Diagnosis and prediction of precursor states in multiple myeloma using cell-free chromatin

Abstract Multiple myeloma (MM) is a hematological malignancy that, in many cases, is preceded by monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM), pre-MM conditions with a lifelong risk of progression to symptomatic MM. Accurate diagnosis, prognostication, and adequate progress monitoring for these conditions require repeated invasive bone marrow biopsies. Additionally, current prognostic models for identifying higher-risk MGUS and SMM patients misclassify 20-30% of patients. Here, we apply chromatin immunoprecipitation of cell-free chromatin (cfChIP-seq) from peripheral blood, which provides a genome-wide map of active promoters in the cells contributing to the circulating DNA, on a cohort of individuals with MGUS (n=65), SMM (n=27), patients with overt MM (n=47), and healthy controls (n=82). We identified an increased representation of specific regions containing promoters specific to B-cells, plasma cells, and erythroblasts in MM blood samples, including promoters of clinically relevant genes such as IGLL5, FCRL5, GYPA, HBD, and POU2AF1. We used these regions to generate an MM score comprising two distinct subscores: a plasma cell-enriched subscore (MM-P), reflecting plasma cell activity, and an erythroblast-enriched subscore (MM-E), indicating erythroid lineage activity. The MM score showed significant differentiation between healthy individuals, pre-MM stages, and overt MM (p-value 0.01). Importantly, the MM-P subscore exhibited strong predictive capability, significantly identifying pre-MM patients at a higher risk of biochemical progression within two years (AUC = 0.71) compared to current clinical prognostic models (AUC = 0.61). This represents a non-invasive alternative to current diagnostic practices. Moreover, integration of the MM-P subscore with clinical risk factors (free light chain ratio 20, serum M-protein 2 g/dL) into a combined MM-P/20/2 score further improved prognostic accuracy, achieving an AUC of 0.81 for biochemical progression and 0.76 for clinical progression. Kaplan-Meier analyses demonstrated robust stratification of high-risk versus low-risk individuals for both biochemical and clinical progression (log-rank p 0.0008, and 0.008, respectively). Our study highlights the clinical potential of cfChIP-seq to non-invasively differentiate plasma cell disease stages, improve risk stratification accuracy, and monitor disease progression, providing a valuable tool to complement existing prognostic assessments and reduce reliance on invasive bone marrow biopsies. Moreover, our results highlight increased erythropoiesis in pre-MM stages, which can be of potential clinical and diagnostic value. This approach also opens a practical path toward personalized medicine in the management of plasma cell disorders.