Unmasking a Composite Lymphoma: Follicular Lymphoma Emerging After Bispecific Antibody Treatment of Marginal Zone Lymphoma Authors: Angel Luis Orosco‐Ttamina, Jiehao Zhou, Juliana Biondo, Farah Hossain, David Viswanatha, Patricia T Greipp, Natasha Lewis, Javier Munoz Published: 2026 ABSTRACT ABSTRACTIntroductionComposite indolent B‐cell lymphomas may remain unrecognized when one component dominates the clinical presentation. Apparent relapse after therapy may, therefore, represent the emergence of a biologically distinct lymphoma rather than the recurrence of the original disease.Case PresentationAn 82‐year‐old woman presented with systemic low‐grade B‐cell lymphoma that was most consistent with marginal zone lymphoma (MZL). Staging demonstrated peripheral blood involvement and multifocal FDG‐avid disease. Treatment with the CD20×CD3‐bispecific antibody mosunetuzumab resulted in complete metabolic remission. During surveillance, a new, isolated cervical lymph node developed despite sustained systemic response. Biopsy revealed follicular lymphoma (Grade 1–2) with a germinal center phenotype and a BCL2 rearrangement, findings discordant with recurrence of MZL. Retrospective review of the original specimen identified in situ follicular neoplasia within the initial biopsy. The disease course was, therefore, reinterpreted as a synchronous composite lymphoma in which treatment of the dominant MZL unmasked a previously clinically silent B‐cell neoplasm with follicular lineage. Local radiotherapy achieved remission of the follicular component.ConclusionsNew disease following treatment of an indolent lymphoma should not automatically be interpreted as relapse. Discordant clinical behavior, particularly focal progression after systemic response, warrants repeat biopsy and integrated pathologic reassessment. This case illustrates how therapy may reveal preexisting clonal heterogeneity and emphasizes the importance of temporality in distinguishing relapse from composite lymphoma.Trial RegistrationThe authors have confirmed clinical trial registration is not needed for this submissionKeywords: composite lymphoma, follicular lymphoma, in situ follicular neoplasia, BCL2 rearrangement, marginal zone lymphoma Introduction Indolent B‐cell lymphomas arise from different stages of B‐cell differentiation and may share overlapping clinical and morphologic features despite distinct biological origins. Marginal zone lymphoma (MZL) derives from post‐germinal center B cells, whereas follicular lymphoma (FL) represents a germinal center‐derived neoplasm typically associated with BCL2 rearrangement and a follicular growth pattern [1, 2, 3]. Although each entity is well characterized individually, diagnostic complexity arises when more than one lymphoma lineage is present in the same patient [4]. In such settings, the clinically dominant component may dictate presentation, staging, and treatment response, thereby obscuring a concurrent, biologically distinct process and producing a composite lymphoma whose true nature may not be immediately apparent [2, 3]. Composite lymphomas may, therefore, remain unrecognized until disease behavior diverges from expectations. Apparent relapse after treatment of an indolent lymphoma is frequently presumed to represent persistence of the original disease; however, changes in distribution or clinical behavior warrant tissue confirmation because a second lymphoma lineage may instead become clinically apparent [4, 5]. We report a case in which treatment response of systemic MZL unmasked a previously occult in situ follicular neoplasia, ultimately establishing a synchronous composite lymphoma [5, 6, 7, 8]. This case highlights how temporality and integrated pathologic reassessment can redefine disease identity and prevent misclassification as relapse or transformation [2, 3]. Case Presentation An 82‐year‐old woman presented in September 2023 with a midline neck mass that was initially suspected clinically to represent a thyroglossal duct cyst and underwent excisional biopsy. Integrated histopathologic and immunophenotypic evaluation established a diagnosis of low‐grade B‐cell lymphoma that was most compatible with MZL (Figure 1). Morphologically, the lesion was composed of small, mature lymphocytes with partial effacement of the nodal architecture and a predominantly diffuse growth pattern. Initial diagnostic lymph node biopsy at presentation (2023). (A) H&E staining shows lymph node effaced by a predominantly diffuse infiltrate of small mature lymphocytes with only focal nodular organization. (B) CD20 highlights a widespread B‐cell population replacing the nodal parenchyma. (C) CD10 staining is largely absent in the infiltrate, arguing against a germinal center phenotype. (D) BCL2 demonstrates diffuse expression within the infiltrating lymphocytes. Overall, in conjunction with the clinicopathologic context of the case presentation, the findings supported an initial diagnosis of MZL. The neoplastic B cells were CD20‐positive and lacked CD5, CD10, and BCL6 expression by immunohistochemistry, and flow cytometry demonstrated a kappa‐restricted CD5‐negative/CD10‐negative B‐cell population. Immunoglobulin gene rearrangement analysis using a polymerase chain reaction (PCR)‐based assay demonstrated a clonal rearrangement. Fluorescence in situ hybridization (performed retrospectively) showed no BCL2 rearrangement. Targeted next‐generation sequencing analysis using the MayoComplete B‐cell Lymphoma panel (Table S1)—which interrogates 46 genes recurrently altered in B‐cell lymphomas including KMT2D, TNFAIP3, MYD88, NOTCH2, KLF2, CREBBP, EZH2, TNFRSF14, and BCL2—detected no pathogenic mutations and revealed only an ATM variant of uncertain significance (allele frequency 53%), suggestive of a likely benign variant. Staging demonstrated systemic involvement, although disease burden was limited. Bone marrow biopsy showed normocellular marrow without morphologic lymphoma, although flow cytometry detected a minute, kappa monotypic, CD5‐negative/CD10‐negative B‐cell population (< 1% cellularity). Peripheral blood flow cytometry identified peripheral blood involvement with immunophenotypically similar monotypic B cells comprising approximately 7% of the white blood cells. FDG‐PET/CT demonstrated a hypermetabolic anterior neck lesion with additional splenic and focal osseous uptake involving the right scapula and humerus, consistent with systemic indolent lymphoma. The patient received systemic therapy on the Phase II MorningSun clinical trial (NCT05207670) with mosunetuzumab, a CD3×CD20 bispecific antibody that engages T cells against B cells, beginning in November 2023. Serial PET/CT imaging showed progressive metabolic improvement, with complete metabolic response achieved by January 2024 (Deauville Score 2) and sustained absence of FDG‐avid disease in April 2024 (Deauville Score 2X). The patient continued mosunetuzumab treatment through October 2024, completing 17 cycles total. At the time of complete metabolic response, the disease was interpreted as treated MZL in complete metabolic remission. Surveillance imaging in July 2024 demonstrated a new, left Level II cervical lymph node with increased uptake (Deauville Score 4). Follow‐up imaging in October 2024 showed interval enlargement and greater avidity of this newly identified left cervical lymph node (Deauville Score 4), raising concern for relapse and prompting tissue evaluation. Core biopsy revealed altered nodal architecture with atypical follicles composed predominantly of centrocytes with occasional centroblasts (< 15 per high‐power field). The neoplastic cells expressed CD20, PAX5, CD10, and BCL6, with a low follicular Ki‐67 proliferation index of approximately 30% without polarization, while BCL2 immunohistochemistry was negative. Flow cytometry identified a CD20‐strongly positive, CD10‐positive, kappa‐restricted B‐cell population comprising approximately 45% of white blood cells, and fluorescence in situ hybridization demonstrated a BCL2 rearrangement in 65% of nuclei. No morphologic or immunophenotypic features of MZL were present. The integrated diagnosis was FL, Grade 1–2 (Figure 2). PCR (PCR‐based capillary electrophoresis) immunoglobulin gene rearrangement analysis detected a clonal rearrangement with a pattern that appeared distinct from that identified in the prior lymph node, supporting the presence of a biologically distinct B‐cell population (Figure 3). Targeted next‐generation sequencing using the same MayoComplete B‐cell Lymphoma panel showed the same ATM variant of uncertain significance (allele frequency 42%) as that identified in the prior sample and no pathogenic mutations. Posttreatment biopsy of a newly enlarged cervical lymph node demonstrating follicular lymphoma (2024). (A) H&E staining, at low power, shows replacement of nodal architecture by closely spaced atypical follicles. (B) High‐power examination demonstrates follicles composed predominantly of centrocytes with occasional centroblasts. (C) CD20 highlights B‐cells with nodular growth pattern. (D) CD10 highlights the follicular center B‐cells. (E) BCL2 immunostaining is negative in the neoplastic follicles. (F) Ki‐67 demonstrates an atypically low proliferative fraction within follicles (approximately 30%). Overall, in the clinicopathologic context described in the case presentation, the findings established follicular lymphoma. PCR‐based immunoglobulin gene rearrangement analysis demonstrating distinct clonal populations in FL and MZL. Capillary electrophoresis tracings of PCR‐amplified immunoglobulin heavy chain (IGH) and kappa light chain (IGK) gene rearrangements from the follicular lymphoma (FL, left panels) and marginal zone lymphoma (MZL, right panels) specimens. The x‐axis represents fragment size in base pairs, and the y‐axis represents fluorescence intensity. Red bars indicate the expected size ranges for framework regions (FR1, FR2, and FR3). Both specimens demonstrate clonal IGH rearrangements with distinct peak positions, indicating different VDJ recombination events. The IGK analysis reveals an additional clonal peak in the MZL specimen (right, lower panel) that is absent in the FL specimen (left, lower panel), further supporting the presence of biologically distinct B‐cell populations. Because these findings were biologically discordant with recurrence of MZL, the original excisional specimen was re‐reviewed. In addition to the dominant MZL pattern, rare small follicular structures with a germinal center phenotype along with strong CD10 and strong BCL2 expression were identified by immunohistochemistry, consistent with in situ follicular neoplasia. This review was prompted by the findings in the subsequent cervical lymph node biopsy and, therefore, was not performed in a blinded fashion. The disease course was therefore reinterpreted as a synchronous composite lymphoma consisting of a dominant MZL at presentation with a concurrent, clinically silent follicular lineage that became clinically apparent after treatment response (Figure 4). Review of the original diagnostic lymph node biopsy revealing in situ follicular neoplasia (2023). (A) H&E staining shows small lymphocytes without significant atypia and with a focally, vaguely nodular architecture. (B) CD20 highlights B‐cells in focal nodules. (C) CD10 demonstrates small germinal centers with aberrantly strong expression within these nodules. (D) BCL2 shows aberrantly strong co‐expression within the germinal centers. Overall, on retrospective review, the findings established the presence of in situ follicular neoplasia at the time of initial diagnosis. The FL component was treated with involved‐site radiotherapy (ISRT), delivered as 4 Gy in two fractions in December 2024. Given the isolated cervical lymph node involvement, local radiotherapy was selected rather than additional systemic therapy. Follow‐up PET/CT in May 2025 demonstrated complete metabolic response with resolution of the biopsied cervical lymph node (Deauville Score 2). Subsequent bone marrow evaluation showed no evidence of lymphoma. Discussion This case illustrates how temporality can clarify lymphoma biology when posttreatment disease behavior diverges from expectations [4]. The initial presentation fulfilled clinicopathologic features typical of MZL, including a CD5‐negative/CD10‐negative immunophenotype, peripheral blood involvement, minimal marrow disease, and multifocal FDG‐avid sites that responded to systemic therapy. The later appearance of a solitary, enlarging cervical node with a germinal center phenotype (CD10+, BCL6+) and a BCL2 rearrangement was biologically incompatible with relapse of the original lymphoma and instead established a second lineage, FL [1, 2, 3]. Retrospective review of the original specimen identified a follicular precursor lesion, reframing the clinical course as a synchronous composite lymphoma unmasked by treatment response rather than relapse or transformation, a biologically plausible scenario given the clinical activity of bispecific antibodies such as mosunetuzumab and epcoritamab in B‐cell lymphomas [9, 10, 11], and it is consistent with reported scenarios in which a clinically dominant indolent lymphoma obscures a concurrent, germinal center‐derived process until re‐biopsy and reassessment [5, 6, 7, 8]. This case also illustrates how retrospective molecular and cytogenetic evaluation of previously obtained tissue may occasionally reveal clinically occult lymphoma components that were not recognized at the time of initial diagnosis. Composite lymphomas represent the coexistence of two morphologically and immunophenotypically distinct lymphomas within the same patient and may remain clinically unrecognized when one component predominates clinically [2, 3, 4]. In indolent lymphomas, observed behavior reflects the dominant clone rather than the total clonal architecture, and evolutionary studies demonstrate that divergent lymphoid neoplasms may arise from either independent clones or branching evolution from an earlier precursor population [5, 6, 7]. In this case, systemic MZL dictated the initial presentation and treatment response, while the follicular lineage remained clinically silent until therapy reduced the dominant clone, a biologic unmasking rather than true sequential lymphoma development [5, 6, 7]. The identification of a BCL2 rearrangement exclusively in the FL specimen, together with distinct immunoglobulin gene rearrangement patterns between the MZL and FL, supports the presence of biologically distinct lymphoma populations. Clonality assessment was performed using conventional PCR‐based immunoglobulin gene rearrangement analysis with capillary electrophoresis. Although the observed differences support separate lymphoma populations, this methodology has recognized limitations in assessing clonal relatedness, particularly in the setting of somatic hypermutation; therefore, definitive clonal independence cannot be established from the available data. NGS‐based clonality approaches with full VDJ sequence reconstruction may provide a more definitive assessment in future composite lymphoma cases. The shared ATM p.Leu1590Phe variant, detected at near‐heterozygous allele frequencies in both specimens, may represent a constitutional variant; however, germline testing was not performed, and a shared ancestral precursor clone cannot be entirely excluded. Although several genes commonly altered in MZL, including KMT2D, TNFAIP3, MYD88, NOTCH2, and KLF2, were interrogated by the NGS panel, no pathogenic variants were identified. While this finding does not exclude the diagnosis of MZL, it provides limited molecular support for subtype‐specific characterization and should be interpreted in the context of the morphologic, immunophenotypic, and clinical findings. Notably, although both MZL and FL express CD20, their differential response to mosunetuzumab likely reflects differences in tumor microenvironment and T‐cell accessibility rather than antigen expression alone. The systemic, blood‐involved MZL was highly accessible to circulating, activated T cells, whereas the localized in situ FL component may have been protected by microenvironmental factors that limit T‐cell infiltration and effector function [12, 13]. The identification of a follicular precursor lesion in the original biopsy is therefore central to the clinical interpretation. Precursor follicular proliferations confined to germinal centers can be present at diagnosis yet clinically inapparent, and they only become relevant once overt FL emerges [6, 7]. The chronological sequence thus reflects change in relative clonal prominence rather than biological progression [5, 6, 7]. An important consideration is that the ISFN component was identified retrospectively in the original biopsy. Because molecular and cytogenetic studies were not performed on the isolated ISFN foci, direct confirmation of BCL2 rearrangement within the ISFN component and formal assessment of clonal relatedness to the subsequent FL were not feasible. This distinction is critical because transformation implies evolution of one lymphoma into another, whereas composite lymphoma reflects coexistence of separate processes arising from different stages of B‐cell differentiation [2, 3]. The second biopsy also demonstrated discordant BCL2 testing, with negative BCL2 immunohistochemistry despite a BCL2 rearrangement by FISH. A lack of BCL2 expression by immunohistochemistry (particularly with clone 124, which was used in this case) despite the presence of a BCL2 rearrangement uncommonly occurs in FL. This most often results from epitope‐altering BCL2 mutations; somatic mutations in the translocated BCL2 gene can cause amino acid substitutions within the epitope region (residues 41–54) recognized by clone 124, resulting in false‐negative immunohistochemistry while alternative antibodies, such as E17, remain positive [14]. Although our next‐generation sequencing analysis interrogated only exons 2–3 of BCL2 and did not identify a BCL2 mutation, alterations outside the covered regions cannot be excluded. Given the strong BCL2 expression in the preexisting in situ follicular neoplasia, acquisition of an epitope‐altering BCL2 alteration remains a possible explanation for this discordance, although this cannot be confirmed from the available data. Such a mechanism would be consistent with documented ongoing intraclonal diversification during progression from ISFN to FL [6]. Reliance on immunohistochemistry alone could therefore misclassify the lesion, emphasizing the need for integrated morphologic, immunophenotypic, and cytogenetic interpretation in germinal‐center lymphomas [2, 3]. Imaging also contributed to recognizing discordant biology. PET‐based response criteria reliably assess treatment response in lymphoma. However, new focal uptake after systemic remission should prompt tissue confirmation rather than assumption of relapse [15]. The shift from disseminated disease to isolated nodal progression was inconsistent with relapse of treated MZL and instead suggested emergence of a different biological process, reinforcing the need for biopsy in cases of unexpected imaging patterns and discordant disease behavior [15]. This consideration that becomes particularly relevant following targeted immune therapies, in which antigen expression may change and repeat tissue evaluation is recommended [11, 16]. Clinically, the principal lesson is interpretive rather than therapeutic. A new disease following treatment of an indolent lymphoma should not automatically be attributed to relapse. Instead, discordant behavior, particularly focal progression after systemic response, should trigger re‐biopsy and clinicopathologic reassessment [4]. In this case, therapy did not generate a second lymphoma; it revealed one that was already present, illustrating how treatment can expose preexisting clonal heterogeneity within composite lymphomas [5, 6, 7, 8]. Conclusions This case underscores a diagnostic pitfall in indolent B‐cell lymphomas: posttreatment progression does not necessarily represent relapse but may instead reflect shifting dominance among preexisting clonal populations within a composite process. Recognition requires attention to discordant biologic behavior rather than reliance on anatomic continuity alone, with repeat biopsy performed to clarify unexpected findings rather than depending exclusively on imaging. The presence of a germinal center lymphoma with a BCL2 rearrangement despite negative BCL2 immunohistochemistry further illustrates that accurate classification requires integrated morphologic, immunophenotypic, and genetic assessment. More broadly, therapy may function as a biologic “unmasking” event, revealing previously silent disease and demonstrating that indolent lymphoma behavior can reflect shifts in clonal prominence rather than linear progression, a distinction with direct implications for prognostic interpretation and clinical decision‐making [5, 6, 7, 8]. Author Contributions A.L.O.‐T. designed the case report, collected and interpreted clinical data, performed the literature review, and drafted the manuscript. J.Z. contributed to pathologic interpretation, data analysis, and critical revision of the manuscript. J.B., F.H., D.V., P.T.G., and N.L. contributed to data interpretation and critical revision of the manuscript. J.M. provided supervision, expert guidance, and critical revision of the manuscript. All authors reviewed and approved the final version of the manuscript prior to submission. Funding The authors have nothing to report. Ethics Statement The authors have nothing to report. Consent Informed consent was obtained from the patient for publication of this case report. Conflicts of Interest Juliana Biondo and Farah Hossain are employees of Genentech, Inc. The other authors declare no conflicts of interest. Supporting information Supporting Information File 1: jha270370‐sup‐0001‐SuppMat.pdf Data Availability Statement The data supporting the findings of this case report are included within the article. Additional data is not publicly available because of patient confidentiality but may be available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials Supporting Information File 1: jha270370‐sup‐0001‐SuppMat.pdf Data Availability Statement The data supporting the findings of this case report are included within the article. Additional data is not publicly available because of patient confidentiality but may be available from the corresponding author upon reasonable request.