|Home | About | Journals | Submit | Contact Us | Français|
Objective: To evaluate the management of patients with malignant tumors of the ear and temporal bone. Design: Retrospective analysis of data. Setting: Radcliffe Infirmary, Oxford, United Kingdom. Participants: Twenty-seven patients were classified into two groups according to the site of origin of the tumor: (1) superficial (17 tumors): tumors arising from the skin of the pinna, parotid, and temporomandibular joint area; (2) deep (10 tumors): tumors arising in the ear canal and temporal bone. Main Outcome Measures: Treatment modality, complications, recurrence rate, disease-free interval, and survival. Results: The mean follow-up period was 25 months (0 to 60), and the median overall survival 46 months (0 to 102). Complications occurred in 6 patients (22%). The 3-year survival was 38% (95% confidence interval [CI], 19 to 58%), and the 5-year survival 19% (95% CI, 3 to 35%). Conclusions: There were insufficient data to demonstrate any difference in survival or disease-free interval related to the site of tumor origin (superficial versus deep tumors). There were independent differences in survival in favor of both performing parotidectomy and using postoperative radiotherapy, but neither reached significance at the 0.05 level.
Malignancies of the ear and temporal bone are rare and aggressive types of tumors. The incidence has been reported to be between one and six cases per million population years, which is less than 0.2% of all tumors of the head and neck.1,2,3
These tumors spread mainly by direct invasion into the temporal bone and neighboring structures6 (parotid, infratemporal fossa, dura, brain). Lymphatic metastases are uncommon (10%) and distant spread extremely rare.7
Currently neither the International Union Against Cancer (UICC) nor the American Joint Committee on Cancer (AJCC) has produced a staging system for ear and temporal bone tumors. The most common staging system being used in publications is from the University of Pittsburgh11,12,13,14 (Table 1).
Surgical treatment of malignant tumors of the ear and temporal bone is not universally agreed upon. In general lesions localized to the external auditory canal (EAC) (T1 lesions) are treated with a limited resection (i.e., local resection, radical mastoidectomy), whereas more advanced lesions (T2 to T4) are treated by en bloc resection (partial temporal bone resection, subtotal petrosectomy, total petrosectomy10,12,13,15).
Between 1997 and 2005, 27 patients undergoing treatment for malignant tumors of the ear and temporal bone were identified from the Otolaryngology and Skull Base Department database at the Radcliffe Infirmary, Oxford, UK. The medical records of these patients were reviewed for age, sex, diagnosis, treatment (surgery, radiotherapy, and/or chemotherapy), complications, follow-up, recurrence, disease-free interval, and survival rate.
Patients were classified into two groups:
They were staged according to the University of Pittsburgh TNM system (Table 1) for the deep tumor group (n=9) and according to the UICC system for the superficial group (n=17). One patient (with a deep angiosarcoma) was excluded from classification by either system.
Five types of surgical procedures were performed, in two broad categories:
Kaplan-Meier survival analyses were used, and tests were conducted on the survival effects of tumor origin, the inclusion of superficial parotidectomy in the operative procedure, and the use of postoperative radiotherapy.
Data were obtained for 27 patients, 18 male and 9 female, a ratio of 2 to 1 (sex distribution did not differ by group; chi square p=0.696). The median age was 66.5 years (range, 34 to 81). The deep group median age was 62.6 (34 to 77 yrs) and the superficial group 74.5 (64 to 81 yrs). This age difference was statistically significant (independent samples t-test and Mann-Whitney U tests, both p<0.005). Diagnosis, staging, and management data are provided in Table Table22.
Histological diagnoses were squamous cell carcinoma (SCC) (20 cases; 74%), basal cell carcinoma (5 cases; 18%), angiosarcoma (1 case; 4%), and anaplastic carcinoma (1 case; 4%).
Seventeen tumors were classified as superficial and 10 deep, although the angiosarcoma case (deep group) was excluded from the statistical analysis as this tumor cannot be staged either with the Pittsburgh (solely for SCC) or the UICC (for superficial skin tumors) staging systems.
The distribution of the tumors (n=26) was T1, 3 cases; T2, 4 cases; T3, 4 cases; and T4, 15 cases. Clinically evident node-positive disease was present in 3 of these cases, and distant metastases in 1. Classification in stages was Stage I, 3 cases; Stage II, 4 cases; Stage III, 10 cases; and Stage IV, 9 cases.
Twenty-five patients underwent surgical treatment, with 17 receiving adjuvant radiotherapy postoperatively. One patient received radiotherapy (45 Gy in 10 fractions) as the only form of treatment and another patient had radiotherapy and chemotherapy (cisplatinum and 5-fluorouracil) for palliation.
The surgical procedures performed were local canal resection (2 cases), radical mastoidectomy (1 case), lateral temporal bone resection (16 cases), subtotal petrosectomy (3 cases), and total petrosectomy (3 cases).
Also, 15 patients underwent superficial parotidectomy, 3 patients had a neck dissection performed (2 modified radical and 1 selective neck dissection), and 9 patients had tissue transfer reconstruction (2 pedicle flaps, 5 free flaps, 1 local flap, and 1 facial nerve graft). One patient had resection of the TMJ and another had resection of intracranial disease.
Six patients suffered complications: one patient died postoperatively due to respiratory distress syndrome. One patient developed osteoradionecrosis of the external ear canal, which was managed conservatively. Two patients suffered facial nerve palsies (one temporary which recovered fully and one permanent, Grade IV House-Brackmann classification18). Two flaps required revision of their vascular anastomoses.
The overall mean follow-up period was 25 months (0 to 60 mos), and for surviving cases was 37 months (2 to 60 mos).
Of the 26 staged patients (excluding the angiosarcoma), 9 were never disease free and 3 had recurrence. The proportion uncured by stage at presentation was Stage I, 0%; Stage II, 29%; Stage III, 50%, and Stage IV, 60%.
The estimated median overall survival was 46 months (0 to 102 mos).
Of 27 patients, 13 remained alive without disease, 1 alive with disease, and 13 died (1 in the postoperative period, 1 of an unrelated cause, 10 from effects of local disease, and 1 from metastatic disease).
The overall 3-year survival was 38% (95% CI, 19 to 58%), and the 5-year survival 19% (95% CI, 3 to 35%).
Estimated median (95% CI) survivals by stage from the Kaplan-Meier analyses were Stage I, no deaths; Stage II, >24 months; Stage III, 16 months (10 to 22 mos), and Stage IV, 9 months (7 to 11 mos).
There were insufficient data to demonstrate any difference in survival or disease-free interval related to the site of tumor origin (superficial versus deep groups). A Kaplan-Meier survival plot by site of origin is given in Fig. Fig.11.
Similarly, although there was a trend toward improved survival when a parotidectomy was performed, the study was underpowered to demonstrate statistical significance. Estimated median overall survival (95% CI) for both groups with parotidectomy was 46 months (0 to 102 mos), and without parotidectomy 9 months (0 to 24 mos). A Kaplan-Meier survival plot comparing inclusion or omission of parotidectomy in the operative procedure is given in Fig. Fig.2.2. The plot is reproduced and stratified by tumor origin (superficial or deep) in Fig. Fig.33.
Also, estimated median overall survival for the superficial group with the administration of radiotherapy was 46 months versus 12 months when radiotherapy was not administered (Fig. 4). The data regarding radiotherapy for the deep group were insufficient for analysis.
To exclude the possibility of a confounding relationship between treatment with radiotherapy or stage of disease and the effect of parotidectomy, a further survival analysis (Cox regression) was conducted using survival models for all factors. Consistent with the findings in general, disease stage was the only factor to affect survival (p<0.05), but both radiotherapy (p=0.29) and parotidectomy (p=0.12) led to improvements that did not reach statistical significance.
Malignant tumors of the ear and temporal bone are very rare, and the lack of prospective trials (randomized or not) poses a problem in establishing the correct management of this type of tumor. The management of these tumors relies only on the available evidence of clinical experience with small isolated or grouped series. SCC has been the most studied and reported of malignant tumors of the ear and temporal bone, but most of the other tumors behave in a similar manner and share the generally poor outcome of this disease.10
One of the main problems in the management of these tumors has been the lack of consensus in staging the disease, a process made difficult due to the close anatomical relationship to structures in the infratemporal fossa and skull base.
The University of Pittsburgh TNM staging system11 has been widely accepted within the published literature in the field. The importance of this classification is that a T2 or T3 N1 tumor is considered to be stage IV, in contrast with the AJCC staging system (sixth edition) in which a T2 or T3 N1 “skin” tumor would be a stage III.
We think that part of the staging problem is that some tumors originate in the skin and then invade deeper structures, whereas others originate in the ear canal and temporal bone and spread laterally to the overlying appendage and adjacent skin. However, we have been unable to demonstrate a statistically significant difference in the survival and recurrence patterns of the two groups, so this hypothesis remains the authors' experience only.
In terms of treatment options, the available evidence supports surgery for primary treatment with curative intent and radiotherapy for adjuvant or palliative treatment.16,17 A literature search found no studies on the effect of chemotherapy in these tumors.
A systematic review performed by Prasad and Janecka2 in 1994, reviewing 26 series with 144 patients treated for temporal bone SCC, suggested that for T3 and T4 tumors (those involving the middle ear and beyond), subtotal petrosectomy (preserving the petrous apex and carotid artery) seems to improve survival.
Also in this review, resection of involved dura did not appear to improve survival and there were insufficient data to comment on survival after treatment in patients with involvement of the petrous apex, internal carotid artery, or brain.2
With regard to the treatment of stage I and II (T1 and T2, N0) tumors, there are three main management options available in the published literature:
In our series, one stage I tumor, and most of the stages II, III, and IV tumors, were treated by en bloc resection. Parotidectomy was included as part of the surgical procedure (as reported in other series3,11,13,15) in cases of confirmed or suspected involvement of the parotid gland.
Although the disease is rare, and numbers have been inadequate to demonstrate statistically significant differences, we did observe improved survival for both superficial and deep tumors when parotidectomy was performed. We therefore recommend a low threshold to include this as part of the surgical treatment of all temporal bone tumors.
Guidelines from the American Society for Head and Neck Surgery and the Society of Head and Neck Surgeons20 recommend postoperative radiotherapy when: (A) resection margins are close (less than 5 mm) or when proximity of tumor to important structures such as internal carotid artery or facial nerve precludes wide margins; (B) tumor is demonstrated at resection margins; (C) postoperative histology reveals perineural invasion. These conditions apply to the majority of resections for temporal bone carcinoma, thus indicating postoperative radiotherapy in most cases,21 although the above-mentioned systematic review (Prasad and Janecka2) showed no statistical significant difference in survival in patients who underwent lateral temporal bone resection and radiotherapy (48% overall 5-year survival) compared with patients who underwent lateral temporal bone resection alone (44% 5-year survival).
Our study shows that there is a trend toward increased overall survival in the superficial group of tumors if radiotherapy is used postoperatively, although this result was not statistically significant (p=0.18). There were insufficient data to analyze the effect of radiotherapy in the deep group.
Finally, the survival of these tumors has been variously reported in 2-, 3-, or 5-yearly intervals. Stage I and II tumors have been reported, depending on the series, to have a 50%2 to 100%10,11 survival. Stage III and IV survival varies from 0%10 to 50%.13,15 Our overall 3-year survival was 38%, and the 5-year survival 19%, perhaps due to the greater number of large (T3 to T4) tumors in this study.
Recurrence has been reported to be anywhere from 25% for stages I and II to 100% for stages III and IV.10 Our study achieved lower recurrence rates for stages III and IV (50 to 64%).
Malignant tumors of the ear and temporal bone are rare and there is no consensus in their staging and management.We were unable to demonstrate any difference in survival or recurrence for superficial or deep tumors of the temporal bone.Parotidectomy seemed to improve survival for both superficial and deep tumor groups, but statistical significance was not reached. Radiotherapy also seemed to improve survival for the superficial group, but again, statistical significance was not reached.