P-438: Castleman disease: histopathologic reclassification, clinical characteristics and endpoints in a reference center in Mexico

Introduction: Castleman's Disease (CD) circumscribes several clinicopathologic disorders at the intersection of hematology, oncology, rheumatology and virology, with convergence in histopathologic and clinical features. Information about CD in Latin America is scarce.

Methods: Patients with CD diagnosis from January 2006 to July 2022 were included. A retrospective review of the clinical charts and examination of the biopsies by a trained hematopathologist was performed. The immunohistochemical panel was extended to exclude differentials and PCR to confirm infection by HHV8 and classification of the cases.

Results: Twenty-seven patients were included; 37% were unicentric Castleman disease (UCD) and 63% multicentric Castleman disease (MCD); median age at diagnosis was 37 and 41 years respectively; 80% of women in UCD and 76% of men in MCD (p=0.018).
In UCD the most frequent nodal areas affected were cervical in 40% and mesenteric in 40%. Media nodal size was 7.1 cm; the 80% hyaline vascular (HV) variety and 20% mixed variant (MV). Successful surgical removal was performed in all cases and none of the patients have relapsed.
The distribution of MCD was 59% idiopathic MCD (iMCD), 35% associated to POEMS syndrome and 6% to HHV8. Fifty-nine percent experienced B symptoms, 70% presented fluid retention and 35% experienced skin involvement. Hepatomegaly was present in 88% and splenomegaly in 53%, mean 18.3 cm and 14.6 cm respectively.
The treatment received by 16 patients with MCD was only chemotherapy in 37%, rituximab as monotherapy in 19%, rituximab in combination with corticosteroids 19% and 25% in combination with CHOP-like chemotherapy. Response to treatment was evaluable in 13 patients: 38% achieved a complete response, 38% a partial response, 15% stable disease and one patient progressive disease.
The revision of the cases derived in the reclassification of 13 %. Cases were more frequently reclassified in MCD.
At 66 months follow-up, median overall survival, and progression free survival have not been reached. Patients with MCD associated to POEMS Syndrome relapsed more frequently 83% vs 17% (p=0.005). Uni and multivariate analysis revealed that the association to POEMS Syndrome increased relapse risk OR 50 (IC 95% 2.559 - 976), p=0.010 and OR 31.64 (IC 95% 1.48-676), p=0.027, respectively. Median PFS in this group was 53 months vs non reached in patients with iMCD or associated to HHV8 (p=0.031).

Conclusions: This is the first series of CD cases in the Latin population that includes a complete description of the clinical and histopathological characteristics and reclassification according to current diagnostic criteria.
Opposed to data reported in the literature, the MV corresponded to 20% of UCD and showed a predominance of cervical nodal involvement. Different from data published, in our cohort we found less mediastinal involvement in UCD (10% vs 16-40%).
In MCD associated to POEMS Syndrome, the PFS was inferior, the poor outcome of this association had not been previously described.