Imaging Feature Analysis of HHV8-Associated Multicentric Castleman Disease: A Comparative Study With HIV-Related Diseases and Idiopathic MCD

OBJECTIVE: The purpose of this study was to evaluate differences in the imaging features of human immunodeficiency virus (HIV)-associated Kaposi sarcoma (KS), HIV-associated lymphoma (HAL), and human herpesvirus 8-associated multicentric Castleman disease (hMCD) and clarify the differences between idiopathic MCD-not otherwise specified (iMCD-NOS) and hMCD. METHODS: This single-center retrospective study included 39 episodes of KS, 13 of HAL, 11 of hMCD, and 25 of iMCD-NOS. Pulmonary lesions were divided into 5 patterns: pulmonary nodules/masses; perihilar opacity; centrilobular nodules with or without ground glass opacity (GGO)/consolidation; focal GGO; and pleural masses. The frequencies of pulmonary computed tomography patterns and the prevalence of lymph node enlargement, hepatomegaly, and splenomegaly were compared between the HIV-associated groups (KS, HAL, and hMCD) and between the hMCD and iMCD-NOS groups. RESULTS: The most frequent imaging patterns of lung involvement in KS, HAL, and iMCD-NOS were perihilar opacity (64.3%), pulmonary nodules/masses (83.3%), and centrilobular nodules with or without GGO/consolidation patterns (60.0%), respectively. No hMCD-derived pulmonary lesions were observed in any patient. The prevalence of lymphadenopathy, hepatomegaly, and splenomegaly significantly differed between the 3 HIV-associated groups (P<0.05). The incidence of lesions occurring in 4 or more sites was significantly higher in the iMCD-NOS group with pulmonary lesions than in the group without pulmonary lesions (P<0.01). The incidences of imaging findings were comparable between the hMCD and iMCD-NOS with pulmonary lesions groups; however, the prevalences of enlarged abdominal lymph nodes (P<0.01), hepatomegaly (P<0.01), and splenomegaly (P<0.05) were significantly higher in patients with hMCD than in those with iMCD-NOS without pulmonary lesions. CONCLUSIONS: The frequency of pulmonary lesions is lower in hMCD than in KS and HAL. Observation of masses in the lungs should prompt suspicion of the concurrent presence of HAL, even when hMCD has been confirmed. While iMCD-NOS primarily affects the lymph nodes and, as the condition worsens, tends to cause lesions in the liver, spleen, and lungs, hMCD tends to cause lesions in the lymph nodes, liver, and spleen over a short period of time; however, the frequency of pulmonary lesions in hMCD is generally lower than in iMCD-NOS.