Биотипная антителная терапия в множестве миеломы и риск прогрессивной многофукальной лейкоэнцефалопатии.

TO THE EDITOR:

Bispecific antibodies (bsAbs) targeting B-cell maturation antigen (BCMA) or G protein–coupled receptor family C group 5 member D (GPRC5D) have reshaped the treatment landscape of multiple myeloma (MM), achieving high response rates and durable efficacy in patients with relapsed or refractory disease. 1 , 2 However, these agents are associated with substantial infectious complications, warranting careful monitoring and prophylactic strategies. 3 , 4 Progressive multifocal leukoencephalopathy (PML), a life-threatening opportunistic infection of the central nervous system caused by reactivation of JC polyomavirus (JCPyV), 5 has recently been reported in bsAb-treated patients with MM. 3 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 We investigated a potential safety signal for PML in bsAb-treated MM using a multimodal approach combining description of patients identified from the French prospective immunotherapeutic approaches for treating PML (THEOREM) registry and the French Pharmacovigilance Database (FPVD), a systematic literature review, and disproportionality analyses in VigiBase, the World Health Organization pharmacovigilance database.

The THEOREM registry includes all consecutive patients with suspected JCPyV-related central nervous system diseases presented to the dedicated national multidisciplinary board since October 2023, with systematic collection of clinical, biological, radiological, management, and outcome data. The FPVD captures all adverse drug reactions (ADR) reported to the national network since 1985. Patients were included in our cohort if they had MM, received bsAbs, and fulfilled American Academy of Neurology diagnostic criteria for definite PML 17 ; FPVD reports additionally required reporting by a health care professional. Patients from both sources were deduplicated using key variables. We also performed a systematic review of PML in patients with MM treated with bsAbs, reported in PubMed and Embase through March 2026 (search query available in supplemental Material ). Finally, we conducted a disproportionality analysis in VigiBase, using a case-noncase design comparing PML reporting with bsAbs vs other MM therapies (immunomodulatory drugs, proteasome inhibitors, anti-CD38 monoclonal antibodies, selinexor, chimeric antigen receptor T cells, and belantamab mafodotin). We included all individual case safety reports in adults with documented sex and age between January 2022 and March 2026. Sensitivity analyses excluded non–health care professional reports, non-MM indications, or low-completeness reports. Secondary analyses restricted noncases to infectious ADR. Potential duplicates were identified using the VigiMatch algorithm. 18 The study complied with the Comission Nationale de l'Informatique et des Libertés (CNIL) Reference Methodology MR-004 and was approved by Toulouse University Hospital (RnIPH 2024-96; CNIL number 2206723 v0). Informed consent or nonopposition form was obtained when applicable. The FPVD is registered with the CNIL (number 2014-032) and all data recorded in the base are anonymous. The disproportionality analysis used fully anonymized data and complied with VigiBase Guidelines and Caveats.

Up to March 2026, 14 patients with definite PML occurring during bsAb treatment for MM were identified after deduplication through the THEOREM registry (n = 13/138 patients [9.4%]) and the FPVD (n = 1 additional patient). All patients had heavily pretreated, triple-class refractory MM, with a median disease duration of 13 years (interquartile range [IQR], 10-14). Median exposure to bsAbs before PML diagnosis was 22 months (IQR, 12-24), and 10 of 14 patients (71%) were in very good partial response or better ( supplemental Table 1 ). The median time from symptom onset to PML diagnosis was 25 days (IQR, 19-45; Figure 1 ). Neurologic manifestations included motor deficits, cognitive impairment, and visual symptoms (57% each; supplemental Figure 1 ). The median CD4 count was 152 cells per μL, and the median cerebrospinal fluid white cell count was 1 cell per μL (IQR, 0-2; supplemental Table 2 ). Despite bsAb discontinuation, outcomes were poor, with 12 of 14 patients (86%) dying from PML after a median of 59 days (IQR, 42-91; supplemental Table 3 ). The systematic review identified 14 additional patients with similar characteristics and high mortality ( supplemental Tables 1 , 3 , and 4 ; supplemental Figure 2 ). These patients had received more previous lines of therapy; among patients with available data, all had undergone autologous stem cell transplant. Median bsAb exposure was shorter (10 months; IQR, 8-18), but exceeded 6 months in all but 1 report, in whom PML was diagnosed 5 days after first exposure. In VigiBase, 166 895 ADR reports involving MM therapies were identified, including 5304 involving bsAbs and 112 patients with PML ( supplemental Table 5 ). Among bsAb-exposed patients, 41 PML reports were retrieved ( supplemental Table 6 ). The reporting odds ratio (ROR) of PML was significantly higher with bsAbs compared with other MM therapies (ROR, 17.7; 95% confidence interval, 12.1-26.1), whereas RORs of PML with other MM therapies were low or not significant ( Figure 2 ). These findings were consistent across calendar years and in sensitivity analyses ( supplemental Tables 7-9 ). Figure 1. Swimmer plot of patients treated with bsAbs who developed PML in the THEOREM registry and FPVD. PD-1, programmed cell death-1; VGPR, very good partial response; VST, virus-specific T cells. Figure 2. Reporting odds ratio with 95% confidence intervals for PML associated with bsAbs vs other MM therapies. Data for 2026 were available through 9 March. Solid lines show main analysis, with noncases defined as any ADR, whereas dashed lines show secondary analysis, with noncases defined as infectious ADR only. (A) Reporting odds ratio (ROR) for bsAbs, stratified by year and overall. (B) ROR for different MM therapy classes. Belantamab mafodotin was not included due to the small number of ADR reported and the expected lack of statistical power. Anti-CD38, anti-CD38 monoclonal antibody; CART, chimeric antigen receptor T-cell therapy; IMiD, immunomodulatory drug; PI, proteasome inhibitor.

Swimmer plot of patients treated with bsAbs who developed PML in the THEOREM registry and FPVD. PD-1, programmed cell death-1; VGPR, very good partial response; VST, virus-specific T cells.

Reporting odds ratio with 95% confidence intervals for PML associated with bsAbs vs other MM therapies. Data for 2026 were available through 9 March. Solid lines show main analysis, with noncases defined as any ADR, whereas dashed lines show secondary analysis, with noncases defined as infectious ADR only. (A) Reporting odds ratio (ROR) for bsAbs, stratified by year and overall. (B) ROR for different MM therapy classes. Belantamab mafodotin was not included due to the small number of ADR reported and the expected lack of statistical power. Anti-CD38, anti-CD38 monoclonal antibody; CART, chimeric antigen receptor T-cell therapy; IMiD, immunomodulatory drug; PI, proteasome inhibitor.

PML is a rare, potentially underrecognized, life-threatening opportunistic infection occurring in patients with heavily pretreated MM. Across data sources, almost all patients had received effective BCMA-directed immunotherapies for >6 months, irrespective of MM disease status. Diagnostic delays, reflected by a median of nearly 1 month between symptom onset and diagnosis, likely relate to the subacute presentation and diagnostic challenges, consistent with previous reports. 19 Visual field defects related to occipital and optic radiations demyelination may represent an important warning sign for hematologists, as they were commonly reported. Progressive neurologic symptoms, even in patients in complete response or after bsAb discontinuation, should prompt brain magnetic resonance imaging and quantitative JCPyV DNA testing in the cerebrospinal fluid. 20 Acute and typically self-limited neurotoxicity occurring shortly after bsAb initiation, consistent with immune effector cell–associated neurotoxicity syndrome, should be distinguished from PML, in which corticosteroids may be harmful. 4 Overall mortality remained extremely high, consistent with previous reports in hematologic malignancies. 21 Emerging immunotherapeutic strategies aiming to restore antiviral immunity showed uncertain efficacy in our cohort, despite their compelling biological rationale and increasing clinical use. 20 , 22

From a pathophysiological perspective, the risk of PML in bsAb-treated MM arises from profound and cumulative immunosuppression driven by the underlying disease, previous therapies, and bsAb-induced immune dysfunction. BsAbs activate T cells to kill myeloma cells, but chronic stimulation may lead to T-cell exhaustion, reduced proliferative capacity, and impaired cytokine production, collectively weakening antiviral immune surveillance and increasing susceptibility to opportunistic infections such as PML. 23 , 24 BCMA-targeting agents may induce deeper and more prolonged hypogammaglobulinemia due to the role of BCMA in plasma cell survival, potentially compounding infection risk. In contrast, GPRC5D-targeting bsAbs may have a distinct infection profile, reflecting differences in antigen expression and tissue distribution. 25 However, data on PML risk with GPRC5D-targeting bsAbs remain scarce, given the frequent coexposure or previous exposure to BCMA-targeting therapies and the limited clinical experience.

Owing to the descriptive design, small sample size, and absence of a reference population, these findings should be interpreted with caution. Importantly, disproportionality analyses are not intended to establish causal relationships, but reflect an excess reporting of PML associated with bsAbs, and do not allow estimation of incidence or risk. In addition, this pharmacovigilance analysis could be affected by: (1) reporting bias related to the recent introduction of bsAbs, which we addressed through calendar year stratification; (2) indication bias, as patients receiving bsAbs are often more heavily pretreated and may have greater cumulative immunosuppression, although therapies used in advanced disease were included in the reference group (belantamab, selinexor, and chimeric antigen receptor T cells); and (3) an increasing diagnostic awareness among clinicians and improved access to JCPyV testing, leading to higher PML detection over time.

In conclusion, PML may represent an emerging complication in patients with MM receiving bsAbs. Importantly, these agents provide substantial clinical benefit in heavily pretreated patients, and our findings should be interpreted within this favorable benefit-risk balance. PML in this setting likely reflects the combined effect of advanced disease, cumulative immunosuppression from previous therapies, and treatment-related immune dysfunction, rather than a drug-specific effect alone. Whether earlier use of bsAbs in less heavily pretreated patients with more preserved immune function modifies this risk remains unknown and warrants prospective evaluation.

Conflict-of-interest disclosure: G.M.-B. reports consulting fees from Pfizer. N.W. reports consulting fees from Alexion and Lucane Pharma, outside the scope of this article. K.B. reports consulting fees from Medinspire startup, outside the scope of this article. V.P. reports consulting fees from Gilead, ViiV Healthcare, Merck Serono, Novartis, and Roche, outside the scope of this article. T.W. reports honoraria from AbbVie, Biogen, Edimark, and Ipsen; research grants from Direction Générale de l'offre de Soins, Novartis, Association des Personnes Concernées par le Tremblement Essentiel, Fondation Planiol, the France Parkinson organization, and Revue Neurologique; prize money from the Société Française de Neurologie; and travel funding from Homeperf, Lübeck, LVL Medical, and the Movement Disorder Society. D.R.-W. reports honoraria from AbbVie, AstraZeneca, BeOne Medicines, Johnson & Johnson, and Lilly. B.S. reports research funding (to the institution) from Roche, Novartis, and Merck; and compensations for lectures from Roche, Novartis, Merck, Alexion, Biogen, and Sanofi, outside the scope of this article. A.P. reports honoraria from AbbVie, Adaptive Biotechnologies, Amgen, Bristol Myers Squibb, GSK, Johnson & Johnson, Menarini Stemline, Pfizer, Sanofi, and Takeda. The remaining authors declare no competing financial interests.

A complete list of the members of the THEOREM Study Group appears in “Appendix.”