Fax: (352) 265-0759
Impact of race on outcome after definitive radiotherapy for squamous cell carcinoma of the head and neck
Article first published online: 22 OCT 2003
Copyright © 2003 American Cancer Society
Volume 98, Issue 11, pages 2467–2472, 1 December 2003
How to Cite
Al-Othman, M. O. F., Morris, C. G., Logan, H. L., Hinerman, R. W., Amdur, R. J. and Mendenhall, W. M. (2003), Impact of race on outcome after definitive radiotherapy for squamous cell carcinoma of the head and neck. Cancer, 98: 2467–2472. doi: 10.1002/cncr.11822
- Issue published online: 17 NOV 2003
- Article first published online: 22 OCT 2003
- Manuscript Revised: 26 AUG 2003
- Manuscript Accepted: 26 AUG 2003
- Manuscript Received: 8 JUL 2003
- absolute survival;
- control rates;
- cause-specific survival;
The objective of the current study was to evaluate the impact of race (black vs. white) on the outcome of patients with invasive squamous cell carcinoma of the head and neck.
Between 1983 and 1997, 686 patients completed definitive, twice-daily radiotherapy (RT) alone or combined with a planned neck dissection; no patients received adjuvant chemotherapy. The minimum follow-up was 2 years, and median follow-up was 7 years for living patients. No patients were lost to follow-up. Fifty-five patients were black (8%).
Although the two groups had similar 5-year local-regional control rates (black patients vs. white patients: 70% vs. 76%, respectively; P = 0.275), black patients had double the risk for distant recurrence compared with white patients (27% vs. 13%; P = 0.012). The 5-year cause-specific and absolute survival rates were lower for black patients (52% vs. 74% [P = 0.001] and 29% vs. 52% [P < 0.001], respectively). Multivariate analyses revealed that race was an independent predictor of freedom from distant metastasis (P = 0.013), cause-specific survival (P = 0.005), and absolute survival (P < 0.001).
Although equal local-regional control rates can be achieved in black patients and white patients with squamous cell carcinoma of the head and neck, the risk of distant recurrence was significantly higher in black patients and resulted in decreased survival. Reevaluation of current strategies for pretreatment metastatic work-ups and development of more effective systemic therapy will be key to improving the survival disparity in this group. Cancer 2003. © 2003 American Cancer Society.
The risk of developing squamous cell carcinoma of the head and neck increases with tobacco use and/or ethanol abuse.1 The most recent report from the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute continues to show that the black population has a higher incidence of head and neck malignancies and lower survival rates compared with the white population.2 The literature contains only a few single-institution studies from the United States that address the impact of race on treatment outcome for patients with head and neck malignancies. One study revealed no significant influence of race on prognosis,3 whereas as the other two studies confirmed the adverse effect of race on survival.4, 5 A major limitation of previous single-institution studies is heterogeneity of the treatments that were applied; thus, survival differences may be related to variability in treatment modalities. The objective of the current study was to analyze the impact of race on outcome in a large population of patients stratified by stage and treated uniformly with definitive, hyperfractionated radiotherapy (RT), either alone or combined with a planned neck dissection.
MATERIALS AND METHODS
Between January 1983 and December 1997, 934 consecutive patients with previously untreated squamous cell carcinoma of the oropharynx, hypopharynx, and larynx (excluding T1–T2 glottis patients) completed continuous-course, definitive RT, either alone or combined with a planned neck dissection. All patients were treated in the University of Florida Department of Radiation Oncology with curative intent. Patient distribution according to race was as follows: white, 847 patients (91%); black, 80 patients (9%), and other, 7 patients (< 1%). For the purpose of the current analysis, a homogeneous subset of 686 patients was defined who were treated with twice-daily RT alone or combined with a planned dissection. Patients were not included if they received adjuvant chemotherapy, had multiple synchronous head and neck primary tumors, or were of an ethnic group other than black or white. Patients with T1 and T2 glottic tumors were not included because of their favorable natural history and because limited fields and once-daily fractionation usually are employed to treat these lesions.6 Of 686 patients who met the criteria described above, 55 patients were black (8%), and the remaining patients were white. Pretreatment staging work-up included a minimum of a computed tomography (CT) scan and/or magnetic resonance image (MRI) of the head and neck and a chest roentgenogram. All patients were staged according to the 1997 American Joint Commission on Cancer (AJCC) staging system.7
Treatment selection usually was based on the recommendations of a weekly multidisciplinary head and neck tumor conference at which patients routinely were presented prior to initiation of therapy. Irradiation techniques were consistent during the study period and have been described previously.1, 8 Patients received 1.20 grays (Gy) per fraction, with a minimum interfraction interval of 4 to 6 hours; a minimum 6-hour interval has been employed since the mid-1980s.9 The median RT dose was similar in black patients and white patients (76.80 Gy vs. 76.50 Gy, respectively; P = 0.397). Two hundred sixty-eight patients (39%) underwent planned neck dissection.10
Living patients had a minimum follow-up of at least 2 years (median, 7 years); no patients were lost to follow-up. The characteristics of the patient population are depicted in Table 1. Metachronous tumors were described as those that occurred > 6 months after treatment. The lower neck was defined as that portion of the neck below the level of the thyroid notch. The distribution of variables was similar for black patients and white patients. There was an insignificant trend toward younger age for black patients compared with white patients.
|Characteristic||Black patients (n = 55)||White patients (n = 631)||Chi-square P value|
|No. (%)||No. (%)|
|≤60||32 (58)||291 (46)|
|>60||23 (42)||340 (54)|
|Male||46 (84)||506 (80)|
|Female||9 (16)||125 (20)|
|Oropharynx||36 (66)||328 (52)|
|Hypopharynx||4 (7)||98 (16)|
|Larynx||15 (27)||205 (32)|
|T1–T2||20 (36)||292 (46)|
|T3–T4||35 (64)||339 (54)|
|Negative||22 (40)||240 (38)|
|Positive||33 (60)||391 (62)|
|I–II||9 (16)||111 (18)|
|III||17 (31)||190 (30)|
|IV||29 (53)||330 (52)|
|WD/MD||36 (65)||441 (70)|
|PD||19 (35)||190 (30)|
|None||34 (62)||384 (61)|
|Unilateral||19 (34)||218 (34)|
|Bilateral||2 (4)||29 (5)|
|Head and neck||2 (4)||16 (3)||0.649|
|All||6 (11)||45 (7)||0.285|
|Lower neck involved||21 (38)||251 (40)||0.886|
All statistical analyses were performed using SAS software.11 The likelihood ratio chi-square test was used to determine the level of balance between the two ethnic groups with regard to selected explanatory variables. Survival estimates for the endpoints of local control, local-regional control, freedom from distant metastasis, cause-specific survival, and absolute survival were obtained using the Kaplan-Meier method.12 The log-rank test was used to detect statistical differences in estimates between strata of selected explanatory variables. Multivariate analyses were performed on selected explanatory variables using a Cox regression backward-selection procedure.13 Certain explanatory variables were regrouped from their original values to eliminate as much estimate bias as possible due to unavoidable imbalance across strata. The following is a description of how these variables were regrouped: age (60 years and younger vs. older than 60 years), primary tumor site (oropharynx vs. hypopharynx vs. larynx), T-stage (T1–T2 vs. T3–T4), T-stage (N0 vs. N positive [N+)]), overall stage (Stage I–II vs. Stage III vs. Stage IV), and planned neck dissection (none vs. dissection).
Time to Recurrence
Overall, 207 patients (30%) experienced disease recurrence after treatment; 23 of 55 black patients (42%) developed recurrent disease, all of them within 5 years of treatment. The median time to recurrence in black patients was 9 months. One hundred eighty-four of 631 white patients (29%) developed recurrent disease after RT; 98% of those recurrences were observed within 5 years of treatment. The median time to recurrence in white patients was 10 months.
The 5-year rates of overall local control, regional control, and local-regional control after RT were 81%, 90%, and 76%, respectively. Table 2 summarizes univariate comparison of the 5-year rates of disease control stratified by race. The probability of local-regional control was not significantly different between black patients and white patients. Salvage therapy after recurrence was equally ineffective in both black patients and white patients. Multivariate analysis of initial local-regional control revealed that the following variables were independent predictors of failure above the clavicles: T3–T4 tumors (P < 0.001), N+ neck disease (P = 0.001), and male gender (P = 0.032); not undergoing neck dissection and increasing overall AJCC stage approached significance (P = 0.069 and P = 0.090, respectively). The variables patient age (P = 0.229), race (P = 0.332), primary tumor site (P = 0.552), and histologic differentiation (P = 0.622) were not statistically significant.
|Endpoint||Black patients (%) (n = 55)||White patients (%) (n = 631)||P value|
|Initial local control||76||81||0.273|
|Initial regional control||89||90||0.795|
|Initial local-regional control||70||76||0.275|
|Ultimate local-regional control||70||76||0.268|
|Freedom from distant recurrence||73||87||0.012|
The 5–year rate of distant recurrence for the entire study population was 14%. Figure 1 shows the likelihood of distant recurrence stratified by race. Distant failure occurred almost twice as often for black patients compared with their white counterparts. Multivariate analysis of freedom from distant recurrence revealed the following independent predictors of disease recurrence: increasing overall AJCC stage (P = 0.003), T3–T4 tumors (P = 0.003), black race (P = 0.013), N+ neck disease (P = 0.016), and hypopharyngeal primary site (P = 0.043). Gender (P = 0.101), histologic differentiation (P = 0.176), neck dissection (P = 0.316), and patient age (P = 0.789) did not reach statistical significance.
The overall cause-specific survival rate at 5 years was 72%. Figure 2 shows cause-specific survival rates by race. The rate for black patients was 30% lower compared with the rate for white patients. Multivariate analysis of cause-specific survival revealed that the following parameters were independently significant for poor survival: T3–T4 primary site (P < 0.001), increasing overall AJCC stage (P < 0.001), black race (P = 0.005), male gender (P = 0.009), hypopharyngeal primary site (P = 0.029), and poor histologic differentiation (P = 0.037). N stage was marginally significant (P = 0.052). Neck dissection and patient age were not statistically significant (P = 0.113 and P = 0.196, respectively).
The 5-year absolute survival rate for the overall study population was 50%. Figure 3 shows absolute survival rates stratified by race. The 5-year survival rate for black patients was 44% lower compared with their white counterparts. Multivariate analysis revealed the following independent predictors for poor survival in order of significance: older age (P < 0.001), black race (P < 0.001), T3–T4 primary tumor (P < 0.001), increasing overall AJCC stage (P = 0.002), hypopharyngeal primary site (P = 0.003), and male gender (P = 0.005). Parameters that did not reach statistical significance included N stage (P = 0.260), neck dissection (P = 0.291), and histologic differentiation (P = 0.354).
Despite a similar distribution of tumor-related variables, black patients had a poorer outcome compared with white patients after undergoing essentially the same treatment. The decreased survival observed in black patients was due to a higher risk of distant recurrence despite similar rates of local-regional control.
Racial disparities in survival after treatment for head and neck carcinoma have been observed previously in population-based studies.2, 14–16 Drawbacks of those studies include patients who often had tumors of different histologies, stage grouping that was imprecise (patients were categorized as having local, regional, distant, or unstated disease extent), treatment intent that varied (curative intent vs. palliative intent), a subset of patients that may have received prior treatment, and treatment modalities that were heterogeneous (surgery, RT, and/or chemotherapy). Although those studies may contain larger numbers of patients, the heterogeneity of the patient population limits their usefulness.
Single-institution studies3–5 have provided additional insight by analyzing more homogeneous patient populations treated under somewhat more uniform conditions. Moore et al.4 from The University of Texas M. D. Anderson Cancer Center (Houston, Texas) reported the largest series comparing outcomes for 909 black patients and white patients with squamous cell carcinoma of the oral cavity and pharynx who were treated with a variety of local and systemic modalities. Tumors were stratified according to the SEER stage grouping; 15% of black patients and 6% of white patients had distant metastasis at presentation. Those authors reported 5-year absolute survival rates for black patients and white patients that were similar to our findings (28% vs. 52%, respectively; P < 0.001). On multivariate analysis, race emerged as a significant independent predictor of overall survival. In a study from the University of Cincinnati (Cincinnati, Ohio), Murdock and Gluckman5 randomly selected and compared the survival of an approximately equal number of black (n = 54) and white patients (n = 52) with squamous cell carcinoma of the oral cavity, pharynx, and larynx who were treated between 1991 and 1996 with surgery and/or RT with or without chemotherapy. Those authors found black patients had an approximately 2-fold greater rate of death compared with white patients (P < 0.01), and race was identified as an independent prognostic factor for survival. In contrast, Roach et al.3 at the Martinez Veterans Administration Hospital (Martinez, California) analyzed 214 patients with laryngeal malignancies who were treated with curative intent with surgery and/or RT between 1968 and 1988 and found that after stratifying patients for a variety of factors (including time from onset of symptoms to diagnosis, time from diagnosis to treatment, patient age, overall AJCC stage, laryngeal subsite, and initial therapy), there were no differences in the survival rates between black patients and white patients.
In contrast to the single-institution studies described above, patients who were included in our analysis were treated uniformly with RT using the same dose-fractionation schedule, thus eliminating treatment-related variables.17 Our data indicate that hyperfractionated RT was effective in obtaining local-regional disease control in both black patients and white patients, although black patients had a lower survival rate due to an increased risk of distant recurrence. Although only 8% of our patients were black, this is similar to the 8.5% incidence reported in a series of 295,022 patients with head and neck malignancies who were included in the National Cancer Data Base between 1985 and 1994.18
In an attempt to explain the survival disparity of black patients versus white patients treated for squamous cell carcinoma of the oral cavity and pharynx, Arbes and associates14 analyzed patients from the SEER Program 1973–1993 public-use database. Those authors found that lower socioeconomic status, more advanced disease at diagnosis, and differences in treatment probably were related to the lower survival rates observed for black patients. Other proposed explanations for differences in treatment outcome include patient comorbidites,19 access to health care,20 and quality of care provided.21 The survival differences also may be related to undefined differences in the biologic characteristics of squamous cell carcinoma of the head and neck. Recently, a number of molecular tumor markers have been investigated in relation to the natural history and prognosis for patients with head and neck malignancies.22, 23 To date, no study has clearly identified any specific biologic factor related to differences in outcome between the two racial groups (e.g., tobacco or alcohol use). Whether the underlying mechanism for the racial differences in outcome in patients with squamous cell carcinoma of the head and neck is biological, socioeconomic, or both remains unanswered.24, 25
The decreased survival for black patients in our study probably was due to an increased risk of distant metastases. The reasons for this increased risk remain unclear. A drawback of our study is that we did not have access to reliable information related to lifestyle (including tobacco and alcohol use) or socioeconomic status, and we did not account for potentially important differences in baseline health status of the patients included in our study.26 In addition, the patient population in Florida is relatively mobile, and reliable demographic data were not available. Disentangling such interdependent variables that are not readily measurable can prove challenging, as reflected in the literature.27, 28
Although we have not accounted for all of the variables that may explain our findings, our data underscore practical implications that are worth consideration. With equivalent local-regional control in black patients and white patients, current treatment regimens do not need to be altered, based on race, to improve disease control above the clavicles. Future studies are needed to improve detection of distant metastases prior to treatment29 and to develop more effective systemic therapy to reduce the risk of recurrence below the clavicles.30 Whether current adjuvant chemotherapy regimens result in lower rates of distant metastasis is debatable.31 Finally, we recommend considering stratification by race before randomization in future prospective studies of patients with squamous cell carcinoma of the head and neck if distant metastasis and survival are endpoints.
- 1MillionRR, CassisiNJ, editors. Management of head and neck cancer: a multidisciplinary approach. Philadelphia: JB Lippincott, 1994.
- 7AJCC cancer staging manual (5th edition). Philadelphia: Lippincott-Raven, 1998., , , et al.
- 8Levitt and Tapley's technological basis of radiation therapy: clinical applications. Baltimore: Lippincott Williams & Wilkins, 1999., , , .
- 11SAS Institute Inc. SAS onlinedoc 9. Cary, NC: SAS Institute, 2003.
- 13Regression models and life tables. J R Stat Soc B. 1972; 34: 187–220..