OBJECTIVE
Characterize the CT findings of abdominopelvic Castleman disease, including a new observation involving the perinodal fat.
METHODS
Multi-center search at 5 institutions yielded 76 adults (mean age, 42.1 ± 14.3 years; 38 women/38 men) meeting inclusion criteria of histopathologically-proven Castleman disease with nodal involvement at abdominopelvic CT. Retrospective review of the dominant nodal mass was assessed for size, attenuation, and presence of calcification, and for prominence and soft-tissue infiltration of the perinodal fat. Hypervascular nodal enhancement was based on both subjective and objective comparison with aortic blood pool attenuation.
RESULTS
Abdominal involvement was unicentric in 48.7% (37/76) and multicentric in 51.3% (39/76), including 31 cases with extra-abdominal involvement. Histopathologic subtypes included hyaline vascular variant (HVV), plasma cell variant (PCV), mixed HVV/PCV, and HHV-8 variant in 39, 25, 3 and 9 cases, respectively. The dominant nodal mass measured 4.4 ± 1.9 cm and 3.2 ± 1.7 cm in mean long- and short-axis, respectively, and appeared hypervascular in 58.6% (41/70 with IV contrast). Internal calcification was seen in 22.4% (17/76). Infiltration of the perinodal fat, with or without hypertrophy, was present in 56.6% (43/76), more frequent with hypervascular vs. non-hypervascular nodal masses (80.5% vs. 20.7%; p < 0.001). Among HVV cases, 76.9% were unicentric, 71.1% appeared hypervascular, and 69.2% demonstrated perinodal fat infiltration.
CONCLUSION
Hypervascular nodal masses demonstrating prominence and infiltration of perinodal fat at CT can suggest the specific diagnosis of Castleman disease, especially the hyaline vascular variant.
ADVANCES IN KNOWLEDGE
Abdominopelvic nodal masses that demonstrate hypervascular enhancement and prominent infiltration of the perinodal fat at CT can suggest the diagnosis of Castleman disease, but nonetheless requires tissue sampling.