In/ J Rud/rr/ron Onurk,~,~ B/d Phn Vol. 23. pp. 733.736 Prmted m the US A. All rights reserved.
Copyright
0360-3016/92 $5.00 + .oO CC 1992 Pergamon Press Ltd.
??Clinical Original Contribution
DOES NECK STAGE INFLUENCE LOCAL CONTROL IN SQUAMOUS CARCINOMAS OF THE HEAD AND NECK?
CELL
DEBRA E. FREEMAN, M. D., WILLIAM M. MENDENHALL, M. D., JAMES T. PARSONS, M. D. AND RODNEY R. MILLION, M. D. Department
of Radiation
Oncology,
University
of Florida College of Medicine,
Gainesville,
FL
Recently, reports have suggested that the probability of local control of head and neck cancers treated with radiotherapy alone is inversely related to the extent of neck node disease at presentation. This has led some to conclude that primary lesions in patients with neck node metastases should be treated more aggressively than lesions of the same T stage in patients presenting with a clinically negative neck. We reviewed the records of 607 patients with squamous cell carcinoma of the head and neck to determine the relationship between the extent of neck node disease at diagnosis and the probability of tumor control at the primary site. All patients were treated with continuouscourse irradiation alone to the primary site and have a minimum follow-up of 2 years. Mucosal sites analyzed included the oropharynx (soft palate, tonsillar region, base of tongue), hypopharynx (pharyngeal wall, pyriform sinus), and supraglottic larynx. Patients with simultaneous primary lesions were excluded from the analysis. Parameters tested included T stage, N stage, and fractionation (once daily or twice daily). For statistical analyses, N stage was grouped as NO, Nl, N2A-N3A, and N2B-N3B. Local control was analyzed using a multivariate analysis, the forward stepwise long-rank test of association of covariates. Multivariate analyses revealed that the following parameters significantly influenced local control: oropharynx, T stage only (p = .OOlO); hypopharynx, T stage (p = .OOl) and twice-daily fractionation (p = .0031); and supraglottic larynx, T stage only Cp = .0002). No significant correlation was found between neck stage and primary tumor control on univariate or multivariate analyses. Our data do not support the conclusion that primary lesions are controlled less often in patients with clinically positive neck nodes. Head and neck neoplasms, Neck nodes.
INTRODUCTION
METHODS
AND MATERIALS
There are a number of prognostic factors that may influence local control of primary squamous cell carcinoma of the head and neck. These include primary site, T stage, histologic grade, patient’s gender, Kamofsky performance status, and hemoglobin level (7, 13). Several authors have also noted an inverse relationship between the extent of neck disease at presentation and the likelihood of primary tumor control, leading some to conclude that patients presenting with clinically positive neck nodes should receive more aggressive treatment to their primary site, either with higher doses of radiation or combined surgery and irradiation (4,7, 8, 18, 19). The purpose of this study is to analyze the impact of neck stage on the probability of local control for patients with primary squamous cell carcinoma arising in the head and neck and treated with irradiation alone for cure.
We reviewed the records of 705 patients with previously untreated squamous cell carcinoma arising in six selected head and neck sites who were treated with curative intent at the University of Florida between October 1964 and June 1988 with irradiation alone to their primary lesion. All patients have had at least 2 years of follow-up from initiation of treatment; no patient has been lost to followup. Ninety-eight patients were excluded from analysis for the following reasons: simultaneous head and neck primary tumor, 64 patients; metachronous head and neck primary tumor occurring within 2 years of treatment of the initial lesion, 5 patients; primary tumor stage inevaluable, 7 patients; neck stage inevaluable, 9 patients; and treatment not completed as planned, 13 patients. The remaining 607 patients were available for analysis. The six mucosal sites studied were the soft palate (52 patients),
Presented at the 33rd Annual Meeting ofthe American Society for Therapeutic Radiology and Oncology, Washington, DC, 48 November 199 1. Reprint requests to: Debra E. Freeman, M.D., Department
of Radiation Oncology, University of Florida Health Science Center, P.O. Box 100385, Gainesville, FL 32610-0385. Accepted for publication 2 1 February 1992.
133
734
1. J. Radiation Oncology 0 Biology 0 Physics
base of the tongue (92 patients), tonsillar region ( 180 patients), pharyngeal wall (80 patients), pyriform sinus (66 patients), and supraglottic larynx ( 137 patients). Lesions were staged according to the 1983 AJCC staging system ( 1). The primary sites were grouped as follows: oropharynx (soft palate, base of the tongue, tonsillar region), hypopharynx (pharyngeal wall, pyriform sinus), and supraglottic larynx. Neck stages were grouped as NO, N 1, N2AN3A, and N2B-N3B. All patients were treated with continuous-course irradiation using megavoltage equipment, receiving external beam irradiation alone or combined with an interstitial implant. Patients treated with a split-course technique were not included. The neck was treated with irradiation alone or combined with neck dissection. Approximately one third of the patients were treated with twice-a-day fractionation; the remainder were treated with once-aday fractionation. Details of the irradiation techniques used have been previously described ( 12). The end point of all analyses that follow is local control, which is defined as complete and continuous disappearance of disease at the primary site. For direct calculations of local control, 132 patients who died of intercurrent disease, distant metastases, or neck disease within 2 years of treatment with their primary tumor continuously controlled were excluded from analysis. For the calculation of local control probability, all patients were included and patients with continuous local control were censored at the date of last follow-up or death (6. 17). Patients developing a local recurrence were censored at the date the recurrence was documented. Fisher’s exact test (9) or a test based on a normal approximation of the binomial distribution (3) was used for comparison of proportions. The forward stepwise log-rank test of association of covariates, which is implemented by the SAS PROC LIFETEST, was used for multivariate analyses (5, 17).
RESULTS The results of a direct analysis of local control by T stage and N stage are shown in Table 1. There were no significant differences in local control with increasing neck stage. Specifically, patients presenting with clinically positive neck nodes did not have a higher rate of local recurrence than patients of the same T stage presenting with a clinically negative neck.
Volume 23. Number 4. 1992
The rate of local control as a function of status of the neck for each of the mucosal sites analyzed is summarized in Table 2. The presence of neck node metastases did not adversely influence the probability of primary tumor control for any site or T stage. A series of multivariate analyses was also performed, testing a variety of parameters for their influence on local control. The initial analysis included all 607 patients and tested the parameters T stage, N stage, fractionation (once daily vs twice daily), and primary site. T stage, primary site, and fractionation proved significant (Table 3). Twicedaily fractionation resulted in improved local control rates. Primary tumors of the hypopharynx had poorer rates of local control than tumors of the oropharynx or supraglottic larynx. Neck stage had no apparent influence on local control. A separate multivariate analysis was performed for each mucosal site, grouped as oropharynx, hypopharynx, and supraglottic larynx. Parameters tested in these analyses were T stage, N stage, and fractionation (Table 4). For the oropharynx and supraglottic larynx, T stage was the only parameter found to significantly influence local control: N stage was not significant. For the hypopharynx, both T stage and fractionation were significant, with twicedaily fractionation associated with improved local control. The extent of neck node disease did not have a significant influence on local control. To test whether certain clinical characteristics of neck disease were predictive of local control, a multivariate analysis on node-positive patients was performed, analyzing the parameters of node location (upper neck vs lower neck plus upper neck) and node mobility (mobile or tethered vs fixed) (Table 5). Only T stage was found to significantly influence local control. The N stage, node location, and node mobility did not predict local control. Lastly, separate analyses were performed for patients treated with once-daily and twice-daily fractionation, as fractionation proved an important prognostic variable in the overall analysis. For the 368 patients treated once daily, T stage and primary site were the parameters found to influence local control, with poorer control rates for primary tumors of the hypopharynx than for lesions of the oropharynx or supraglottic larynx (Table 6). For the 239 patients treated twice daily, only T stage significantly predicted local control (Table 7). In neither group was neck stage predictive of primary tumor control. For a
Table I. Local control according to N stage (475 patients*) NO
NI
N2A-N3A
N2B-N3B
Stage
No.
(Q’)
No.
(a)
No.
(“/)
TI T2 T3 T4
25126 79/103 45 164 4111
(96) (77) (70) (36)
6/8 21/28 18/28 7/l I
(75) (75) (64) (64)
515 15/1x 12116 Of7
* Excludes
I32 patients
who died < 2 years from treatment
with primary
No.
(St)
(100)
616
(100)
(83) (75) (0)
29135 40172 13137
(83) (56) (35)
site continuously
controlled.
p values NO vs N+ .38 .34 .I2 .63
735
Does neck stage influence local control? 0 D. E. FREEMANet al.
Table 4. Multivariate analysis according to site
Table 2. Local control according to site Stage Soft palate
Tonsillar region
Base of tongue
Pyriform sinus
Pharyngeal wall
Supraglottic larynx
Tl T2 T3 T4 Tl T2 T3 T4 Tl T2 T3 T4 Tl T2 T3 T4 Tl T2 T3 T4 T1 T2 T3 T4
N+
p value
515
l/l
6/10 415 n.d. 415 23135 14/18 415 l/l 717 516 O/2 313 9110
515 5112 215 515 17/19 23137 6/18 314 14117 19123 6115 516 13/19 213 l/5 n.d. 315 8/17 2/8 212 14117 13123 315
n.d. .15 .18 n.d. .50 .I2 .19 .09 .80 .34 .73 .40 .67 .21
NO
n.d. n.d. 313 14/19 s/10 O/l 919 18/22 17125 O/3
n.d. n.d. n.d. .46 .60 .78 n.d. .65 .30 .18
NO = Clinically negative neck; Nf = Clinically positive neck nodes; n.d. = No data.
given site and T stage, patients treated with twice-daily fractionation received similar total tumor doses at a constant dose per fraction ( 120 cGy), such that there was little, if any, difference in tumor dose to the primary site for node-positive and node-negative patients ( 15). These results suggest that the lack of difference in local control rates for node-positive and node-negative patients is not a result of higher tumor doses being delivered to the primary sites for patients presenting with neck metastases.
DISCUSSION In 1988, we reported the results of a direct analysis of the influence of neck stage on local control for 526 patients with primary squamous cell carcinomas of the head and neck arising from 10 mucosal sites (11). We found no evidence to suggest that advanced neck disease adversely affected local control rates. In the current study, only mucosal sites with an intermediate or high risk of neck disease
Table 3. Multivariate analysis: Overall group (607 patients) Rank 1 2 3 4 5
Variable T stage Twice-daily fractionation Hypopharynx Oropharynx N stage
Rank
Variable
Significance
Oropharynx (324 patients)
1 2
< .Ol .26
Supraglottic larynx (137 patients)
3 1 2
Hypopharynx ( 146 patients)
3 1 2
T stage Twice-daily fractionation N stage T stage Twice-daily fractionation N stage T stage Twice-daily fractionation N stage
Site
3
.98 < .Ol .23 .67 < .Ol < .Ol .62
at presentation were included (10); patients with more than one primary lesion were excluded; and both direct and multivariate analyses were performed. Again, we found no evidence that patients presenting with clinically positive necks were at higher risk for local recurrence than patients with the same T stage with clinically negative necks. A similar conclusion was reached by Richard et neck al., ( 16) who reviewed 1,7 13 patients undergoing dissection at the Institut Gustave Roussy for squamous cell carcinoma of the head and neck. These data differ from those reported in several other series in the literature. In an analysis of 248 patients with supraglottic cancers treated with irradiation alone, Wall et al. ( 19) from the M. D. Anderson Hospital found a 15% to 30% lower local control rate for patients with palpable neck disease at diagnosis compared to patients presenting with a clinically negative neck. The risk of local failure appeared to increase as the bulk of neck disease increased. In another series from the M. D. Anderson Hospital, Ang et al. (2) reported 10% lower primary tumor control for patients with carcinomas of the oropharynx and nasopharynx presenting with neck node metastases compared to patients with a clinically negative neck. For tumors of the supraglottic larynx, Van den Bogaert et al. (18) showed a higher rate of primary tumor failure in patients presenting with nodal disease compared to patients without neck metastases (50%’ vs 30%). Hahn et al. (4) from the University of Virginia also reported a 37% primary recurrence rate for patients with lesions of the supraglottic larynx or pyriform sinus presenting with N2 or N3 necks, compared to a 19% local recurrence if neck stage was NO or Nl. Both Wall et al. and Hahn et al.
Table 5. Multivariate analysis: Patients with clinically positive neck nodes (37 1 patients)
Significance < .Ol < .Ol .Ol .86 .97
Rank
Variable
1 2 3 4
T stage Location Mobility N stage
Significance < .Ol .66 .77 .94
736
I. J.
Table 6. Multivariate
Rank
1 2 3 4
Radiation
Oncology
analysis: Once-a-day (368 patients) Variable
T stage Hypopharynx N stage Oropharynx
0 Biology
0 Physics
Volume
fractionation
Significance < .Ol < .Ol .43 .80
4. 1992
Table 7. Multivariate
analysis: Twice-daily (239 patients)
Rank
Variable
I
T stage N stage Hypopharynx Supraglottic larynx
7 ; 4
recommended higher doses of radiation or combined modality therapy for patients presenting with advanced neck disease. It is difficult to explain these conflicting data. Most of the published series, including our own, are retrospective and therefore subject to variations in patient selection and treatment parameters such as dose, fractionation, and overall treatment time. Some of the reported series did not stratify patients by T stage or grouped T stages (TlT2 vs T3-T4) when analyzing local control for a particular neck stage (4, 16, 19). Given that T stage was the most important predictor of local control in all of the patient groups in this study, it is likely that bias could have been introduced into local control comparisons if patients were not stratified by both T stage and N stage. Also, methods used to calculate local control vary between institutions, which hampers direct comparison of reported data ( 14).
23. Number
fractionation
Significance < .Ol .30 .80 .96
For head and neck carcinoma treated with irradiation alone, we determine the dose to the primary lesion based on T stage and primary site. Within a particular T stage, more advanced lesions receive slightly higher doses than tumors of lesser volume. Involved lymph nodes that fall within the primary irradiation field receive a tumor dose equivalent to, or slightly higher than, the dose received by the primary tumor. Large nodes falling outside or only partly within the primary volume receive additional treatment (a boost dose) with opposed, anterior and posterior, wedged fields that do not deliver further irradiation to the primary site. As we found no evidence that patients presenting with neck node metastases are at higher risk for local recurrence than patients of the same T stage with clinically negative necks, we see no indication for more aggressive treatment to the primary site based solely on the extent of neck disease at diagnosis.
REFERENCES 1. American Joint Committee on Cancer. Manual for staging of cancer, 2nd edition. Philadelphia, PA: J. B. Lippincott Co.; 1983:3 l-42. 2. Ang, K. K.; Peters, L. J.; Weber, R. S.; Maor, M. H.; Morrison, W. H.; Wendt, C. D.; Brown, B. W. Concomitant boost radiotherapy schedules in the treatment of carcinoma of the oropharynx and nasopharynx. Int. J. Radiat. Oncol. Biol. Phys. 19: 1339-I 345; 1990. 3. Colton, T. Statistics in medicine. Boston: Little Brown & Co.; 1974:163-167. 4. Hahn, S. S.; Spaulding, C. A.; Kim, J.-A.; Constable, W. C. The prognostic significance of lymph node involvement in pyriform sinus and supraglottic cancers. Int. J. Radiat. Oncol. Biol. Phys. 13:1143-l 147;1987. 5. Kalbfleisch, J. D.; Prentice, R. L. The statistical analysis of failure time data. New York: Wiley & Sons: 1980. estimation from 6. Kaplan, E. L.; Meier, P. Nonparametric incomplete observations. J. Am. Stat. Assoc. 53:457481;1958. 7. Kramer, S.; Martial, V. A.; Pajak, T. F.: MacLean, C. J.: Davis, L. W. Prognostic factors for loco/regional control and metastasis and the impact on survival. Int. J. Radiat. Oncol. Biol. Phys. 12:573-578:1986. 8. Marks, J. E.; Freeman, R. B.; Lee, F.; Ogura. J. H. Carcinoma of the supraglottic larynx. AJR 132:255-260: 1979. 9. Mendenhall, W.; Ott, L.; Larson, R. F. Statistics: a tool for the social sciences. North Scituate, MA: Duxbury Press: 1974:333-336. 10. Mendenhall, W. M.; Million, R. R. Elective neck irradiation for squamous cell carcinoma of the head and neck: analysis of time-dose factors and causes of failure. Int. J. Radiat. Oncol. Biol. Phys. 12:741-746;1986. 11. Mendenhall, W. M.; Parsons, J. T.; Amdur, R. J.; Spangler.
12.
13.
14.
15.
16.
17. 18.
19.
A. E.: Williams, T. R.; Cassisi, N. J.; Million, R. R. Squamous cell carcinoma of the head and neck treated with radiation therapy: the impact of neck stage on local control. Int. J. Radiat. Oncol. Biol. Phys. 14:249-252;1988. Million, R. R.; Cassisi, N. J. Management of head and neck cancer: A multidisciplinary approach. Philadelphia, PA: J. B. Lippincott Co.: 1984:299-364, 373-39 1. Overgaard, J.; Hansen, H. S.; Jorgensen, K.; Hansen, M. H. Primary radiotherapy of larynx and pharynx carcinoma-an analysis of some factors influencing local control and survival. Int. J. Radiat. Oncol. Biol. Phys. 12:5 15521:1986. Parsons, J. T.; McCarty, P. J.; Rao, P. V.; Mendenhall, W. M.; Million, R. R. On the definition of local control. Int. J. Radiat. Oncol. Biol. Phys. 18:705-706;1990. Parsons, J. T.; Mendenhall, W. M.; Cassisi, N. J.: Issacs, J. H.; Million, R. R. Hyperfractionation for head and neck cancer. Int. J. Radiat. Oncol. Biol. Phys. 14:649-658;1988. Richard, J. M.; Sancho-Gamier, H.; Micheau, C.; Saravane, D.; Cachin, Y. Prognostic factors in cervical lymph node metastasis in upper respiratory and digestive tract carcinomas: study of 1,7 13 cases during a 15-year period. Laryngoscope 97:97-101;1987. SAS user’s guide: Statistics, version 5 edition. Gary. NC: SAS Institute; 1985:529-557. Van den Bogaert, W.; Ostyn, F.; van der Schueren, E. The different clinical presentation, behaviour and prognosis of carcinomas originating in the epilarynx and the lower supraglottis. Radiother. Oncol. 1: 117- 131; 1983. Wall, T. J.; Peters, L. J.; Brown, B. W.; Oswald, M. J.; Milas, L. Relationship between lymph nodal status and primary tumor control probability in tumors of the supraglottic larynx. Int. J. Radiat. Oncol. Biol. Phys. 11: 1895-1902;1985.
Copyright
0360-3016/92 $5.00 + .oO CC 1992 Pergamon Press Ltd.
??Clinical Original Contribution
DOES NECK STAGE INFLUENCE LOCAL CONTROL IN SQUAMOUS CARCINOMAS OF THE HEAD AND NECK?
CELL
DEBRA E. FREEMAN, M. D., WILLIAM M. MENDENHALL, M. D., JAMES T. PARSONS, M. D. AND RODNEY R. MILLION, M. D. Department
of Radiation
Oncology,
University
of Florida College of Medicine,
Gainesville,
FL
Recently, reports have suggested that the probability of local control of head and neck cancers treated with radiotherapy alone is inversely related to the extent of neck node disease at presentation. This has led some to conclude that primary lesions in patients with neck node metastases should be treated more aggressively than lesions of the same T stage in patients presenting with a clinically negative neck. We reviewed the records of 607 patients with squamous cell carcinoma of the head and neck to determine the relationship between the extent of neck node disease at diagnosis and the probability of tumor control at the primary site. All patients were treated with continuouscourse irradiation alone to the primary site and have a minimum follow-up of 2 years. Mucosal sites analyzed included the oropharynx (soft palate, tonsillar region, base of tongue), hypopharynx (pharyngeal wall, pyriform sinus), and supraglottic larynx. Patients with simultaneous primary lesions were excluded from the analysis. Parameters tested included T stage, N stage, and fractionation (once daily or twice daily). For statistical analyses, N stage was grouped as NO, Nl, N2A-N3A, and N2B-N3B. Local control was analyzed using a multivariate analysis, the forward stepwise long-rank test of association of covariates. Multivariate analyses revealed that the following parameters significantly influenced local control: oropharynx, T stage only (p = .OOlO); hypopharynx, T stage (p = .OOl) and twice-daily fractionation (p = .0031); and supraglottic larynx, T stage only Cp = .0002). No significant correlation was found between neck stage and primary tumor control on univariate or multivariate analyses. Our data do not support the conclusion that primary lesions are controlled less often in patients with clinically positive neck nodes. Head and neck neoplasms, Neck nodes.
INTRODUCTION
METHODS
AND MATERIALS
There are a number of prognostic factors that may influence local control of primary squamous cell carcinoma of the head and neck. These include primary site, T stage, histologic grade, patient’s gender, Kamofsky performance status, and hemoglobin level (7, 13). Several authors have also noted an inverse relationship between the extent of neck disease at presentation and the likelihood of primary tumor control, leading some to conclude that patients presenting with clinically positive neck nodes should receive more aggressive treatment to their primary site, either with higher doses of radiation or combined surgery and irradiation (4,7, 8, 18, 19). The purpose of this study is to analyze the impact of neck stage on the probability of local control for patients with primary squamous cell carcinoma arising in the head and neck and treated with irradiation alone for cure.
We reviewed the records of 705 patients with previously untreated squamous cell carcinoma arising in six selected head and neck sites who were treated with curative intent at the University of Florida between October 1964 and June 1988 with irradiation alone to their primary lesion. All patients have had at least 2 years of follow-up from initiation of treatment; no patient has been lost to followup. Ninety-eight patients were excluded from analysis for the following reasons: simultaneous head and neck primary tumor, 64 patients; metachronous head and neck primary tumor occurring within 2 years of treatment of the initial lesion, 5 patients; primary tumor stage inevaluable, 7 patients; neck stage inevaluable, 9 patients; and treatment not completed as planned, 13 patients. The remaining 607 patients were available for analysis. The six mucosal sites studied were the soft palate (52 patients),
Presented at the 33rd Annual Meeting ofthe American Society for Therapeutic Radiology and Oncology, Washington, DC, 48 November 199 1. Reprint requests to: Debra E. Freeman, M.D., Department
of Radiation Oncology, University of Florida Health Science Center, P.O. Box 100385, Gainesville, FL 32610-0385. Accepted for publication 2 1 February 1992.
133
734
1. J. Radiation Oncology 0 Biology 0 Physics
base of the tongue (92 patients), tonsillar region ( 180 patients), pharyngeal wall (80 patients), pyriform sinus (66 patients), and supraglottic larynx ( 137 patients). Lesions were staged according to the 1983 AJCC staging system ( 1). The primary sites were grouped as follows: oropharynx (soft palate, base of the tongue, tonsillar region), hypopharynx (pharyngeal wall, pyriform sinus), and supraglottic larynx. Neck stages were grouped as NO, N 1, N2AN3A, and N2B-N3B. All patients were treated with continuous-course irradiation using megavoltage equipment, receiving external beam irradiation alone or combined with an interstitial implant. Patients treated with a split-course technique were not included. The neck was treated with irradiation alone or combined with neck dissection. Approximately one third of the patients were treated with twice-a-day fractionation; the remainder were treated with once-aday fractionation. Details of the irradiation techniques used have been previously described ( 12). The end point of all analyses that follow is local control, which is defined as complete and continuous disappearance of disease at the primary site. For direct calculations of local control, 132 patients who died of intercurrent disease, distant metastases, or neck disease within 2 years of treatment with their primary tumor continuously controlled were excluded from analysis. For the calculation of local control probability, all patients were included and patients with continuous local control were censored at the date of last follow-up or death (6. 17). Patients developing a local recurrence were censored at the date the recurrence was documented. Fisher’s exact test (9) or a test based on a normal approximation of the binomial distribution (3) was used for comparison of proportions. The forward stepwise log-rank test of association of covariates, which is implemented by the SAS PROC LIFETEST, was used for multivariate analyses (5, 17).
RESULTS The results of a direct analysis of local control by T stage and N stage are shown in Table 1. There were no significant differences in local control with increasing neck stage. Specifically, patients presenting with clinically positive neck nodes did not have a higher rate of local recurrence than patients of the same T stage presenting with a clinically negative neck.
Volume 23. Number 4. 1992
The rate of local control as a function of status of the neck for each of the mucosal sites analyzed is summarized in Table 2. The presence of neck node metastases did not adversely influence the probability of primary tumor control for any site or T stage. A series of multivariate analyses was also performed, testing a variety of parameters for their influence on local control. The initial analysis included all 607 patients and tested the parameters T stage, N stage, fractionation (once daily vs twice daily), and primary site. T stage, primary site, and fractionation proved significant (Table 3). Twicedaily fractionation resulted in improved local control rates. Primary tumors of the hypopharynx had poorer rates of local control than tumors of the oropharynx or supraglottic larynx. Neck stage had no apparent influence on local control. A separate multivariate analysis was performed for each mucosal site, grouped as oropharynx, hypopharynx, and supraglottic larynx. Parameters tested in these analyses were T stage, N stage, and fractionation (Table 4). For the oropharynx and supraglottic larynx, T stage was the only parameter found to significantly influence local control: N stage was not significant. For the hypopharynx, both T stage and fractionation were significant, with twicedaily fractionation associated with improved local control. The extent of neck node disease did not have a significant influence on local control. To test whether certain clinical characteristics of neck disease were predictive of local control, a multivariate analysis on node-positive patients was performed, analyzing the parameters of node location (upper neck vs lower neck plus upper neck) and node mobility (mobile or tethered vs fixed) (Table 5). Only T stage was found to significantly influence local control. The N stage, node location, and node mobility did not predict local control. Lastly, separate analyses were performed for patients treated with once-daily and twice-daily fractionation, as fractionation proved an important prognostic variable in the overall analysis. For the 368 patients treated once daily, T stage and primary site were the parameters found to influence local control, with poorer control rates for primary tumors of the hypopharynx than for lesions of the oropharynx or supraglottic larynx (Table 6). For the 239 patients treated twice daily, only T stage significantly predicted local control (Table 7). In neither group was neck stage predictive of primary tumor control. For a
Table I. Local control according to N stage (475 patients*) NO
NI
N2A-N3A
N2B-N3B
Stage
No.
(Q’)
No.
(a)
No.
(“/)
TI T2 T3 T4
25126 79/103 45 164 4111
(96) (77) (70) (36)
6/8 21/28 18/28 7/l I
(75) (75) (64) (64)
515 15/1x 12116 Of7
* Excludes
I32 patients
who died < 2 years from treatment
with primary
No.
(St)
(100)
616
(100)
(83) (75) (0)
29135 40172 13137
(83) (56) (35)
site continuously
controlled.
p values NO vs N+ .38 .34 .I2 .63
735
Does neck stage influence local control? 0 D. E. FREEMANet al.
Table 4. Multivariate analysis according to site
Table 2. Local control according to site Stage Soft palate
Tonsillar region
Base of tongue
Pyriform sinus
Pharyngeal wall
Supraglottic larynx
Tl T2 T3 T4 Tl T2 T3 T4 Tl T2 T3 T4 Tl T2 T3 T4 Tl T2 T3 T4 T1 T2 T3 T4
N+
p value
515
l/l
6/10 415 n.d. 415 23135 14/18 415 l/l 717 516 O/2 313 9110
515 5112 215 515 17/19 23137 6/18 314 14117 19123 6115 516 13/19 213 l/5 n.d. 315 8/17 2/8 212 14117 13123 315
n.d. .15 .18 n.d. .50 .I2 .19 .09 .80 .34 .73 .40 .67 .21
NO
n.d. n.d. 313 14/19 s/10 O/l 919 18/22 17125 O/3
n.d. n.d. n.d. .46 .60 .78 n.d. .65 .30 .18
NO = Clinically negative neck; Nf = Clinically positive neck nodes; n.d. = No data.
given site and T stage, patients treated with twice-daily fractionation received similar total tumor doses at a constant dose per fraction ( 120 cGy), such that there was little, if any, difference in tumor dose to the primary site for node-positive and node-negative patients ( 15). These results suggest that the lack of difference in local control rates for node-positive and node-negative patients is not a result of higher tumor doses being delivered to the primary sites for patients presenting with neck metastases.
DISCUSSION In 1988, we reported the results of a direct analysis of the influence of neck stage on local control for 526 patients with primary squamous cell carcinomas of the head and neck arising from 10 mucosal sites (11). We found no evidence to suggest that advanced neck disease adversely affected local control rates. In the current study, only mucosal sites with an intermediate or high risk of neck disease
Table 3. Multivariate analysis: Overall group (607 patients) Rank 1 2 3 4 5
Variable T stage Twice-daily fractionation Hypopharynx Oropharynx N stage
Rank
Variable
Significance
Oropharynx (324 patients)
1 2
< .Ol .26
Supraglottic larynx (137 patients)
3 1 2
Hypopharynx ( 146 patients)
3 1 2
T stage Twice-daily fractionation N stage T stage Twice-daily fractionation N stage T stage Twice-daily fractionation N stage
Site
3
.98 < .Ol .23 .67 < .Ol < .Ol .62
at presentation were included (10); patients with more than one primary lesion were excluded; and both direct and multivariate analyses were performed. Again, we found no evidence that patients presenting with clinically positive necks were at higher risk for local recurrence than patients with the same T stage with clinically negative necks. A similar conclusion was reached by Richard et neck al., ( 16) who reviewed 1,7 13 patients undergoing dissection at the Institut Gustave Roussy for squamous cell carcinoma of the head and neck. These data differ from those reported in several other series in the literature. In an analysis of 248 patients with supraglottic cancers treated with irradiation alone, Wall et al. ( 19) from the M. D. Anderson Hospital found a 15% to 30% lower local control rate for patients with palpable neck disease at diagnosis compared to patients presenting with a clinically negative neck. The risk of local failure appeared to increase as the bulk of neck disease increased. In another series from the M. D. Anderson Hospital, Ang et al. (2) reported 10% lower primary tumor control for patients with carcinomas of the oropharynx and nasopharynx presenting with neck node metastases compared to patients with a clinically negative neck. For tumors of the supraglottic larynx, Van den Bogaert et al. (18) showed a higher rate of primary tumor failure in patients presenting with nodal disease compared to patients without neck metastases (50%’ vs 30%). Hahn et al. (4) from the University of Virginia also reported a 37% primary recurrence rate for patients with lesions of the supraglottic larynx or pyriform sinus presenting with N2 or N3 necks, compared to a 19% local recurrence if neck stage was NO or Nl. Both Wall et al. and Hahn et al.
Table 5. Multivariate analysis: Patients with clinically positive neck nodes (37 1 patients)
Significance < .Ol < .Ol .Ol .86 .97
Rank
Variable
1 2 3 4
T stage Location Mobility N stage
Significance < .Ol .66 .77 .94
736
I. J.
Table 6. Multivariate
Rank
1 2 3 4
Radiation
Oncology
analysis: Once-a-day (368 patients) Variable
T stage Hypopharynx N stage Oropharynx
0 Biology
0 Physics
Volume
fractionation
Significance < .Ol < .Ol .43 .80
4. 1992
Table 7. Multivariate
analysis: Twice-daily (239 patients)
Rank
Variable
I
T stage N stage Hypopharynx Supraglottic larynx
7 ; 4
recommended higher doses of radiation or combined modality therapy for patients presenting with advanced neck disease. It is difficult to explain these conflicting data. Most of the published series, including our own, are retrospective and therefore subject to variations in patient selection and treatment parameters such as dose, fractionation, and overall treatment time. Some of the reported series did not stratify patients by T stage or grouped T stages (TlT2 vs T3-T4) when analyzing local control for a particular neck stage (4, 16, 19). Given that T stage was the most important predictor of local control in all of the patient groups in this study, it is likely that bias could have been introduced into local control comparisons if patients were not stratified by both T stage and N stage. Also, methods used to calculate local control vary between institutions, which hampers direct comparison of reported data ( 14).
23. Number
fractionation
Significance < .Ol .30 .80 .96
For head and neck carcinoma treated with irradiation alone, we determine the dose to the primary lesion based on T stage and primary site. Within a particular T stage, more advanced lesions receive slightly higher doses than tumors of lesser volume. Involved lymph nodes that fall within the primary irradiation field receive a tumor dose equivalent to, or slightly higher than, the dose received by the primary tumor. Large nodes falling outside or only partly within the primary volume receive additional treatment (a boost dose) with opposed, anterior and posterior, wedged fields that do not deliver further irradiation to the primary site. As we found no evidence that patients presenting with neck node metastases are at higher risk for local recurrence than patients of the same T stage with clinically negative necks, we see no indication for more aggressive treatment to the primary site based solely on the extent of neck disease at diagnosis.
REFERENCES 1. American Joint Committee on Cancer. Manual for staging of cancer, 2nd edition. Philadelphia, PA: J. B. Lippincott Co.; 1983:3 l-42. 2. Ang, K. K.; Peters, L. J.; Weber, R. S.; Maor, M. H.; Morrison, W. H.; Wendt, C. D.; Brown, B. W. Concomitant boost radiotherapy schedules in the treatment of carcinoma of the oropharynx and nasopharynx. Int. J. Radiat. Oncol. Biol. Phys. 19: 1339-I 345; 1990. 3. Colton, T. Statistics in medicine. Boston: Little Brown & Co.; 1974:163-167. 4. Hahn, S. S.; Spaulding, C. A.; Kim, J.-A.; Constable, W. C. The prognostic significance of lymph node involvement in pyriform sinus and supraglottic cancers. Int. J. Radiat. Oncol. Biol. Phys. 13:1143-l 147;1987. 5. Kalbfleisch, J. D.; Prentice, R. L. The statistical analysis of failure time data. New York: Wiley & Sons: 1980. estimation from 6. Kaplan, E. L.; Meier, P. Nonparametric incomplete observations. J. Am. Stat. Assoc. 53:457481;1958. 7. Kramer, S.; Martial, V. A.; Pajak, T. F.: MacLean, C. J.: Davis, L. W. Prognostic factors for loco/regional control and metastasis and the impact on survival. Int. J. Radiat. Oncol. Biol. Phys. 12:573-578:1986. 8. Marks, J. E.; Freeman, R. B.; Lee, F.; Ogura. J. H. Carcinoma of the supraglottic larynx. AJR 132:255-260: 1979. 9. Mendenhall, W.; Ott, L.; Larson, R. F. Statistics: a tool for the social sciences. North Scituate, MA: Duxbury Press: 1974:333-336. 10. Mendenhall, W. M.; Million, R. R. Elective neck irradiation for squamous cell carcinoma of the head and neck: analysis of time-dose factors and causes of failure. Int. J. Radiat. Oncol. Biol. Phys. 12:741-746;1986. 11. Mendenhall, W. M.; Parsons, J. T.; Amdur, R. J.; Spangler.
12.
13.
14.
15.
16.
17. 18.
19.
A. E.: Williams, T. R.; Cassisi, N. J.; Million, R. R. Squamous cell carcinoma of the head and neck treated with radiation therapy: the impact of neck stage on local control. Int. J. Radiat. Oncol. Biol. Phys. 14:249-252;1988. Million, R. R.; Cassisi, N. J. Management of head and neck cancer: A multidisciplinary approach. Philadelphia, PA: J. B. Lippincott Co.: 1984:299-364, 373-39 1. Overgaard, J.; Hansen, H. S.; Jorgensen, K.; Hansen, M. H. Primary radiotherapy of larynx and pharynx carcinoma-an analysis of some factors influencing local control and survival. Int. J. Radiat. Oncol. Biol. Phys. 12:5 15521:1986. Parsons, J. T.; McCarty, P. J.; Rao, P. V.; Mendenhall, W. M.; Million, R. R. On the definition of local control. Int. J. Radiat. Oncol. Biol. Phys. 18:705-706;1990. Parsons, J. T.; Mendenhall, W. M.; Cassisi, N. J.: Issacs, J. H.; Million, R. R. Hyperfractionation for head and neck cancer. Int. J. Radiat. Oncol. Biol. Phys. 14:649-658;1988. Richard, J. M.; Sancho-Gamier, H.; Micheau, C.; Saravane, D.; Cachin, Y. Prognostic factors in cervical lymph node metastasis in upper respiratory and digestive tract carcinomas: study of 1,7 13 cases during a 15-year period. Laryngoscope 97:97-101;1987. SAS user’s guide: Statistics, version 5 edition. Gary. NC: SAS Institute; 1985:529-557. Van den Bogaert, W.; Ostyn, F.; van der Schueren, E. The different clinical presentation, behaviour and prognosis of carcinomas originating in the epilarynx and the lower supraglottis. Radiother. Oncol. 1: 117- 131; 1983. Wall, T. J.; Peters, L. J.; Brown, B. W.; Oswald, M. J.; Milas, L. Relationship between lymph nodal status and primary tumor control probability in tumors of the supraglottic larynx. Int. J. Radiat. Oncol. Biol. Phys. 11: 1895-1902;1985.
