Ewing's Sarcoma of the Head and Neck: A Retrospective Analysis of 24 Cases

Introduction and purpose. Primary Ewing's sarcoma arising from the bones of the head and neck region is extremely rare representing only 1– 4% of all Ewing's sarcoma cases. Previous reports suggest a better prognosis for that particular anatomic site. The purpose of this study was to analyze the clinico-epidemiologic characteristics of that rare clinical presentation, as well as its patterns of failure and prognosis following treatment. Materials and methods. This study included a retrospective review of the medical records of patients with the diagnosis of Ewing's sarcoma of the head and neck region treated at King Faisal Specialist Hospital and Research Center between 1975 and 1996. Results. Out of a total number of 24 cases analyzed, there were 17 males and 7 females with a ratio of 2.4:1. The median age at diagnosis was 16.5 years. A painful swelling was the most common clinical presentation.The maxilla was the most common site of presentation (9/24 cases). There were 3/24 cases who presented with metastatic disease at diagnosis.The majority of patients (16/24 cases) had a tumor size >10 cm. Most patients were treated with systemic chemotherapy plus localized irradiation following an initial biopsy.With a mean follow up of 3.4 years, the 5-year actual overall survival (OS) for the whole group was 53%, while the 5-year actuarial disease-free survival (DFS) was 30%. These figures were higher than those repor ted from our institution for young patients (≤ 14 years treated for Ewing' s sarcoma in other anatomic locations (30% v 15%). The response to chemotherapy was the only prognostic factor that affected both the OS and DFS. Conclusion. The prognosis of Ewing's sarcoma of the head and neck region is slightly better than that of other anatomic sites.The response to systemic chemotherapy is one of the most important prognostic factors affecting both DFS and OS of Ewing's sarcoma of the head and neck. Multimodality therapy consisting of an initial biopsy, aggressive combination chemotherapy and localized radiotherapy is the treatment of choice for Ewing's sarcoma of the head and neck region and may result in long-term survival.


Introduction
Prim ar y Ew ing's sarcom a arising in the head and neck region is extrem ely rare, com prising 1± 4% of all cases of Ewing's sarcom a. 1,2 M ost authors claim a better prognosis for Ewing's sarcom a of the head and neck region as com pared to that arising in other anatom ic locations. 1± 5 T he m ost comm only affe cted bones in the head and neck region are the skull, the m andible and the m axilla. 1 There has also been case reports of localized Ew ing's sarcom a affecting the orbital roof, 3 the retrophar ynx 4 and the nasal cavity. 5 T he aim of the present study was to review the cases of Ew ing's sarcom a of the head and neck region treated at our institution during the past 20 years, aim ing at a better u nderstand in g of the clinicoepidem iologic characteristics of that rare anatom ic location. The details of treatm ent including surgery, chem otherapy and radiation therapy, as well as the patterns of failure are presented. Also, the various potential prognostic factors affe cting both the diseasefree ( D F S ) a n d th e o ver all su r viva l ( O S ) a re highlighted.

M aterials and m ethods
T he following study included a retrospective analysis of data availab le from the m edical records of 24 cases of prim ary Ewing's sarcom a arising in the head and neck region, treated at the King Faisal Specialist Hospital and Research Center during the period 1975 to 1996. T he King Faisal Specialist Hospital and Research C enter is a tertiar y care center serving the whole kingdom of Saudi Arabia. It is the largest oncology referral center at present receiving patients from all over the kingdom . T he analysis included reviewing various prognostic factors that could affect the treatment outcom e in term s of D FS as well as OS.T he charts were reviewed for age, sex, presenting sym ptom s and signs and duration of sym ptom s. T h e prim ar y tu m o r site w as determ ined for each patient. The tumor size as well as the initial stage of presentation were determ ined for all patients. Staging work-up of patients including com puterized tom ography (CT ) scan of the head and neck, chest X-ray, C T scan of the chest, bone scan, bone m arrow biopsy and serum lactic dehydrogenase (L D H) levels were reviewed.
T h e d i ff er en t tr ea tm en t m o d a liti es a d o p ted inclu din g su rger y, rad iotherapy and chem otherapy were reviewed in d etail. T he radiation therapy charts were reviewed for the total do se given, the nu m ber of fra ction s, the ove rall treatm en t tim e, the energy use d and the u se o f C T plan ing. C hem otherapy details inclu ded tim ing of system ic chem otherapy, agents used, num ber of cycles as well as the response to chem otherapy treatm ent.
Patients were followed up regularly every 3 m onths for the ® rst year, every 6 m onths in the subsequent 2 years and then yearly thereafter. Patients w ho were lost to follow -up were censored from the survival analysis at the time of last follow-up. N one of those patients participated in clinical trials.

Results
O ut of 24 evaluable patients, there were 17 m ales and 7 fem ales w ith a m ale to fem ale ratio of 2.4:1. T he m edian age of the patients was 16.5 years (range 2± 33 years). A painful swelling was the m ost com m on presentation found in m ore than 90% of patients as shown in Table 1. T he m ean duration of sym ptom s was 5.5 m onths (range 2± 12 m onths).
T he m axilla was the m ost com m on site of presentation (9/24 cases). M etastasis at initial presentation was found in only 3/24 cases. M ost of our patients had tum or size >10 cm in diam eter (16/24 cases).
A bone m arrow core biopsy from the iliac crest was perfor m ed in 19 patien ts; and w as fou nd to be in® ltrated by m alignant cells in 1 patient only. The LD H was assayed in 19 patients and was elevated in 12/19 (63%). A bone scan was done in 22/24 patients (92%). It showed localized increased uptake opposite the initial site of bony involvement in all 22 patients with no evidence of m etastatic spread.
T he m ajority of patients had biopsy only (16/24 cases); while the rem aining had either incom plete excision (6/24 cases) or com plete surgical excision (2/24 cases). T he m ost com m on treatm ent m odality was initial biopsy followed by com bined system ic chem otherapy plus radiotherapy to the prim ary site involved. T his m odality was adopted in 14/24 cases (58% ). T here were various other treatment m odalities, the m ost com m on of which was surgical excision followed by post-o perative chem o-ra dio therapy in 5/24 cases (21%).
System ic chem otherapy was delivered in 22/24 cases (92% ). Two patients refused system ic chemotherapy tre atm ent, 1 was treated w ith com plete surgical excision followed by post-operative radiotherapy while the other was treated with incom plete excision and refused post-operative irradiation. T he m ost com m on regim en given was VAC (vincristine, adriamycin, cyclophosph am ide; 10 patients) or VAC altern atin g w ith IE P (ifo sf am id e, etopo sid e an d cisplatinum ) given in 7 patients and VAIA (vincristine, adriamycin, ifosfam ide and actinomycin D ) in 5 patien ts. T h e ch o ice o f ch em o th erap y reg im en depended on the protocol adopted during a given period of time, i.e. patients treated from 1975 to 1980 received VAC alone, w hile those treated from 1980 to 1990 received VAC /IEP. All patients treated from 1990 onwards received VAIA. There were only 13/22 patients (59% ) who com pleted six or m ore cycles of system ic chemotherapy. C om plete rem ission was obtained in 57% of patients, w hile partial rem ission occurred in 43% of patients. N one of the patients had disease prog ression on chemotherapy.
Radiotherapy was delivered in 20/24 cases (83%), with a radical intent in 19/20 patients (95%). T he total radiation dose ranged from 3000 to 5600 cG y with a m edian dose of 5040 cG y. T here were 3/19 patients who received a total radiation dose £ 4000 cGy. The total dose of radiation given depended on the treating physician. It did change to a more radical dose (5400 cGy) over time. The majority of the patients (13/20) were treated with 6 MV energy photons. Six patients were treated with Cobalt 60 and the remaining patient with 8 M V energy photons. CT planing was used in 10/20 cases (50%). T here were 7/24 cases who developed LR (29% ), and 11/24 cases (46%) who developed D M .The lungs were the most com m on single site of D M (27% ); there were 4/11 patients (36%) who had m ore than one site of D M . Five out of 24 patients developed both LR and D M .
W ith a m ean follow-up time of 3.4 years, the 5-year actuarial O S for the whole group of patients was 53% ; w hile the actu arial 5-ye ar D F S w as 30% ( Fig. 1a and b). T here was no statistically signi® cant difference in 5-year O S and D F S between m ales and fem ales (54% vs 50% and 26% vs 36% , respectively). Tum or size did not seem to have any im pact on O S and D FS. T he actuarial 5-year D FS for patients with tum or size of 5± 10 cm was 30% com pared to 29% for those with tum ors >10 cm in diam eter (p = 0.93).
Analysis of the survival results (both D FS and O S) according to the prim ar y tum or site did not show any signi® cant im pact of that factor on the treatment outcom e. T he m andible had the highest 5-year O S (80% ), followed by the m axilla, skull and orbit (71%, 50% and 38%, respectively). T his difference was not statistically signi® cant.
T he response to chemotherapy treatment was the only prognostic factor which affe cted both the O S and D FS. T he 5-year O S for patients with com plete rem ission (CR) was 68% compared to 0% for patients w ith partial rem ission (PR) to chem otherapy (p = 0.015); also the 5-year D FS for CR patients was signi® cantly higher than that of PR patients (27% vs 0% ; p = 0.02).

D iscussion
Ew ing's sarcom as arising from the bones of the head and neck region are extremely rare. M ost cases in the literature are reported sporad ically as case reports. T he largest reported series until now is that of the Intergroup Ewing's Sarcom a Study (IESS) in 1987. Siegal et al. 1 reported in that series on a total of 29 cases of Ewing's sarcom a of the head and neck region w hich represented 4% of all Ew ing's sarcom a cases reported by the IESS group. In the present series, Ew ing's sarcom a of head and neck constituted 9% of all Ew ing's cases treated in this institute (24/259 cases). T his relatively high ® gure m ight be attributed to the fact that our hospital is the m ain referral oncology center in the region. T he m ale to fem ale ratio was 2.4:1, this is in accordance with the literature being a disease where m ales are m ore frequently affe cted. 6 T he m edian age at diagnosis for this study was 16.5 years, which represents an expected ® gure for Ewing's sarcom a, a disease affe cting prim arily patients <20 years of age. 7 The m axilla was the m ost comm on site of presentation in this study (9/24 cases). Although this represents a rare site of prim ar y presentation, yet there has been a total of 22 cases of prim ary Ewing's sarcom a of the m axilla reported in the literature. 8,9 In the present series, m ost patients were treated w ith in i tia l b io p s y fo l lo w e d b y s y s te m i c chem otherapy plu s radical rad iotherapy. T his is d ue to the fa ct that m o st of ou r patien ts have large lesions (>10 cm ) of the head and neck, w here radical su rger y would be m utilating. Radical surger y can , howeve r, play an im por tant role in achievin g local con tro l in cer tain areas (e.g. m and ible and skull), especially in tum or s < 5 cm in d iam eter w here the tr e at m e n t r e la te d m o r b id i ty w o u ld n o t b e increase d. 5,10,11 The local control rate in the present study was 71% .This ® gure is lower than the 90% ® gure reported in o th er ser ies. 12 T h ere a re tw o m ain facto r s affecting the local control rate, these are the initial tum or size an d the total rad iation dose delivered .
T h e m ed ian d o se in th is stu d y (5 04 0 cG y) is con sidered adequate to achieve local control by m ost investigators. 12,13 O f the 3 patients w ho received a total radiation dose £ 4000 cG y, on ly 1 developed local recurren ce. T he lower local con trol rate ® gure obtained in the pre sent study seem s to be attribu ted to the large tum or size (> 10 cm ) in the m ajority o f ou r patien ts (16/24 cases). M ost of o ur patients prese nt relatively late in the cou rse of their disease du e to lon g distan ces, cultural reasons an d lack o f efficient prim ar y care hospitals in rem ote areas o f the kin gd om . The 5-year actuarial D FS as well as O S in the present study (30% and 53% , respectively) are lower than the ® gures reported by IESS (80% survival at 3 years). Again we attribute those lower ® gures m ainly to tw o facto r s: th e larger tu m o r siz e at in itial prese ntation as well as the low com pliance rate to chem otherapy treatm ent, w ith o nly 59% of patien ts com pleting m ore than six cycles. D espite our lower sur vival rates, yet those are still higher than the 15% 5-year relapse free sur vival reported for a group of young patients (£ 14 years) w ith the d iagn osis o f E w ing' s sarco m a in o ther anatom ic locations treated in our institute between 1980 and 1993 (personal com m unication). Sim ilarly, the IESS reported a higher m edian survival for head and neck location in com pariso n to Ew ing's sarcom a in other anatom ic sites. 1 The only prognostic factor signi® cantly affe cting both D FS and O S in this study was the response to chem otherapy. This ® nding is in ag reement with the recent articles that emphasize the im portance of the initial response to chemotherapy treatment as the only signi® cant independent predictor of survival for Ewing's sarcom a patients. 14 It should be noted, however, that the power of subanalysis in this study was low due to the sm all num ber of patients.
In the early phase of the present study, patients w ere treated w ith th e u su a l VAC co m b in atio n reg im en s, how eve r, star ting in 1995, w ith the introduction of m ore aggressive chemotherapy, we are now using VAC A alternating with I/E (ifosfam ide, etoposide) for a total of 52 weeks. T his new alternating chemotherapy com bination regim en has already yielded a statistically signi® cant im provem ent in D FS when com pared to the standard VAC A alone in a phase III randomized study. 15 In conclusion, Ewing's sarcom a of the head and neck region carries a slightly better prognosis than that in other sites. M ultimodality therapy consisting of an initial biopsy, aggressive com bination chemotherapy and localized  radiotherapy is the usual treatm ent of choice and could result in long-term disease-free survival. Radical surgery m ay be perform ed for local control in certain anatomic locations (e.g. mandibular and skull lesions), however, this needs to be followed by an adequate reconstructive surgery. The response to chem otherapy treatm en t rem a in s o n e o f th e m o st im po r tan t progn ostic factors affe cting the D F S as well as the O S of patients with Ewing's sarcom a of the head and neck region.