Retromolar trigone squamous cell carcinoma treated with radiotherapy alone or combined with surgery

Authors


Abstract

BACKGROUND

Treatment outcomes were analyzed for retromolar trigone squamous cell carcinoma.

METHODS

Between June 1966 and August 2003, 99 patients were treated with radiotherapy alone (35 patients) or radiotherapy combined with surgery (64 patients). Followup ranged from 0.2 to 23.8 years (median, 3.3 yrs). All living patients had followup for at least 1 year.

RESULTS

The 5-year local–regional control rates after definitive radiotherapy versus surgery and radiotherapy were as follows: Stages I–III, 51% and 87%; Stage IV, 42% and 62%; and overall, 48% and 71%, respectively. The 5-year cause-specific survival rates after definitive radiotherapy compared with surgery and radiotherapy were as follows: Stages I–III, 56% and 83%; Stage IV, 50% and 61%; and overall, 52% and 69%, respectively. Multivariate analyses revealed that the likelihood of cure was better with surgery and radiotherapy compared with definitive radiotherapy.

CONCLUSIONS

The likelihood of cure after treatment for retromolar trigone squamous cell carcinoma was influenced by the extent of disease and treatment. Patients treated with surgery and radiotherapy had a better outcome than those treated with radiotherapy alone. Cancer 2005. © 2005 American Cancer Society.

Squamous cell carcinoma of the retromolar trigone is relatively uncommon. Patients usually have a long history of tobacco abuse, often combined with heavy ethanol consumption.1 The most common complaints at diagnosis are pain and trismus; the pain is either local or referred to the ear.1 The tumor frequently extends to adjacent sites before it is diagnosed. Byers et al.1 reported on a series of 110 previously untreated patients who were treated with surgery and/or radiotherapy (RT) at the M. D. Anderson Cancer Center (Houston, Texas). Extension to the following adjacent sites before treatment was observed: soft palate, 65 patients (59%); buccal mucosa, 47 patients (43%); anterior tonsillar pillar, 93 patients (85%); and mandibular gingiva, 67 patients (61%). Radiographic evidence of bone invasion was present in 17 patients (15%). Involvement of the regional lymphatics is less often observed than it is for tonsillar or tongue base carcinomas. In Byers et al.,1 33 of 110 patients (30%) had clinically positive nodes at diagnosis. The most commonly involved lymph nodes are in Level II.1 No patient in the M. D. Anderson Cancer Center series presented with bilateral involved nodes and only 3 patients (3%) had multiple levels involved in the ipsilateral neck.1

The treatment of retromolar trigone carcinomas is controversial.2 Patients with early-stage lesions are usually treated with a single modality (either surgery or RT), whereas those with advanced disease are frequently treated with surgery and adjuvant RT. The optimal treatment is often debatable because there are few reported outcomes data and, at times, retromolar trigone cancers are grouped with tumors arising from adjacent sites, such as the anterior tonsillar pillar.3, 4

The purpose of this article is to report outcomes of patients treated with curative intent at the University of Florida with RT alone or RT combined with surgery.

MATERIALS AND METHODS

Between June 1966 and August 2003, 99 previously untreated patients were treated with curative intent at the University of Florida with RT alone or RT combined with surgery. Followup ranged from 0.2 to 22.8 years (median, 3.3 yrs). Followup on living patients ranged from 1.1 to 14.2 years (median, 4.8 yrs). No patient was lost to followup.

Thirty-five patients were treated with definitive RT alone (33 patients) or combined with a planned neck dissection (2 patients). Patients were treated with surgery and adjuvant preoperative RT (9 patients) or postoperative RT (55 patients), composing a total of 64 patients. In general, patients with early-stage tumors were treated with a single modality, whereas those with advanced disease were more likely to undergo surgery and adjuvant RT. Surgery was the preferred treatment for most patients with T1–T2 tumors unless the patient was a poor surgical candidate or declined surgery. A small subset of patients with T1–T2N0 lesions received surgery alone but are not included in this report because it was not possible to reliably identify them and access their medical records. Definitive RT was employed for the occasional patient with advanced disease in whom the tumor was thought to be incompletely resectable or if the medical condition of the patient precluded surgery.

The median RT dose was as follows: RT alone, 70 Gy (range, 57–82 Gy); preoperative RT, 46 Gy (range, 33–50 Gy); and postoperative RT, 65 Gy (range, 22–75 Gy). In general, the dose of postoperative RT varied with margin status: negative margins, 60–65 Gy; microscopically positive margins, 65–70 Gy; and macroscopic residual disease, 70–75 Gy. Once-daily fractionation was used on 79 patients, and twice-daily fractionation was used on 20 patients. Continuous-course RT was used to treat 89 patients, 9 patients were irradiated with planned split-course technique, and 1 patient received unplanned split-course RT. Planned split-course RT was used between 1970 and 1974 and has since been abandoned. Seven patients who received definitive RT received either induction (1 patient) or concomitant (6 patients) adjuvant platinum-based chemotherapy. Induction chemotherapy has since been abandoned. Six patients who received adjuvant chemotherapy were in the definitive RT group, and 1 patient was in the surgery and adjuvant RT group.

Patients were staged according to the 2002 American Joint Committee on Cancer (AJCC)5 staging system (Table 1). The current study is part of an Institutional Review Board-approved Radiation Oncology Outcome Tracking Project (RADTRAC).

Table 1. Patient Population
ParameterDefinitive RT no. (%)Surgery + RT no. (%)Total no. (%)
No. Patients35 (35)64 (65)99 (100)
Gender   
 Male27 (77)48 (75)75 (76)
 Female8 (23)16 (25)24 (24)
Race   
 White29 (83)56 (88)85 (86)
 African American4 (11)6 (9)10 (10)
 Other2 (6)2 (3)4 (4)
T Stage   
 T17 (20)3 (5)10 (10)
 T215 (43)17 (27)32 (32)
 T34 (11)11 (17)15 (15)
 T49 (26)33 (52)42 (43)
N Stage   
 N020 (57)35 (55)55 (56)
 N110 (29)12 (19)22 (22)
 N2A2 (6)1 (2)3 (3)
 N2B2 (6)15 (23)17 (17)
 N31 (3)1 (2)2 (2)
Overall Stage   
 I5 (14)2 (3)7 (7)
 II11 (31)9 (14)20 (20)
 III7 (20)14 (22)21 (21)
 IVA8 (23)38 (59)46 (47)
 IVB4 (11)1 (2)5 (5)

Complications were coded as severe if they resulted in an unplanned treatment split during RT, necessitated hospitalization, or surgical intervention, or resulted in death.6

All statistical calculations were carried out using SAS statistical software (SAS Institute, Cary, North Carolina).7 The probabilities of local control, local–regional control, distant metastasis-free survival, cause-specific survival, and overall survival were calculated by the Kaplan–Meier product-limit method.8 The log rank test statistic was used to determine statistical significance between strata of selected explanatory variables. Multivariate analyses were performed using Cox regression.9 Parameters included in multivariate analyses included: T stage, (T1–T2 vs. T3–T4), N stage (N0 vs. N1–N3), overall stage (I–II vs. III–IV), treatment group (RT alone or with a neck dissection vs. surgery and adjuvant RT), gender, and ethnicity (African American vs. Other). We have previously observed that African Americans treated with definitive twice-daily RT have a worse prognosis compared with similarly treated patients of other ethnic groups.10 Although performance status is probably a prognostic factor, it was not reliably documented in many patients, and, thus, was not included. Patient age is likely not a prognostic factor and was not included in this analysis.11

RESULTS

Local Control

The 5-year local control rates are depicted in Table 2. Multivariate analysis of local control revealed only treatment group (P = 0.0031) to be a significant predictor of this endpoint. Patients treated with surgery and RT had a better local control rate than those treated with RT alone.

Table 2. Local Control
T StageDefinitive radiotherapySurgery + RadiotherapyTotal
No. patientsNo. local recurrencesPercentage 5-yr local controlNo. patientsNo. local recurrencesPercentage 5-yr local controlNo. patientsNo. local recurrencesPercentage 5-yr local control
T1–T326134931678571965
T49644331165421761
Total351948641771993663

Ultimate Local Control

The 5-year ultimate local control rates after definitive RT versus surgery and RT were as follows: Stage T1–T3, 64% and 78%; Stage T4, 49% and 65%; and overall, 59% and 71%.

Local–Regional Control

The 5-year local–regional control rates are shown in Table 3. The ultimate 5-year local–regional control rates after definitive RT versus surgery and RT were as follows: Stage I–III, 68% and 87%; Stage IV, 42% and 62%; and overall, 59% and 71%.

Table 3. Local—Regional Control
StageDefinitive radiotherapySurgery + RadiotherapyTotal
No. patientsNo. LR recurrencesPercentage 5-yr LR controlNo. patientsNo. LR recurrencesPercentage 5-yr LR controlNo. patientsNo. LR recurrencesPercentage 5-yr LR control
  1. LR: local—regional.

I–III23115125387481470
IV12842391462512257
Total351948641771993663

Multivariate analysis revealed that overall stage (P = 0.0145) and treatment group (P = 0.0003) was significant predictors of local–regional control. Patients who were treated with surgery and RT had a better outcome compared with those who received RT alone.

Distant Metastasis-Free Survival

The 5-year distant metastasis-free survival rates were as follows: Stage I–III, 84%; Stage IV, 93%; and overall, 88%.

Multivariate analysis of distant metastasis-free survival revealed that only the treatment group was a significant predictor (P = 0.0371). Patients who received surgery and RT had a lower risk of developing distant metastases.

Cause-Specific Survival

The 5-year cause-specific survival rates are depicted in Table 4. Multivariate analysis of cause-specific survival revealed that overall stage (P = 0.0300) and treatment group (P = 0.0059) were significant predictors of this endpoint. Patients who received surgery and RT had improved cause-specific survival compared with those treated with definitive RT.

Table 4. Cause-Specific Survival
StageDefinitive radiotherapySurgery + RadiotherapyTotal
No. patientsNo. CS deathsPercentage 5-yr CS survivalNo. patientsNo. CS deathsPercentage 5-yr CS survivalNo. patientsNo. CS deathsPercentage 5-yr CS survival
  1. CS: cause-specific.

I–III23105625483481470
IV12750391461512157
Total351752641869993564

Overall Survival

The 5-year overall survival rates are shown in Table 5. Multivariate analysis of overall survival revealed overall stage (P = 0.0476) and treatment group (P = 0.0098) were significant predictors. Patients who were treated with surgery and RT had a better overall survival than those who received definitive RT.

Table 5. Overall Survival
StageDefinitive radiotherapySurgery + RadiotherapyTotal
No. patientsNo. deathsPercentage 5-yr survivalNo. patientsNo. deathsPercentage 5-yr survivalNo. patientsNo. deathsPercentage 5-yr survival
I–III232242251772483958
IV12838393045513842
Total353040644756997750

Complications

Three of 9 patients (33%) treated with preoperative RT and surgery experienced severe complications. One patient had a postoperative orocutaneous fistula that required postoperative surgical intervention. One patient developed a postoperative orocutaneous fistula that necessitated a gastrostomy tube. One patient was rehospitalized for a postoperative wound infection and required a permanent gastrostomy tube.

Thirteen (24%) of 55 patients treated with surgery and postoperative RT experienced severe complications. Nine patients had a severe osteoradionecrosis necessitating surgery. Postoperative gastrointestinal bleed and aspiration pneumonia developed in 1 patient and necessitated a 7-week hospitalization. One patient had a postoperative orocutaneous fistula and carotid artery exposure necessitating flap reconstruction. One patient had a severe soft tissue necrosis of the oral tongue necessitating surgical intervention. Two patients had an orocutaneous fistula necessitating gastrostomy tube placement.

Five (14%) of 35 patients treated with definitive RT experienced a severe complication. One patient was hospitalized during RT for dehydration. An abscess developed in 1 patient after a planned neck dissection that required incision and drainage. Three patients experienced severe osteoradionecrosis.

DISCUSSION

The treatment of squamous cell carcinoma of the retromolar trigone is controversial. Determining the optimal management strategy is all the more difficult because it is a relatively uncommon tumor, and there is a paucity of data on outcomes. The natural history and response to surgery and/or RT for head and neck cancers vary significantly depending on primary site and extent of disease. Oropharyngeal carcinomas have a higher probability of local control after definitive RT than similarly staged oral cavity cancers.12–14 T1–T2 squamous cell carcinomas of the anterior tonsillar pillar have lower local control after RT than tonsillar fossa carcinomas, and it is likely that retromolar trigone lesions fare less well than anterior tonsillar pillar tumors.13

Byers et al.1 reported on 110 previously untreated patients with squamous cell carcinomas of the retromolar trigone who received surgery and/or RT with curative intent at the M. D. Anderson Cancer Center between 1965 and 1977. All patients had a 2-year minimum followup. Surgery was usually selected for patients with associated leukoplasia, poor dentition, large infiltrating tumors, bone destruction, and/or trismus. Treatment consisted of surgery alone, 46 patients; preoperative RT and surgery, 3 patients; surgery and postoperative RT, 11 patients; and RT alone, 50 patients. The crude local control rates after surgery alone or combined with adjuvant RT versus RT alone were as follows: Stage T1, 7 (100%) of 7 patients versus 5 (83%) of 6 patients; Stage T2, 24 (92%) of 26 patients versus 26 (84%) of 31 patients; and Stage T3, 12 (92%) of 13 patients versus 6 (86%) of 7 patients. The 5-year overall survival rate was 26%. Of 50 patients treated with definitive RT who were locally controlled, 7 (14%) required a partial mandibulectomy for osteoradionecrosis.

Huang et al.15 reported on 65 patients treated with RT alone or combined with surgery at Washington University (St. Louis, Missouri) between 1971 and 1994. Patients had followup for 5 years or more. Sixteen patients received definitive RT, and the remaining patients were treated with surgery and preoperative RT (10 patients) or postoperative RT (39 patients). The 5-year disease-free survival rates were as follows: preoperative RT and surgery, 90%; surgery and postoperative RT, 63%; and definitive RT, 31%. Multivariate analysis of disease-free survival revealed that treatment modality (P = 0.002) and N Stage (P = 0.012) significantly influenced this endpoint. Multivariate analysis showed that the only parameter that significantly influenced local–regional control was the treatment modality (P = 0.046). Patients treated with surgery and adjuvant RT had a more favorable outcome than those treated with definitive RT. The rates of severe complications (osteoradionecrosis, soft tissue necrosis, and severe trismus) were as follows: preoperative RT and surgery (0%); surgery and postoperative RT, 12%; and RT alone, 11%. The authors of the current study concluded that surgery and adjuvant RT resulted in improved local–regional control and disease-free survival compared with RT alone.

Kowalski et al.16 reported on 114 patients who underwent a hemimandibulectomy and neck dissection for retromolar trigone squamous cell carcinoma between 1960 and 1991 at the Hospital A. C. Camargo (São Paulo, Brazil). Sixty-one patients received adjuvant postoperative RT to a median dose of 50 Gy (range, 10–70 Gy) for positive margins and/or involved nodes. Median followup was 25 months; 19 patients were lost to followup from 1 to 47 months (mean, 15 mos) after treatment. Complications were observed in 52% of patients and included wound dehiscence (20%), wound infection (18%), flap necrosis (12%), fistula (12%), seroma (7%), pneumonia (5%), hematoma (4%), and/or carotid rupture (2%). There was one postoperative death due to a pulmonary embolus. The 5-year disease-free and overall survival rates were 49% and 55%, respectively. Advanced stage and hard palate involvement were significantly associated with decreased disease-free and overall survival.

Our data and those reported by Huang et al.15 indicate that the probability of cure is influenced by the extent of disease and treatment. Patients treated with surgery and RT appear to have a more favorable outcome than those treated with RT alone. In contrast, Byers et al.1 observed similar local control rates after either surgery alone or combined with adjuvant RT compared with definitive RT. The reasons for these differing observations are unclear. A small subset of early-stage patients treated with surgery alone at our institution was not included in our analysis, which would have biased the comparison in favor of those who received definitive RT. Conversely, treatment of the occasional patient with advanced, unresectable disease would have biased the comparison in favor of surgery. Based on the limited data available, it is likely that surgery and RT offer a better probability of cure than definitive RT.

Huang et al.15 observed similar complication rates after RT alone compared with surgery and RT, whereas we observed more severe complications in those treated surgically. Kowalski et al.16 also observed a relatively high rate of complications after surgery alone or combined with RT. Although this should be considered when determining the optimal treatment for a particular patient, the risk of severe postoperative complications is probably offset by the increased likelihood of dying of cancer after definitive RT, particularly for patients with advanced disease.

CONCLUSION

The optimal treatment for patients with squamous cell carcinoma of the retromolar trigone is surgery. Although definitive RT may result in a similar cure rate for T1–T2 cancers, the probability of severe complications (particularly osteoradionecrosis) is probably higher. Postoperative RT should be employed for those with a significant risk of residual disease, particularly patients with positive margins and/or extracapsular extensions.17 High-risk patients should be considered for adjuvant chemotherapy to be administered concomitantly with postoperative RT.18, 19 Patients with T3–T4 tumors that appear to be completely resectable are best treated with surgery and postoperative RT. Those with advanced incompletely resectable T4 cancers are optimally treated with definitive RT employing an altered fractionation schedule combined with concomitant chemotherapy.20 Patients with advanced T4 cancers and extensive nodal disease, as well as those in very poor medical condition who are unable to undergo aggressive treatment, are treated palliatively.

Ancillary