Parosteal osteosarcoma: Value of MR imaging and CT in the prediction of histologic grade

James S. Jelinek, Mark D. Murphey, Mark J. Kransdorf, Barry M. Shmookler, Martin M. Malawer, Regina C. Hur

Research output: Contribution to journalArticlepeer-review

63 Scopus citations


PURPOSE: To evaluate the use of magnetic resonance (MR) imaging and computed tomography (CT) for predicting the histologic grade of parosteal osteosarcomas. MATERIALS AND METHODS: Sixty parosteal osteosarcomas were analyzed for tumor size and location, presence of a cleavage plane, intramedullary extension, soft-tissue mass (distinct from ossified mass), and the presence and pattern of ossification. Axial and longitudinal views were evaluated for specific osseous sites within the bone. Tumors were classified as low grade (grade 1) or high grade (grades 2-3). RESULTS: There were 32 low-grade lesions and 28 high-grade lesions. Average maximal lengths of low- and high-grade tumors were 7.7 and 15.0 cm, respectively. A cleavage plane was present in 20 (62%) low-grade and 19 (68%) high-grade lesions. On cross- sectional images, intramedullary extension was present in 13 (41%) low-grade and 14 (50%) highgrade lesions. A focal soft-tissue mass distinct from the ossific matrix was identified in 25 (89%) high-grade lesions and in only two (6%) lowgrade lesions. All 17 high-grade lesions evaluated with MR imaging were of predominantly high signal intensity on T2-weighted images. CONCLUSION: A poorly defined soft-tissue component distinct from the ossific matrix is the most distinctive feature of high-grade parosteal osteosarcoma and may be an optimal site for biopsy.

Original languageEnglish (US)
Pages (from-to)837-842
Number of pages6
Issue number3
StatePublished - Dec 1996


  • Bone neoplasms, CT
  • Bone neoplasms, MR
  • Osteosarcoma

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging


Dive into the research topics of 'Parosteal osteosarcoma: Value of MR imaging and CT in the prediction of histologic grade'. Together they form a unique fingerprint.

Cite this