Periosteal Ewing's Sarcoma: Report of Two New Cases and Review of the Literature

Background. The origin of Ewing's sarcoma in a periosteal location is rare and not clearly documented. Other malignant bone tumors appear to have a somewhat better prognosis when confined between periosteum and bone. Is it the same for periosteal Ewing's sarcoma? Methods. We describe two new cases and comprehensively review the literature consisting of 18 documented cases since the condition was first described in 1986 (S.M. Bator.Cancer 58:1781– 4). Results. Periosteal Ewing's sarcoma differs from the other forms of Ewing's sarcoma in terms of sex predominance, location of tumor, surgical stage at presentation and typical imaging studies. Eighteen out of the 20 patients were reported to be alive with no evidence of disease. Conclusions. It seems that the prognosis of this rare variant of Ewing's sarcoma family of tumors might be better but the small number of cases precludes such a firm conclusion.


Introduction
T he classi® cation of prim ary m alignant bone tum ors is still expanding to include new subtypes based upon th e clin ical pr esen ta tio n , rad io g ra ph ic fe atu res (in clu d in g m o d er n im ag in g m o d a lities su ch as com puted tom ography (CT ), m agnetic resonance im ag ing (M RI), positron emission tomography scan (PET scan) and others, 1± 4 anatom ical localization an d n ew so p histicated m o d er n cyto genetic an d m olecular biology techniques. 5± 9 I t is o b v io u s th at th e classical h isto lo g ical/ m orphological m ethods of classifying bone tum ors are crude and not of suffic ient accuracy to distinguish subtypes. Subtyping is m ost probab ly of prognostic signi® cance, but m ight have therapeutic im plications. M any think that bone tum ors arising in a periosteal location (surface lesions) have a som ewhat better prognosis than those arising in the m edullary cavity of the sam e bones. 10± 11 Ewing's sarcom a (ES) is one of the best exam ples of a distinctive tum oral disease which, based upon clinical radiological and cytogenetic param eters, 1,9 which, is considered today to be a gro u p of d iffe re n t d ise ases. A cco rd in g to th e anatomical site today we recognize three subtypes: . T he ® r st description of periosteal Ewing's sarcom a (PES) was probab ly that published in 1956 by Sherm an and Soong in a com prehensive radiological review of 111 cases of ES of bone which included a roentgen clas-si® cation. 20 They described three cases of PES am ong 12 other cases, which they de® ned as ª cortical Ewing's sarcom a of long bones,º but without m entioning the nam e and obviously based only upon plain X -ray ® lm s and classical histological criteria. T he ® rst wellestablished case report of PES was published in 1986 by Bator et al. 15 H e actually de® ned PES: ª . . . in a periosteal location w ithout extension into either the bon e or ad jac en t soft tissu esº . Since th en fou r additional papers, describing a total of 18 cases, have been published. 16± 19 We add to this list two new cases and review the literature.

Case repor ts
(1) A 16-year-o ld m ale patient was referred to the Orthopedic Oncology U nit in our center on July 1994. He com plained of a grow ing, large (>10 cm ) painful mass in the postero-lateral aspect of the distal half of his right thig h, w hich he had experienced for 3 months. H is general condition was good except for low fever for the last few m onths. Physical examination revealed a tender longitudinal m ass along the biceps m uscle in the right distal thigh, 12.5 cm in size. N o palpable lym ph nodes were noted in the groin or other places. Blood tests showed an elevated erythroc yte sedim entation rate (ESR) of 100/120; no r m al w h ite b lo od cell (W B C ) co u nt; n or m al alkaline phosphatase (AP) blood levels; and a slight increase of lactate dehydrogenase (LDH) blood levels. Plain X-ray ® lm s of the thigh and knee region showed priosteal elevation and thickening. O ur differential diagnosis was of an infectious disease, such as prim ary osteom yelitis, or secondary to a soft tissue process or some form of a m alignant surface bone neoplasm . Protocol staging studies included: plain chest X -ray ® lm ; total body bone scan; CT of the lesion and chest; and M RI of the lesion. The CT and MRI of the distal fem ur showed a periosteal/surface lesion, as the m edullary canal and the endosteal surface of the distal fem ur were intact. The system ic bone scan and chest C T were norm al. T he patient underwent an open incisional biopsy under general anesthesia. T he histopathological results indicated a classical ES in a periosteal location. According to AM ST S (Enneking's) surgical stag ing system the patient was in stage II-B. 21 He received neoadjuvant chem otherapy and, in D ecem ber 1994, underwent a lim b-sp aring operation where the distal half of the right fem ur and knee joint were resected including the lateral ham string muscles and biopsy scar. The defect was replaced by a m odular endoprosthesis. H istopathological evaluation of the specim en showed 100% necrosis and practically no tum or m ass was found. H e continued the sam e chem otherapy until July 1995. T he area was not irradiated and during the follow -up period of 3 years since then he has been free of disease.
(2) A 27-year-old m ale patient was referred to our unit on O ctober 1995 because of a painful growing mass in the m edial aspect of his right thigh for 2½ months. His general condition was good. Physical exam ination showed a tender longitudinal lesion in the right m id-m edial asp ect of his thigh. T here was no evidence of palp able lym ph nodes in any locations. Blood tests, including ESR, W BC count, AP and LD H, were all norm al. Conventional radiograph of the fem ur showed generalized periosteal thickening, with an area of bulging periosteum and a slight hypo dense region within it. A soft tissue com ponent was noted (Fig. 1). O ur differential diagnosis was either a soft tissue tum or encroaching upon the bone or a m alignant surface bone neoplasm. Staging studies included plain chest X-ray ® lm ; total body bone scan; CT of the lesion and chest and M RI of the lesion. As in the previous case, the C T and M RI of the thigh showed a periosteal/surface lesion without any involvement of the m edullary canal (Fig. 2). System ic bone scan and chest CT were normal. O n O ctober 1995 he underwent a core needle biopsy. The histopathological result was classical ES in a periosteal location. According to AMST S the patient was in stage II-B.  one in the ® bula and one involving the scapula. In 16 cases the tumor was diaphyseal and in three it was m etadiapheaseal. All patients were at stage II-B at presentation (according to AMSTS), 21 meaning that no metastases were detected in baseline staging studies. N on-surgical treatment included chemotherapy for all 20 patients. For 14 patients details were noted about the method by which chemotherapy was given. In 13 of the 14 cases neoadjuvant chemotherapy was given and in one case adjuvant chemotherapy was given. Precise documentation about the timing of radiation therapy was also noted for 14 of the 20 patients. Of these, ® ve patients received external beam radiation therapy (two patients received only preoperative radiation therapy and three patients received combined pre-and postoperative radiation therapy). All of the remaining six patients received radiation therapy, but whether it was pre-or post-operative was not stated. Hence, 11 of the 20 patients received radiotherapy. N ineteen of the 20 patients had tum ors located in long bones (one was at the scapula). Of these 19 patients, 18 underwent lim b-sp aring surgery and one underwent am putation.
Follow-up periods stated in the papers at publication were between 2 m onths and 10 years, w ith a m ean of 3 years. For 11 patients follow-up periods of m ore than 2 years were noted. Two patients were dead, at 1 and 2 years after end of therapy, and the rem aining 18 patients were alive with no evidence of disease. The fate of the patients after the publication of the papers is not know n to us.

Discussion
Concerning the clinical presentation and age distribution, no differences between PES and other form s of ES 1,12,22 have been show n. There are differences, however, between PES and the other form s of ES with regard to sex predom inance, location of tum or an d surgical stage at pre sentation . A clear m ale predom inance is noted in PES (the m ale to fem ale ratio is 5.7:1), 15± 19 while in m edullary E S only a slight m ale predom inance, w ith a m ale to fem ale ratio of about 1.5:1, is know n 23, 24 and there has been no male predom inance in extraskeletal ES reported. 13,14 . M ed u llar y 12,23± 24 an d extrao sseo u s 13,14,25 E S develop in both proxim al and distal long bones and ax ial¯at bon es. PE S show s a pred om in an ce in proxim al extrem ities and axial PES is uncom m on. Only one case of PES in an axial¯at bone (scapula) was reported by Kolar et al. 16 The sm all num ber of cases precludes a real statistical signi® cance to these clinical observations w hich show only trends. In PES, all 20 patients were diagnosed at stage II-B of the disease, with no evidence of distant m etastases. In comparison , m etastases at presentation occur in about 25% of cases of m edullary ES 12,23 and in about 10% of cases of extraskeletal ES. 13 Im aging studies help to con® rm the diagnosis of PES which is de® ned when there is no tumor invasion of the medullary cavity. 10,11,26± 29 . A subperiosteal mass with a periosteal thickening and a Codm an triangle are diagnostic. 10,11,26± 30 These radiological signs appear both in PES and m edullary ES, but PES usually shows a uninterrupted periosteal reaction compared with thè onion skin' periosteal reaction observed medullary ES. 20 A subperiosteal location and the absence of medullary bone involvement help to distinguish PES from the other types of ES. 3,4,15± 17,20,31,32 Although conventional radiography provides the most useful inform ation for diagnosis and for gauging biological aggressiveness of the tum or, it has som e lim itations in estimating the extent of intram edullary disease in m edullary ES or in soft tissue involvem ent. 1  tissue) ES tum ors w hich grew enough to invade the periosteum will be de® ned as PE S, so there m ight be an overlap between the two subtypes. Still, the entity of PE S is quite well established, and there is a difference between the periosteal form and the soft tissue form in term s of sex, anatom ical location in bones and staging at diagnosis. At histopathological exam ination, all subtypes of ES, w hether m edullary, extraskeletal or periosteal, appear the sam e. In general, E S consists of uniform , sm all, round or oval highly undifferentiated cells with a pale appearance and scanty cytoplasm . It contains glycogen-positive granules with positive periodic acid± Schiff stain. 12± 17,25 It is not understood w hy PES seem s to have a better progn osis than the other two for m s. T his obser vation is sim ilar to that of a better prognosis in periosteal osteosarcom a and periosteal chondrosarcoma than in their m edullary counterparts. 10,11 One possible explanation is that the location at the periosteum causes such pain that the patients seek medical help earlier. A nother possible explanation can be found in the cytogenetic pro® le of the patients. N one of the 18 patients in the ® ve articles reviewed, together with our own two patients, underwent cytogenetic analysis. T he reason for this favorable prognosis may be the latter. It is strongly recom m ended that such an analysis is perform ed for PES patients in the future. After review ing the literature it seem s to us that this rare entity should be considered in the differential diagnosis of the ES fam ily of tum ors since there is a possibility that it has a better prognosis than m edullary or soft tissue ES.