Radiotherapy for patients with early-stage glottic carcinoma

Univariate and multivariate analyses in a group of consecutive, unselected patients

Abstract

BACKGROUND

Radiotherapy (RT) has a remarkable success rate in the treatment of patients with glottic carcinoma. The objectives of the current study were to assess the results in a group of consecutive patients with comparable characteristics who were treated with RT (6-megavolt photon linear accelerator) and to determine the prognostic factors that may influence local control in patients with early-stage glottic carcinoma. The impact on local control of tobacco smoking and second primary malignancies also was investigated.

METHODS

Four hundred ten patients with T1–T2 squamous cell carcinoma of the glottis who were treated between 1986 and 2001 were analyzed retrospectively with regard to local control and overall survival. Potential prognostic factors for local control were evaluated with univariate and multivariate models. The impact of technologic advances also was evaluated.

RESULTS

The 5-year and 10-year overall survival rates were 83% and 63.5%, respectively. The overall 10-year local control rate for patients with T1–T2 glottic carcinoma was 89%. The median time to recurrence was 7 months. Univariate analysis showed that tumor category, tumor size, macroscopic appearance of the lesion, RT fraction size, persistent edema, year of RT treatment, unchanged dysphonia, and surgical option all had a significant influence on local control; whereas multivariate analysis showed that only persistent dysphonia and year of RT treatment were significantly associated with increased local control. A 22.2% rate of second primary malignancies was reported: second primary tumors were the major cause of death in the patients studied. Only 2 patients died of laryngeal carcinoma; 304 patients were alive with their disease in complete remission, 1 patient was alive with recurrent laryngeal carcinoma after undergoing salvage surgery, and 103 patients died of either intercurrent disease or a second primary tumor.

CONCLUSIONS

The use of a 6-megavolt photon linear accelerator achieved a high rate of local control in patients with T1–T2 glottic carcinoma. Dysphonia and the year of RT treatment were the most important prognostically significant factors for patient outcome. The occurrence of a second primary tumor was the most frequent cause of death, especially among patients who did not stop smoking after a diagnosis of glottic carcinoma. Cancer 2003;98:765–72. © 2003 American Cancer Society.

DOI 10.1002/cncr.11575

Stage I–II squamous cell carcinoma of the glottis (SCCG) is curable by either surgery or radiotherapy (RT). The advantages of RT in terms of preservation of structure and function of the larynx have been documented extensively, and RT currently is the treatment of choice in most institutions.1–7 Comparable results in terms of control rates can be obtained with laser surgery in selected patients, but that procedure may result in a worse quality of voice.8, 9

In Italy, the guidelines in current use for the treatment of patients with T1 and T2 SCCG were developed in 1999 by the Italian National Research Council (CNR). The CNR guidelines indicate that both surgical options and RT options are valid alternatives and defer the choice between those two options to the experience and policy of the institution to which patients are referred for treatment.10

At the Aviano Institute, patients with Stage I–II SCCG are evaluated for treatment by the relevant Head and Neck Committee. It has been our policy to use RT for the treatment of patients with Stage I–II SCCG, excluding patients with lesions that are suitable for subligamentous cordectomy (type II cordectomy).11 Specifically, patients with glottic carcinoma in situ and with superficial infiltrating carcinoma of the vocal cord underwent laser cordectomy, whereas patients with a histologic diagnosis of infiltrative Stage I–II SCCG were given RT as first-choice treatment. In the RT-treated group, surgery was reserved for the management of persistent disease at the end of treatment and/or disease recurrences. However, before initiation of RT, the patients were given an unbiased explanation of both the advantages and the disadvantages of RT and surgical alternatives, with the patients ultimately deciding on therapy. All patients decided in favor of RT.

The objective of the current study was to evaluate local control and the rate of preservation of laryngeal function in a group of consecutive patients with Stage I–II SCCG who were treated with a 6-megavolt (MV) photon linear accelerator and were followed for a long period.The current study also examined 1) the prognostic factors that may have an impact on the outcome of patients after RT; 2) the impact on local control of pretreatment and posttreatment smoking habits; 3) the risk of developing a second primary tumor and its impact on survival; and 4) the impact of technologic developments on the results of RT treatment.

MATERIALS AND METHODS

A retrospective review of 410 patients who were treated with radical RT to the larynx for Stage I–II SCCG (T1–T2,N0 according to the 2002 International Union Against Cancer classification system) was performed.12 Patients were treated at the Radiotherapy Division of the Centro di Riferimento Oncologico Institute, Aviano, between June 1985 and December 2001. All received RT as first-choice treatment.

Patient inclusion criteria were a histologic diagnosis of infiltrative squamous cell carcinoma and no previous RT for head and neck neoplasms. In all patients, a biopsy specimen was obtained through microlaryngoscopy. Staging procedures included a laryngeal computed tomography (CT) scan, bronchoscopy, and upper endoscopy. Twenty-eight patients (6.8%) were women, and 382 patients (93.2%) were men. The median age was 61 years (range, 31–85 years); 151 patients were older than age 70 years, and 123 patients were younger than age 60 years. Three hundred eleven patients had a Karnofsky performance status (PS) of 90, and only 40 patients had a PS ≤ 70. Disease was staged retrospectively according to the 1997 criteria of the American Joint Commission on Cancer. Table 1 summarizes the clinical characteristics of the 410 patients who were included in the study. Seventy patients had tumors classified as T2a, and 17 patients had tumors classified as T2b because of impaired vocal cord mobility. Sixty-six patients had a histologic Grade 3 tumors according to the World Health Organization grading system. In Table 1, anterior commissure includes both primary tumors that originated at this site and tumors that originated in the anterior one-third of the vocal cord but involved the adjacent commissure.

Table 1. Characteristics of the 410 Patients Studied

Characteristic	No. of patients (%)
^a Vc: vocal cord.
Tumor category
T1a	267 (65.0)
T1b	56 (13.7)
T2a (no impaired Vc mobility)	70 (17.1)
T2b (impaired Vc mobility)	17 (4.1)
Tumor site
Anterior Vc	99 (23.4)
Median Vc	170 (41.9)
Posterior Vc	17 (4.2)
Anterior commissure	124 (30.5)
Tumor grade
Grade 1	60 (17.4)
Grade 2	214 (62.9)
Grade 3	66 (19.4)
Unknown	70 (—)
Tumor size
< 1 cm	213 (51.9)
≥ 1 cm	197 (48.1)
Macroscopic appearance
Exophytic	259 (62.9)
Infiltrative	151 (37.1)

The microscopic appearance and size of tumors were recorded at each patient's first clinical examination. One hundred fifty-one patients (37.1%) had infiltrative lesions, and 197 patients had lesions with a greatest dimension > 1 cm.

CT scans of the larynx were obtained from all patients. The CT scan findings were negative for laryngeal glottic carcinoma in 210 patients, positive in 172 patients, and dubious in 28 patients. The dubious scans were excluded from the statistical analysis of prognostic factors.

Before the start of RT, the Head and Neck Committee assigned patients to the surgical procedure that would have been required if the initial treatment had been an operation on the basis of local tumor extent and patient PS. According to the committee evaluation, a subtotal reconstructive laryngectomy was the advisable procedure for patients with large glottic tumors who had either supraglottic extension or impaired vocal cord mobility, when their cardiopulmonary status allowed of it. A total laryngectomy was the advisable surgical procedure in patients with a poor PS and/or cardiovascular insufficiency and/or deep subglottic extension. A cordectomy would have been required in 219 patients, more extended procedures would have been required in 151 patients, and a total laryngectomy would have been required in 40 patients.

Smoking habits were recorded, and 85 patients (20.7%) had never smoked or had stopped smoking 12 years or more before the diagnosis of glottic carcinoma, 195 patients stopped smoking at the time of diagnosis, and 130 patients (31.7%) continued smoking after diagnosis. All patients were followed until death or for a minimum of 8 months (range, 8–198 months).

Treatment

Patients were treated with a continuous course of RT delivered by a 6 MV linear accelerator after either a traditional simulation (patients treated before 1996) or a CT simulation (patients treated from 1996 onward). During treatment, all patients were immobilized using a thermoplastic mask.

A portal film verification was obtained during the first week of treatment and at any time changes occurred (such as edema of the neck) in the baseline parameters thereafter. The isodose distribution was tailored to the individual patient. For patients with Stage I disease, the inferior and superior borders of the radiation field extended from the bottom of the cricoid cartilage to the superior thyroid notch. The planning target was included at 95% of the dose indicated by the International Commission on Radiation Units and Measurements. Among patients with Stage I disease, 70.7% were treated with a 5.0-cm port, with the anterior border 1.5–2.0 cm in air. When the anterior commissure was involved, a bolus was employed to cover the prethyroid space adequately. Only in patients with reduced vocal cord mobility were lymph node levels II and III included in the RT field up to a dose of 54 grays (Gy); the remaining 1125–1600 centigrays (cGy) were given as a boost to the larynx.

Until June 1990, patients with Stage I disease were treated with an isocentric, 4-field technique (2 laterolateral fields and 2 anterior oblique fields, with or without wedge filters) for a total dose of 66.0–68.4 Gy given in 33–38 fractions over a period of 50–56 days (Fig. 1). From July 1990 onward, the dose per fraction was increased by up to 225 cGy, to a total dose of 63.0–65.2 Gy in 28–29 fractions over 38–39 days. The 95% isodose curve was tangential to the anterior arythenoid cartilage limit. The margin fell off anteriorly. Patients with Stage II disease and normal vocal cord mobility were treated with laterolateral fields and wedge filters when indicated. After RT, patients were evaluated at 2-month intervals for the first year, at 3-month intervals during the second year, and every 4–6 months thereafter.

Figure 1. Graphic illustrations of radiotherapy modality treatment (top) and isodose distribution (bottom).

Statistical Analysis

Overall survival (OS) and disease-free survival (DFS) were assessed from the end of RT until evidence of recurrence or until death. Survival curves were calculated using the method of Kaplan and Meier,13 and log-rank tests were used to determine differences between groups.14 Univariate and multivariate survival and DFS analyses were performed using Cox proportional hazards models with corresponding 95% confidence intervals.15 Covariates to be tested in the multivariate analysis were selected among factors that were significant in the univariate analysis. The results were considered statistically significant at the level of P ≤ 0.05 (two-tailed P value). The variables analyzed included patient characteristics (age, PS, T category, histologic grade, tumor site, tumor size, macroscopic appearance of the lesion, smoking and drinking habits, and positive or negative pre-RT CT scan findings), treatment characteristics (RT fraction size, time between diagnosis and RT, year of RT treatment, persistence of edema and/or dysphonia post-RT, and surgical option), and second primary tumor.

RESULTS

Five-year and 10-year OS and DFS rates by T category (T1 vs. T2) are shown in Figures 2 and 3. Sixty-six percent of patients with T1 tumors and 55% of patients with T2 tumors survived at 10 years with 90% and 78% probability of local control, respectively. All patients underwent the full course of RT. The median follow-up for all 410 patients was 63 months (range, 8–193 months).

Figure 2. Overall survival of 410 patients with laryngeal carcinoma according to tumor status. Solid line: T1 (n = 323); dashed line: T2 (n = 87).

Figure 3. Specific disease-free survival of 410 patients with laryngeal carcinoma according to tumor status. Solid line: T1 (n = 323); dashed line: T2 (n = 87).

The median time between a histologic diagnosis of SCCG and the start of RT was 37 days (range, 7–91 days). One hundred forty-two patients started RT ≤ 31 days after undergoing laryngeal biopsy, 135 patients began RT 32–44 days after undergoing laryngeal biopsy, and 133 patients started RT 45 days after undergoing laryngeal biopsy.

Local Control

The first evaluation of response to treatment was performed 45 days after the end of RT. Three hundred ninety-eight patients achieved a complete remission (49 were pathologically confirmed complete responses), and 12 patients (3.4%) had locally persistent disease. Five of those 12 patients underwent salvage surgery with total laryngectomy (3 patients with T1a tumors and 2 patients with T2a tumors), and the other 7 patients were treated with a conservative approach (3 patients with T1a tumors and 4 patients with T2a tumors).

Thirty-three patients developed recurrent disease at the primary site, and 2 patients had lymph node recurrences on the neck; the median time to local recurrence was 7 months (range, 2.5–36.0 months). Two patients had lymph node recurrences 11 months and 13 months after the end of RT, respectively; both patients underwent radical neck dissection. One of those 2 patients is alive in complete remission, and the other died in complete remission of a second primary tumor (lung). All but one patient with local persistence of disease or disease recurrence underwent salvage surgery, and all but two patients benefited from the procedure. Table 2 summarizes the characteristics of the patients who developed recurrent disease and provides details of the salvage surgical procedures employed.

Table 2. Tumor Category, Salvage Procedure, and Outcome in 47 Patients with Persistent/Recurrent Glottic Carcinoma

Characteristic	No. of patients
a One patient was alive with disease in the larynx. b Two patients died of laryngeal carcinoma.
Tumor category
T1a	25
T1b	5
T2a	15
T2b	2
Salvage surgery
Conservative laryngectomy	26
Total laryngectomy	18
Radical neck dissection	2
No surgery	1
Outcome
Alive	26a
Dead	21b

Survival

Two patients died of laryngeal carcinoma. One patient died of an extensive subglottic recurrence 11 months after the end of RT, and the other patient died of a local recurrence after undergoing salvage surgery 63 months after the end of RT. At the time of data analysis, one patient who developed a local recurrence after undergoing salvage surgery (total laryngectomy) was receiving chemotherapy. One hundred three patients died of causes unrelated to laryngeal carcinoma: Forty-seven patients died from intercurrent disease, and 56 patients died from a second primary tumor. No patients had clinical evidence of recurrence in the glottic site.

Ninety-one patients developed a secondary primary tumor: 37 in the lung, 19 in the head and neck region, 6 in the esophagus, 10 in the gastrointestinal tract, and 11 in the genitourinary region; 5 patients developed lymphoma, 1 patient developed a melanoma, and 2 patients developed skin carcinoma. Only 33 of those 91 patients are alive and without evidence of disease in the larynx. Figure 4 reports OS according to second primary tumors. Patients who developed a second primary tumor had a 32% 10-year survival rate compared with a 77% 10-year survival rate in patients who were without a second primary tumor.

Figure 4. Survival of 410 patients with early-stage laryngeal carcinoma according to whether patients ever had a second tumor. Solid line: never had a second tumor (n = 319); dashed line: had a second tumor (n = 91).

Three-hundred twenty-five patients (79.2%) were smokers, and only 85 patients had never smoked or had stopped smoking 12 years or more before their diagnosis of laryngeal carcinoma. One hundred thirty patients continued smoking during RT; of these, 74 patients are alive (57 patients are in complete remission, and 17 patients have local recurrences that require salvage surgery), and 56 patients have died: Thirty-one patients died as a consequence of the second primary tumor, 24 patients died of intercurrent disease, and 1 patient died of local recurrence. Of 130 patients who continued smoking, 45 patients (34%) developed a second primary tumor, compared with only 16.4% of patients who never smoked or who stopped smoking after diagnosis.

After completing RT, 20 patients were required to undergo total laryngectomy: as salvage therapy after a local recurrence in 18 patients and for chronic edema in 2 patients who were in complete remission. The organ preservation rate was 50%.

RT was tolerated well, with only 13 patients experiencing Grade 3 toxicity (according to the World Health Organization criteria) and only 24 patients requiring an interruption of treatment (range, 7–12 days). One hundred thirty patients (31%) had laryngeal edema at 2 months after completing RT; and, in 58 of those 130 patients (14.1%), the edema persisted for as long as 6 months. In 323 patients, there was an improvement in dysphonia compared with their pre-RT status; dysphonia was stabilized in 72 patients (17.5%), and only 10 patients experienced a worsening of voice quality after RT.

Before 1996, RT dose distribution was calculated manually at the central axis of four angular, coplanar beams. January 1996 represents the date that a CT simulator and a treatment planning system became available at our department. With the introduction of these technologies, the dose distribution to the tumor was calculated on more than 1 section (range, 5–7 sections), and the fields were tailored individually on the beam eye view of the planning target volume outlined by the radiation oncologists on CT slices (183 of the 410 patients in the current series).

Univariate Analysis

The variables patient age, Karnofsky PS, tumor site, T category, histologic grade, tumor size, macroscopic appearance of the lesion, positive findings at CT scan, RT fraction size, persistence of edema, improvement of dysphonia, surgical option, smoking habits before and after RT, time between diagnosis and RT, and year of RT treatment were tested for all 410 patients. T category (P = 0.02), tumor size (P = 0.002), macroscopic appearance of the lesion (P = 0.03), RT fraction size(P = 0.01), edema at 2 months and 6 months (P = 0.02 and P = 0.08, respectively), improvement of dysphonia (P = 0.001), positive CT scan findings (P = 0.0002), surgical option (P = 0.001), and year of RT treatment (P = 0.002) were statistically significant variables and significantly influenced local control (Table 3).

Table 3. Univariate Analysis: Disease-Free Survival

Factor	No. of patients (%)	HR (95% CI)a	Chi-square (df)	P value
HR: hazard ratio; 95% CI: 95% confidence interval; df: degrees of freedom; RT: radiotherapy; cGy: centigrays; CT: computed tomography. a Cox proportional hazards model. b Reference category.
Tumor category
T1a	267 (65.1)	1b	5.58 (1)	0.02
T1b	56 (13.7)	1.0 (0.4–2.6)	—	—
T2a	70 (17.1)	2.6 (1.4–5.0)	—	—
T2b	17 (4.1)	1.3 (0.3–5.7)	—	—
Tumor size
< 1 cm	213 (52.0)	1b	9.84 (1)	0.002
≥ 1 cm	197 (48.0)	2.7 (1.5–5.1)	—	—
Macroscopic appearance
Exophytic	258 (62.9)	1b	4.77 (1)	0.03
Infiltrative	152 (37.1)	1.9 (1.1–3.4)	—	—
RT fraction size
≥ 225 cGy	297 (72.4)	1b	6.49 (1)	0.01
< 225 cGy	113 (27.6)	2.1 (1.2–3.8)	—	—
Edema at 2 mos
Never	280 (68.3)	1b	5.30 (1)	0.02
Ever	130 (31.7)	2.0 (1.1–3.5)	—	—
Edema at 6 mos
Never	352 (85.9)	1b	3.16 (1)	0.08
Ever	58 (14.1)	1.8 (0.9–3.6)	—	—
Dysphonia
Improved	323 (79.8)	1b	46.49 (1)	≤ 0.001
Unchanged	82 (20.2)	7.9 (4.4–14.3)	—	—
Unknown	5	—	—	—
CT scan findings
Negative	210 (55.0)	1b	13.77 (1)	0.0002
Positive	172 (45.0)	3.4 (1.8–6.5)	—	—
Dubious	28	—	—	—
Surgical options
Cordectomy	219 (59.0)	1b	10.22 (1)	0.001
Other	152 (41.0)	3.0 (1.5–5.8)	—	—
Unknown	39	—	—	—
Y of RT treatment
≥ 1996	183 (44.6)	1b	9.80 (1)	0.002
< 1996	227 (55.4)	3.2 (1.5–6.6)	—	—

Multivariate Analysis

A multivariate analysis was performed on the statistically significant variables that were identified in the univariate analysis. Among the variables tested, only no change in the status of dysphonia and year of RT treatment significantly affected local control. The macroscopic appearance of the lesion had only a modest impact on local control (P = 0.07) (Table 4).

Table 4. Multivariate Analysis of Prognostic Factors

Factor	HR (95% CI)a	Chi-square (df)	P value
HR: hazard ratio; 95% CI: 95% confidence interval; df: degrees of freedom; RT: radiotherapy. a Cox proportional hazards model. b Reference category.
Macroscopic appearance
Exophytic	1b	3.35 (1)	0.07
Infiltrative	2.2 (0.9–5.0)	—	—
Dysphonia
Improved	1b	34.88 (1)	≤ 0.001
Unchanged	11.4 (5.1–25.6)	—	—
Y of RT treatment
≥ 1996	1b	10.93 (1)	0.001
< 1996	5.4 (2.0–14.7)	—	—

DISCUSSION

In the current series of 410 patients with T1–T2 SCCG who received RT at our institute as first-choice treatment and who were followed for a long period, excellent local control rates were achieved. The rates reported in the current study were similar to the excellent rates (as high as 89%) reported in similar, previously published series.16–22

At our institute, no changes in the RT approach for patients with T1 and T2 SCCG have been introduced for the last 16 years, although there has been a change in RT dose. Specifically, the dose per fraction (180–200 cGy) was increased to 225 cGy starting in 1990. Since 1996, however, RT treatments have been tailored individually according to tumor extension as well as laryngeal and patient anatomy by means of a CT simulator and a computerized treatment planning system that generates treatment plans from a minimum of five CT slices. All patients were treated using a thermoplastic mask as an immobilization device.

Univariate and multivariate analyses showed that patients who were treated before 1996 had hazard ratios for the probability of local recurrence of 3.2 and 5.4, respectively. The comparison between patients who were treated before 1996 (hazard ratio, 1) and patients who were treated from 1996 onward was highly significant (P = 0.001). These data are consistent with reports by other authors23 and confirm the value of a state-of-the-art technology that has proven ability to increase local control significantly in patients with early-stage SCCG.

The univariate analysis showed that local control was affected adversely by T category, macroscopic appearance of the lesion at diagnosis, tumor size, and (to a greater extent) tumor size (hazard ratio for a greatest tumor dimension ≥ 2.7, P = 0.002). In contrast, the multivariate analysis showed that only exophytic lesions versus infiltrative lesions had borderline significance (hazard ratio, 2.2; P = 0.07); whereas bulky tumor and involvement of the anterior commissure had no impact on local control. The results of our univariate and multivariate analyses conflict with those reported from studies on similar patient series24–26 and provide evidence that RT is a valid therapeutic option for patients with early-stage glottic carcinoma irrespective of the characteristics of the lesion.

Pretreatment CT scan results have been identified as a predictor of local control in patients with SCCG who are treated with RT.27 We found that this parameter was highly predictive of local control on univariate analysis (hazard ratio, 3.4; P = 0.002) but not on multivariate analysis, which is in disagreement with reports by others.28

The 410 patients who were included in the current analysis were treated using a 6 MV photon linear accelerator. In agreement with other reports, this energy achieved better results compared with lower energies (e.g., 4 MV) and cobalt 60.24 The patients who were treated after 1990 received a dose of 225 cGy per fraction, which led to a significant improvement in local control (hazard ratio, 2.1; P = 0.01), although the multivariate analysis did not confirm this improvement (P = 0.96). In addition, in agreement with the results published by other authors,29 the interval between diagnosis and the start of RT (< 32 days, 32–34 days, or ≥ 45 days) did not have an adverse effect on outcome among the patients in our series (P = 0.39).

The persistence of edema at 2 months and 6 months after the end of RT and the status of dysphonia were tested in the univariate analysis. The presence of edema at 2 months after the end of RT and an unchanged quality of voice were correlated with poor local control. The persistence of dysphonia maintained prognostic significance for local control (0.001) in multivariate analysis. These results indicate that patients with persistent dysphonia should be followed closely and should be evaluated with a CT scan and/or nuclear magnetic resonance imaging whenever possible. The subgroup of patients who were eligible to undergo cordectomy had better local control (hazard ratio, 3.0; P = 0.001), as reported in the study by Lee et al.24 however, multivariate analysis did not confirm this result.

The great majority of our patients were heavy smokers (80%). Only 85 patients had never smoked or had stopped smoking 12 years before the diagnosis of SCCG. Although 130 patients continued smoking during RT, this did not appear to influence the outcome of the treatment (P = 0.56).

During follow-up, 91 patients (22.2%) developed a second primary tumor (mostly sited in the upper aerodigestive tract). Similar rates have been reported in other studies;30, 31 therefore, a correlation can be established between the smoking habit and a second primary tumor. Indeed, 45 patients (50.5%) who were heavy smokers developed a second primary tumor, which led to death in 31 patients (32%).

Of 410 patients with SCCG who were included in our study, only 17 (4%) were diagnosed with T2b tumors, which is a small number compared with similar cohorts of patients.32 It appears that this was the result of the sensibilization campaign carried out by our Head and Neck Committee among local general practitioners, who were invited to refer to the committee all smoking patients with dysphonia symptoms whether or not the patients sought medical aid. In addition, of the 17 patients with T2b tumors, only 2 (11%) experienced disease recurrence, and both patients underwent salvage surgery (total laryngectomy).These data emphasize the need for an educational program to induce lifestyle changes in our patients in favor of abandoning cigarette smoking and alcohol consumption.

In summary, RT is a satisfactory method of treatment for patients with early-stage SCCG. In fact, RT achieved an overall local control rate as high as 89% and achieved good preservation of voice quality. In addition, patients who fail to achieve tumor control with RT still may undergo salvage surgery.

In our experience, there was no correlation between patient characteristics and tumor characteristics. Factors that had a significant influence on local control were persistent dysphonia after RT, accuracy of and individually tailored treatments, year of RT, and (marginally) the clinical appearance of the lesion. The survival of these patients was influenced strongly by the occurrence of second primary tumors, which were more frequent in patients who continued smoking after diagnosis. In the current study, 61.5% of patients who smoked and developed a second primary tumor died, compared with only 14.7 of nonsmokers.

Currently, no major developments may be expected in RT technologies that may improve on the current result in patients with T1 SCCG. However, for patients with T2 SCCG, more sophisticated treatments, such as intensity-modulated RT, may improve local control by delivering higher doses without increasing toxicity. We conclude that promoting a healthier lifestyle and checking at regular intervals for possible second primary tumors may have a favorable impact on the survival of patients with T1–T2 SCCG.

Acknowledgements

The authors thank Daniela Michilin for her careful assistance with the preparation and translation of the article.