AUTHORS
Joshua Kassner, MD, Theresia Akhlaghi, MD2, David Jayabalan, MSc3, Christian A Gordillo, MSc4, Sarah Wesley, MD5, Victor Zepeda, MD6, Roger Pearse, MD7, Ruben Niesvizky, MD8, Ran Reshef, MD, MSc9, Mateo Mejia Saldarriaga, Assistant Professor10, Mark Bustoros, MD11
AFFILIATIONS
1 Weill Cornell Medical College, New York, New York, United States
2 Weill Cornell Medical College, New York, New York, United States
3 Weill Cornell Medical College, New York, New York, United States
4 Columbia Herbert Irving Comprehensive Cancer Center, New York, New York, United States
5 Columbia University Medical Center, New York, New York, United States
6 Weill Cornell Medical College, New York, New York, United States
7 Weill Cornell Medical College, New York, New York, United States
8 Weill Cornell Medical College, New York, New York, United States
9 Columbia Herbert Irving Comprehensive Cancer Center, New York, New York, United States
10 Weill Cornell Medical College, New York, New York, United States
11 Weill Cornell Medical College, New York, New York, United States
ABSTRACT
Background. Most patients awaiting BCMA-directed CAR T-cell therapy receive bridging therapy, yet its objectives are not well defined. ICANS and especially delayed neurologic events such as cranial nerve palsy and parkinsonism are a significant source of morbidity after BCMA CAR T-cell therapy, particularly cilta-cel, and no pre-infusion variable reliably predicts them. We aimed to evaluate whether response to bridging therapy predicts subsequent toxicity, depth of response and survival.
Methods. We analyzed patients receiving ide-cel or cilta-cel for relapsed/refractory myeloma at two centers. Of 106 patients, 2 with incomplete records and 7 with oligosecretory disease (not evaluable for bridging response) were excluded, leaving 97 evaluable patients. Best response to bridging was assessed by International Myeloma Working Group criteria. Cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) were graded by American Society of Transplant and Cellular Therapy (ASTCT) consensus criteria; delayed neurotoxicity comprised cranial nerve palsy, parkinsonism or other delayed neurologic events. Comparisons used Fisher’s exact test, logistic regression and Cox proportional hazards regression adjusted for age and prior lines (continuous), ECOG 0–3, and binary terms for non-paraskeletal extramedullary disease, high-risk cytogenetics, CAR-T product and peak ALC (median split, 1.66 K/µL).
Results. Median age was 66 years, patients had a median of 4 prior lines of therapy, extramedullary disease in 27% and high-risk cytogenetics in 34%; 84% received cilta-cel . Sixteen patients received no bridging; of the 81 bridged, 59 achieved at least stable disease (≥SD) and 22 progressed (PD). Median follow-up was 24.8 months
Outcomes were similar in the 23 patients achieving ≥PR and the 36 with stable disease: grade ≥2 ICANS 0% versus 3%, delayed neurotoxicity 9% versus 6%, CR to CAR-T 87% versus 91%, and progression-free survival did not differ (HR 0.89, p=0.83); no endpoint separated them (all p≥0.10). Both differed from the 22 patients progressing on bridging (grade ≥2 ICANS 23%, delayed neurotoxicity 32%, CR 59%). These groups were therefore combined as ≥SD, and all subsequent comparisons are ≥SD versus PD.
Any CRS was more frequent with progression during bridging therapy (73% versus 95%; OR 7.81, p=0.032), as was tocilizumab use (53% versus 82%; OR 4.06, p=0.021). CRS severity trended the same way: grade ≥2 in 20% versus 41% (OR 2.71, p=0.087), grade ≥3 in 2% versus 9% (p=0.18). No CRS endpoint was significant after adjustment (all adjusted OR ≤3.35, p≥0.11).
Neurologic complications were substantially more frequent with PD during bridging therapy. ICANS grade ≥2 occurred in 2% versus 23% (OR 17.1, p=0.005) and delayed neurotoxicity in 7% versus 32% (OR 6.42, p=0.007). Both remained significant after adjustment and in cilta-cel recipients alone (all p≤0.002). Peak ALC was similar (median 1.68 versus 1.98 K/µL, p=0.78).
Depth of response to CAR-T also differed between ≥SD and PD responses to bridging therapy. Best response was CR in 90% versus 59% (OR 0.17, p=0.004) and ≥VGPR in 95% versus 77% (OR 0.19, p=0.033), while overall response was preserved (97% versus 91%, p=0.30).
Progression on bridging also predicted survival. Median progression-free survival was 34.7 versus 12.2 months (HR 3.58, p<0.001; adjusted HR 2.63, p=0.028) as did overall survival (HR 7.40, p=0.003; adjusted HR 7.72, p=0.007). A ≥PR versus <PR dichotomy was null for every endpoint (ICANS grade ≥2 p=0.18, delayed neurotoxicity p=0.72, progression-free survival p=0.36), indicating disease control rather than depth of response defines the risk group.
Talquetamab-bridged patients (n=11) achieved ≥PR in 11/11 (100%) versus 12/70 (17%) of other bridged patients (p<0.0001), with no grade ≥2 CRS and no ICANS.
Conclusions. Failure to achieve disease control on bridging identified patients at 6- to 17-fold risk of high-grade neurotoxicity, 3.6-fold risk of progression and 7-fold risk of death after BCMA CAR-T. Grade ≥2 CRS was numerically higher with progression (41% versus 20%) but did not reach significance. PR or better conferred no PFS or OS benefit over stable disease, suggesting any degree of disease control appears sufficient. Limitations are the retrospective design and small event counts (6 grade ≥2 ICANS, 11 delayed neurotoxicity events) with need for prospective validation.
DISCLOSURES
Joshua Kassner, MD: No relevant conflicts of interest to declare.