The American Joint Committee on Cancer (AJCC) 2002 staging system (AJCC 2002) suggested that squamous cell carcinoma of the oral cavity (OSCC) with T4b is unresectable. The current retrospective results show that selected T4b patients were resectable with favorable outcomes.
From January 1996 to December 2000, 103 consecutive untreated T4 OSCC patients (reclassified by AJCC 2002) without carotid artery encasement and skull base extension were eligible for radical treatment. All received head-and-neck magnetic resonance imaging (MRI) and/or computed tomography (CT) scans before operation. The surgical principles were safety margins of ≥1 cm for primary tumors, modified/radical neck dissections for clinical lymph node-positive disease, and supraomohyoid neck dissection for lymph node-negative disease. In all, 95.1% of patients (98 of 103 patients) underwent free-flap reconstructions. Adjuvant radiotherapy or concomitant chemoradiotherapy was administed to those with pathological T4 (AJCC 1997), cervical lymph node metastasis, or close margins (≤4 mm). Survivals were calculated according to the method of Kaplan and Meier.
In all, 58 patients were classified as having T4a disease and 45 were classified as having T4b disease. No statistical difference was observed in the 5-year local control, neck control, disease-free survival, and overall survival rates between the T4a and T4b groups. In multivariate analyses, pathologic lymph node status (pN0-1 vs. pN2) was found to be the sole independent predictor for T4b for local control (P = .012), disease-free survival (P = .005), and overall survival (P = .008).
According to the American Joint Committee on Cancer (AJCC) 2002 staging system (AJCC 2002),1 T4 squamous cell carcinomas of the oral cavity (OSCCs) are further divided into T4a (resectable) and T4b (unresectable) categories. T4a denotes a tumor that invades through the cortical bone, into the deep (extrinsic) muscle of the tongue, the maxillary sinus, or the skin of the face. In contrast, T4b denotes a tumor that involves the masticator space (MS), pterygoid plates (PP), skull base, or an encasement of the internal carotid artery.
Cancer of the oral cavity adjacent to the MS and/or PP can deeply invade into the PP or MS components (MS components include the ramus bone of the mandible, masseter muscle, medial pterygoid muscle, lateral pterygoid muscle, and temporalis muscle) simply because of the anatomic vicinity.2–4 Unfortunately, the MS and PP are deep anatomic sites; hence, a surgical approach is often difficult to pursue. In our experience, neither a clear preoperative tumor infiltration by imaging nor adequate surgical margins could be achieved during surgical procedures involving MS and/or PP tumors, resulting in an unsatisfactory survival at this site. Although local and neck control of the tumor and the patient's survival are affected largely by tumor biology and tumor control, the patient's response may also be affected by the mode of radical surgery. Performing radical surgery provides sufficient margins. However, this treatment means that postoperative reconstruction is necessary. Free-flap reconstruction versus local-flap-based reconstruction is the recent mode of change for a large and difficult surgical defect. With collaboration between our group, who provide the free-flap reconstruction, and the surgical team, who perform resection of the OSCC, positive results may be obtained for instances in which the tumor invades the MS and/or PP.
Reviewing our patient database for cases involving T4b tumors between the years 1996 and 2000, we learned that not all T4b tumors were unresectable. We therefore retrospectively report the results for selected patients with T4b tumors (exclusion criteria included patients with involvement of the skull base or internal carotid artery) who underwent radical surgery performed by our group. Also, we determined the survival and prognostic factors for selected T4b OSCC cases and compared with the results for T4a patients.
MATERIALS AND METHODS
Between January 1996 and December 2000, 103 consecutive and previously untreated OSCC patients who underwent radical surgery were retrospectively pooled from our databank. Ninety-seven patients (94.2%) were male and 6 patients (5.8%) were female. Of the 103 patients (median age, 51 years), 58 had T4a disease (median age, 54 years) and 45 had T4b disease (median age, 49 years). Of the 58 T4a patients, 6.9% (4 of 58) of the tumors involved deep (extrinsic) muscle of the tongue, 58.6% (34 of 58 tumors) involved only the bone (maxilla and/or mandible), 29.3% (17 of 58 tumors) involved only the skin, and 5.2% (3 of 58 tumors) involved both the bone and the skin. The MS is composed of 5 components. Including the PP, there are 6 anatomic components involved in T4b tumors. Of the 45 T4b patients, 37.8% (17 of 45) of the tumors involved 1 anatomic component, whereas the other 62.2% (28 of 45 tumors) involved multiple components of MS and/or PP. Five patients had PP invasion, 28 had ramus of the mandible invasion, 26 had masseter muscle invasion, 26 had medial pterygoid muscle invasion, 4 had lateral pterygoid muscle invasion, and 3 had temporalis muscle invasion. The evaluation before surgery included a medical history, clinical examination, flexible fiberoptic pharyngoscopy, complete hemogram and blood biochemistry, head and neck magnetic resonance imaging (MRI) or computed tomography (CT) scan, chest X-ray, bone scan, and liver echography. The disease extent was staged according to AJCC 2002.
Surgery and Postoperative Radiation Therapy Regimen
The primary tumors were excised with ≥ 1 cm safety margins (both peripheral and deep margins) and a modified/radical neck dissection (ND) was performed in patients with clinically positive lymph node disease and a supraomohyoid ND was performed in patients with clinically negative lymph node disease. A mandibulectomy and/or maxillectomy was performed based on tumor extension or margin space. Tumor margin tissue was sent for cryosectioning. If a margin was positive, then additional tissue was excised and sent for cryosectioning to ensure that the margin was free of tumor. Surgical defects were immediately reconstructed by a free-flap or local-flap by plastic surgeons.
Postoperative radiotherapy (RT) was performed on patients with pT4 tumors (AJCC 1997),5 positive lymph nodes, or close margins (≤4 mm). RT was initiated within 4 to 8 weeks after surgery. The initial RT treatment field included the whole tumor bed area with 1 to 2 cm margins and the regional lymphatics using 6-megavolt (MV) X-ray beams via bilateral-opposed fields with either a matched anterior supraclavicular field or the 3-dimensional conformal RT technique based on postoperative CT simulator imaging. The prescribed dose was 1.8 to 2 grays (Gy) per fraction per day, given 5 days per week. When the spinal cord dose reached 46 to 46.8 Gy, the tumor bed area with close margins was irradiated with a 6-MV photon beam and the electron beam was boosted to the posterior and lower neck area, or a 3-dimensional conformal RT technique was used during the boost period. The total radiation dose was 66 Gy for patients with multiple positive neck lymph nodes and/or extracapsular spreading (ECS) and 60 Gy to the remainder of the patients. Concomitant chemoradiotherapy (CCRT) with cisplatin-based agents was given to patients with ECS or pathologic multiple lymph node metastases. The cisplatin-based regimen was cisplatin at a dose of 30 mg/m2 weekly6 or a biweekly regimen comprised of cisplatin, tegafur, and leucovorin.7
The clinical-pathologic risk factors evaluated for T4 patients included sex, age (≤40 years vs. >40 years), treatment modalities (operation vs. operation + RT/CCRT), differentiation (well/moderate vs. poor), skin invasion, bone invasion, nerve invasion, vessel invasion, lymph invasion, tumor depth (best cutoff depth: <15 mm vs. ≥15 mm), close margin (≤4 vs. >4 mm), pathologic lymph node disease (pN0-1 vs. pN2), ECS (with vs. without), and tumor status (T4a vs. T4b).
Follow-up and Statistical Methods
The follow-up study continued until November 2005. All patients received follow-up examinations for at least 60 months after surgical treatment or until death. The survival time and time intervals for locoregional recurrence or distant metastases were calculated from the operation date. Descriptive statistics were summarized using frequencies, percentages, medians, standard deviations, and ranges. The Kaplan–Meier method was used to calculate the survival rate. Statistical significance was evaluated using the log-rank test. Univariate and multivariate analyses were used to define the independent risk factors. Multivariate analyses of the prognostic factors were performed using the Cox logistic regression method with forward selection.
All but 1 patient (n = 102) underwent ND. Supraomohyoid ND was performed in 38 patients, whereas modified or radical ND was performed in 64 patients. One patient underwent tumor resection without ND. A total of 101 patients underwent a mandibulectomy and/or maxillectomy. Reconstructions with flap were performed in all patients. Ninety-eight patients had free-flaps and 5 patients had local-flaps. In all, 61.2% of the patients (63 of 103 patients) received postoperative RT or CCRT. Thirty-four patients had T4a tumors and 29 had T4b tumors. At the time of analyses (November 2005), 45 patients (43.7%) were alive and 58 patients (56.3%) were dead.
The clinicopathologic characteristics of OSCC including age, sex, mode of ND, mode of treatment, cell differentiation, nerve invasion, vessel invasion, lymph invasion, tumor depth (15 mm by best cutoff), close margin, lymph node status, and presence of absence of ECS were found to be similar between patients with T4a and T4b tumors, although 53.3% of the T4b tumors (24 of 45 tumors) were buccal cancer (Table 1).
Table 1. Clinical Pathological Characteristics of T4a (n = 58) and T4b Patients (n = 45)
OP indicates operation; RT, radiation therapy; CCRT, concurrent chemoradiation therapy; AJCC, American Joint Committee on Cancer.
Modified or radical
No neck dissection
Tumor depth (mm)
Close margin (mm)
Unknown but free margin
AJCC pathological lymph node status
Clinical Course of T4a Patients
At the time of analyses (November 2005), 41.3% of the T4a patients (24 of 58 patients) were alive and 58.7% (34 of 58 patients) were dead (17 deaths were due to primary disease, 4 deaths were the result of a second primary tumor, 1 death was the result of a third primary tumor, and 12 deaths were related to other causes). In all, 6.9% of the patients (4 of 58 patients) had postoperative mortalities (within the first 3 months). Of the 4 patients, 2 deaths were due to a myocardial infarction, 1 resulted from pneumonia, and 1 resulted from massive oral bleeding. In all, 25.9% of the patients (15 of 58 patients) developed local recurrences (median, 8 months; range, 2–69 months), 10.3% (6 of 58 patients) developed a neck recurrence (median, 10.5 months; range, 8–26 months), and 5.2% (3 of 58 patients) had distant metastases (median, 4 months; range, 3–11 months). Local and/or neck recurrence was observed in 27.6% of the patients (16 of 58 patients). Two patients underwent salvage surgery (1 was alive without disease after surgery and 1 died of disease) and 14 received palliative treatment.
Second primary cancer affected 17.2% of the patients (10 of 58 patients) (median, 28 months; range, 0–69 months). Of the 10 patients, 6 had oral cancers, 3 had lung cancers, and 1 had esophageal cancer. Six patients underwent curative intent radical surgery (5 oral cancers and 1 lung cancer) and the remaining 4 patients received palliative treatment. Four patients were alive without disease (3 second primary oral cancers and 1 second primary lung cancer), 4 died due to a second primary tumor, 1 died due to a third primary of oral cancer, and 1 died due to another cause.
Clinical Course of T4b Patients
At the time of analyses (November 2005), 46.7% of the patients (21 of 45 patients) were alive and 53.3% (24 of 45 patients) were dead (17 died due to primary disease, 2 died due to a second primary tumor, and 5 deaths were related to other causes). There were no surgery-related mortalities (within the first 3 months). Of the 45 patients, 28.9% (n = 13 patients) developed local recurrences (median, 6 months; range, 5–26 months), 24.4% (n = 11 patients) developed a neck recurrence (median, 5 months; range, 1–20 months), and 8.9% (n = 4 patients) had distant metastases (median, 9 months; range, 6–13 months). In total, 16 patients had local and/or neck recurrence. Three patients underwent salvage surgery (1 died of heart disease 61 months after successful salvage therapy and 2 died of disease) and 13 received palliative treatment.
Second primary cancer affected 11.1% of the patients (n = 5 patients). Of the 5 patients that had second primary cancer, 3 were oral cancers and 2 were gastrointestinal tract carcinomas. All 5 patients received curative intent radical surgery, 2 of whom were alive without disease and 3 of whom had died due to disease at the time of last follow-up.
In T4 patients, univariate analysis showed that nerve invasion (P = 0.028), vessel invasion (P = 0.041), tumor depth ≥15 mm (P = 0.048), and pN2 disease (P < 0.001) were significant adverse factors for 5-year local control rates. Female gender (P = 0.016), nerve invasion (P = 0.029), and pN2 disease (P < 0.001) were determined to be significant adverse factors for 5-year neck control rates. Treatment by operation + RT/CCRT (P = 0.047), nerve invasion (P = 0.005), close margin (P = 0.024), pN2 disease (P < 0.001), and ECS (P = 0.003) were determined to be significant adverse factors for 5-year disease-free survival. Age >40 years (P = 0.050), nerve invasion (P = 0.021), lymph invasion (P = 0.011), pN2 disease (P < 0.001), and ECS (P = 0.005) were determined to be significant or marginal adverse factors for 5-year overall survival.
In T4a and T4b patients, univariate analysis revealed that vessel invasion (P = 0.024) and pN2 disease(P = 0.022) for T4a tumors and nerve invasion (P = 0.009) and pN2 disease (P = 0.006) for T4b tumors were determined to be significant adverse factors for 5-year local control rates. Female gender (P = 0.002), vessel invasion (P = 0.050), pN2 disease (P = 0.049), and ECS (P = 0.043) were found to be significant for T4a tumors and skin invasion (P = 0.005) and pN2 disease (P = 0.013) were found to be significant for T4b tumors for 5-year neck control rates. Significant adverse factors for 5-year disease-free survival included pN2 disease (P = 0.023) for T4a tumors and operation + RT/CCRT (P = 0.048), nerve invasion (P = 0.002), skin invasion (P = 0.017), close margin (P = 0.017), pN2 disease (P < 0.001), and ECS (P = 0.008) for T4b tumors. Significant or marginal adverse factors affecting 5-year overall survival included age >40 years (P = 0.030), vessel invasion (P = 0.004), lymph invasion (P = 0.011), pN2 disease (P = 0.001), and ECS (P = 0.035) for T4a tumors and nerve invasion (P = .006), skin invasion (P = .054), pN2 disease (P < 0.001), and ECS (P = 0.049) for T4b tumors.
Pathological lymph node status (pN0-1 vs. pN2) was found to be a significant prognostic factor for T4, T4a, and T4b OSCC with regard to local control, neck control, disease-free survival, and overall survival. Additional significant prognostic factors included nerve invasion in disease-free survival and age >40 years for overall survival for T4 patients, female gender for neck control and age >40 years for overall survival for T4a patients, and skin invasion for neck control for T4b patients (Table 2).
Table 2. Multivariate Analyses of Local and Neck Controls, Disease-free and Overall Survivals in Patients Enrolled
The 5-year local control, neck control, disease-free survival, disease-specific survival, and overall survival rates for all T4 patients (AJCC 2002; n = 103) were 70%, 81%, 62%, 63%, and 47%, respectively. No significant difference was noted between T4a and T4b OSCC with regard to local control (71% vs. 69%; P = 0.843), neck control (87% vs. 74%; P = 0.066), disease-free survival rate (66% vs. 57%; P = 0.348), disease-specific survival rate (65% vs.60%; P = 0.397), and overall survival rate (45% vs. 49%; P = 0.776) (Figs. 1 and 2).
According to the criteria from the AJCC in 2002, T4b OSCC is unresectable. Before 2002, a T4b classification did not exist. Thus, some of our OSCC patients (staged before 2002) who were classified as having T4b disease by AJCC 2002 had been treated by radical surgery. Therefore, we reviewed our cohort results of T4a patients (n = 58 patients) and T4b patients (n= 45 patients) who were treated surgically between January 1996 and December 2000 (n = 103). Our aims were to analyze the therapeutic outcome of T4 OSCC according to AJCC 2002 and the risk factors of therapeutic failure. To the best of our knowledge, there has been no previous outcome report for surgical treatment for T4b OSCC.
Anatomically, the MS is located behind the retromolar trigone. The surgery performed in the MS may be more difficult than in the retromolar trigone. In the MS, there are at least 2 possible ways for tumor spreading to occur. One is an upward invasion along the neurovascular bundle,3, 4, 8 whereas the other is through the medial side of the MS through the loose buccal space, then to the lateral wall of the maxilla. To decrease the recurrence rate, sufficient margin is essential. However, obtaining an adequate deep margin from the skull base is a difficult challenge. To achieve this goal, either the coronoid process or the condylar process should be excised (for neurovascular bundle spreading) or the lateral wall of the maxillary sinus should be excised as high as possible. To obtain a better surgical field-of-view for the infratemporal fossa and the lateral wall of the maxillary sinus, temporary removal and repositioning of the zygomatic arch was sometimes necessary.
A sufficient pathologic margin was the most important concern for patient outcome in our group. For radical excision of OSCC, our policy was to have at least a 1-cm safety margin. Hence, patients who had a tumor involving the skull base or encasement of the internal carotid artery were excluded because these areas are clearly unresectable. We considered a tumor that invaded the MS and/or PP as resectable if adequate safety margins could be achieved. In the past 20 years, free tissue graft techniques have led the field of head and neck surgery to a new era. We have a high free-flap reconstruction success rate (approximately 96–98%).9 With the help of the reconstructive team, we can perform radical surgery more confidently.10, 11 In our series, free-flap reconstruction was performed in the majority of T4a patients (54 of 58 patients; 93.1%) and resected T4b patients (44 of 45 patients; 97.8%). The 5-year disease-free survival between T4a patients (66%) and resected T4b patients (57%) was not found to be significantly different (P = 0.348). The higher rate of free-flap reconstruction may further contribute to our favorable outcome than those reported for the retromolar OSCC, which was considered a more simple surgery than that for a tumor that invades the MS.3, 4 Despite the efforts of the surgical team, second primary cancers developed in 14.6% of the patients (17.2% in T4a patients and 11.1% in T4b patients). Approximately 50% of these patients (6 of 11 patients) were alive at the time of last follow-up without disease due to effective treatment of the second primary tumor.
To gain a better control for advanced OSCC patients, postoperative RT was applied in our patients once there was a positive lymph node, close margins (≤4 mm), or pT4. For those patients with ECS or pathologic multiple lymph node metastases, they received a cisplatin-based regimen CCRT. Our therapeutic results focused on T4 tumors were also comparable to those reported for T3-T4 tumors with regard to the 5-year overall survival (45% in T4a patients and 49% in resected T4b patients vs. 29-39%), disease-free survival (66% in T4a patients and 57% in resected T4b patients vs. 44-60%), and disease-specific survival (65% in T4a patients and 60% in resected T4b patients vs. 65%) (Table 3).3, 8, 12–14
Table 3. Survivals of Previously Untreated T3 and T4 Oral Carcinoma Patients
Patient no. (Years)
Metastatic lymph node (Clinical or pathological) (%)
Primary treatment mode
5-Year survival (%)
OP indicates operation; RT, radiation therapy; CCRT, concurrent chemoradiation therapy.
Initially, all patients were treated with operation, but the study did not mention the postoperative treatment.
In this study, we found that lymph node (pN2) rather than tumor status is the sole significant adverse factor in T4a and T4b patients for 5-year local control (P = 0.032 vs. P = 0.012), disease-free survival (P = 0.032 vs. P = 0.005), and overall survival (P = 0.002 vs. P = 0.008). In addition, T4b lymph node status (pN2) (P = 0.041) and skin invasion (P = 0.014) were found to be significantly adverse factors for 5-year neck control (Table 2). We expected that radical tumor excision with free-flap reconstruction might render more favorable local control regardless of lymph node status; however, pN2 still compromised all parameters of tumor control, including local control. Its significance was evident even when adjuvant CCRT with cisplatin-based regimens were given to our patients.6, 7 This implies that T4 OSCC with pN2 has more aggressive tumor behavior and requires novel therapies.
T4a OSCC patients with an onset of disease at age >40 years were found to have a significantly lower overall survival compared with patients with a disease onset at age ≤40 years (hazard ratios [HR] = 4.404; 95% confidence interval [95% CI], 1.320–14.686) (Table 2). This result indicates that such patients (T4a OSCC with disease onset at age > 40 years) may be candidates for more aggressive adjuvant therapy. Although gender (female) was found to be a significant risk factor for neck control in T4a OSCC patients in the current study (HR = 8.927; 95% CI, 1.629–48.917), there were 54 males versus only 4 females. Therefore, the question of gender being a significant risk factor for neck control must be answered in future studies.
In the current study, we demonstrated a flaw in the staging system and further established that aggressive surgery for local and regional disease combined with adjuvant therapy, even in cases of advanced local disease, can offer a very reasonable survival rate. Those patients who would be deemed “unresectable” given this new staging system would certainly have no survival rate, if not treated at all or treated with another modality. There were 2 limitations of this report. One was that it was a retrospective study and the other was that no information was given regarding treatment in a comparable group of patients by alternative means. In such a case, we cannot be absolutely certain that our scheme is the best one.