Prognostic value of tumor regression during radiotherapy for head and neck cancer: a prospective study.

OBJECTIVE

Prospective evaluation of tumor regression during external irradiation for head and neck squamous cell carcinomas and its association with long-term local control.

METHODS AND MATERIALS

Two hundred twenty-eight patients with histologically confirmed squamous cell carcinoma [oral cavity: 59 (26%), oropharynx: 65 (29%), hypopharynx: 37 (16%), larynx: 67 (29%)] were included between January 1986 and December 1990. Curative intent external irradiation delivered 65-70 Gy over a period of 7 weeks (five 2 Gy fractions per week). Tumor regression was evaluated clinically and endoscopically every week.

RESULTS

Tumor regression, assessed at 2 weeks, was as follows: no response: 62 (30%), 25% response: 121 (59%); 50% response: 23 (11%). At 5 weeks, 9 (4%) patients showed 0-25% regression, 75 (33%) showed 50% regression, 115 (50%) showed 75% regression, and 29 (13%) showed complete regression. Median follow-up was 79 months (range: 6-96 months). The local control probability was 68% (62-74%) at 2 years, 65% (59-70%) at 5 years. Univariate analysis showed that, at 2 weeks, local control was significantly different between the nonresponders and the patients with 25% or greater response (p < 0.025) and that, at the fifth week, local control was very different between the major responders (75 and 100%) and the minor responders (0-50%) (p < 0.0001). Multivariate analysis (Cox Proportional Hazards Model) showed that the probability of local relapse was significantly and independently increased for minor regression at 5 weeks [Relative risk (RR) of failure was 2.3 (1.4-3.7)], for nonlaryngeal tumors [RR: 2.4 (1.3-4.5)], and for Stage T3-T4 [RR:2.4 (1.4-4)]. Three prognostic groups can, therefore, be proposed: 1) low risk of recurrence when regression > or = 75% and laryngeal tumor or T1-T2 tumors in other sites: 106 (46.5%) patients, 2-year local control probability: 84% (77-92%); 2) high risk of recurrence: regression < or = 50% and T3-T4 nonlaryngeal tumors: 44 (19%) patients, 2-year local control probability: 27% (13-41%); 3) intermediate risk of recurrence: 78 (34.5%) patients, 2-year local control probability: 69% (58-80%).

CONCLUSION

The present study suggests that tumor regression during external radiotherapy is an independent predictive factor of local control in head and neck carcinomas.