Prophylactic dexamethasone at the time of absolute lymphocyte expansion does not mitigate risk or event-free survival for immune effector cell associated delayed neurotoxicity.

Introduction

Ciltacabtagene autoleucel (cilta-cel) has significantly improved survival in patients with relapsed or refractory Multiple Myeloma (RRMM) [ 1 , 2 ] However, it is also associated with a 10-15% risk of immune effector cell-associated delayed neurotoxicity (IEC-DNT) which include immune-effector cell associated cranial nerve palsies (IEC-NP), parkinsonism (IEC-PKS) and other rarer neuropathies (i.e., IEC associated Acute Inflammatory Demyelinating Polyneuropathy; IEC-AIDP) seen at a rate of 5–10%, ~3% and <1% respectively [ 3 – 5 ] Whilst, the use of effective bridging therapy and tumor cytoreduction has reduced the risk, IEC-DNT continue to be reported in patients receiving cilta-cel as standard of care.

We have previously demonstrated that a high post-infusion peak absolute lymphocyte count (ALC peak ) of ≥3×10^9/L, a surrogate for robust in vivo CAR-T cell expansion, was associated with an increased risk of IEC-DNT at 31%, IEC-NP at 21% and IEC-PKS at 12% [ 3 ]. In contrast, the risk was negligible (~1-2%) when ALC peak fell under that threshold. Notably, autopsy studies have previously shown that IEC-PKS could possibly be result of an on-target, off tumor effect of highly active CAR-T cells that have crossed the blood-brain barrier [ 6 ] In contrast, autoreactive polyclonal CD8+ non-CAR-T cell expansion have been implicated in the pathogenesis of IEC-associated peripheral neuropathies [ 7 ] Regardless, It is hypothesized that high CAR-T expansion leads to excessive secretion of inflammatory cytokines that activate vascular endothelium and leads to break down of the blood-brain barrier [ 8 ] It also creates a persistent inflammatory environment and subsequent chronic T-cell activation [ 8 , 9 ] This combination likely facilitates cross-reactivity with healthy neuronal antigens or trigger a persistent immune attack on the cranial nerves in IEC-NP or basal ganglia in the case of IEC-PKS.

In this context, attenuating CAR-T cell over-expansion following infusion may be a possible strategy to reduce the risk of IEC-DNT. Dexamethasone, a potent corticosteroid, exerts immunosuppressive effects by inhibiting T-cell proliferation, reducing cytotoxicity and promoting apoptosis [ 10 – 12 ]. Accordingly, we hypothesized that a short course of dexamethasone prophylaxis at the time of rapid ALC expansion could mitigate the risk of IEC-DNT.

Methods

Selection of the historical control groups

To evaluate the impact of this intervention, a historical control cohort of patients with high ALC was identified from those treated between February 2022 and November 2024 across the three Mayo Clinic centers in Rochester, Arizona, and Florida, prior to the implementation of prophylactic dexamethasone [ 3 ].

Intervention study design and statistical analysis

Given the substantial morbidity and mortality associated with IEC-PKS, we designed a prospective observation single-arm study with a binomial framework to evaluate whether dexamethasone (DEX) administered at the time of ALC expansion could reduce the risk of IEC-PKS compared with historical controls. Based on prior data demonstrating an IEC-PKS incidence of at least 10% among patients with ALC peak  ≥ 3 × 10⁹/L, we aimed to detect a reduction to 1%. Accordingly, the null hypothesis was defined as an IEC-PKS event rate of ≥10%, and the alternative hypothesis as an event rate of ≤1%. Assuming a one-sided alpha of 0.10 and 90% power, a sample size of 38 patients with high ALC was required for the intervention cohort. Under this design, the occurrence of ≤1 IEC-PKS event would lead to rejection of the null hypothesis, supporting further investigation of DEX as a mitigation strategy. Conversely, the occurrence of ≥2 events would preclude rejection of the null hypothesis, indicating insufficient evidence to support the efficacy of DEX prophylaxis.

The primary endpoints were incidence of IEC-DNT, IEC-NP, and IEC-PKS and event-free survival (EFS), defined as the time from cilta-cel infusion to the occurrence of IEC-DNT. Secondary endpoints included progression-free survival (PFS), defined as the time from infusion to disease progression or death. Disease progression and response assessments were defined by the 2016 International Myeloma Working Group criteria [ 13 ]. Distributions of time-to-event outcomes were estimated using the Kaplan–Meier method and compared using the log-rank test. Hazard ratios and corresponding 95% confidence intervals were calculated using Cox proportional hazards models. Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate. Continuous variables were compared using the Wilcoxon rank-sum test.

Intervention cohort for ALC-based mitigation treatment with dexamethasone

The use of prophylactic oral dexamethasone (PPX DEX) was implemented across three Mayo Clinic centers (Rochester, Arizona, and Florida) between December 2024 and August 2025 in a new cohort. Patients receiving standard-of-care cilta-cel underwent serial monitoring of absolute lymphocyte count (ALC) during the first two weeks after CAR-T cell infusion. PPX DEX (10 mg twice daily for a minimum of 3 days) was initiated when ALC exceeded or rapidly approached 3 × 10⁹/L (defined as high-ALC). DEX was stopped or rapidly tapered after 3 days if serial ALC trended down. DEX was resumed when subsequent ALCs re-expanded to a level higher than the previous peak. The intervention group included all patients who received PPX DEX. Additionally, patients receiving therapeutic dexamethasone (TX DEX) at doses exceeding the PPX regimen for management of other toxicities were included in the intervention group. A CONSORT diagram outlining cohort selection is presented in Fig. 1 .

CONSORT diagram of cohort.

CONSORT diagram of cohort (Created in https://BioRender.com ).

Results

Patient characteristics

The historical cohort included 229 patients (patient characteristics shown in Supplementary Table 1 ). Overall incidence of IEC-DNT in the historical cohort was 11% (IEC-NP, 7% and IEC-PKS, 4%). Ninety-six patients (42%, 96/229) in the historical cohort had high-ALC and served as the control group for this analysis. Only 1% (2/133) of patients with historical low-ALC experienced IEC-DNT and were not included.

From December 2024 to August 2025, there were 122 patients who received cilta-cel after the initiation of the PPX DEX protocol (Supplementary Table 1 ). The new cohort exhibited a 14% overall incidence of IEC‑DNT (17/122), comprising 11% IEC‑NP (13/122) and 4% IEC‑PKS (5/122); notably, one patient developed both. Around half of the patients (63/122, 52%) demonstrated high-ALC. The intervention cohort consisted of 57 patients who received DEX: 47 received PPX DEX (44 for ALC ≥ 3 × 10⁹/L and 3 for rapid ALC rise) and an additional 10 received TX DEX during the ALC peak period. There were 13 patients with high-ALC who did not receive PPX DEX and were excluded from the intervention cohort.

Baseline characteristics of the control and intervention cohorts are summarized in Table 1 . Median follow-up was 17.6 months (95% CI, 15.9 to 18.9) in the control group and 5.8 months (95% CI, 4.8 to 6.0) in the intervention group. Patients in the intervention group were older and comprised a significantly higher proportion of patients having received only 1–3 prior lines of therapy (66% vs 35%, p < 0.001). There were, however, no significant differences in high-risk cytogenetic features (defined as the presence of t(4;14), t(14;16), t(14;20), gain or amplification of 1q, deletion 17p, or deletion 1p), response to bridging therapy, pre-lymphodepletion disease burden, tocilizumab use, anakinra use or incidence of CRS, ICANS, or IEC-HS (Table 1 ).

Table 1 Patient characteristics of the control and intervention groups. Control Group ( n  = 104) Intervention Group ( n  = 57) Demographics N (%) N (%) p -value Age, median (IQR) 65 (56–71) 69 (64–74) 0.004 Male Sex 66 (63%) 35 (61%) 0.87 Race (Black or Hispanic) 10 (10%) 3 (5%) 0.38 Pre-LD BMPC ≥ 20% 24 (25%) 21 (39%) 0.10 Pre-LD LC ≥ 20 mg/dL 31 (33%) 19 (34%) 1.00 HRCA † 66 (67%) 36 (77%) 0.33 2 or more HRCA † 29 (29%) 15 (32%) 0.70 1–3 Prior lines of therapy 36 (35%) 37 (66%) <0.001 Received bridging therapy 91 (88%) 51 (89%) 0.80 Response to bridging therapy‡ 21/91 (23%) 13/51 (26%) 0.84 Clinical course CRS 92 (88%) 53 (93%) 0.42  ≥Grade 3 2 (2%) 0 (0%) 0.54 Tocilizumab use 90 (87%) 52 (91%) 0.45  ≥2 doses 43 (41%) 25 (44%) 0.87 ICANS 15 (14%) 9 (16%) 0.82  ≥Grade 3 1 (1%) 3 (5%) 0.13 IEC-HS or MAS-L 9 (9%) 5 (9%) 1.00 Anakinra use 11 (11%) 9 (16%) 0.45 Infection  All grade 38 (37%) 16/44 (36%) 1.00  ≥Grade 3 22 (21%) 12/44 (327%) 0.52 Key: CRS: Cytokine Release Syndrome; HRCA: High-risk cytogenetic abnormalities; ICANS: Immune Effector Cell-associated Neurotoxicity; IEC-HS: Immune Effector. Cell-associated Hemophagocytic Lymphohistiocytosis-like Syndrome; MAS-L: macrophage activation syndrome-like; LD: Lymphodepletion. † High-risk cytogenetic abnormalities (HRCA) were defined as the presence of t(4;14), t(14;16), t(14;20), gain or amplification of 1q, del(17p), or del(1p). ‡ Response to bridging therapy is defined as a partial response of better (as defined by the International Myeloma Working Group in 2016).

Patient characteristics of the control and intervention groups.

Key: CRS: Cytokine Release Syndrome; HRCA: High-risk cytogenetic abnormalities; ICANS: Immune Effector Cell-associated Neurotoxicity; IEC-HS: Immune Effector. Cell-associated Hemophagocytic Lymphohistiocytosis-like Syndrome; MAS-L: macrophage activation syndrome-like; LD: Lymphodepletion.

† High-risk cytogenetic abnormalities (HRCA) were defined as the presence of t(4;14), t(14;16), t(14;20), gain or amplification of 1q, del(17p), or del(1p).

‡ Response to bridging therapy is defined as a partial response of better (as defined by the International Myeloma Working Group in 2016).

Patterns of Dexamethasone use in the intervention cohort

The median cumulative dexamethasone-equivalent dose administered in the 47 patients who received PPX DEX was 60 mg (range, 60–180 mg). Of these, 6/47 patients had ALC re-expansion after corticosteroid cessation and 5/6 patients received a second course of dexamethasone. The ALC trends and intervention of these 6 patients are displayed in Supplementary Fig. 1 . In the 10 patients who received TX DEX, the cumulative dexamethasone-equivalent dose administered was 360 mg (range, 130–1540 mg). The reasons for TX DEX included refractory cytokine release syndrome (CRS), presence of IEC-associated neurotoxicity syndrome (ICANS), or IEC-associated Hemophagocytic Lymphohistiocytosis-Like Syndrome (IEC-HS).

Clinical Outcomes

The use of dexamethasone did not significantly reduce the incidence of IEC-DNT, IEC-NP, or IEC-PKS

The incidence of IEC-DNT in the control group was 22% (23/104), including IEC-NP in 14% (15/104) and IEC-PKS in 9% (9/104). In the intervention cohort, the cumulative incidence of IEC-DNT was 21% (12/57), with IEC-NP observed in 16% (9/57) and IEC-PKS in 7% (4/57). Among the first 38 patients treated with dexamethasone, 3 cases of IEC-PKS were observed, which did not meet the predefined threshold for efficacy. At 120 days, estimated EFS for IEC-DNT was 80% in the intervention group versus 79% in the control group (HR 0.94; 95% CI, 0.46–1.89; Fig. 2A ). For IEC-NP, 120-day estimated EFS was 84% versus 87% (HR 1.10; 95% CI, 0.48–2.52; Fig. 2B ). For IEC-PKS, 120-day EFS was 93% versus 91% (HR 0.80; 95% CI, 0.25–2.61; Fig. 2C ). There was no significant difference in PFS between the intervention and control group (Fig. 2D ).

Survival Curves comparing the control and intervention groups.

A Event-free survival(EFS) for immune effector cell-associated delayed neurotoxicity (IEC-DNT). B EFS for immune-effector cell-associated nerve palsy (IEC-NP). C EFS for immune-effector cell-associated parkinsonism. D Progression-free survival (PFS).

To account for the higher proportion of patients with fewer prior lines of therapy in the intervention cohort, and the possibility that IEC- NP may be more frequent in less heavily pretreated MM [ 5 ], a sensitivity analysis was conducted according to prior treatment exposure (1–3 vs ≥4 prior lines; Supplementary Table 2 ). The cumulative incidence of IEC-DNT, IEC-NP, and IEC-PKS did not differ significantly between the intervention and control cohorts across either subgroup.

Impact on clinical course of IEC-DNT

We next compared the severity of IEC-DNT between groups (Fig. 3 ). The timing of toxicities and the instituted treatment for all the IEC-NP and IEC-PKS cases are shown in Fig. 4 .

Severity of symptoms in patients with Immune effector cell-associated nerve palsy (IEC-NP).

A The distribution of patients with single cranial neuropathy, multiple cranial neuropathies, and multiple cranial neuropathies and peripheral neuropathies between the control and intervention groups. B The distribution of moderate-severe facial nerve palsy between the control and intervention groups defined by the House-Brackmann Score.

Individual patient timeline of specific CAR-T associated toxicity and management.

Swimmer plots displaying the timing of CAR-T associated toxicities and treatment in ( A ) patients with immune effector cell-associated nerve palsy (IEC-NP) and ( B ) Immune effector cell-associated parkinsonism (IEC-PKS) separated by intervention vs control groups. In ( A ), patient 6 in the control group was excluded as they did not have a high absolute lymphocyte count. Patients 18-20 in intervention group were excluded as patients did not receive prophylactic dexamethasone. In ( B ) patient 5 in the control group was excluded as they did not have a high absolute lymphocyte count. Key: CNP, cranial nerve palsy ; CRS, cytokine release syndrome; ICANS, immune effector cell-associated neurotoxicity syndrome; IVIG, intravenous immunoglobulin. White bars. CRS (purple bars), ICANS (blue bars), duration with CNP symptoms (dark green bars), duration with other severe neuropathies (light green bars), duration with moderate-severe IEC-PKS symptoms (dark orange bars), duration with mild IEC-PKS symptoms (light orange bars), duration without toxicity (white bars), corticosteroid treatment (yellow triangle) with error bar reflecting duration, IVIG treatment (orange circle), high-dose cyclophosphamide (green hexagon), carvidopa/levadopa (pink diamond), intrathecal chemotherapy (blue star), death (red cross), alive at last follow-up (red circle).

Among IEC-NP cases, there were 15 in the control group and 9 in the intervention group. The distribution of single versus multifocal nerve involvement was similar between groups (Fig. 3 ). Facial nerve palsy was by far the most common nerve involved in both groups (7/9 in intervention vs 14/15 in control). Using the House–Brackmann scale [ 14 ], moderate-to-severe dysfunction (grade ≥4) was observed less frequently in the intervention group (13% vs 57%, p  = 0.04), suggesting a potential reduction in severity despite similar incidence rates. There was no statistically significant difference in the duration of symptoms between the intervention and control groups (median time to complete resolution 70 vs 48 days; p  = 0.63).

For IEC-PKS, 9 cases occurred in the control group and 4 in the intervention group. Symptom severity at presentation and nadir were assessed using the Hoehn and Yahr (H&Y) scale [ 15 ]. The majority of patients (11/13) had H&Y stage ≥3 disease, indicating mild-moderate bilateral involvement with some functional impairment. The two mild (stage ≤2) cases of IEC-PKS occurred in one patient from the control group and one patient from the intervention group. At the time of data cut, 2/9 cases in the control group had died from IEC-PKS, 4/9 had improved to milder symptoms (H&Y stage 2 or lower) after a course of high-dose cyclophosphamide, 2/9 had improved to milder symptoms with high-dose corticosteroids, whilst 1/9 had self-limiting symptoms. In the 4 IEC-PKS cases from the intervention group, the mild case of IEC-PKS had early symptom resolution after corticosteroids. Of the 3 severe cases, 1 had improved to milder symptoms post-cyclophosphamide whilst the other 2 died from sepsis after high-dose cyclophosphamide without symptom improvement or resolution to earlier corticosteroids or intrathecal chemotherapy. All four patients who died in relation to IEC-PKS had achieved deep responses, including MRD negativity, and had no evidence of progressive disease at the time of death.

There was no difference in all-grade or high-grade (≥Grade 3) infections between the control and intervention groups. Given our prior observation that patients with IEC-DNT have increased odds of developing IEC-enterocolitis (IEC-EC) [ 3 , 16 ], we evaluated the role of DEX in reducing the incidence of IEC-EC. No statistically significant difference in IEC-EC rates was observed between groups, with an incidence of 8.3% (8/104) in the control group compared to 3.5% (2/57) in the DEX intervention group ( p  = 0.50).

Impact on ALC kinetics

Given the absence of a significant difference in EFS, we evaluated the impact of dexamethasone on ALC kinetics over the first 28 days after CAR-T infusion (Fig. 5 ). There was no significant difference in median ALC peak between the intervention and control groups (5.46 vs 5.53 × 10⁹/L; p  = 0.69). ALC at day 28 did not differ significantly between groups (0.74 vs 0.84 × 10⁹/L; p  = 0.94). The area under the curve for ALC levels over 28 days post CAR-T infusion (AUC) was not significantly different between the intervention and control groups (37 vs 42 × 10⁹ cells·day/L; p  = 0.14).

Analysis of Absolute Lymphocyte Count (ALC) in the Control and Intervention Group.

A Absolute lymphocyte count (ALC; median, interquartile range) expansion of the control and intervention groups is plotted from cilta-cel infusion to day 28. B Comparisons of ALC at different time points in patients from the control and intervention group. C ALC expansion over time, from day 0 to day 28, is calculated as the area under the curve for 28 days (AUC 0-28 ). D Comparisons of ALC at different time points in dexamethasone-treated patients who developed immune effector cell-associated nerve palsy (IEC-NP) or parkinsonism (IEC-PKS) vs those who did not, and ( E ) shows AUC 0-28 in the same three groups. * p  < 0.05, NS, not significant.

Within the intervention cohort, patients who developed IEC-DNT exhibited higher ALC expansion despite dexamethasone. Median ALC peak trended higher in patients with IEC-DNT compared to those without (8.31 vs 5.07 × 10⁹/L; p  = 0.09). Additionally, AUC (71 vs 31 × 10⁹ cells·day/L; p  = 0.04) and ALC day28 (3.10 vs 0.70 × 10⁹/L; p  = 0.03) were significantly higher in patients who developed IEC-DNT compared to those who did not. Notably, subdivision of DNT cases into IEC-NP and IEC-PKS revealed that patients who subsequently developed IEC-PKS despite DEX prophylaxis had the highest median ALC peak and AUC among all groups (Fig. 5 D, 5E ).

Discussion

Herein, our multi-site study has found that the use of dexamethasone prophylaxis did not significantly reduce the risk of developing IEC-DNT, including IEC-NP and IEC-PKS. Our findings differ from a recent study by Forsberg et al. of 16 patients with high-ALC (>5 × 10 9 /L) treated with dexamethasone, where the use of dexamethasone prophylaxis improved overall survival [ 17 ] However, the Forsberg study may be underpowered to detect EFS related to IEC-PKS based on their comparison of 9 patients in the historical cohort versus 7 in the PPX DEX intervention cohort where no IEC-PKS was seen. The larger sample size in our study provides more power to detect a significant risk reduction in IEC-DNT. Additionally, we did not find a statistically or biologically significant reduction in ALC expansion kinetics with DEX intervention. This suggests that the dose and/or duration of corticosteroid exposure may be insufficient to meaningfully alter the ALC trajectory, a surrogate for CAR-T cell expansion, or downstream inflammatory processes.

While dexamethasone did not reduce the incidence of IEC-DNT, we observed a statistically significant reduction in severity of presentations for IEC-NP, with significantly fewer cases of moderate to severe facial nerve palsies. This observation may be clinically relevant, as it demonstrates that whilst DEX did not alter T-cell proliferation, it may have an impact on suppressing T-cell function, possibly through the suppression of cytokine production or alteration of differentiation mechanisms [ 18 ]. However, we unfortunately did not observe similar benefit in the IEC-PKS cases. Given the historical and overall high morbidity and mortality of IEC-PKS, we have ceased the use of PPX DEX and are currently exploring alternative mitigating approaches, including an intermediate dose of cyclophosphamide for its lymphodepleting effect.

As no biologically or statistically significant decrease in ALC was identified with DEX, we were unable to determine if decreasing ALC expansion has any impact on reducing the incidence of IEC-DNT. However, it is important to consider that IEC-DNT may also be driven by other interacting mechanisms beyond CAR-T activation and expansion, such as blood-brain barrier disruption, antigen-specific cross-reactivity, T-cell homing and local CNS immune responses, which may not be impacted by a short course of systemic corticosteroids. In addition, a threshold-based approach using ALC may not fully capture the relationship between CAR-T dynamics, where the risk of neurotoxicity is not solely dependent on the magnitude of CAR-T expansion but may also be influenced by qualitative aspects such as the host’s immune landscape, CAR-T cell composition, activation and persistence, or compartmental trafficking into the central nervous system. Therefore, future interventional strategies could select medication with additional immune-modulating effects.

Our study has several limitations, including the use of a historical control group, which introduces the potential for temporal and unmeasured confounding factors. However, we ensured that the shift in practice towards providing more effective bridging and cytoreduction prior to CAR-T infusion was not a major factor in our analysis, with no differences in measurable tumor burden between the two groups. Management of early toxicities such as CRS, ICANS, and IEC-HS was also not significantly different. Additionally, our modest sample size limits the power to detect moderate to mild reduction in risk. Because IEC-PKS is severe, a modest risk reduction is unlikely to be acceptable; therefore, research should prioritize strategies that achieve a more meaningful decrease in incidence.

Despite these limitations, our study provides further insight into the prevention of IEC-DNT. Our findings suggest that a short-course prophylactic strategy of dexamethasone may be associated with a modest reduction in the severity of facial nerve palsies. However, this effect does not adequately translate to a substantially reduced incidence of IEC-DNT, IEC-NP or IEC-PKS. These results highlight the need for further mechanistic studies to identify additional biomarkers for improved risk stratification and to guide future alternative strategies with greater T-cell modulatory effects or targeting alternative inflammatory pathways.

Conclusion

In conclusion, the use of prophylactic dexamethasone during ALC expansion did not reduce the incidence and EFS of IEC-DNT, IEC-NP or IEC-PKS. This highlights the need for more mechanistically informed immunosuppressive or targeted strategies to mitigate this challenging BCMA-directed CAR-T complication.

Supplementary information

Supplementary Figure 1

Supplementary Tables 1 and 2

Acknowledgements

We would like to thank the Immune Effector Cell Compliance Program managers and data coordinators for sharing the CAR-T outcome data, and patients and families for participating in the study.

Author contributions

K.J.L., Y.L.: Conceptualization, data curation, writing–original draft. K.J.L. performed the statistical analyses. K.D reviewed the statistical analysis. R.P., S.C., C.S., K.D., A.D., Ma.G., D.C., Mo.G., L.H., H.S., P.K., M.T., T.K., R.W., J.C., M.B., S.Z., N.A., P.L.B., U.Y., E.W., S.G., S.K., S.A., R.F., A.Z., Y.L.: Data curation, critical appraisal and writing-editing the manuscript. All authors made substantial contributions to acquisition of data, critically revised the manuscript, and gave final approval of the manuscript to be submitted.

Funding

No relevant funding to disclose for this study.

Data availability

No datasets were generated or analyzed during the current study.

Competing interests

K.J.L., M.T., K.D., A.D., Ma.G., D.C., L.H., H.S., P.K., T.K., R.W., J.C., M.B., N.A., S.Z., U.Y., E.W., S.G.: no relevant disclosures; R.P.: advisory board role for Sanofi Aventis and Astra Zeneca, research funding from Bristol Myers Squibb Foundation and GlaxoSmithKline; S.C.: honoraria from Sanofi, Ascentage Pharma, Sobi, Legend Biotech, and research funding from Johnson & Johnson, Takeda, C4 Therapeutics, Abbvie, Ascentage Pharma, AstraZeneca; C.S.: consultancy for Astra Zeneca, GSK, Janssen-Cilag, Menarini Stemline, oncopeptides, Pfizer; honoraria from Abbvie, Astra Zeneca, GSK, Janssen-Cilag, Sanofi, Takeda; other (travel expenses): Astra Zeneca, BeOne, Janssen-Cilag, Menarini Stemline, oncopeptides, Sanofi; Mo.G.: personal fees from Ionis/Akcea, honorarium from Alnylym, personal fees from Prothena, personal fees from Sanofi, personal fees from Janssen, personal fees for Data Safety Monitoring board from Abbvie & Arcellex, fees from Johnson & Johnson, Honoraria for Astra Zeneca, Medscape, Dava Oncology. Alexion; P.K.: Honoraria: Pharmacyclics, Sanofi, BeiGene, MustangBio, AstraZeneca, AbbVie. Consulting or Advisory Role: Sanofi. Research Funding: Amgen, Takeda, Sanofi, AbbVie, GlaxoSmithKline, Sorrento Therapeutics, Karyopharm Therapeutics, Regeneron, Ichnos Sciences, Bristol Myers Squibb/Celgene; P.L.B.: Consultant: Oncopeptidfes. Salarius, Radmetrix, Omeros, CellCentric, AbbVie, Pfizer; S.K.: Consulting or Advisory Role: Takeda, Janssen Oncology, Genentech/Rocher, Abbvie, BMS/Celgene, Pfizer, Regeneron, Sanofi, K36 Therapeutics; travel, accommodation and expenses: Abbvie, Pfizer; Research funding: Takeda, Abbvie, Novartis, Sanofi, Janssen Oncology, MedImmune, Roche/Genentech, CARsgen Therapeutics, Allogene Therapeutics, GSK, Regeneron, BMS/Celgene; S.A.: Consulting or Advisory Role: Takeda, BeiGene, GlaxoSmithKline, Sanofi, Pharmacyclics, BMS, Amgen, Janssen, Regeneron, Cellectar. Research Funding: Pharmacyclics, Janssen Biotech, Cellectar, BMS, Amgen, GlaxoSmithKline, AbbVie, Ascentage Pharma, Sanofi; R.F.: consultancy for AbbVie, Adaptive, Amgen, Apple, BMS/Celgene, GSK, Janssen, Karyopharm, Pfizer, RA Capital, Regeneron, Sanofi. Scientific advisory board: Caris Life Sciences. Board of directors: Antengene. Patent for FISH in multiple myeloma. A.Z.: patents submitted for DACH1-IgG, PDE10A-IgG, CAMKV-IgG, Tenascin-R-IgG as biomarkers of neurological autoimmunity, received research funding from Roche/Genetech and Center for MS and Autoimmune Neurology at Mayo Clinic not relevant to this study, Consulted without personal compensation for Alexion pharmaceuticals, steering committee for the KYSA-8 study for autologous CAR-T-19 treatment for SPS; no personal compensation; Y.L.: advisory board role for Janssen, Sanofi, BMS, Regeneron, Genentech, Tessera, Legend, NexT Therapeutics, steering committee for Janssen, Kite/Gilead, research funding from Janssen, BMS, scientific advisory board for NexImmune, Caribou and data safety monitor board for Pfizer.

Ethics approval, consent to participate and inclusion statement

The study received approval from the Mayo Clinic Institutional Review Board (16-008261). All procedures adhered to applicable guidelines and regulations, and given the retrospective nature of this research, patient informed consent was not required.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41408-026-01598-5 .

References

Associated Data

Supplementary Materials

Supplementary Figure 1

Supplementary Tables 1 and 2

Data Availability Statement

No datasets were generated or analyzed during the current study.