failure (local disease, distant metastasis, or intercurrent death). RESULTS No significant differences were observed between the four treatment groups. Overall results showed a 2-year cumulative incidence rate of failure of 75%. When analyzed, the first cause of failure was local recurrence only, 33%; distant only, 25%; distant and local simultaneously, 9%; and intercurrent death, 8%. CONCLUSIONS The methodology of competing risks allowed an unequivocal description of first events in limited SCLC. The extent of the local problem has been relatively overshadowed by the use of conventional descriptive methods.