Computed tomography in the analysis of calcification patterns in pediatric bone tumors of the hip: a new approach

Abstracts

OBJETIVO: No grupo pediátrico, o diagnóstico radiológico dos tumores dos ossos ilíacos, ísquios e púbis apresenta dificuldades e peculiaridades próprias, mas a literatura revisada não trata especificamente desse tema. Este trabalho pretende investigar a existência de padrões radiológicos confiáveis para o diagnóstico diferencial desses tumores. MATERIAIS E MÉTODOS: Foram revistos os achados radiológicos de tumores dos ossos do quadril em dez pacientes com idades entre 8 e 19 anos. RESULTADOS: Reação óssea (esclerose ou lise), reação periosteal (lamelar em camada única, múltiplas camadas ou radial), extensão do tumor no osso e grau de invasão das partes moles revelaram baixa especificidade. As calcificações nas partes moles, consideradas em conjunto, foram inespecíficas. Contudo, separando as próximas do osso comprometido, que apresentam formas e tamanhos variados - padrão I -, daquelas afastadas do osso, finas e amorfas - padrão II -, observamos que o padrão I foi totalmente inespecífico e o padrão II foi identificado nos três casos de osteossarcoma (100%) e em apenas um dos casos de Ewing (16,6%). CONCLUSÃO: Neste material, as calcificações padrão II revelaram sensibilidade de 100% e especificidade de 90% para osteossarcoma. Contudo, sua importância pode não se limitar ao diagnóstico radiológico. As calcificações padrão II indicam, provavelmente, os sítios ideais para biópsia.

Tumores ósseos; Ilíaco; Quadril; Pediatria; Tomografia computadorizada

ORIGINAL ARTICLE

Computed tomography in the analysis of calcification patterns in pediatric bone tumors of the hip: a new approach^*

Gabriel Antônio de Oliveira^I; Henrique Zambenedetti Werlang^II; Pedro Martins Bergoli^II; Madalena Frechiani^III; Fernão Oliveira^IV

^IMD, Radiologist at Hospital Infantil Nossa Senhora da Glória, Preceptor in Pediatric Radiology – Medical Residence of Diagnostic Imaging Center at Hospital Universitário Cassiano Antônio de Morais/Hospital Infantil Nossa Senhora da Glória

^IIMD, Radiologists, ex-Residents at Diagnostic Imaging Center – Hospital Universitário Cassiano Antônio de Morais/Hospital Infantil Nossa Senhora da Glória

^IIIMD, Pediatric Oncologist at Hospital Infantil Nossa Senhora da Glória

^IVMD, Radiologist at Hospital Infantil Nossa Senhora da Glória

^{Mailing address}

ABSTRACT

OBJECTIVE: In the pediatric group, the radiological diagnosis of bone tumors of the hip is difficult and presents some peculiarities, but reviewed literature does not approach this specific problem. The objective of the present study was to investigate the existence of reliable radiological patterns for the differential diagnosis of these tumors.

MATERIALS AND METHODS: Radiological findings of bone tumors of the hip in ten patients in the age range between 8 and 19 years have been reviewed.

RESULTS: Bone reaction (sclerosis or lysis), periosteal reaction (lamellar with single or multiple layers, or radial), tumor extent in the bone and level of soft tissues invasion have presented low specificity. Soft tissue calcifications, when considered as a whole, were non-specific. However, when those calcifications with varied shapes and sizes, nearby the affected bone (pattern I) were separated from those, thin and amorphous, away from the bone (pattern II), we have observed that the pattern I was totally non-specific, and the pattern II was found in the three cases of osteosarcoma (100%) and in only one case of Ewing's sarcoma (16.6%).

CONCLUSION: In the present study, pattern II calcifications have shown a 100% sensitivity and 90% specificity for osteosarcoma. However, their importance may be not limited to the radiological diagnosis. Pattern II calcifications indicate probably ideal sites for biopsy.

Keywords: Bone tumors; Ilium; Hip; Pelvis; Pediatrics; Computed tomography.

INTRODUCTION

In long bones, orthogonal plain films are very useful in the formulation of diagnostic hypotheses. Different lesions tend towards presenting higher incidence in epiphyses, metaphyses or diaphyses⁽¹⁾. Osteosarcomas, for example, use to be metaphyseal; the Ewing's sarcoma, diaphyseal or metadiaphyseal. In the hip bones (ilium, ischium and pubis), on the contrary, orthogonal incidences are not feasible and the lesions can not be classified according to their longitudinal disposition. Additionally, the overlapping of pelvic viscus and its contents images complicate the analysis of plain films, so computed tomography (CT) becomes essential for the diagnosis. The hematopoiesis maintenance is another feature of the hip bones — in long bones, the red bone marrow is replaced by yellow marrow. As a result of these particularities, iliac tumors present some distinct radiological aspects.

In the reviewed pediatrics literature, we have not found any specific study on the radiological semiology of bone tumors of the hip.

MATERIALS AND METHODS

We have retrospectively analyzed the dossiers of ten patients with histologically confirmed bone tumors of the hip documented by means of plain films and CT. The CT studies have been analyzed in the non-contrast phase to avoid the confusion between tumor vessels and calcifications. Magnetic resonance imaging (MRI) is poorly appropriate for diagnosis, and has not been presented in this study, despite its fundamental role in the staging of this kind of lesions. Post-chemo-radiotherapy images have not been utilized, as well as post-contrast images. In the first case, because of a significant calcification increase in the soft parts of the tumor, probably resulting from tissular necrosis. In the second case, because the tissular, neovascular enhancement could be confused with tumor calcifications.

The sacrum and coccyx, although included in the bone structure of the hip, have been excluded from this study for presenting peculiar characteristics, besides other in common with vertebrae.

The casuistic has been limited to the first decades of life because of the several diseases affecting these bones in the other age ranges. Six cases had histological diagnosis for Ewing's sarcoma, three for osteosarcoma, and one for chondrosarcoma. Ages ranged between 8 and 19 years. Six patients were male and four, female.

Other criteria employed in the analysis are shown in Table 1.

RESULTS

All of the patients presented with pain as their initial symptom, reported as restricted to the hip in seven patients; in the hip irradiating to the coxa in one; and in the coxa and/or knee in two. All of them presented with a palpable mass at diagnosis.

The lesion involved less than 25% of the bone in two cases of Ewing's sarcoma, and in one case of chondrosarcoma. Between 25% and 50% in one case of Ewing's sarcoma, and in one of osteosarcoma, and more than 50% in three cases of Ewing's sarcoma, and in two of osteosarcomas.

All of the three cases of osteosarcomas had predominantly sclerotic bone lesions (, and ). In the cases of Ewing's sarcoma, there was a predominance of sclerotic lesions in three (, ⁴ and ⁶), lytic in two ( and ⁵) and a balance between sclerosis and lysis in one (^{case 3}). In the chondrosarcoma, there was a lytic lesion (^{case 10}). Lesions presented with a permeative pattern in all of the cases, with ill-defined limits between the normal and affected areas.

Cortical involvement had occurred in all of the cases.

Periosteal reaction occurred in the three osteosarcomas, in one of them associated with single lamellar reaction (). In the Ewing's sarcomas, two have not presented any periosteal reaction, two presented with single lamellar reaction ( and ), and two, mixed reaction (^{cases 3} and ⁶). In the chondrosarcoma, the reaction was of lamellar nature.

An extra-osseous mass was demonstrated in eight cases whose CT and/or MRI were available, always occurring from both faces and exceeding the major transverse axis of the bone.

Calcifications in soft parts were evaluated every time CT was available for analysis; pattern I calcifications were found in all of the cases (Ewing's sarcoma, osteosarcoma, chondrosarcoma); pattern II, in osteosarcomas and in one Ewing's sarcoma (^{case 6}). The majority of calcifications near the bone were pattern I; and the distant ones were pattern II. Pattern III calcifications have not been found.

Contiguity invasion into the sacrum has occurred in only one case of primary iliac osteosarcoma ().

Metastasis have been found at diagnosis in only one case of Ewing's sarcoma (^{case 6}, pulmonary metastasis).

DISCUSSION

Iliac involvement occurs in 18% of Ewing's sarcoma, and in 12% of osteosarcomas. Chondrosarcomas, most frequently, affect this bone (60%), but they are rare in the childhood and adolescence⁽²⁾. Therefore, the casuistic presented in this study may be considered as representative.

Seven patients presented with a severe bone involvement by the tumor, most probably related to delay in diagnosis. Small alterations on the first x-ray of the hip had been neglected in . In ^{case 6}, the patient reported pain in the knee, so the initial imaging investigation — including a MRI — was restricted to this site. Although it is a well known fact, it will never be enough to remind that pain in the coxa and/or knee may be related to a hip bone lesion. Therefore, in cases where knee or coxa x-ray images are normal, the necessity of extending the radiological investigation to the hip must be considered.

Other factors may be taken into consideration for the existence of so extensive lesions. Differently from the long bones, in the ilium, ischium and pubis, the barriers constituted by conjugation cartilages and yellow bone marrow — hypovascularized — fat are absent. This absence makes the tumor cells dissemination easier.

Sunburst or lamellated periosteal reactions occurred in all the types of tumors, although the first one has predominated in osteosarcomas. Cortical involvement and extra-osseous tumor mass have been present in all of the cases.

Tumor cells have the capacity to stimulate osteoblasts to produce sclerosis, and osteoclasts, lysis. Mixed lesions occur when these two functions coexist. In osteogenic sarcomas, additionally to this stimulus, tumor cells also produce osteoid matrix (an intercellular substance produced by normal or tumor osteoblasts). Only mineralized matrices can be detected in images⁽¹⁾. Also it is important to note that the amount of osteoid produced by osteosarcomas depends on the predominant differentiation of tumor cells: osteoblastic, chondroblastic and fibroblastic. However, it is impossible to differentiate the reactional sclerosis from that resulting from the production of osteoid matrix by the tumor osteoblasts.

Predominance of sclerosis has occurred in all of the cases of osteosarcoma, but also in half of cases of Ewing´s sarcoma. Predominance of osteolysis has only occurred in Ewing's sarcomas, but it may also occur in osteosarcomas, mainly the telangiectatic ones, and in osteomyelitis.

With regard to calcification in soft parts, they may occur in areas of tissular necrosis, tumor-induced bone metaplasia, osseous fragments dislocated by the tumor, periosteal reaction, or resulting from calcium deposition on the tumor matrix (osteoid or chondroid). These calcifications present different patterns according Table 1. Pattern I calcifications were totally non-specific, probably because of the impossibility of distinguishing calcifications resulting from the osteoid matrix (in osteosarcomas) from those resulting from periosteal reaction and dislocated osseous fragments. Pattern II calcifications have been evidenced in all the osteogenic sarcomas and in one of the Ewing's sarcomas (^{case 6}). In osteosarcomas, presumably, they have resulted from the tumor osteoid matrix calcification (ossification). Since Ewing's sarcomas do not produce osteoid matrix, when present, these calcifications could result from necrosis, or, eventually, from tumor-induced bone metaplasia⁽³⁾. Resnick says that 9% of Ewing's sarcomas calcify, but he neither specify which bones are affected nor the types of calcifications. The same author reports two cases of Ewing's sarcoma of the ilium with calcifications in soft parts. In this situation, a mistaken radiological diagnosis of osteosarcoma may be induced⁽⁴⁾.

Despite the recent developments of histopathology, immunohistochemistry, cytogenetics and electronic microscopy, the diagnosis of Ewing's sarcoma remains as a challenge, and it is difficult to differentiate them from other tumors, especially small cell osteosarcomas^(1,3,5–15), neuroectodermal tumors, mainly the primitive ones, and lymphomas^(1,4). It is known that bone tumors may present different histological aspects, depending on the biopsy area. On CT, areas with pattern II calcifications suggest the presence of osteoid matrix. In a tumor, provided the likelihood of an induced bone metaplasia is remote, the presence of osteoid matrix is usually considered as an indication of osteosarcoma. However, the likelihood of such calcifications being a result of tissular necrosis should not be disregarded.

Therefore, in the hypothesis of the biopsy does not reach areas with pattern II calcifications, there is a possibility of the osteoid tissue - existing only in these areas - not being included in the biopsy slide. The result would be a diagnostic mistake. CT-guided biopsy aimed at areas with pattern II calcifications could avoid this mistake.

In summary, in iliac bone tumors, soft parts calcifications seem to be of significance for the presumed radiological diagnosis, and even more for biopsy guidance. A specific protocol adopting this approach could be employed in prospective studies, especially the multicentric ones.

REFERENCES

Received September 8, 2005.

Accepted after revision March 29, 2006.

• 1. Baunin C, Rubie H, Sales de Gauzy J. Sarcoma d'Ewing. Ecycl Méd Chir Radiodiagnostic Appareil Locomoteur 2001;31:520-A-50,9p.
• 2. Schajowicz F. Neoplasias ósseas e lesões pseudotumorais. 2Ş ed. Rio de Janeiro: Revinter, 2000.
• 3. Schajowicz F. Current trends in the diagnosis and treatment of malignant bone tumors. Clin Orthop Relat Res 1983;180:220252.
• 4. Resnick D. Tumor and tumor like lesions of bone: image and pathology of specific lesions. In: Resnick D, editor. Diagnosis of bone and joint disorders. 4th ed. Philadelphia: WB Saunders, 2002; 37634128.
• 5. Sim FH, Unni KK, Beabout JW, et al Osteosarcoma with small cell simulating Ewing's tumors. J Bone Joint Surg [Am] 1979;61:207215.
• 6. Martin SE, Dwyer A, Kissane JM, et al Small-cell osteosarcoma. Cancer 1982;50:990996.
• 7. Roessner A, Immenkamp M, Hiddemann W, et al Case report 331. Small cell osteosarcoma of the tibia with diffuse metastatic disease. Skeletal Radiol 1985;14:216225.
• 8.Edeiken J, Raymond AK, Ayala AG, et al Small-cell osteosarcoma. Skeletal Radiol 1987;16:621628.
• 9. Ayala AG, Ro JY, Raymond AK, et al A clinico-pathologic study of 27 cases. Cancer 1989;64: 21622173.
• 10. Kyriakos M, Gilula LA, Becich MJ, et al Intracortical small cell osteosarcoma. Clin Orthop 1992;279:269280.
• 11. Sanjay B, Raj GA, Vishwakarma G. A small cell osteosarcoma with multiple skeletal metastasis. Arch Orthop Trauma Surg 1988;107:5860.
• 12. Ayala AG, Ro JY, Papadopoulos NK, et al Small cell osteosarcoma. Cancer Treat Res 1993;62: 139149.
• 13. Mawad JK, MacKay B, Raymond AK, et al Electron microscopy in the diagnosis of small round cell tumors of bone. Ultrastruc Pathol 1994;18: 263268.
• 14. Nakajima H, Sim FH, Bond JR, et al Small cell osteosarcoma of bone: review of 72 cases. Cancer 1997;79:20952106.
• 15. Mulligan ME, Lewis DR Jr, Resnick CS, et al Small cell osteosarcoma of the ulna: a case report and review of the literature. J Hand Surg [Am] 1999;24:417420.

*

Study developed at Service of Diagnostic Imaging – Hospital Infantil Nossa Senhora da Glória, Vitória, ES, Brazil.

Publication Dates

• Publication in this collection
  07 Feb 2007
• Date of issue
  Dec 2006

History

• Received
  08 Sept 2005
• Accepted
  29 Mar 2006