Pre-operative chemoradiotherapy in non-small cell lung cancer stage III patients. Feasibility, toxicity and long-term results of a phase II study.

The aim of this study was to evaluate the feasibility, the response rate and the effect on survival of full dose polychemotherapy delivered concurrently with bifractionated radiotherapy at a radical dose, in a subset of patients with marginally resectable or unresectable stage IIIA-B non-small cell lung cancer (NSCLC). Treatment consisted of two courses of cisplatin 100 mg/m2 for 1 day plus etoposide 120 mg/m2 for 3 days delivered from day 1 to day 22, plus radiotherapy delivered in two cycles of 2560 cGy each from day 3 to day 12 and from day 24 to 33 (total dose 5120 cGy in 31 days). The daily dose was 320 cGy in two equal fractions. After surgery, three additional courses of cisplatin plus etoposide were planned. From February 1988 to June 1991, 39 patients with stage III NSCLC (19 were judged as having marginally resectable, 20 as having unresectable disease) were entered into the study. Out of 39 patients (22 squamous cell carcinoma, 17 adeno/large cell carcinoma), 24 had stage IIIa (62%) and 15 stage IIIb (38%). Median PS was 80 (70-90). A total of 78 (74 evaluable) concurrent cycles of pre-operative chemoradiotherapy were delivered. The prominent side-effect was leucopenia: leucopenia > or = grade 3 at nadir occurred in 20 cycles (27%), thrombocytopenia > or = grade 3 at nadir in seven cycles (9%), 19 patients (54%) had a treatment delay of 1 week between the two cycles. Other important toxicities were sepsis in 5 patients (13%), oesophagitis > grade 2 in 9 patients (23%) and pneumonitis in 5 patients (13%). The response rate was 67% (6 CR (complete response), 16%; 19 PR (partial response), 51%). A resection was subsequently performed in 20 (51%) patients: 14 out of 19 marginally resectable (74%) and 6 out 20 initially unresectable (30%) patients. One other patient had an exploratory thoracotomy. Surgical specimens were tumour-free in 3 patients (14%); in 8 patients (38%) only microscopic tumour was found, and in 10 (48%) macroscopic residual tumour was found. Out of 23 patients attaining a CR, 5 relapsed locally and 11 only distantly. At present, with a follow-up ranging from 64 to 90 months, 34 patients have died, 1 is alive with recurrent disease and 4 (17%) are alive without evidence of disease. Median survival was 16 months, with 18% 3-year survivors and 13% 5-year survivors. Resected patients had a median survival of 21 months, versus 10 months for unresected patients (P = 0.01). No significant difference was evident between stage IIIa and stage IIIb patients.