Summary Mechanisms of prolonged cytopenia (PC) after chimeric antigen receptor (CAR) T‐cell therapy, an emerging therapy for relapsed or refractory diffuse large B‐cell lymphoma, remain elusive. Haematopoiesis is tightly regulated by the bone marrow (BM) microenvironment, called the ‘niche’. To investigate whether alterations in the BM niche cells are associated with PC, we analysed CD271+ stromal cells in BM biopsy specimens and the cytokine profiles of the BM and serum obtained before and on day 28 after CAR T‐cell infusion. Imaging analyses of the BM biopsy specimens revealed that CD271+ niche cells were severely impaired after CAR T‐cell infusion in patients with PC. Cytokine analyses after CAR T‐cell infusion showed that CXC chemokine ligand 12 and stem cell factor, niche factors essential for haematopoietic recovery, were significantly decreased in the BM of patients with PC, suggesting reduced niche cell function. The levels of inflammation‐related cytokines on day 28 after CAR T‐cell infusion were consistently high in the BM of patients with PC. Thus, we demonstrate for the first time that BM niche disruption and sustained elevation of inflammation‐related cytokines in the BM following CAR T‐cell infusion are associated with subsequent PC.

A 71-year-old Japanese man presented with severe thrombocytopenia. A whole-body CT at presentation showed small cervical, axillary, and para-aortic lymphadenopathy, leading to suspicion of immune thrombocytopenia due to lymphoma. Biopsy was difficult to perform because of severe thrombocytopenia. Thus, he received prednisolone (PSL) therapy and his platelet count gradually recovered. Two and a half years after PSL therapy initiation, his cervical lymphadenopathy slightly progressed without other clinical symptoms. Hence, a biopsy from the left cervical lymph node was performed, and he was diagnosed with nodal peripheral T-cell lymphoma (PTCL) with T follicular helper (TFH) phenotype. Due to various complications, we continued treatment with prednisolone alone after the diagnosis of lymphoma; however, there was no further increase in lymph node enlargement and no other lymphoma-related symptoms for one and a half years after diagnosis. Although immunosuppressive therapy has been reported to produce a response in some patients with angioimmunoblastic T-cell lymphoma, our experience suggests that a similar subset may exist in patients with nodal PTCL with TFH phenotype, which has the same cellular origin. Immunosuppressive therapies may constitute an alternative treatment option even in the era of novel molecular-targeted therapies, especially for elderly patients who are ineligible for chemotherapy.

It is difficult to histologically differentiate extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma) from chronic gastritis (CG)/ reactive lymphoid hyperplasia (RLH). To determine whether immunohistochemistry for IRTA1 and MNDA can differentiate gastric MALT lymphoma from CG/RLH, we investigated 81 stomach biopsy specimens [Wotherspoon grade (WG) 1, 11 cases; WG 2, 9 cases; WG 3, 20 cases; WG 4, 31 cases; and WG 5, 10 cases]. According to a previously reported algorithm involving PCR for immunoglobulin heavy (IgH) chain locus rearrangement, all 81 cases were divided into three groups: CG/RLH (55 cases), MALT lymphoma (19 cases) groups, and IgH undetectable group (7 cases). We analyzed the CG/RLH and MALT lymphoma groups. The median percentage of IRTA1-positive cells was 0% (range 0%–90.6%) in the CG/RLH group and 43.5% (range 0%–97.6%) in the MALT lymphoma group (p < 0.0001). The median percentage of MNDA-positive cells was 32.4% (range 0%–97.6%) in the CG/RLH group and 55.1% (range 0%–97.6%) in the MALT lymphoma group (p = 0.0044). These results indicate that immunohistochemistry for IRTA1 and MNDA can help differentiate gastric MALT lymphoma from CG/RLH.

The distribution and clinical impact of cell-of-origin (COO) subtypes of diffuse large B-cell lymphoma (DLBCL) outside Western countries remain unknown. Recent literature also suggests that there is an additional COO subtype associated with the germinal center dark zone, that warrants wider validation to generalize clinical relevance. Here, we assembled a cohort of Japanese patients with untreated DLBCL and determined the refined COO subtypes, that include the dark zone signature (DZsig), using the NanoString DLBCL90 assay. To compare the distribution and clinical characteristics of the molecular subtypes, we used a dataset from the cohort of BC Cancer (BCC) (n = 804). Of the 1050 patients where DLBCL90 assay was successfully performed in our cohort, 35%, 45%, and 6% of patients were identified to be germinal center B-cell-like (GCB)-DLBCL, activated B-cell-like (ABC)-DLBCL, and DZsigpos-DLBCL, respectively, with the highest prevalence of ABC-DLBCL differing significantly from that of BCC (P &lt; 0.001). GCB-DLBCL, ABC-DLBCL, and DZsigpos-DLBCL were associated with two-year overall survival rates of 88%, 75%, and 66%, respectively (P &lt; 0.0001), with patients of the DZsigpos-DLBCL having the poorest prognosis. In contrast, GCB-DLBCL without DZsig showed excellent outcomes following rituximab-containing immunochemotherapy. DZsigpos-DLBCL was associated with the significant enrichment of tumors with CD10 expression, concurrent MYC/BCL2 expression, and depletion of microenvironmental components (all P &lt; 0.05). These results provide evidence of the distinct distribution of clinically relevant molecular subtypes in Japanese DLBCL and that refined COO, as measured by the DLBCL90 assay, is a robust prognostic biomarker that is consistent across geographical areas.

No abstract