Return To Abstract Listing


E2340. Rosai-Dorfman Disease: Cross-Sectional Scan Findings in 32 Pathologically Proven Cases

Rayan J1,  Elshikh M2,  Wei P3,  Schellingerhout D2,  Garg N.2 1. Baylor College of Medicine, Houston, TX; 2. The University of Texas MD Anderson Cancer Center, Houston, TX; 3. University of Texas Health Science Center at Houston, Houston, TX

Address correspondence to M. Elshikh (

Objective: Rosai-Dorfman disease (RDD) is a rare, idiopathic nonneoplastic disease characterized by lymphadenopathy and extranodal disease. Our objective was to document imaging findings in various organs affected by RDD.

Materials and Methods: We conducted a retrospective review of pathologically confirmed cases seen at our institution from 1998 to 2015. We found 32 patients with RDD; 18 were women and 14 were men. These cases were analyzed for anatomic locations and radiologic appearance.

Results: In our patients, there were 51 separate sites of involvement in nine different anatomic locations. The most commonly affected locations were lymph nodes and the maxillofacial area. Cervical lymph nodes were the most commonly affected lymph node group. Only 4 (12.5%) patients had pure nodal involvement, while 20 (62.5%) had pure extranodal disease and 8 (25%) had mixed nodal and extranodal disease. Radiographically, RDD imaging findings vary greatly from one tissue to another; however, disease presentation on PET/CT and T1-weighted MRI was similar in most of the affected organs. All lesions on PET/CT were hypermetabolic, whether nodal or extranodal, except in one patient. On MRI, all lesions except for two were isointense on T1-weighted series, irrespective of the anatomic site. T2 signal intensity was variable from isointense to hyperintense between different anatomic locations. Also, enhancement pattern varied form homogenous to heterogeneous. Nodal disease presented as pathologic homogenous lymph node enlargement except in one case, in which there were necrotic areas. Maxillofacial RDD did not show a specific imaging pattern. Subcutaneous RDD presented as ill-defined subcutaneous masses with homogeneous enhancement that was isointense on T1-weighted imaging and hyperintense on STIR sequences. Bone disease appeared as a lytic mass with associated soft tissue component that was isointense on T1-weighted images and hyperintense on T2-weighted images with evident homogeneous enhancement. All CNS lesions except for one were dural-based masses that were isointense on T1-weighted images and T2-weighted images and showed homogeneous enhancement.

Conclusion: RDD represents a wide-spectrum disease not limited to the lymph nodes. RDD presentation is extremely variable on cross-sectional scans; however, most RDD lesions are hypermetabolic on PET/CT and isointense on T1-weighted MRI.