Analysis of prognostic factors in 146 patients with anterior skull base sarcoma: An international collaborative study

Authors

  • Ziv Gil MD, PhD,

    1. Department of Surgery, Head and Neck Service, Memorial Sloan-Kettering Cancer Center, New York, New York
    Search for more papers by this author
  • Snehal G. Patel MD,

    1. Department of Surgery, Head and Neck Service, Memorial Sloan-Kettering Cancer Center, New York, New York
    Search for more papers by this author
  • Bhuvanesh Singh MD,

    1. Department of Surgery, Head and Neck Service, Memorial Sloan-Kettering Cancer Center, New York, New York
    Search for more papers by this author
  • Giulio Cantu MD,

    1. Department of Cranio-Maxillo-Facial Surgery, National Cancer Institute, Milan, Italy
    Search for more papers by this author
  • Dan M. Fliss MD,

    1. Department of Otolaryngology, Head and Neck Surgery, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel
    Search for more papers by this author
  • Luiz P. Kowalski MD,

    1. Department of Otolaryngology, Head and Neck Surgery, Cancer Hospital A. C. Camargo, Sao Paulo, Brazil
    Search for more papers by this author
  • Dennis H. Kraus MD,

    1. Department of Surgery, Head and Neck Service, Memorial Sloan-Kettering Cancer Center, New York, New York
    Search for more papers by this author
  • Carl Snyderman MD,

    1. Department of Otolaryngology, Head and Neck Surgery, University of Pittsburgh, Pittsburgh, Pennsylvania
    Search for more papers by this author
  • Jatin P. Shah MD,

    Corresponding author
    1. Department of Surgery, Head and Neck Service, Memorial Sloan-Kettering Cancer Center, New York, New York
    • Head and Neck Service, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021
    Search for more papers by this author
    • Fax: (212) 717-3304

  • for the International Collaborative Study Group

    Search for more papers by this author
    • Other members of the International Collaborative Study Group: Simon G. Talbot, MD, and Ashok Polluri, MD (Memorial Sloan-Kettering Cancer Center, New York, NY); Patrick G. Bridger, MD (Prince of Wales Hospital, Sydney, Australia); Anthony D. Cheesman, FRCS (the Royal National Throat, Nose, and Ear Hospital, London, United Kingdom); Geraldo De Sa, MD (Istituto Nacional de Cancer, Rio de Janeiro, Brazil); Paul Donald, MD (University of California-Davis, Sacramento, Calif); Patrick Gullane, MD (Toronto General Hospital, Toronto, Ontario, Canada); Ivo Janecka, MD (Massachusetts Eye and Ear Hospital, Boston, Mass); Shin-etsu Kamata, MD (Cancer Institute Hospital, Tokyo, Japan); Paul A. Levine, MD (University of Virginia Health System, Charlottesville, Va); Luiz R. Medina, MD (Hospital das Clinicas, Sao Paulo, Brazil); Sultan Pradhan, MD (Tata Memorial Hospital, Mumbai, India); Victor Schramm, MD (Center for Head and Neck Surgery, Denver, Colo); William I. Wei, MD (University of Hong Kong, Hong Kong, China)


  • See editorial on pages XXX-XXX, this issue.

Abstract

BACKGROUND.

Single-institutional studies lack sufficient power to assess the role of surgery and radiotherapy in the management of sarcomas involving the anterior skull base. For this study, an international collaborative study group analyzed a large cohort of patients who underwent surgery for the treatment of skull base tumors.

METHODS.

A subset of 146 patients who had a histologic diagnosis of skull base sarcoma (SBS) formed the basis of this report. Most patients were aged ≥21 years (77%) and had stage IV disease (56%). Adjuvant radiotherapy was received by 35% of patients, and chemotherapy was received by 10% of patients.

RESULTS.

Orbital involvement was encountered in 53% of patients, involvement of the orbital wall was encountered in 46% of patients, and intracranial extension was encountered in 28% of patients. Positive microscopic margins were reported in 43% of patients (51 of 118 patients). Treatment-related complications were reported in 27% of patients, and postoperative mortality was reported in 1.4% of patients. With a median follow-up of 34 months, the 5-year overall, disease-specific, and recurrence-free survival rates were 62%, 64%, and 57%, respectively. Tumor grade and adjuvant radiotherapy were not significant predictors of survival. Prior radiotherapy, intraorbital extension, positive margins, and postoperative complications were significant predictors of reduced disease-specific survival on univariate analysis. The presence of positive/close margins, however, was the only independent predictor of poor overall, recurrence-free, and disease-specific survival on multivariate analysis (relative risk, 2.4; P = .006). The 5-year disease-specific survival rate was 77%, 43%, and 36% for patients with negative, close, and positive margins, respectively.

CONCLUSIONS.

The current results indicated that wide craniofacial resection with negative margins is an independent prognostic predictor of better outcome. Patients with positive margins have a high risk for tumor recurrence independent of tumor grade. Cancer 2007. © 2007 American Cancer Society.

Soft tissue sarcomas (STS) of the head and neck represent a heterogeneous group of tumors with distinct clinical implications. STS are found infrequently in the mature population, in which they account for 10% of all sarcomas1–3 and only 1% of head and neck tumors.4 These tumors are more common in childhood, and up to 33% occur in the head and neck region.5, 6 The skull base is one of the common anatomic regions involved in head and neck sarcomas.7, 8 The majority of these patients present with symptoms, the most frequent of which is a painless mass.4 Other common symptoms are epistaxis, nasal obstruction, pain, and cranial nerve deficits.9

Among all patients with sarcoma, those who have tumors originate in the skull base region have the lowest overall survival (OS) rates.10 The limited number of studies on skull base sarcomas (SBS) have reported an 5-year OS rate of 75% and a progression-free survival rate at 5 years of 60%.11 Most series indicated that approximately 50% of patients who fail will have done so within 2 or 3 years after the completion of treatment.12, 13 Unlike sarcomas in other anatomic sites, in which distant metastases are the most common cause of death, the great majority of patients with SBS will succumb to local recurrence at the skull base that involves vital structures, such as the carotid arteries, cavernous sinus, and brain.

The primary treatment modality for most patients with SBS remains wide surgical resection. A major determinant of outcome for these patients is the adequacy of surgical resection of the primary tumor. This concept, along with the development of new surgical and reconstructive techniques, has led to the development of a variety of radical surgical approaches for malignant tumors originating in the anterior skull base.14 An evolution in the therapy has resulted in the use of postoperative adjuvant radiation, which may improve local control and disease-free survival for patients with high-grade tumors.15 Adequate surgical resection has remained a cornerstone to the successful treatment of patients with SBS.16

Although head and neck sarcomas have been described well in the surgical and oncologic literature, there is a lack of clinical data on sarcomas involving the anterior skull base. The clinical behavior, staging, treatment, and outcomes have been characterized poorly for patients with SBS. A significant contributor to the debate regarding this disease originates from the difficulty in diagnosis of skull base soft tissue tumors based on conventional pathologic and cytologic measures. A major factor in the behavior of SBS is that soft tissue sarcomas belong to a spectrum of diseases with high and low grade. Overall, a significant debate remains whether combined-modality treatment, including surgical resection and radiation therapy, provides the best results in terms of disease control and survival.

Craniofacial resection (CFR) is a well-established technique for the surgical excision of tumors involving the anterior skull base and paranasal sinuses.17, 18 This technique requires a frontal craniotomy along with a transfacial approach to allow broad exposure of the anterior cranial fossa and subcranial compartment. The use of CFR and other surgical approaches to the anterior skull base (ie, the subcranial approach) has been documented for a variety of malignant and benign tumors.19, 20 However, because of the rarity of these neoplasms, little is known about the surgical results and clinical outcomes of patients undergoing CFR specifically for SBS. All of the reports on anterior SBS are based on cohorts of patients with head and neck sarcoma who received varying treatment regimens; therefore, identification of independent prognostic predictors using multivariate analysis has not been possible. To address this dilemma, data from an international collaborative study with clinical information on >1300 patients who underwent CFR were used, and a subset of patients with SBS was analyzed. The objective of this study was to assess the efficacy of surgical treatment and adjuvant therapy, in a multi-institutional cohort and to identify independent predictors of outcome after treatment using multivariate analysis. To our knowledge, this is the only study published to date on sarcomas isolated to the anterior skull base.

MATERIALS AND METHODS

A multi-item questionnaire was sent to 17 participating investigators from various parts of the world to obtain information on patient, tumor, and treatment variables and on outcomes of treatment. Completed data sheets on patients who were treated between October 1956 and January 2000 were submitted by each investigator to the central analyzing office for data entry, compilation, and statistical analysis.

Data Entry, Patient Exclusions, and Statistical Methods

Data were entered into a commercially available spreadsheet (Microsoft Excel 2000; Microsoft Corporation, Seattle, Wash), and statistical analysis was performed using a computerized software package (JMP version 4.0; SAS Institute Inc., Cary, NC). We previously reported the results from CFR in 1307 patients who had malignant tumors of the skull base,18 and the current report focuses on a subset of 146 patients from that group who had a histologic diagnosis of sarcoma (11%).

The follow-up interval was calculated in months from the date of CFR to the date of last follow-up or death, and the recurrence-free interval was calculated from the date of CFR to the date of first recurrence. For disease-specific survival (DSS), patients who reportedly died of noncancer-related causes (7 of 146 patients; 4.8%) and patients who were alive with disease at last follow-up (12 of 146 patients; 8.2%) were censored. Recurrence-free survival (RFS), OS, and DSS rates were calculated using the Kaplan-Meier method, and univariate comparisons between groups were performed using the log-rank test. All P values ≤.05 were considered significant, and significant factors were entered into multivariate analysis using the Cox proportional-hazards model. Nonparametric qualitative and quantitative comparisons were performed using the Fisher exact test and the Mann-Whitney U test, respectively.

Patient Demographics

Patients ranged in age from 1 year to 75 years (median age, 39 years). Table 1 presents the demographic data on these patients. Most patients were aged >21 years (n = 112 patients; 77%), and 5 patients were aged ≥70 years (3.4%).

Table 1. Patient Demographics
CharacteristicNo. of patients%
Total cohort146100
Age, y
 Range1–75 
 Median39 
 Age group
  ≤203222
  21–404531
  41–706343
  >7053.4
Data not present10.6
Sex
 Men5235.5
 Women9263
 Data not present21.5
Medical comorbidity
 None11881
 Present64
 Data not present2215

Primary Tumor

The types of sarcomas encountered in the study population are indicated in Table 2. The anatomic extent of the primary tumor is listed in Table 3.

Table 2. Tumor Types Among 146 Patients with Anterior Skull-Base Sarcomas
TumorNo. of patients%
  1. NOS indicates not otherwise specified.

1. Low-grade sarcoma, NOS5235.6
2. Osteosarcoma1913
3. Rhabdomyosarcoma1711.6
4. Fibrosarcoma149.6
5. High-grade sarcoma, NOS117.5
6. Leiomyosarcoma106.8
7. Malignant fibrous histiocytoma85.5
8. Angiosarcoma53.4
9. Neurofibrosarcoma42.7
10. Epithelioid sarcoma21.4
11. Ewing sarcoma10.7
12. Dermatofibrosarcoma10.7
13. Carcinosarcoma10.7
Table 3. Tumor, Lymph Node, Metastasis (TNM) Classification and Staging
ClassificationNo. of patients%
Tumor classification
 T185.5
 T242.7
 T32819
 T48155
 Data not present2517
Lymph node classification
 N014498
 N122
 N200
 N300
Metastasis classification
 M0146100
 M100
Stage
 I85.5
 II42.7
 III2718
 IV8256
 Data not present2517

The common tumor (T) classification of soft tissue sarcomas is based on the staging system revised by the American Joint Committee on Cancer (AJCC), which takes into consideration only tumor depth (above or deep to the superficial fascia) and tumor dimension (<5 cm or >5 cm). Because most of our tumors were deep and measured ≤5 cm, and in the absence of a conventional classification for SBS, we adopted the AJCC staging criteria for nasal and paranasal epithelial tumors. The same classification system was used previously for other nonepithelial skull base neoplasms, such as melanomas and soft tissue sarcomas.18, 19, 21–23 Staging for 121 patients (83%) was based on reported data and available information on orbital and intracranial extension (see Table 3).

Treatment

Most patients had received some form of treatment before CFR (59%). Previous treatment modalities included surgery (38%), radiotherapy (35%), and chemotherapy (24.5%). The details of prior radiation therapy were available in 80 patients, and the dose ranged from 28 grays (Gy) to 82 Gy (median dose, 50 Gy).

Adjuvant postoperative radiotherapy was received by 50 patients (34%) at a median dose of 60 Gy (range, 22–70 Gy). Adjuvant chemotherapy was received by 9 patients (6%), and chemoradiation was received by 5 patients (3.5%).

Follow-up

The follow-up interval ranged from 1 months to 171 months (median, of 34 months). The length follow-up was >3 years in 70 patients (45%).

RESULTS

Low-grade sarcoma occurred in 52 of our patients (35.6%), and 94 patients had high-grade sarcoma (64.4%). The most common types of sarcoma were osteosarcoma (n = 19 patients; 13%), rhabdomyosarcoma (n = 17 patients; 11.6%), and fibrosarcoma (n = 14 patients; 9.6%). Table 2 summarizes the histologic diagnosis of the study patients. The majority of patients had stage IV disease (81 of 146 patients; 55%). Only 2 patients had neck metastases. None of the patients had distant metastases at the time of surgery, because those with distant metastases received chemotherapy and radiation instead of surgery. Orbital involvement was reported in 53% of patients (n = 78 patients), and the tumor involved the orbital bone in approximately 50% of the patients in that group (36 of 78 patients). Intracranial extension of the tumor (ie, infiltration of the dura or brain) was reported in 28% of patients (n = 41 patients). Table 3 summarizes the location, extent, and stage of the tumors.

Details on surgery, reconstruction, and surgical margins status are listed in Table 4. Close or microscopically positive margins were reported in 51 of 118 pathologic specimens (43%). The likelihood of negative margins was 48% (38 of 79 patients) among patients with high-grade tumors and 33% (13 of 39 patients) among patients with low-grade tumors. It is noteworthy that there was no significant difference in the risk of positive margins between the 2 groups (P = .3).

Table 4. Details of Surgical Treatment
TreatmentNo. of patients%
Tracheostomy
 No9565
 Yes3625
Reconstruction
 No106.8
 Yes11377.4
 Data not present2315.8
Type of reconstruction, n = 113
 Locoregional flaps5940.4
 Free flaps2617.8
 Autologous nonvascularized2114.4
 Nonvascularized bone32.1
 Cadaver/bovine pericardium21.4
 Multiple flaps10.7
 Titanium mesh10.7

Seventy-seven percent of patients underwent reconstruction, and the most frequently used method was a locoregional flap (ie, galeal pericranial and temporalis muscle flaps), which was used in 59 of 113 patients (40%). Free flaps (rectus abdominis, radial forearm, and lateral thigh) were used in 26 of 113 patients (17.8%).

Data on postoperative complications were recorded in 132 patients (Table 5). Overall, 40 patients (27%) developed a postoperative complication. The most common complications were related to the local wound (17%) or the central nervous system (13.5%). The postoperative mortality rate was 1.4% (n = 2 patients).

Table 5. Postoperative Complications After Craniofacial Resection
Complication typeNo. of patients%
  1. CNS indicates central nervous system.

Complication
 No9263
 Yes4027.4
 Data not present149.6
Wound complications
 No11075
 Yes2215
 Data not present1410
CNS complications
 No11486.5
 Yes1813.5
Systemic complications
 No13098.5
 Yes21.5
Orbital complications
 No12796.3
 Yes53.7
Postoperative mortality
 No13098.5
 Yes21.5

With a median follow-up of 34 months (range, 1–171 months), the 5-year OS and DSS rates calculated with the Kaplan-Meier method were 62% and 64%, respectively (Fig. 1). The median time to recurrence after CFR was 17 months, and the 5-year RFS rate was 57%. The majority of patients (60 of 77 patients; 78%) who were recurrence-free had been followed for an interval longer than the median time to recurrence (median, 42 months).

Figure 1.

With a median follow-up of 34 months, the 5-year overall (black), disease-specific (dotted line), and recurrence-free (gray line) survival rates calculated by the Kaplan-Meier method were 62%, 64%, and 57%, respectively.

On univariate analysis, prior radiotherapy, intraorbital invasion, and positive/close surgical margins were predictors of worse DSS (Table 6). Tumor grade and adjuvant radiotherapy were not significant predictors of survival. On multivariate analysis, a positive/close surgical margin was the only independent predictor of worse DSS (relative risk, 2.4; P = .006). The 5-year DSS was 77% in patients who had negative margins and 34% in patients who had positive/close margins (Fig. 2). There was no statistical difference in survival between the patients with positive and close surgical margins.

Figure 2.

Multivariate analysis identified a positive surgical excision margin as the only independent predictor of worse disease-specific survival (relative risk, 2.4; P = .006). The 5-year disease-specific survival rate was 77% in patients who had negative margins and 34% in patients who had positive or close margins.

Table 6. Prognostic Factors for Disease-Specific Survival
VariableNo. of patients5-year DSS, %P*
  • DSS indicates disease-specific survival.

  • *

    Univariate analysis (log-rank test).

  • Multivariate analysis (P = .006).

Age, y  .8
 <404162 
 >419770 
Sex  .06
 Women5253 
 Men9270 
Medical problems  .6
 None11858 
 Yes655 
Previous surgical manipulation  .5
 No8964 
 Yes5663 
Previous radiation therapy  .03
 No5972 
 Yes8659 
Intraorbital extension  .045
 None6667 
 Yes7848 
Intracranial extension  .68
 None6470 
 Bone3952 
 Brain/dura4156 
Tumor grade  .15
 Low5374 
 High9358 
Surgical margins  .0006
 Negative6777 
 Close2643 
 Positive2536 
Postoperative complications  .0035
 No8768 
 Yes4039 
Adjuvant radiotherapy  .32
 No9269 
 Yes5052 

DISCUSSION

Sarcomas constitute a heterogeneous group of tumors that can originate in connective tissue, muscle, skeleton, vessels, or nerves. In adult patients with sarcoma originating from all anatomic sites, a poor prognosis is associated with large tumor size; high histologic grade; invasion to bone, nerves, and vessels; and positive surgical margins.13, 24 Patients who have SBS have a significantly worse outcome compared with the outcome of patients who have sarcomas of the extremities or superficial trunk.12 This probably is related to the late detection of tumors, but it also represents the surgical challenges in resecting tumors at this location. The route of spread for tumors that originate in the anterior skull base and paranasal sinuses is determined by the complex anatomy of the craniofacial compartments. These tumors may invade laterally to the orbit and middle cranial fossa, inferiorly to the maxillary antrum and palate, posteriorly to the nasopharynx and pterygopalatine fossa, and superiorly to the cavernous sinus, carotid artery, dura, and brain. The proximity of the primary tumor to vital organs complicates tumor resection and adjuvant radiotherapy and, thus, places these patients at risk for local recurrence.

In this multicenter study, we combined data from 17 different institutions on 146 patients with soft tissue and bone sarcoma of the anterior skull base and paranasal sinuses. Approximately 87% of the tumors originated in soft tissue; the remaining tumors were osteosarcomas. The 5-year actuarial RFS rate was 57%, and the 5-year DSS and OS rates were 64% and 62%, respectively. Compared with other tumors of the anterior skull base, our data indicated that patients who had sarcoma had a better prognosis than patients who had adenocarcinoma, squamous cell carcinoma, undifferentiated carcinoma, and malignant melanoma and a survival similar to that of patients who had minor salivary gland tumors.18

Among adults, the most common sarcoma of head and neck origin is osteosarcoma, followed by malignant fibrous histiocytoma; whereas, in children, rhabdomyosarcoma is the most common.7 In our study, osteosarcoma and low-grade sarcoma were the most common neoplasms. We identified only 1 previous publication that addressed the issue of sarcomas of various origins along the cranial fossa. Among 63 patients in that study, it was reported that chordoma and chondrosarcoma were the most common neoplasms of the middle cranial fossa.11 None of our patients had these tumor types. It may be argued that the previous series focused on a distinct population of tumors (ie, distinct anatomic sites) with different biologic behavior than the tumors presented in our current series.

Among patients with head and neck sarcoma, those with chondrosarcoma or dermatofibrosarcoma have the best prognosis, whereas those with osteosarcoma, angiosarcoma, and rhabdomyosarcoma have the poorest outcome.25 It is noteworthy that, among patients who had tumors in the anterior cranial fossa, we did not observe differences in survival between those with various tumor types. This may be because most of our patients had advanced-stage disease, and their tumors were located in proximity to critical structures. Alternatively, the small number of patients with each subtype of sarcoma may have reduced the statistical power of the study.

The International Consensus Symposium on the diagnosis and treatment of adult soft tissue sarcomas in the head and neck determined that wide surgical resection with negative margins is the treatment of choice for patients with SBS.26 Nevertheless, it is difficult to achieve en bloc tumor resection in patients who have sarcomas that infiltrate the anterior skull base. Although tumor resection should be the objective of surgery, our series, along with other reports, indicates that negative surgical margins are achieved in <50% of patients.1 This high positive margin rate may be related to the biologic behavior of SBS and its proximity to the brain, dura and carotid artery, cavernous sinus, and orbit. However, this rate is higher than that reported for other anterior skull base malignancies.18, 19 A high rate of positive margins is of particular interest in view of our observation that this pathologic finding is associated with a significantly poor outcome (positive histologic resection margin status can reduce DSS by 43%). Various studies that examined patients with soft tissue and bone sarcomas of the head and neck consistently reported an association between positive margins and a low survival rate.4, 25, 27, 28 It also was reported that positive surgical margins predicted poor survival in patients with other malignant skull base tumors18, 19 and in patients with sarcomas at other anatomic sites.10 In keeping with our results, the results from a study at Princes Margaret Hospital in Toronto indicated that the most common cause of death in patients with head and neck sarcoma was uncontrolled local recurrence caused by residual disease after surgery.13 For that reason, it was inferred that microscopic resection margins should be included in the staging system for soft tissue sarcomas.29 In agreement with those findings, we clearly demonstrated the importance of the complete clearance of all overt tumor extensions, achieving both macroscopic and microscopic clean margins. Unfortunately, the head and neck in general and the anterior skull base in particular are among the most difficult anatomic regions for the establishment of negative margins.29 Many of these patients reportedly harbor residual tumor at the primary site, causing a significant risk for local recurrence. These patients potentially may benefit from undergoing re-excision of the tumor.30 However, the benefit of re-excision in patients with head and neck sarcomas or SBS is undetermined.

It has been demonstrated that postoperative radiotherapy improves local control of soft tissue sarcomas of various origins,10, 31, 32 although the benefit of radiation therapy for treating SBS has not been evaluated to date. Regardless of the lack of prospective studies on the benefit of radiotherapy for head and neck sarcomas or SBS, there has been a significant increase in the use of adjuvant radiation therapy for the treatment of this disease, especially in patients who have high-grade tumors.25 Nonetheless, our multivariate analysis failed to reveal differences in survival between patients who underwent surgery alone and patients who received adjuvant radiotherapy. One possible contributor to the failure of radiotherapy to improve survival may be the presence of positive margins after surgery. Colville et al. reported a benefit from radiotherapy for head and neck sarcomas only in patients who had negative surgical margins.16 Given our results after treatment of SBS, any modifications in standard treatment recommendations should be made judiciously. Adjuvant radiotherapy generally is recommended for patients who have undergone a marginal surgical resection or for patients with high-grade sarcoma. Whether patients with advanced-stage, low-grade sarcoma (ie, infiltrating the orbit, cribriform plate, pterygoid plates, or dura) should be treated with adjuvant radiotherapy remains an open question. The long-term side effects of radiation become a consideration in these patients, who generally are young or may be at risk for radiation-induced sarcoma of the head and neck.

There also is significant debate regarding the advantage of chemotherapy for the treatment of head and neck sarcomas.26 A variety of chemotherapeutic agents, as either adjuvant or neoadjuvant therapy, has been used in the treatment of head and neck sarcomas based on the treatment of patients with sarcomas involving the extremities.11, 33 Unfortunately, with the exception of rhabdomyosarcoma, most authors reported no advantage for chemotherapy in patients with head and neck sarcomas13 or SBS.11 In general, sarcomas that occur in children and young adults tend to be chemotherapy-sensitive, particularly osteogenic sarcoma, rhabdomyosarcoma, and Ewing sarcoma.34 Conversely, in the more common situation of adult sarcomas, failure of chemotherapy occurs because of a high incidence of multidrug resistance.35 To date, the use of chemotherapy as postoperative adjuvant therapy or in a preoperative setting in an effort to decrease the extent of locoregional treatment is not established completely as a standard of care.36 Hence, except for certain sensitive tumors, chemotherapy with or without radiotherapy should be reserved for clinical trials or should be applied in a palliative setting for patients who have inoperable tumors or distant metastases.26

Our analysis suggests that patients with SBS have a high risk for tumor recurrence regardless of the tumor grade. In the high-grade group, 43% of patients had tumor recurrence compared with 40% in the low-grade group. In agreement with our data, a study from the University of Texas M. D. Anderson Cancer Center on 799 patients with sarcoma of various origins also failed to demonstrate a survival benefit in patients with low-grade tumors.10 It is noteworthy that, whereas some studies on head and neck sarcomas have suggested that grade is a prognostic indicator for survival,1, 13 other studies on sarcomas of all anatomic sites and on sarcomas isolated to the head and neck reported no significant association between local recurrence and tumor grade.11, 37 In the current study, we demonstrated that, in the skull base in particular, patients have a high risk for tumor recurrence regardless of tumor grade. This probably is because of the close proximity of the tumor to critical structures (ie, cranial nerves, orbit, and brain). The same anatomic constraints also curtail the ability to obtain wide surgical resection (with negative margins) and the delivery of high-dose radiotherapy (because of the risk of radiation-associated complications).

Head and neck sarcomas rarely metastasize to neck lymph nodes. In our series, the rate of locoregional metastases from SBS was 2%, similar to that reported for head and neck sarcomas in general.4 Because neck lymph nodes rarely are encountered in patients with SBS, neck dissection should only be performed if regional metastases are identified based on clinical and radiologic evaluation.

Our study raises a few questions in view of the high rate of residual disease in both low-grade and high-grade SBS: 1) Should patients with low-grade sarcoma undergo radical surgery similar to patients with high-grade sarcoma? 2) Should surgical margins be evaluated using frozen sectioning after gross tumor extirpation? 3) Should these patients undergo re-resection in case of unplanned excision or if gross positive margins are reported? 4) Finally, should postoperative adjuvant radiotherapy be administered to patients who have positive/close surgical margins or for high-grade tumors? Because of the lack of randomized data, it is difficult to delineate treatment guidelines for patients with sarcomas of the anterior cranial fossa. Because the only prospective, randomized trials that recommend adjuvant radiotherapy were based on a population of patients with soft tissue sarcomas of the extremities,29, 32 whether postoperative adjuvant radiotherapy will improve the survival of patients who have SBS (both low-grade and high-grade) with or without positive surgical margins remains a matter of debate and awaits further studies.

The current study presents some obvious weaknesses that arise in this type of retrospective data collection and analysis. Similar to other studies that involved patients who were treated over a relatively long time, our study is susceptible to potential contamination of the data with tumors of other histologic types rather than soft tissue or bone sarcoma. In addition, accurately subtyping sarcoma on the basis of histology and immunohistochemistry alone often can be difficult. Although molecular methods, such as karyotyping, have improved the ability of pathologists to classify sarcomas, these techniques are not feasible on retrospective analysis. In our study, 63 of 146 patients had tumors that were diagnosed as sarcoma not otherwise specified, and it is conceivable that a proportion of them would be reclassified if the histologic material were examined prospectively, especially with newer techniques. Given the limitations of this multicenter study, many of the controversies regarding SBS will have to await further study.

Nevertheless, the relatively large group of patients, along with the detailed set of data available for analysis, enabled us to analyze independent predictors of outcome. Such a multivariate analysis approach was not feasible previously because of the infrequent occurrence of sarcomas of the anterior cranial fossa.

In conclusion, this collaborative analysis of patients who received treatment at various institutions across the world demonstrated the efficacy of CFR in the management of anterior skull base soft tissue and bone sarcomas. Regardless of tumor grade, total surgical resection with negative margins is crucial to achieve acceptably disease-free survival, with or without adjuvant radiation therapy. Morbidity and mortality after CFR for SBS is relatively low, suggesting that this surgical approach is appropriate for the treatment of patients with anterior SBS. Further study is needed to define the value of radiation therapy as adjuvant treatment after surgery.

Acknowledgements

We thank Esther Susan N. Perez, MD, Nancy Bennett, Arlene Cooper, and Chan-Bene Lin for their assistance