AUTHORS
Ehsan Malek, MD, James Driscoll II, MD, PhD2, Brian C. Betts, MD3, Marco Davila, MD, PhD4, Paul Wallace, PhD5, Teresa Hahn, Phd6, Han Yu, PhD, MBBS7
AFFILIATIONS
1 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
2 University Hospitals Case Medical Center, Cleveland, Ohio, United States
3 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
4 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
5 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
6 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
7 Roswell Park Comprehensive Cancer Center, Buffalo, New York, United States
ABSTRACT
Background: The aging immune system undergoes progressive remodeling marked by thymic involution, contraction of naïve lymphocyte reserves, and loss of adaptive plasticity. In multiple myeloma (MM), disease biology and prior therapies further accelerate this process, yet chronological age remains a remarkably poor surrogate for host functional immune fitness. While high-dimensional "immune-age" metrics exist, most rely on complex, opaque machine learning algorithms that lack clear biological interpretability, require extensive specialized profiling panels, and lack anchoring to healthy control cohorts. To overcome these barriers to clinical translation, we developed a transparent, healthy control-referenced immune aging framework based on the canonical triad of naïve lymphocyte subsets and evaluated its prognostic value for long-term survival independent of chronological age.
Methods: We analyzed prospectively collected immunophenotyping and clinical outcome data from patients with MM post-induction and prior to autologous hematopoietic cell transplantation (HCT) alongside a reference cohort of healthy cell donors. Using cubic spline regression on healthy control data, we modeled expected age-associated -transformed absolute counts for naïve CD4+ T cells (), naïve CD8+ T cells (), and naïve B cells (). Standardized -residuals (observed minus expected, divided by healthy control residual SD) quantified relative immune deviation. Residuals were combined into a composite ImmuneAgingScore using predefined, biologically informed weights (0.50 CD4, 0.25 CD8, 0.25 CD19). Cox proportional hazards models evaluated associations between ImmuneAgingScore and overall survival (OS) or progression-free survival (PFS) before and after adjusting for chronological age. Primary survival analyses evaluated patients after excluding four with uncertain transplant dates.
Results: The primary baseline cohort comprised 220 patients with MM (median age 61.0 years, IQR 54.0–67.0; 53.2% male) and 152 healthy donors (median age 49.0 years, IQR 44.0–56.0; 50.7% male). Patients with MM demonstrated profound, highly significant depletion across all three absolute naïve compartments compared with healthy controls: naïve CD4+ T cells (median 28.85 vs 195.74 cells/L; ), naïve CD8+ T cells (median 36.32 vs 99.16 cells/L; ), and naïve CD19+ B cells (median 27.86 vs 152.63 cells/L; ). Compartment-specific dynamics revealed that naïve CD4+ T-cell depletion accelerated sharply in MM at ages >60 years, whereas naïve CD8+ T-cell deficits remained consistently depressed across the full age spectrum. Over a median OS follow-up of 119 months (; 103 deaths, 163 progression events), increasing continuous ImmuneAgingScore significantly predicted inferior OS (hazard ratio [HR] 1.17 per unit increase; 95% CI, 1.01–1.36; ). This prognostic association persisted after adjusting for chronological age (HR 1.24; 95% CI, 0.96–1.59; ), confirming non-redundancy with calendar age. Quartile analysis confirmed that the highest immune-aging quartile (Q4) experienced significantly worse OS compared to the youngest (Q1) (HR 1.78; 95% CI, 1.01–3.14; ). ImmuneAgingScore demonstrated a non-significant trend toward inferior PFS (HR 1.08; 95% CI, 0.96–1.21; ).
Conclusions: A healthy control-referenced ImmuneAgingScore derived from three routinely quantified naïve lymphocyte subsets reveals profound age-adjusted immune depletion in MM and captures independent prognostic information for overall survival. By replacing opaque computational models with a transparent, biologically grounded framework, this metric provides a practical, scalable biomarker of host immune fitness to refine risk stratification in the era of immune-based myeloma therapies.
DISCLOSURES
Ehsan Malek, MD: Regeneron - Honoraria; J&J - Research Funding; Medpacto Inc. - Research Funding; Legend - Research Funding; BMS - Consultancy (Includes expert testimony); Regeneron - Consultancy (Includes expert testimony); J&J - Consultancy (Includes expert testimony); Legend - Consultancy (Includes expert testimony)