Beyond Extranodal Site Count: Clinical and Anatomical Characterization of Primary Multifocal Extranodal Diffuse Large...

Extranodal (EN) involvement is common in diffuse large B-cell lymphoma (DLBCL), but current prognostic models inadequately capture the anatomical complexity of EN dissemination. We evaluated primary multifocal extranodal DLBCL (PMEN-DLBCL) as a distinct clinical and anatomical phenotype using a standardized lesion-level classification framework.

We retrospectively analyzed 393 adults with newly diagnosed DLBCL treated with rituximab-based chemoimmunotherapy between 2003 and 2025. Patients were classified as primary nodal (PN; n = 231), primary focal extranodal (PFEN; n = 115), or primary multifocal extranodal (PMEN; n = 47) DLBCL according to predefined anatomical criteria. Clinical characteristics, organ involvement patterns, treatment response, survival outcomes, CNS relapse, and patterns of extranodal dissemination were evaluated. Exploratory co-occurrence analyses and unsupervised hierarchical clustering were performed within the PMEN cohort.

PMEN accounted for 12.0% of all DLBCL cases and was enriched for adverse clinical features, including impaired performance status, elevated lactate dehydrogenase, and high-risk IPI categories. PMEN demonstrated preferential involvement of axial and appendicular bones, liver, muscle, adrenal glands, lung, and kidney, whereas PFEN was enriched for gastric and head-and-neck disease. Although overall response rates were comparable across groups, PMEN exhibited lower complete response rates (60.5% vs. 78.3% in PFEN) and more frequent early treatment failure. Five-year progression-free survival was 52.3% in PMEN compared with 71.2% in PFEN. In univariable analysis, PMEN was associated with inferior PFS relative to PFEN (HR 1.86, 95% CI 1.08-3.20), but this association lost significance after adjustment for established prognostic factors. Within PMEN, adrenal involvement independently predicted inferior PFS and overall survival. Exploratory clustering identified skeletal-dominant and visceral-dominant phenotypes, with markedly lower complete response rates in visceral-dominant disease.

PMEN-DLBCL represents a distinct anatomical and clinical phenotype characterized by multifocal extranodal dissemination, preferential skeletal and deep visceral organ involvement, adverse baseline features, and inferior response quality. Specific organ involvement and spatial dissemination patterns appear more informative than extranodal site count alone, supporting further investigation of anatomical disease architecture in risk stratification of extranodal DLBCL.