The multi-center study analyzed 761 relapsed multiple myeloma patients receiving standard-of-care ciltacabtagene autoleucel (cilta-cel) across 15 U.S. centers from May 2022 to December 2024 to identify modifiable risk factors for Parkinsonism and non-relapse mortality. Non-response to bridging therapy emerged as a critical risk factor, associated with 10-fold increased risk of Parkinsonism (5% vs 0.5% in responders, p<0.05), with 95% of Parkinsonism cases occurring in bridging non-responders. Peak absolute lymphocyte count (ALC) ≥3000/μL demonstrated strong predictive value, identifying 68% of Parkinsonism patients and conferring 12.7-fold increased risk (95% CI: 5.1-34.8, p<0.001) compared to lower ALC levels. The overall Parkinsonism rate was 2.9% with median onset at 30 days, while one-year non-relapse mortality reached 9%, predominantly driven by infections (56%) and immune-mediated toxicities (22%), with bridging non-response, ECOG PS ≥2, high-risk cytogenetics, and age ≥70 years identified as independent NRM predictors. These findings support implementing effective tumor debulking strategies and using peak ALC ≥3000/μL as a biomarker for preemptive risk mitigation interventions.

Key Points

• Bridging Response Impact: Non-response to bridging therapy conferred 10-fold increased Parkinsonism risk (5% vs 0.5%, p<0.05) and 2.41-fold increased NRM risk (95% CI: 1.07-6.19, p=0.046); 95% of Parkinsonism cases occurred in bridging non-responders
• ALC as Predictive Biomarker: Peak ALC ≥3000/μL identified 68% of Parkinsonism patients with 12.7-fold increased risk (p<0.001); median peak ALC in Parkinsonism patients was 5.88 vs 1.17 x 10³/μL in non-affected patients (p<0.001)
• Parkinsonism Incidence and Timing: Overall Parkinsonism rate of 2.9% (22/761 patients) with median onset at 30 days (range: 17-214 days); 91% response rate and 68% CR rate maintained in affected patients post-CAR-T
• NRM Patterns: One-year NRM estimate of 9%, two-year estimate of 10%; infections accounted for 56% of NRM events (35/63), acute immune-mediated toxicities 22%, delayed immune-mediated toxicities 9.5%
• Independent NRM Risk Factors: Multivariable analysis identified ECOG PS ≥2 (OR 3.97, p<0.001), age ≥70 years (OR 2.65, p=0.003), high-risk cytogenetics (OR 2.39, p=0.007), and bridging non-response as independent predictors
• Overall Safety and Efficacy: Grade ≥3 CRS in 3%, grade ≥3 ICANS in 5%, IEC-HS in 5%; 92% overall response rate, 70% CR rate, 72% one-year PFS estimate with 10.1 months median follow-up
• Delayed Neurotoxicity Spectrum: 10% overall delayed neurotoxicity rate including 4.6% cranial nerve palsy, 2.9% Parkinsonism, 2.4% other delayed neurotoxicity events
• Intervention Threshold: ALC threshold of 2500-3000/μL can identify majority of at-risk patients; number needed to treat of 9 if intervention effectively reduces risk to baseline 1%

Conclusion:
This large real-world cohort study of 761 patients establishes non-response to bridging therapy and peak ALC ≥3000/μL as critical, potentially modifiable risk factors for Parkinsonism and non-relapse mortality following ciltacabtagene autoleucel therapy in relapsed multiple myeloma. The striking 10-fold increased Parkinsonism risk and 2.4-fold increased NRM risk associated with inadequate bridging response emphasizes the paramount importance of effective tumor debulking strategies prior to CAR-T infusion. Peak ALC ≥3000/μL emerges as a practical, measurable biomarker that identifies 68% of patients who will develop Parkinsonism with 12.7-fold increased risk, providing a clear threshold for implementing preemptive interventions with a number needed to treat of only 9 patients. These findings directly inform risk mitigation strategies through optimized bridging therapy selection, early ALC monitoring for intervention triggers, and targeted management of additional NRM risk factors including advanced age, poor performance status, and high-risk cytogenetics, ultimately supporting the development of the CITADEL observational study for early intervention protocols.

Authors:
Surbhi Sidana (contributed equally), Brett Reid (contributed equally), Danai Dima (contributed equally), Lauren C. Peres, Mahmoud Gaballa, Rahul Banerjee, Oren Pasvolsky, Aimaz Afrough, Christen Dillard, Christopher Ferreri, Shebli Atrash, Cindy Varga, Andrew Portuguese, Masooma Rana, Hitomi Hosoya, Lekha Mikkilineni, Vanna Hovanky, Saurabh Zanwar, Nilesh Kalariya, Damian Mikulski, Charlotte Wagner, Christopher R. Cahoon, Omar Castaneda Puglianini, Gabe De Avila, Christian Gordillo, Eli Zolotov, Jenny Bhurtel, Ariel Grajales-Cruz, Utkarsh Goel, Aishwarya Sannareddy, Jeries Kort, Rafaella Cassano, Shonali Midha, James Davis, Rebecca Gonzalez, Megan Herr, Zhuoer Xie, Hamza Hassan, Sneha Purvey, Marcus Geer, Kimberly Green, Fabiana Perna, Hien Liu, Taiga Nishihori, Jack Khouri, Shahzad Raza, Faiz Anwer, Susan Bal, Omar Nadeem, Ciara L. Freeman, Leyla Shune, Ran Reshef, Kenneth Shain, Melissa Alsina, Rachid Baz, Doug Sborov, Saurabh Dahiya, Frederick L. Locke, David Miklos, Peter Voorhees, Larry Anderson, Luciano Costa, Noa Biran, Shaji Kumar, Yi Lin (contributed equally), Krina K. Patel (contributed equally), and Doris Hansen (contributed equally).

ASH Abstract#1034