Computed tomography assessment of bone lesions in patients with POEMS syndrome

K. Glazebrook, Francis L. Guerra Bonilla, Adam Johnson, S. Leng, A. Dispenzieri

Research output: Contribution to journalArticlepeer-review

21 Scopus citations


Objectives: To describe the imaging findings on computed tomography (CT) and skeletal survey (SS) in patients with POEMS syndrome.

Methods: We retrospectively reviewed, with institutional review board approval, the dysproteinemia database at our institution for patients with new diagnosis of POEMS syndrome between January 1998 and December 2008. Twenty-four patients were identified with PET/CT or CT and had skeletal survey (SS) available for review.

Results: Twenty-four patients were included in the study group with median age of 47 years. All CTs demonstrated at least one sclerotic lesion. The most common pattern was multiple small lesions, with 18 patients (75 %) having at least 5 lesions less than 1 cm. The larger lesions had a central lytic component and were FDG avid. SS had a false negative rate of 36 % (8 patients). Serial CT after treatment showed a decrease in size and number of sclerotic lesions in 53 % of cases (13 patients), the majority showing increased sclerosis. Two patients had complete resolution of sclerotic lesions.

Conclusions: CT identified sclerotic lesions in all study patients with POEMS syndrome, the majority being less than 1 cm in size, which were not identified radiographically. CT may demonstrate increased sclerosis or even resolution of sclerotic lesions corresponding to treatment response.


Original languageEnglish (US)
Pages (from-to)497-504
Number of pages8
JournalEuropean radiology
Issue number2
StatePublished - Feb 2015


  • Computed tomography
  • POEMS syndrome
  • Radiography
  • Sclerosis
  • Skeletal survey (SS)

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging


Dive into the research topics of 'Computed tomography assessment of bone lesions in patients with POEMS syndrome'. Together they form a unique fingerprint.

Cite this