Background: Clonal hematopoiesis of indeterminate potential (CHIP) has been linked to impaired hematopoietic recovery and adverse outcomes in patients with malignancy, though prior studies in anti-BCMA CAR T-cell therapy have not demonstrated strong associations between CHIP status and survival. However, emerging risk models such as the Clonal Cytopenia Risk Score (CCRS, Blood 2024) offer a refined approach by integrating CHIP mutations and cytopenias to stratify hematopoietic vulnerability. In this study, we applied CCRS in a cohort of patients with relapsed/refractory multiple myeloma (RRMM) treated with idecabtagene vicleucel (ide-cel) or ciltacabtagene autoleucel (cilta-cel) to evaluate its association with inflammatory markers, hematologic recovery, and survival outcomes—testing whether CCRS provides additive prognostic value beyond CHIP status, systemic inflammation, and tumor-intrinsic risk features.

Methods: We sequenced and analyzed 113 paired samples collected within 30 days before lymphodepletion and 90 days post CAR-T infusion from patients treated with ide-cel (n=80) or cilta-cel (n=33) at our institution. CHIP was assessed using targeted next-generation sequencing (Archer Myeloid Panel, 3000x median depth) on peripheral blood, with a variant allele frequency (VAF) threshold of ≥1%. CHIP risk was stratified by CCRS. Outcomes included inflammatory markers, hematologic recovery, progression-free survival (PFS), and overall survival (OS). Survival was analyzed using Kaplan–Meier and multivariable Cox regression models adjusted for ferritin (>400 ng/mL), MM high-risk genomic features (del17p, t(4;14), t(14;16)), and extramedullary disease (EMD).

Results: The median age of the cohort was 66 years (range 38-84), with a slight male predominance (67 men vs. 46 women). Median follow-up was 17.7 months (range 0.4-40.8), during which 38 patients died and 2 developed therapy related myeloid malignancy. At baseline, CHIP was detected in 41 of 113 patients (36%), most commonly involving DNMT3A (n=18), TET2 (n=10), and PPM1D (n=5). Post-infusion, CHIP was detected in 27 patients, with PPM1D emerging as the second most frequent mutation. A total of 22 patients had overlapping mutations at both timepoints. Most variants were stable post CAR T infusion, with median VAF changes of < 1%. Baseline CHIP status was not significantly associated with age, MM genomic risk features, EMD, ferritin, CRP, PFS or OS (all p>0.05).

Using the CCRS, 80 patients were classified as low risk, 31 as intermediate, and 2 as high risk; intermediate and high groups were combined for analysis (CCRS int-high). CCRS int–high patients had significantly elevated baseline ferritin (p=0.009) and CRP (p=0.035) compared to the CCRS-low group, although peak values during follow-up did not differ (p=0.235 and p=0.98, respectively).

CCRS int–high patients exhibited persistently impaired hematologic recovery. Hemoglobin levels were significantly lower from baseline through Day 90 (all p ≤ 0.001), and differences remained at Day 180 (p ≈ 0.017) and Day 360 (p ≈ 0.00015). Similarly, platelet counts were markedly lower at baseline and remained significantly depressed across all follow-up timepoints (all p≤0.001). ANC was also lower at baseline in the CCRS int-high group (p=0.026), but differences attenuated by Day 30 and were no longer significant thereafter (p > 0.10).

In univariable analyses, CCRS int–high status was associated with significantly worse outcomes, including OS (HR 2.24, 95% CI: 1.16–4.29; p=0.016) and PFS (HR 2.23, 95% CI: 1.35–3.71; p=0.0019). In multivariable models adjusting for high ferritin, MM high-risk genomic features, and EMD, CCRS remained the only variable consistently associated with inferior PFS (HR 2.00, 95% CI: 1.20–3.39; p=0.008) and OS (HR 2.00, 95% CI: 1.04–3.92; p=0.008). In contrast, ferritin, genomic risk, and EMD each trended toward worse outcomes but did not reach statistical significance (all p > 0.05).

Conclusion: CCRS robustly stratifies clinical outcomes, identifying patients at significantly increased risk for prolonged cytopenias, disease progression, and death—independent of systemic inflammation and tumor-intrinsic high-risk features. These findings suggest that CCRS captures a distinct axis of hematopoietic vulnerability not reflected by CHIP status alone. Integrating CHIP-derived risk metrics into post–CAR-T surveillance may enhance risk-adapted monitoring and inform supportive care strategies for long-term survivorship.

This content is only available as a PDF.
2025
Sign in via your Institution