MRD and Biomarkers

Category: MRD and Biomarkers

Subclonal and clonal progression of previously characterized mutations and variants of unknown significance (VUS) across a cohort of plasma cell dyscrasia patients

(PA-171) Subclonal and Clonal Progression of Previously Characterized Mutations and Variants of Unknown Significance (VUS) Across a Cohort of Plasma Cell Dyscrasia Patients

Friday, September 19, 2025

Kristin Ezell, MD

Resident Physician
University of Connecticut

Introduction: Plasma cell dyscrasias (PCD) are genetically diverse, and key oncogenic drivers have been identified. In multiple myeloma (MM), subclonal RAS mutations, in 50–61% of cases, are known to promote progression. Alterations in TP53, DIS3, FAM46C, and NF-κB regulators add complexity, and IGH translocations remain central to risk stratification. Yet, few markers predict MGUS or smoldering MM progression. Current research targets mutational burden, cytogenetic shifts, and clonal diversity. Variant allele frequency (VAF) serves as a clonality proxy (Boscolo Bielo et al 2023), with higher VAF indicating founder clones and lower VAF (commonly < 10%) suggesting subclones. Subclonal diversity may signal progression and resistance, as clonal shifts under treatment correlate with poor prognosis. Understanding clonal evolution is key to defining progression in PCDs.

Methods:   We analyzed genomic variants from 252 PCD samples using the Tempus xT® 648-gene NGS panel. Samples included bone marrow (n=228), tissue (n=5), and blood (n=19), collected from Jan 2020 to May 2025. Diagnoses confirmed by chart review included MGUS (n=84), SMM (n=23), MM (n=102), AL amyloidosis (n=14), and other PCDs (n=29). NGS reports were digitized as sample-specific CSVs and analyzed using Python in Jupyter (v7.3.2) and Excel (v16.96.1). COSMIC (v102) web platform identified cancer-associated single nucleotide variants (SNVs).

Results:
We identified 332 VUS and 81 oncogenic variants in 252 PCD samples. The malignant cohort had higher oncogenic mutations per sample than the premalignant cohort (Welch’s t-test, p< 0.0001), though VUS counts did not differ significantly (p=0.2276).

Among VUS, most frequently observed mutations were in KMT2C, ARID1B, NOTCH1, ZFHX3, and ATM. Notably, all ARID1B, NOTCH1, and ATM SNVs were clonal (VAF >10%). Of 19 ARID1B SNVs, 12 have not been previously associated with any cancer; only one (c.1016_1021dup) was previously linked to hematologic malignancy. Seven of 12 NOTCH1 and 3 of 11 ATM SNVs were not previously reported. Among clonal VUS, the malignant cohort harbored 217 unique genes (574 SNVs; 399 unique). Mean SNVs per gene were higher in malignant samples (1.44 vs. 1.09, p< 0.0001).

Most frequent oncogenic mutations occurred in DNMT3A, TP53, MYD88, TET2, and CXCR4, with most DNMT3A, MYD88, and CXCR4 variants being subclonal. Mean VAF was higher in VUS versus oncogenic groups across malignancy status (One-way ANOVA p< 0.0001 then Tukey’s test with Benjamini-Hochberg adjusted p-values). VAF did not differ between oncogenic premalignant and malignant samples (p=0.9848).

Conclusions:   Our analysis reveals distinct clonal patterns and mutational profiles across PCDs. Subclonal oncogenic mutations were more common in malignant states, while VUS were often clonal. Further studies on longitudinal clonal dynamics are needed.