T1/T2 glottic cancer managed by external beam radiotherapy: The influence of pretreatment hemoglobin on local control

Purpose: Pretreatment hemoglobin (Hb) level has been reported to be an important prognostic factor for local control and survival in various malignancies. However, in many settings, the adverse effect of a low Hb may be related to more advanced disease. The purpose of this analysis was to assess the influence of pretreatment Hb on local control in a large series of patients with a localized cancer (T1/T2 glottic cancer, AJCC 1992) treated in a standard fashion. Materials and Methods: Between January 1981 and December 1989, 735 patients (median age 63; 657 males, 78 females) with T1/T2 glottic cancer were treated with radiation therapy (RT). The standard RT prescription was 50 Gy in 20 fractions over 4 weeks (97% of patients). Factors studied for prognostic importance for local failure included pretreatment Hb, age, sex, T category, anterior commissure involvement, subglottic extension, and tumor bulk (presence of visible tumor vs. subclinical disease). Results: With a median follow up of 6.8 years (range 0.2-14.3), 131 patients have locally relapsed for an actuarial 5-year relapse-free rate of 81.7%. The 5-year actuarial survival was 75.8%. The mean pretreatment hemoglobin level was 14.8 g/dl and was similar in all prognostic categories. On multivariate analysis, using the Cox proportional hazards model, pretreatment Hb predicted for local failure after RT. The hazard ratio (HR) for relapse was calculated for various Hb levels. For example, the HR for a Hb of 12 g/dl vs, a Hb of 15 g/dl was 1.8 (95% confidence interval 1.2-2.5). Previously established factors, including gender, T category, subglottic extension, as well as tumor bulk, were also prognostically important for local control. Conclusions: This analysis, in a large number of similarly treated patients, indicates that pretreatment Hb is an independent prognostic factor for local control in patients with T1/T2 carcinoma of the glottis treated with RT. The underlying biology of this observation needs to be explored, and using this information, it may be possible to develop strategies to improve treatment outcome. (C) 1998 Elsevier Science Inc.