PAROSTEAL (JUXTACORTICAL) OSTEOSARCOMA

Transcription

1 PAROSTEAL (JUXTACORTICAL) OSTEOSARCOA A CLINICAL AND HISTOPATHOLOGICAL STUDY O 1 1 CASES AND A REVIEW O THE LITERATURE R. LORENTZON, S.E. LARSSON, L. BOQUIST rom the Departments of Orthopaedic Surgery and Pathology, University of Ume#{226} At reexamination of all osteosarcomata recorded in the Swedish Cancer Registry during the years 195 to 196, 1 1 cases ofparosteal osteosarcoma were found. No case ofsocalled periosteal osteosarcoma was identified. The tumours constituted 1.6 per cent ofall proved primary malignant bone tumours. The ages of the 11 patients (six women and five men) ranged from 17 to 62 years (average 33 years). The clinical and histopathological findings of this study and of those collected from a review of the literature suggest the occurrence oftwo different types ofparosteal osteosarcoma: the predominant type is originally benign but has a definite malignant potential, causing metastases after long symptomfree intervals; the other type is highly malignant from the beginning. Primary amputation is recommended for the latter category oftumours, and compartmental, radical en bloc resection followed by regular review is recommended for the former. Parosteal (juxtacortical) osteosarcoma is a rare neoplasm (Dahlin ; Spjut et al ). Since the original description by Geschickter and Copeland (1 95 1), other reports have appeared (Dwinnell, Dahlin and Ghormley 1954; Jacobsson 195; Aakus, Eide and Stokke 1 960; Scaglietti and Calandriello ; Harkess 1964; Sirsat and Doctor 1965; van der Heul and von Ronnen 1967; Aegerterand Kirkpatrick 196; Edeiken et al ; arr and Huvos ; Unni et al a) and this neoplasm is now generally considered to be a distinct type of osteosarcoma with welldefined radiographical and histopathological features. Although most parosteal osteosarcomata have a relatively favourable prognosis, published series have also included cases with a clinical course characteristic of aggressive and highly malignant tumours. In fact, it has been suggested that two basically different types of parosteal osteosarcoma exist, one originally benign but potentially malignant and the other primarily malignant (Scaglietti and Calandriello 1962; arr and Huvos 1972; Lichtenstein 1972). oreover, an extremely rare periosteal variant of osteosarcoma has been reported, differing from parosteal osteosarcoma clinically, radiographically and histologically (Unni, Dahlin and Beabout 1 976b). ore information was therefore needed on the correlation between the clinical course and the histopathological characteristics of primary and recurrent tumours, and on the indications and contraindications for different s. The present study is a clinical and histopathological analysis of 1 1 cases of parosteal osteosarcoma recorded in the Swedish Cancer Registry during the 1 iyear period from 195 to i96. ATERIAL AND ETHODS Since 1 95 all cases of malignant tumours diagnosed in Sweden have been reported to the Cancer Registry by a system of double reporting from clinicians on the one hand and pathologists and cytologists on the other hand. The reliability of the information has previously been discussed by the Swedish National Board of Health and Welfare (1972) and by us (Larsson and Lorentzon 1974; Larsson, Lorentzon and Boquist 1975, 1976; Larsson et a!. 1979). During the years 195 to 1 96 a total of 242 cases was recorded in the Cancer Registry as osteosarcoma. At review of all available complete records, radiographs and histopathological material, I 53 cases were definitely classified as genuine osteosareoma (Larsson et a!. 1979), four as extraosseous osteosarcoma (Lorentzon, Larsson and Boquist I 979) and 1 1 as parosteal osteosarcoma. These lastmentioned 1 1 cases formed the basis of the present study. Information about survival was obtained from the Register Office. The histopathological specimens were reviewed, and in most cases additional sections were cut and stained with haematoxylin and eosin, Van Gieson s stain, periodic acidschiff, and Laidlaw s silver impregnation. RESULTS The clinical details of all 1 1 patients are summarised in Table I. Age and sex. The ages of the 1 1 patients ranged from 17 to 62 years (average 33 years). There were six women and five men. R. Lorentzon, D, Departmentof Orthopaedic Surgery S.E. Larsson, D, Department of Orthopaedic Surgery University of Ume#{226}, Ume#{227},Sweden. Professor L. Boquist. D, Department of Pathology J Requests for reprints should be sent to Dr Lorentzon. 6 THE JOURNAL O BONE AND JOINT SURGERY

2 PAROSTEAL (JUXTACORTICAL) OSTEOSARCOA 7 Site of the tumour. In five cases the site of the tumour was the posterior surface of the distal, and in one case it was the proximal ; the other five tumours had arisen on the proximal fibula, distal humerus, scapula, ischial tuberosity and calcaneus respectively. Symptoms. The most frequent sign on admission was a localised swelling, present in nine of the patients; this swelling was completely painless in five patients, while four had had progressively increasing pain. Two patients complained of pain but no swelling was apparent. Three patients had had symptoms for one year or longer, and the patient with a tumour on the scapula had noticed a swelling in this region for 1 5 years; the remaining seven patients had had a shorter duration of symptoms, ranging from two to six months. Radiographical findings In five cases the radiographs showed a densely ossified and lobulated mass on the posterior metaphysial cortex of the distal (ig. 1); similar dense tumours were Table I. Clinical details of the 1 1 patients Case number Sex Age (years) Site of tumour Duration of symptoms Symptoms Treatment etastasis Outcome 20 2 years 3 months Pain High thigh amputation etastasised to lungs and spine 3 years after disease 4 years after months Local resection. id None thigh amputation 3 years after first (local recurrence of tumour) Alive and well 9 years after months High thigh amputation None Alive and well 15 years after months High thigh amputation None Alive and well 7 years after year Radiotherapy, cobalt None 60 (600 rads, 36 days) No recurrence of tumour. Died from myocardial infarction 10 years after 6 45 Proximal 4 months and pain Radiotherapy, cobalt None 60 (7300 rads, 60 days). Hip disarticulation 7 months after first Alive and well 9 years after 7 17 Proximal fibula 2 months and pain Radiotherapy, xray Lung metastasis I 0 (6700 rads, 41 days) years after milligrams cyclophosphamide intraarterially and leucovorin rescue. Resection of 14 centimetres of proximal fibula 1 month after first disease 10 years after 19 humerus 1 month and pain Local resection twice None (1 month interval). Radiotherapy (dose unknown) Alive and well 1 1 years after first 9 62 Scapula 15 years 4 months. Progressive growth of mass Radiotherapy, cobalt Lung metastasis (5400 rads, 52 years after days). Resection of the scapula 4 months after irradiation disease 1 7 years after Ischial tuberosity 6 months Pain Local resection. Local Lung metastasis 3 recurrence treated by years after resection and hemipelvectomy 2 and 3 years after first disease 3 years after Calcaneus 6 months Pain Local resection. etastases in femoral Belowknee amputa and fascial soft tissue tion 3 months after months after first. Lung metastasis 1 9 months after disease 2 years after VOL. 62B, No. I, EBRUARY 190

3 R. LORENTZON, S.E. LARSSON, L. BOQUIST showed as a large ossified mass, four to five centimetres thick, which had almost completely embedded the scapula; another tumour had arisen on the medial surface of the calcaneus (ig. 2) and another was on the inner cortex of the ischial tuberosity (ig. 3). Case 1. Radiograph of a parosteal osteosarcoma on the posterior surface of the distal shows a dense lobulated mass attached to the bone by a broad base. A pathological fracture with surrounding bone resorption can be seen on the anterior aspect of the. seen at the other six sites. In most cases the tumours were attached to the cortex of the bone by a broad base. The tumours on flat bones had the same radiographical characteristics as those on long bones: one tumour Pathological findings The tumours were heavily ossified, but occasional soft areas could be encountered. Infiltration was seen in surrounding soft tissues. Osteoid was found in all tumours (ig. 4), although the amount varied considerably, and chondromatous foci were observed in five. The appearance of the stroma varied, being rather acellular in some tumours while in others it was moderately or highly cellular. A few tumours had a fibrosarcomatous appearance (ig. 5), and areas of necrosis or bleeding were found in four. Variations in degree of atypism and in mitotic frequency were recorded. Correlation between the course of the disease and the lightmicroscopic findings disclosed a poor degree of differentiation (igs 6 and 7) for the tumours in those patients who died in metastasising disease soon after (Cases 1, 1 0 and 1 1) whereas a high degree of differentiation (ig. ) was noted for the tumours in those patients who were still alive in December 197 (Cases 2, 3, 4, 5, 6 and ). Patients who died from their disease 10 years or more after exhibited tumours with a moderate or rather high degree of differentiation (Cases 7 and 9). Appraisal of the degree of differentiation was based mainly upon cellularity, atypism and the frequency of mitoses Recurrent lesions usually showed less differentiation than the primary tumours, whereas the microscopic appearance of metastases was generally similar to that of the primary tumour. Tumours fulfilling the criteria of ig.2 igure 2Case 1 1. A radiograph showing a lobulated dense mass localised medially on the calcaneus. igure 3Case I 0. This tumour on the ischial tuberosity has extended into the pelvis. THE JOURNAL O BONE AND JOINT SURGERY

4 PAROSTEAL (JUXTACORTICAL) OSTEOSARCOA 9 igure 4Photomicrograph of a section of parosteal osteosarcoma (Case 6) showing osteoid and areas of partly spindleshaped stromal cells. (Haematoxylin and eosin, x 300.) igure 5Section of a parosteal osteosarcoma (Case I 1) with a fibrosarcomatous appearance. Irregular or spindleshaped cells are closely packed and polymorphous. One mitotic figure is arrowed. (Haematoxylin and eosin, x 300.) igure 6The highly anaplastic stromal cells of a poorly differentiated parosteal osteosarcoma (Case 10). (Haematoxylin and eosin, x 260.) igure 7Another poorly differentiated parosteal osteosarcoma (Case 1 ) has rather closely packed and highly anaplastic, partly spindleshaped stromal cells. (Haematoxylin and eosin, x 250.) igure A parosteal osteosarcoma of rather high differentiation (Case 2) is composed of partly calcified osteoid and stromal cells with slight or moderate polymorphism. (Haematoxylin and eosin, X 220.) (Note. igures 4 to have been reduced to 90 per cent of original scale.),,if , d..: V.. :. I ;e j..., H y :. ;4,: 1. ; ;... 4 L S c,.... r pj....,. ;.. s.., ;....,,.. k#.,., 7 W//#{231} f(i I T... w ig. 6 ig. 7 I e I. I..,. f,.., : ig. 4.. s :t socalled periosteal osteosarcomata were not observed in this series. Treatment and course Table I contains details of given to the i 1 patients, and the course of their disease. DISCUSSION Parosteal osteosarcoma is a rare neoplasm: including the 1 6 cases originally described by Geschickter and Copeland (1951), only 206 cases have hitherto been recorded in the world literature (Dwinnel et al. 1954; Jacobsson 195; Aakus et a!. 1960; Scaglietti and Calandriello 1962; Harkess 1964; Sirsat and Doctor 1965; van der Heul and von Ronnen 1967; Aegerter and Kirkpatrick 196; Edeiken et a!. 1971; arr and Huvos 1972; Unni et al. 1976a). The 1 1 cases in this report constituted 1.6 per cent of all histopathologically proved primary malignant bone tumours recorded in the Swedish Cancer Registry from 195 to 196 (Larsson and Lorentzon 1974) and 6.3 per cent of all unequivocally proved osteosarcomata during the same period (Larsson et al. 1979). The annual incidence in Sweden, as calculated from the findings in the present study, corresponds to one case per eight million inhabitants. These figures agree well with those of the Netherlands Committee on Bone Tumours where parosteal osteosarcomata accounted for about two per cent of all primary malignant bone tumours and for 6.2 per cent of all VOL. 62B, No. 1. EBRUARY 190

5 90 R. LORENTZON, S.E. LARSSON, L. BOQUIST osteosarcomata (Netherlands Committee on Bone Tumours 1966; van der Heul and von Ronnen 1967); in comparison, parosteal osteosarcomata constituted only 3.7 per cent of all osteosarcomata reported from the ayo Clinic (Dahlin 1 967). The average age of our patients was 33 years which confirmed the earlier recognition that this tumour tends to occur in an older age group than intramedullary osteosarcoma (arr and Huvos 1972; Unni et a!. 1976a; Larsson et al. 1979). No case of socalled periosteal osteosarcoma was found in our series. The periosteal tumours are extremely rare, are more common in males than in females, usually affect patients in the second decade, and are most often localised to the upper part of the tibial shaft (Unni et a! b); in contrast, in our series there was a slight female predominance, an average age of 33 years, and no tibial localisation at all. Radiologically, periosteal osteosarcomata typically exhibit a slight thickening of the cortex, producing a mass which blends into the soft tissues (Unni et a!. 1976b); the tumours in our series consisted of dense lobulated masses attached to the cortex usually by a broad base. icroscopically, periosteal osteosarcomata are characterised by a seemingly intact cortex, masses of abnormalappearing cartilage and osteoid extending into the soft tissues, peripheral chondroid areas arranged perpendicular to the cortex, central malignantappearing osteoid and little tendency to infiltrate adjacent soft tissues (Unni et a! b); in contrast, our osteosarcomata were heavily ossified lesions with areas of osteoid, usually only a small amount of cartilage, occasionally a fibrosarcomatous appearance, and an infiltrative growth into surrounding soft tissues. ive of the I 1 tumours in the present study involved the distal, which is the predominant site in all published series. Remarkably, three of the tumours had arisen on the scapula, ischium and calcaneus, sites which have not previously been recorded. These three tumours had the same radiographical and histopathological features as the tumours on long bones, and thus fulfilled the diagnostic criteria of parosteal osteosarcoma. The duration of symptoms is generally reported to be considerably longer for patients with parosteal osteosarcoma than for those with intramedullary osteosarcoma (van der Heul and von Ronnen 1967; Unni et a!. 1976a). This is confirmed by the present study in which 55 per cent of the patients had symptoms for periods ranging from six months to several years; in comparison, we found that 62 per cent of patients with intramedullary osteosarcoma sought medical advice within two months and an additional 26 per cent within six months (Larsson et a! ). The symptoms of parosteal osteosarcoma also differ from those of intramedullary osteosarcoma. ive of the 1 1 patients had a completely painless swelling, while with intramedullary osteosarcoma local pain was the predominant complaint (Larsson et a!. 1979). These facts support the opinion put forward by Unni and coworkers (1976a), that a slowgrowing, painless and hard mass in the popliteal region should arouse suspicion of parosteal osteosarcoma. Radiographs are usually very characteristic (Dahlin 1 967). Nevertheless, lesions such as myositis ossificans, osteocartilaginous exostosis, extraosseous osteosarcoma and intramedullary osteosarcoma should be considered as differential diagnoses. Histopathologically, although some of the tumours were rather acellular and exhibited a slight anaplasia and few mitotic figures, they could all be classified as osteosarcomata, mainly because of the appearance of the stroma and the occurrence ofosteoid. In general, the tumours in this study were highly differentiated compared to previously studied intramedullary (Larsson et a! ) and extraosseous (Lorentzon et a!. 1979) osteosarcomata, and there was an interesting correlation between the degree of differentiation and the clinical course ofthe disease. Some patients were still alive nine years or more after and the tumours in these cases had features of osteosarcoma with a high degree of differentiation. any authors have emphasised that the malignancy of parosteal osteosarcoma is low in comparison with that of intramedullary osteosarcoma (Scaglietti and Calandriello 1962; Dahlin 1967; van der Heul and von Ronnen 1 967). Longterm followup is essential for assessing the real malignant potential of parosteal osteosarcoma since recurrence can be considerably delayed. In the present series the fiveyear survival rate was 73 per cent, and the 10year survival rate up to that time was 36 per cent. Similar survival rates were obtained from a review of five previously published series on parosteal osteosarcoma (Table II), whereas the figures were completely different for intramedullary osteosarcoma (Larsson et al ). Of the five patients in our series who died from their tumours, two succumbed with metastases after remarkably long symptomfree intervalsthat is, after 10 and 17 years respectivelyand most other published series have recorded deaths after similarly long intervals (Table II). Besides these longterm survivors, all reported series of parosteal osteosarcoma have also included a number of patients who died within a few years irrespective of the given. The results of various types of therapy are always difficult to evaluate in retrospect. A review of the literature shows (Table III) that local excision of the tumour has almost invariably resulted in recurrence. This was the case also in five out of six patients in the present series treated by resection (either alone or in combination with radiotherapy). Similar high rates of recurrence after resection were recorded in most other series (Geschickter and Copeland I 95 1 ; Scaglietti and Calandriello ; van der Heul and von Ronnen 1967; Edeiken et a!. 1971; arr and Huvos 1972). An THE JOURNAL O BONE AND JOINT SURGERY

6 PAROSTEAL (JUXTACORTICAL) OSTEOSARCOA 91 Table II. Survival rates in different series Series Total number cases of Number of cases allowing at kast a 5year analysis Survivals after 5 years Number Per cent Number of deaths from the disease The present study (2, 3, 4, 10, 17 years)* Geschickter and Copeland (1951) (3, 4, 5, 9 years) Scaglietti and Calandriello (1962) (7 months, 7 months, 2 years) VanderHeulandvonRonnen(1967) (1,2,9, 11 years) Edeiken et a!. (1971) (1, 3 years) arr and Huvos (1972) (1, 1.5, 3, 4, 5, 7 years) Total (39 per cent) *The figures in brackets denote the timelag between and death Table III. Results of various types of primary for parosteal osteosarcomata in different series. The figures indicate the number of recurrences and metastases in relation to the total number of cases subjected to the specified therapy. Only cases allowing a fiveyear analysis have been included Primary Series Curettage Excision Resection Amputation RadiatiOn alone Present series. 5/6 1/3 1/2 Geschickter and Copeland ( ) 6/6 2/3 0/2 Scaglietti and Calandriello ( I 962) 4/5 0/2 4/4 Van der Heul and von Ronnen ( 1 967) 4/5 1/2 3/5 Edeikeneta/. (1971) 1/1 1/1 2/2 3/5 arr and Huvos (1972) 1/1 3/5 3/3 2/4 Unnieta/. (l976)* 27/31 3/10 4/27 *Concerning this series, an accurate fiveyear analysis was not possible on the basis of the published details. exception was the report by Unni and coworkers ( 1 976a) in which recurrences were found in only three out of 10 patients; the followup for nine of their patientsranged from only nine months up to seven years and therefore longer periods might have shown an increase in the recurrence rate in that series too. Primary amputation did not prevent the occurrence of late metastases in one out of three patients in the present study and in a considerable number of patients in the other reviewed series. Patients treated with modern highdose irradiation were too few to allow any definite conclusions as to the radiosensitivity of parosteal osteosarcoma. Nor have there been any reports of modern chemotherapy or immunotherapy for this tumour. The findings of this study substantiate the theory that two distinct types of parosteal osteosarcoma exist: one type which is primarily highly malignant and one which is originally benign but with an inherent malignant potential. or welldelineated tumours with a welldifferentiated histological appearance, radical en bloc excision of the tumour including all the surrounding soft tissues is strongly recommended. When there is the risk of spread of the tumour along muscle or fascia, en bloc resection should be performed including whole muscles from origin to insertion or entire muscle compartments. In patients with advanced tumours, or tumours containing pleomorphic areas, primary amputation should be undertaken. This work was supported by grants from the Swedish Cancer Society (No. 71 lb706x) and the Research oundation, Department of Oncology, University of Ume#{226},Sweden. VOL. 62B, No. 1. EBRUARY 190

7 92 R. LORENTZON, S.E. LARSSON, L. BOQUIST REERENCES Aakus T, Eide 0, Stokke T. Parosteal osteogenic sarcoma. Acta Radiol 1960;54:2940. Aegerter E, Kirkpatrick J, eds. Orthopaedic diseases. 3rd ed. Philadelphia: WB Saunders Company, 196. Da.hlin DC. Bone tumors: general aspects and data on 3,97 cases. 2nd ed. Springfield, Illinois: Charles C Thomas, Dwlnnell LA, Dahlln DC, Ghormley RK. Parosteal (juxtacortical) osteogenic osteosarcoma. J Bone Joint Surg [Am] 1 954;36A: Edeiken J, arrell C, Ackerman LV, Spjut HJ. Parosteal sarcoma. AJR 1971;111:5793. air GH, Huvos AG. Juxtacortical osteogenic sarcoma: an analysis of fourteen cases. J Bone Joint Surg [Am] 1972;54A: Geschlckter C, Copeland. Parosteal osteoma of bone: a new entity. Ann Surg ;133: Harkess JW. Parosteal osteosarcoma. Am Surg 1 964;30:7306. Jacobsson SA. Early juxtacortical osteosarcoma. J Bone Joint Surg [Am] 195;40A: Larsson SE, Lorentzon R. The geographic variation of the incidence of malignant primary bone tumors in Sweden. J Bone Joint Surg [Am] 1974;56A: Larsson SE, Lorentzon R, Boquist L. Giantcell tumor of bone: a demographic, clinical, and histopathological study of all cases recorded in the Swedish Cancer Registry for the years 195 through 196. J Bone Joint Surg [Am] 1975;57A: Larsson SE, Lorentzon R, Boquist L. ibrosarcoma of bone: a demographic, clinical, and histopathological study of all cases recorded in the Swedish Cancer Registry from 195 to 196. J Bone Joint Surg [Br] 1976;5B:4127. Larsson SE, Lorentzon R, Wedr#{233}nH, Boquist L. Osteosarcoma: a multifactorial clinical and histopathological study with special regard to therapy and survival. Acta Orthop Scand 1979;49:5711. Lkhtenstein L. Bone tumors. 4th ed. Saint Louis: CV osby Company, Lorentzon R, Larsson SE, Boquist L. Extraosseous osteosarcoma: a clinical and histopathological study of four cases. J Bone Joint Surg [Br] 1979;61B:205. NatiOnaIBOard ofhealth and Welfare, The Cancer Registry. Cancerincidence in Sweden Stockholm: GOteborgs Offsettryckeri AB, Netherlands Committee on Bone Tumours. Radiologkal atlas ofbone tumours. The Hague and Paris: outon and Company, 1966;1. Scagiletti 0, Calandriello B. Ossifying parosteal sarcoma. I Bone Joint Surg [Am] 1962;44A: Sii#{149}sat V, Doctor V. Parosteal (juxtacortical) osteogenic sarcoma: emphasis on histopathology. J Postgrad ed ;1 1:1917. Spjut Hi, Dorfman IW, echner RE, Ackerman LV. Tumors of bone and cartilage. Atlas oftumorpathology. Washington, DC: Armed orces Institute of Pathology, 1970; Series 2: asc 5. Unni KK, Dublin DC, Beabout JW, Ivins JC. Parosteal osteogenic sarcoma. Cancer 1976a;37: Uniil KK, Dublin DC, Beabout JW. Periosteal osteogenic sarcoma. Cancer 1976b;37: vander Heal RO, von RonnenJR. Juxtacortical osteosarcoma:, differential,, and an analysis ofeighty cases.j Bone Joint Surg [Am] 1967;49A: THE JOURNAL O BONE AND JOINT SURGERY