B cell-driven refractory hemocytopenia in immune checkpoint inhibitor therapy: a report of two cases.

Introduction
The integration of immune checkpoint inhibitors (ICIs) with chemotherapy has transformed the therapeutic landscape for advanced lung cancer, significantly improving survival outcomes compared to chemotherapy alone (1). However, ICIs enhance activation and survival of T cells, leading to some ICIs-related adverse events (irAEs) becoming more frequent and severe (2). The overall incidence of hematologic toxicities is higher with immunochemotherapy than with chemotherapy alone, suggesting that ICIs may contribute to their occurrence. In advanced lung cancer patients, the rate of all-grade hematologic toxicities is less than 1% for immunotherapy (3), approximately 30–40% for chemotherapy (4,5), and up to 70% for immunochemotherapy (6,7). Although high-grade hematological irAEs (heme-irAEs) are rare, occurring in less than 1% of patients, they are often refractory to conventional therapies such as glucocorticoids and hematopoietic growth factors, and can be fatal due to complications including severe infections, hemorrhage, and shock (8). Despite emerging reports of thrombocytopenia, neutropenia, aplastic anemia, autoimmune hemolytic anemia, and pure red cell aplasia (9-15), the underlying mechanisms remain unclear, and evidence-based management strategies are lacking in current guidelines.
A critical knowledge gap persists: why do certain patients develop refractory heme-irAEs that are resistant to standard therapies like glucocorticoids and hematopoietic growth factors? B-cell-driven hematologic irAEs such as immune thrombocytopenia and autoimmune hemolytic anemia have been reported, and intravenous immunoglobulin (IVIG) or rituximab are already recommended in guidelines as second- or third-line options for steroid-refractory immune cytopenias. Both T- and B-cell–mediated mechanisms have been implicated in irAEs. While T-cell hyperactivation is well described in myocarditis and hepatitis (16,17). The contribution of B-cell dysregulation and autoantibody production—particularly in hematologic irAEs—remains incompletely defined. Clarifying this distinction is crucial, as it could inform both early risk stratification and the choice of targeted interventions.
Here, we present two rare cases of grade-IV refractory hemocytopenia following ICIs-chemotherapy, characterized by (I) profound resistance to glucocorticoids, hematopoietic growth factors, and transfusions; (II) the exclusive presence of anti-Sjögren’s-syndrome-related antigen A (Ro) (SSA) antibodies in refractory patients, which were absent prior to treatment; and (III) subsequent hematologic recovery accompanied by serologic clearance of anti-SSA antibodies following B cell-targeted therapy with IVIG and rituximab. Here, we present two cases of glucocorticoid and hematopoietic growth factors resistant-refractory hemocytopenia after ICI-chemotherapy, both showing the emergence of anti-SSA antibodies and hematologic recovery after B-cell-targeted therapy. These observations are hypothesis-generating and highlight the need to explore whether anti-SSA antibodies may serve as a biomarker of refractory heme-irAEs. We present this article in accordance with the CARE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-939/rc).

Case presentation

Patient 1
An elderly female with extensive-stage small cell lung cancer (ES-SCLC) initiated first-line therapy with atezolizumab (1,200 mg), etoposide (60 mg/m2 ×3 days), and carboplatin (300 mg/m2) every 3 weeks, combined with concurrent thoracic radiotherapy. During cycles 1–4, no heme-irAEs were observed. At week 1 post-cycle 5, grade-II neutropenia [neutrophil (NEUT) count: 0.85×103/µL] resolved within 48 hours of granulocyte colony-stimulating factor (G-CSF) administration.
One week after the 6th cycle of ICIs plus chemotherapy, routine blood examination indicated grade-III heme-irAE, with white blood cells (WBC) count of 1.6×103/µL, NEUT count of 0.85×103/µL, hemoglobin (HGB) count of 118 g/L, and platelet (PLT) count of 130×103/µL (Figure 1A). Despite treatment with G-CSF at a dose of 1.2×107 IU/day for 7 consecutive days, the NEUT count remained unchanged, and the PLT count further decreased to 3×103/µL. Secondary causes of hemocytopenia were systematically evaluated. Concomitant medications were reviewed and no new myelosuppressive or hemolytic drugs had been introduced before the onset. Other immune diseases were assessed by clinical history and autoimmune serology [e.g., antinuclear antibody (ANA), extractable nuclear antigen (ENA) panel including SSA, rheumatoid factor], without evidence of an alternative systemic autoimmune disorder. Infection-related cytopenia was evaluated with physical examination, chest imaging, repeated bacterial and fungal blood cultures, urine cultures, and viral tests [e.g., hepatitis B and C, human immunodeficiency virus (HIV), cytomegalovirus (CMV), Epstein-Barr virus (EBV)], all of which were negative or non-diagnostic. Taken together, these findings supported an immune-related adverse event associated with ICI-based therapy as the most likely cause of hemocytopenia.
The patient received methylprednisolone (80 mg/day), recombinant human thrombopoietin (rh-TPO) (15,000 U/day), recombinant human granulocyte stimulating factor (rhG-CSF) (75 µg/day), interleukin 11 (IL-11) (1.5 mg/day), inosine injection (500 mg/day) and irregular PLT transfusion for 6 days. The WBC and NEUT returned to normal, but PLT showed no recovery (Figure 1A).
On day 12, we performed a bone marrow aspiration. Bone marrow morphology showed an almost absence of megakaryocytes and decrease of granulocyte and erythroid cells lineage (Figure 1B). Peripheral immunophenotyping revealed preserved Treg/B-cell counts but depleted NK/T-cell subsets (Table S1). IgG and IgM antibodies decreased and CH50 antibody increased (Table S2). Additionally, we detected autoimmune antibodies in the peripheral blood, and the patient was positive for anti-SSA antibodies, which are commonly seen in patients with autoimmune diseases but unreported in prior ICI-related heme-irAEs (Figure 1C). The patient had no history of autoimmune disease and anti-SSA antibody was negative before treatment. We speculated that ICIs may aberrantly activate B cells to produce anti-SSA antibodies, which might be involved in the destruction of blood cells. The patient was then given IVIG (10 g/day) for 5 days. The patient’s PLTs gradually returned to normal. The anti-SSA antibody was still positive on day 22 (Figure 1C), but there was no decrease in PLTs within 4 months of follow-up.

Patient 2
A middle-aged female with ES-SCLC received 4 cycles of sintilimab (200 mg) plus etoposide (60 mg/m2 for 3 consecutive days) and carboplatin (300 mg/m2) every 3 weeks, along with concurrent chest radiotherapy. The patient achieved a partial response (PR) based on the Response Evaluation Criteria in Solid Tumors (RECIST), and subsequently underwent anti-programmed cell death protein 1 (PD-1) monotherapy maintenance therapy for 5 cycles without experiencing any irAEs. However, day 14 post-last treatment, the patient developed a fever. Blood tests revealed a low WBC count of 1.73×103/µL, NEUT count of 1.06×103/µL, HGB count of 90 g/L, and PLT count of 181×103/µL (Figure 1D). Despite 21-day G-CSF (1.2×107 IU/day) in other hospital, cytopenias worsened, indicating grade-IV refractory heme-irAE. Patient 2 underwent the same comprehensive exclusion workup as Patient 1. No bacterial or viral infections were detected, and autoimmune diseases, hematologic malignancies, chemo-drug toxicity, and radiation-induced hemocytopenia were ruled out. Anti-PD-1 antibody-induced hemocytopenia was considered as a potential cause.
On day 22, a bone marrow aspiration was performed, revealing amounts of fibrous tissue infiltration and extremely inhibited proliferation of the granulocyte and erythroid lineages, but normal megakaryocytes (Figure 1E). The phenotype of each cell subtype was normal. The counts of T, B, and NK cells in the blood all showed a decrease (Table S1). IgG and β2MG antibodies in peripheral blood were found to be elevated (Table S2). Similar to Patient 1, this patient also tested positive for anti-SSA antibody (Figure 1F). To investigate further, we conducted antinuclear antibodies tests in other patients with small cell lung cancer and in those who developed grade III–IV heme-irAE but recovered within 1 week after receiving glucocorticoids/rhG-CSF/rh-TPO/erythropoietin (EPO) treatment. However, the anti-SSA antibody was negative in these patients, suggesting that anti-SSA antibodies are not secreted by small cell lung cancer and only manifest in patients with refractory hemocytopenia (Table S3).
The patient was treated with a combination of rhG-CSF (300 µg/day), dexamethasone (15 mg/day), IVIG (2.5 g/day), and red blood cells and plasma infusion for another 11 days, but no increase in WBC and NEUT was observed (Figure 1D). The patient developed refractory hemocytopenia and did not respond to rhG-CSF, glucocorticoids and IVIG. We hypothesized that B cell abnormalities induced by ICIs could lead to autoantibody production, resulting in the destruction of blood cells. Rituximab, which binds the CD20 antigen on the surface of B lymphocytes and causes depletion of abnormal B lymphocytes, may reduce antibody-dependent hematologic toxicities. Therefore, intravenous rituximab was administered (at a dose of 100 mg) on day 30 after onset. After 9 days of injection, the WBC and NEUT counts returned to normal levels, and the anti-SSA antibody became negative (Figure 1F). Rituximab was well tolerated; no infusion reactions, renal impairment, or serious infections were observed during treatment or follow-up. The patient subsequently received anlotinib treatment and had normal peripheral blood cells within 4 months of follow-up.
Based on these two cases, we present a conceptual, antibody-guided framework (Figure 2) that may help generate hypotheses for future studies on ICIs-related hemocytopenia. This schematic is not intended as a practice-changing algorithm, but rather as an illustration of how serologic information might be integrated into decision-making in highly selected, life-threatening refractory cases in the context of existing guidelines: (I) step 1: exclude secondary causes (e.g., infections, autoimmune disorders, drug/radiation toxicity); (II) step 2: confirm ICIs-associated hemocytopenia and test for anti-SSA antibodies; (III) step 3: for anti-SSA⁻ patients, initiate conventional therapy [glucocorticoids + hematopoietic growth factors (rhG-CSF/rh-TPO/EPO)], achieve favorable outcomes. For anti-SSA⁺ patients with severe ICIs-related cytopenias who remain refractory to an adequate trial of high-dose glucocorticoids, B cell-targeted salvage therapy (IVIG/rituximab) may be considered within a conceptual, antibody-guided framework. This algorithm is a theoretical, hypothesis-generating construct that provides a diagnostic and therapeutic opinion for this specific setting. By integrating anti-SSA as a biomarker, patient outcomes may improve, underscoring its potential clinical relevance.
All procedures performed in this study were in accordance with the ethical standards of the institutional and/or national research committee(s) and with the Declaration of Helsinki and its subsequent amendments. Written informed consent was obtained from the patients for publication of this case report and accompanying images. A copy of the written consent is available for review by the editorial office of this journal.

Discussion
The incidence of immune-related hematologic toxicities (heme-irAEs) remains low (3.6%) but poses life-threatening risks in refractory cases, particularly when driven by aberrant B cell activity. In our two cases, anti-SSA antibodies were observed in association with refractory hemocytopenia, whereas they were not detected in several small comparison groups. These observations suggest a possible contribution of B cell- and antibody-mediated mechanisms in a subset of heme-irAEs, complementing rather than challenging existing T cell-centred models. Given the small sample size and retrospective nature, these findings are exploratory and cannot establish causality or predisposition. Importantly, in our cases, the emergence of anti-SSA antibodies was documented after the onset of hematologic toxicity. Instead, our findings suggest that its presence may identify a subgroup of patients predisposed to refractory disease once hematologic toxicity develops.

Mechanistic insights: beyond T cells to B cell-driven pathology
With the increasing use of ICIs, there is a growing interest in heme-irAE. However, the underlying mechanism remains largely unknown. The pathogenesis of refractory hemocytopenia in our cases diverges from classic T cell-mediated cytotoxicity (18). Instead, we propose a B cell-centric model wherein ICIs disrupt immune tolerance, leading to clonal expansion of autoreactive B cells and anti-SSA antibody production. These abnormal antibodies may directly target hematopoietic precursors (e.g., CD34⁺ cells) or mature blood cells via two potential pathways: (I) antigen-specific destruction: analogous to autoimmune disease, abnormal antibodies may bind Ro/La-like epitopes on erythroid or megakaryocytic progenitors, triggering apoptosis or differentiation arrest (19,20); (II) Fc-mediated effector mechanisms: complement activation (C3/C4 consumption) or antibody-dependent cellular phagocytosis (ADCP) by macrophages could amplify cell destruction, as seen in autoimmune cytopenias (20).
Notably, the persistent anti-SSA antibody in Patient 1, despite hematologic recovery, suggests incomplete immune modulation by IVIG, which only blocks Fc receptors but fails to eliminate pathogenic B cell clones. In contrast, rituximab’s efficacy in Patient 2 underscores CD20⁺ B cell depletion as a critical step to terminate autoantibody production.

ICIs re-shape B cell biology: a missing link in heme-irAEs
Emerging evidence indicates that ICIs not only activate T cells but also reprogram B cell metabolism and differentiation (21-23). PD-1/programmed cell death ligand 1 (PD-L1) axis inhibition may enhance B-cell clonality, promoting plasma cell formation and autoantibody secretion (24,25). For instance, PD-1 deficiency in mice leads to spontaneous germinal center expansion and anti-nuclear antibody generation, mirroring autoimmune phenomena observed in ICI-treated patients (26). Clinically, elevated levels of circulating autoreactive B cells and preexisting antibodies have been documented in patients developing severe irAEs (27). These findings align with our observation of anti-SSA antibody emergence in refractory hemocytopenia, suggesting that ICIs may license autoreactive B cell clones to target hematopoietic precursors or mature blood cells. Critically, rituximab’s efficacy in our cases supports the centrality of B cells, as CD20⁺ B cell depletion terminates pathogenic antibody production and restores hematopoiesis. This paradigm shift underscores the need to redefine heme-irAE mechanisms beyond T cell-centric models.

Contrasting existing literature: why anti-SSA antibody matters
Previous studies predominantly attribute irAEs to CD8⁺ T cell-mediated cytotoxicity or cytokine-driven organ injury (28). These mechanisms fail to fully explain the subset of refractory cases unresponsive to conventional therapies. Our study identifies a distinct B cell-driven pathology, where anti-SSA antibody production emerges as a hallmark of glucocorticoid resistance, a phenomenon previously unrecognized in ICIs-related hematotoxicity. The de novo emergence of anti-SSA antibodies in ICI-treated patients without prior autoimmune disease, coupled with their serologic clearance following CD20-targeted therapy, suggests a unique, treatment-induced etiology distinct from classic autoimmune disorders.
We propose that ICIs-induced immune hyperactivation may disrupt B cell tolerance, leading to the aberrant production of anti-SSA antibodies. This could result from the direct activation of autoreactive B cell clones, which are normally suppressed by immune checkpoint pathways such as PD-1/PD-L1. The inhibition of these pathways by ICIs may remove critical regulatory signals, allowing autoreactive B cells to escape peripheral tolerance and produce pathogenic autoantibodies. This mechanism aligns with the acute onset of hematologic toxicity following ICIs therapy, contrasting with the chronic immune dysregulation seen in classic autoimmune diseases.
This hypothesis not only provides a plausible explanation for the emergence of anti-SSA antibodies in our patients but also highlights the need for further research into the role of B cell tolerance breakdown in ICIs-related toxicities. Understanding the precise mechanisms underlying this process could uncover novel therapeutic targets to prevent or mitigate ICIs-related hematologic toxicities.

Clinical translation: anti-SSA antibody as a decision-making tool
Establishing evidence-based strategies for refractory heme-irAEs remains challenging due to their rarity. For patients with high-grade immunotoxicity, high-dose glucocorticoids therapy has been recommended as a first-line treatment (29). However, this approach is associated with poorer prognosis of immunotherapy, and guidelines are gradually reducing the glucocorticoid dose. IVIG is recommended for patients who do not improve with glucocorticoids. If previous treatment with glucocorticoids and/or IVIG is unsuccessful, immunosuppressants, such as T cell inhibitors (e.g., cyclosporine, mycophenolic acid) or B cell inhibitors (e.g., rituximab), may be considered (29). However, current guidelines emphasize glucocorticoids as first-line therapy but neglect antibody screening, potentially delaying targeted interventions.
Our findings support a theoretical, anti-SSA-guided management concept for severe, steroid-refractory heme-irAEs. In this concept, anti-SSA antibodies are viewed as a candidate biomarker that might help identify patients in whom earlier consideration of B cell-targeted therapy could be appropriate. Autoimmune serology, particularly anti-SSA antibody testing, should be performed in all patients with ICIs-related heme-irAE. We propose an antibody-guided algorithm (Figure 2 revised): (I) anti-SSA⁺ patients: immediate escalation to IVIG (to neutralize antibodies) or rituximab (to eradicate autoreactive B cells), bypassing prolonged steroid trials; (II) anti-SSA⁻ patients: patients negative for anti-SSA antibody are at lower risk of progressing to refractory hematologic toxicity and typically exhibit favorable responses to standard first-line therapies, including glucocorticoids and hematopoietic growth factors (e.g., G-CSF, TPO agonists), which may achieve adequate disease control without escalation to advanced immunosuppressive agents.

Limitations and future directions
However, there are some limitations in this study. First, the incidence of immune-related refractory hemocytopenia remains unclear. Our analysis is based on only two small cell lung cancer patients, making the proposed diagnostic framework and related algorithms highly preliminary and not yet validated in a larger patient cohort. Although we observed a potential association between anti-SSA antibody and refractory hemocytopenia, further research is needed to confirm the clinical relevance of this finding. Additionally, the reinitiation of immunotherapy in such patients has not been addressed. We acknowledge that making broad recommendations for evaluation and treatment based on just two cases is indeed premature and is one of the limitations of our study. Our intention, however, was not to establish definitive guidelines, but rather to propose a preliminary diagnostic and therapeutic approach based on these cases. Future studies will validate anti-SSA antibody as a biomarker and characterize antibody subtypes (e.g., IgG1 vs. IgG4) and their pathogenic mechanism in multicenter cohorts. We hope that this suggestion will help in better identifying patients with immune-related refractory hematologic toxicity and in initiating timely treatment.

Conclusions
In conclusion, we describe two cases of glucocorticoid-refractory hemocytopenia after ICI-based therapy in which emergent anti-SSA antibodies and response to B cell-directed therapies were observed. These findings raise the hypothesis that a B cell-mediated toxicity axis may contribute to a subset of heme-irAEs and that anti-SSA antibodies could be explored as a biomarker in future studies.

Supplementary
The article’s supplementary files as