Outcomes of relapse after teclistamab therapy in multiple myeloma

Teclistamab is a first-in-class CD3xBCMA bispecific T-cell engager (TCE), approved by the FDA in October 2022 for the treatment of triple-class exposed relapsed/refractory multiple myeloma (RRMM) patients after ≥4 prior lines of therapy, a population that, prior to the advent of TCEs and CAR T-cell therapies, had a median overall survival (OS) of less than one year [1, 2]. The MajestTEC-1 trial demonstrated an overall response rate (ORR) of 63% and median progression-free survival (PFS) and OS of 11.3 and 18.3 months, respectively. However, despite high rates of deep initial responses, most patients relapse, and no standard of care has been established following teclistamab failure. In this study, we report the outcomes of relapse after teclistamab therapy. We retrospectively analyzed 179 consecutive patients who completed step-up dosing and received at least one full dose of teclistamab at the University of Pennsylvania. All were triple-class exposed RRMM. Nine (5%) received 1 full dose, 10 (6%) received 2, 7 (4%) received 3, and 153 (85%) received ≥4 doses. Patients were treated either in clinical trials (n = 55; first dose: Dec 2018–Dec 2023) or commercially (n = 122; first dose: Dec 2022–Apr 2024) (Fig. 1a). Data cutoff was February 15th, 2025, with a median follow-up of 21.3 (95% CI: 19.5–24.9) months from teclistamab initiation. The International Myeloma Working Group (IMWG) criteria were used for response assessment [3] and risk stratification [4].

Fig. 1. Outcomes of relapse after teclistamab therapy.

a CONSORT flow diagram depicting the design of this study. b Pie chart summarizing the outcomes of 179 patients treated with teclistamab. c Pie chart summarizing the relapse patterns in 104 patients who progressed on teclistamab. d Overall survival from the time of relapse in the 104 patients who progressed after teclistamab. e Overall survival (left) and progression-free survival (right) from first salvage therapy in 67 teclistamab-treated patients who relapsed and received salvage therapy. Tick marks indicate censored data. NE denotes not estimable. f Overall response rate (≥partial response) in 67 teclistamab-treated patients who relapsed and received salvage therapy, stratified by type of salvage therapy received. g Progression-free survival from first salvage therapy in all patients who received salvage (left) and penta-exposed patients who received salvage (right), stratified by type of salvage therapy (top) and cellular therapies (i.e., CAR T-cell or T-cell engager therapy) or not (bottom). h Univariable Cox regression analysis of clinical factors associated with progression-free survival from first salvage therapy in penta-exposed patients who received salvage. Patients who received temporizing infusional chemotherapy without definitive salvage therapy were excluded. Among patients who received temporizing therapy, progression free survival was measured from the start of definitive salvage therapy. i Sankey diagram illustrating response trajectories in 19 teclistamab-treated patients who received prior BCMA-directed therapies.

The ORR among all treated patients was 64.8% (Supplementary Fig. 1a), with a 1-year OS of 69.9%, and median PFS of 9.8 months (Supplementary Fig. 1b), consistent with recently published large real-world cohorts [5–7]. Forty-two patients (23.5%) had died, including 31 from MM progression and 9 from teclistamab-related toxicity (Table 1). Nineteen patients (10.6%) discontinued teclistamab due to infections, including 8 (4.5%) who died as a result (Supplementary Table 1). At last follow-up, 64 patients (35.8%) remained in remission (30 on teclistamab, 22 off-treatment under observation, and 12 discontinued due to toxicity), while 104 (58.1%) had progressed (44 with biochemical progression, 23 with CRAB criteria, 37 with extramedullary disease (defined as soft tissue plasmacytomas not contiguous with bone). Among the 60 patients with clinical progression (CRAB or extramedullary relapse), 31 died of complications related to MM progression (13 with CRAB, 18 with extramedullary relapse; Fig. 1b, c, Table 1). Notably, oligo/nonsecretory disease (defined as M-spike <0.5 g/dL and involved free light chain <100 mg/L) was present in 3/23 (13%) patients with CRAB progression and 11/37 (29.7%) patients with extramedullary relapse (Fig. 1c). Of the 104 patients who progressed after teclistamab, median OS from relapse was 6.6 months (95% CI: 4.3–15.3; Fig. 1d); 73/104 (70%) were alive, 71 were eligible for salvage therapy, 67 had received treatment, and 4 had not yet started salvage therapy at last-follow-up.

Table 1.

Outcomes at teclistamab relapse or last follow-up.

Outcomes at teclistamab relapse or last follow-up (n = 179)	n (%)
Dead	42 (23.5)
MM progression	31 (17.3)
• CRAB progression	13 (7.3)
• Extramedullary progression	18 (10.1)
MM in remission	11 (6.1)
• Teclistamab toxicity	9 (5)
• Infection	8 (4.5)
• Neurotoxicity	1 (0.6)
• Death from causes unrelated to MM	2 (1.1)
• Recurrent small bowel obstruction	1 (0.6)
• Metastatic prostate cancer	1 (0.6)
Alive	137 (76.5)
Ongoing remission	64 (35.8)
• Ongoing teclistamab	30 (16.8)
• Teclistamab discontinued	34 (19)
• Observation	19 (10.6)
• Comorbidities (melanoma, fracture x 2)	3 (1.7)
• Toxicity (weakness, infection x 11)	12 (6.7)
Relapsed	73 (40.8)
Stratified by type of relapse
• Biochemical	51 (28.5)
• CRAB progression	10 (5.6)
• Extramedullary progression	12 (6.7)
Stratified by salvage therapy status
• Salvage therapy started	67 (37.4)
• Salvage therapy pending	4 (2.2)
• Ineligible for salvage therapy (poor PS or lack of effective options)	2 (1.1)

MM multiple myeloma, PS performance status.

Baseline characteristics of these 67 patients who received salvage therapy are summarized in Supplementary Table 2. The median age was 66 years. Twenty-two (33%) patients had extramedullary disease (EMD), including 3 (5%) who also developed secondary plasma cell leukemia (PCL) at some point during their disease course. Nineteen of these patients had extramedullary progression at time of teclistamab relapse. Among patients with available data, 22% (15/39) met IMWG 2024 high-risk criteria at diagnosis and 30% (20/35) at relapse. Before starting teclistamab, median of 6 prior lines of therapy (IQR: 5–9) had been administered; all were triple-class exposed (immunomodulatory drug/IMiD, proteasome inhibitor/PI, and CD38 mAb), and 85.1% were penta-exposed (2 IMiD, ≥2 PIs, and CD38 mAb). Nineteen (28.4%) patients had received prior BCMA-directed therapies. At teclistamab initiation, 31 patients (46.3%) had cytopenias, defined as hemoglobin < 10 g/dL or platelets < 75,000/µL, or high disease burden, defined as or ≥ 60% bone marrow plasma cells, and 57 (85.1%) met IMWG frailty criteria. Cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) occurred in 33 (49.3%) and 5 (7.5%) patients, respectively; all events were grade ≤2. Baseline characteristics were largely similar between cohorts, except that commercially treated patients were older (median 69 vs. 64 years; P = 0.0043), more likely to have received prior BCMA-directed therapy (50% vs. 6.1%; P < 0.001), and more frequently met IMWG frailty criteria (94.1% vs. 75.8%; P = 0.027).

A swimmer’s plot detailing the clinical trajectories of the 67 patients treated with salvage therapy is provided in Supplementary Fig. 2. The 1-year OS and PFS from initiation of first salvage were 57% (95% CI: 44.9–72.4) and 17.5% (95% CI: 10.2–30.2), and median OS and PFS were 14.6 (95% CI: 8.3–NE) and 2.8 (95% CI: 1.9–4.3) months, respectively (Fig. 1e). Salvage therapies were grouped as follows and detailed in Supplementary Table 3(1): high-dose melphalan and autologous stem cell rescue (ASCT; n = 5) (2), talquetamab (n = 15) (3), BCMA-directed therapies (n = 4) (4), elotuzumab- or selinexor-based regimens (n = 16) (5), clinical trials (n = 10), and (6) others (n = 17). Twenty-two (32.8%) patients received temporizing therapy, defined as short-term treatment administered between teclistamab discontinuation and intiation of definitive salvage therapy; in the context of CAR T-cell therapy, this is referred to as “holding” (prior to T-cell apheresis) or “bridging” (after T-cell apheresis, prior to lymphodepletion) therapy. The ORR was 38%, with specific ORRs by regimen group shown in Fig. 1f. There was a nonsignificant trend toward higher ORR in patients receiving CART or TCE-based salvage therapy compared to others (55% vs. 32%; P = 0.13). Temporizing therapy was administered in 6/22 (27%) patients treated with CART or TCE (holding/bridging therapy in 4/4 CART patients and infusional chemotherapy in 6/22 TCE patients) compared with infusional chemotherapy in 10/44 (23%) patients in the non-CART/TCE group. PFS curves were calculated from the start of definitive salvage; six patients who received only temporizing therapy and died before initiation of definitive salvage were excluded (n = 61, Fig. 1g). There was a trend toward longer median PFS in patients who received CART or TCE (4.4 months) compared to others (2.1 months; P = 0.14). Given that most patients treated with teclistamab after ≥4 prior lines are also penta-exposed, Cox regression was used to identify clinical factors associated with PFS within this subgroup (Fig. 1h). Achieving a partial response or better (≥PR) to salvage therapy was associated with reduced risk of progression or death (HR: 0.1; P < 0.001), as was receiving CART or TCE therapy (HR: 0.5; P = 0.073).

Finally, 19 patients received ≥1 prior BCMA-directed therapies before teclistamab (CART: 6, belantamab: 13), including 8 who received both. We tracked reasons for discontinuation, treatment duration (i.e., belantamab duration or duration of CART response), interval between therapies, and sequential responses across BCMA therapies (Fig. 1i, Supplementary Table 4). Three patients (16%) discontinued their first BCMA therapy (all ADC) due to toxicity, while the remaining 16 progressed. The ORRs were 63% for both the first and second prior BCMA-directed therapies, and 64% for teclistamab. There was no significant correlation between response to prior BCMA-directed therapy and subsequent response to teclistamab (Spearman rho/ρ: −0.1; P = 0.69; Supplementary Fig. 3a). Similarly, no correlation was observed between the interval from prior BCMA-directed therapy and response to teclistamab (ρ: −0.25; P = 0.3; Supplementary Fig. 3b). Only four patients received BCMA-dt after teclistamab relapse (belantamab after failed ide-cel manufacture: 1, ide-cel: 2, cilta-cel: 1), all with holding and/or bridging therapy. The ORR was 100%, with a median PFS was 8.9 months (Fig. 1g).

This study is limited by its retrospective design. Selection bias may have influenced treatment choices and outcomes, particularly among patients who received CART or enrolled in clinical trials, both of which require adequate fitness and sufficiently controlled disease to allow for apheresis or trial participation. The small size and heterogeneity of the subgroups further limited statistical power to detect differences between salvage strategies. Nonetheless, in this study, relapse after teclistamab was common and associated with poor OS, highlighting the urgent need for effective post-relapse options. Notably, 23% of patients with clinical progression had oligo/nonsecretory disease, emphasizing the importance of comprehensive monitoring beyond routine serum markers. Sensitive imaging modalities (e.g., PET/CT, CT skeletal survey, or whole body MRI) and bone marrow evaluation should be considered in patients with bone pain or unexplained symptoms, or clinical suspicion for extramedullary disease. While responses to salvage therapy were generally modest, exploratory signals suggest that cellular immunotherapies (e.g., CART and TCEs) may be associated with longer PFS in select patients. These findings should be interpreted with caution and validated in larger cohorts, but they underscore the need for broader access to novel therapies and prospective studies to define optimal sequencing after teclistamab failure.

Author contributions

K.L.Y. and S.S.A. conceptualized the project. K.L.Y. and M.H. collected the clinical data and drafted the manuscript. L.P., F.S., H.P., Z.H., S.K., A.J.W., D.T.V., E.A.S., J.A.F., M.R., A.D.C., A.L.G., S.S.A., provided detailed patient data, feedback, and edited the manuscript.

Competing interests

D.T.V. has received research funding from Takeda and Active Biotech and consulting fees from Takeda, Karyopharm, GSK, Genentech, and Sanofi. E.A.S. declares research funding from AbbVie and K36; consultancy for Amgen, BMS, Celgene, and Janssen. M.R. has patents related to CD19 CAR T-cells; consults for GLG, Guidepoint, AbClon, Acera, Vittoria Bio; research funding received from AbClon, Oxford NanoImaging, Vittoria Biotherapeutics, CURIOX, and Beckman Coulter; scientific founder of Vittoria Biotherapeutics. A.D.C. has received research support from Novartis, GSK, Genentech, and Janssen; consulting fees from Janssen, BMS, GSK, Genentech, Legend, Sanofi, Pfizer, AbbVie, Regeneron, Moderna, AstraZeneca, iTeos, Prothena, Kite, Novartis, Ichnos; and has patents related to CAR T cells. A.L.G. declares research support from Johnson & Johnson, Novartis, Tmunity, and CRISPR Therapeutics; consultancies/honoraria from Johnson & Johnson, Novartis, BMS, Regeneron, AbbVie, AstraZeneca, Smart Immune, and Gracell Bio; and DSMB membership for Janssen. S.S.A. discloses research funding from Johnson & Johnson and consulting from Johnson & Johnson and Sanofi. The other authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41408-025-01408-4.