Morphological, immunophenotypic and neuroradiological characteristics of primitive B-large cell diffuse lymphoma of the central nervous system: A retrospective cohort analysis.

Primitive diffuse large B-cell lymphoma (PCNSL-DLBCL) accounts for 2-3% of all central nervous system (CNS) tumors, predominantly affecting supratentorial sites in adult patients. In the present study, the immunohistochemical (IHC) profile was analyzed in a cohort of 23 cases of CNS DLBCL to determine whether B-cell markers [CD20, CD79a and paired box 5 (PAX5)], BCL2, CD10, BCL6, MUM1 and c-Myc, as well as the Ki67 labeling index, were associated with the clinicopathological (age, sex and location) and neuroradiological (T1, T2, post-contrast and diffusion-weighted imaging) parameters of the patients. The Hans algorithm was applied to subtype DLBCLs into germinal center B-cell (GCB) and non-germinal center B-cell (non-GCB). Univariate statistical analysis was performed using Fisher's exact test to identify statistically significant associations between clinicopathological, IHC and radiological variables. Multivariate analysis was then conducted to assess the independent effects of these variables on overall survival (OS). The most frequent site of PCNSL-DLBCL was confirmed in supratentorial areas, with invasion of midline structures; 7 cases presented with a single lesion, while 11 exhibited multifocal disease. All cases exhibited positive expression of CD20, PAX5 and CD79a; 21 cases exhibited positive expression of BCL2. The mean Ki67 labeling index was 80%. A total of 17 cases were classified as non-GCB and 6 as GCB. A statistically significant association was documented between the site of the disease and the number of lesions (P=0.04), the site of disease and subtype (P=0.03), and the IHC subtype and mean enhancement value (P=0.008). The present study demonstrated that DLBCL with a non-GCB immunohistochemical profile was predominantly observed in patients with unfavorable prognostic factors, such as age >60 years, a high Ki67 labeling index and shorter OS. Statistically significant associations were found between disease location and the number of lesions, disease location and IHC subtype, and between IHC subtype and the mean enhancement value, which may serve as predictors for the IHC subtype. Multivariate analysis and Kaplan-Meier curves revealed that the IHC subtype and enhancement value were independent prognostic factors, with the GCB profile and higher enhancement values being associated with longer OS times and improved prognosis (P=0.0158). In conclusion, the identified associations suggested that the IHC subtype and apparent diffusion coefficient neuroimaging parameters were useful validated prognostic indicators.