Clone wars: Evidence of clonal stability in a longitudinal prospective cohort of cancer survivors with serial NGS analysis Free

Introduction: Hematopoietic stem and progenitor cells with preleukemic mutations (mutⁿs), collectively termed clonal hematopoiesis (CHIP) serve as the cellular origin of myeloid neoplasms (MN). While risk of MN development varies, its natural history and transformation potential remain poorly defined. Additionally, uncertainty about progression can cause significant anxiety for patients (Tal Sella et al., Blood Advances, 2022). We present findings from our ongoing 10-year CHIP protocol using serial NGS to monitor clonal evolution and MN in cancer survivors.

Methods: Within this prospective cohort, we conducted a retrospective analysis of 210 participants who underwent serial NGS from March 2020 to June 2024. Mutⁿs were classified as CHIP+ if they had a variant allele frequency (VAF) ≥2% (or ≥1% for IDH1, IDH2, and JAK2). Clonal evolution, categorized as stability, regression, or progression was evaluated by annualized VAF changes, adjusted for age-related increases (e.g. DNMT3A increases by ~2%/yr based on published data). Mutⁿs were classified as slow (≤5%/yr: DNMT3A, TET2, ASXL1), intermediate (5–10%/yr: TP53, PPM1D, IDH1/2, KRAS, NRAS, SF3B1), or fast-growing (>10%/yr: JAK2, SRSF2). Growth kinetics for CBL, BCORL1, NF1, and GNAS remain undefined. Biostatistics comparisons used Pearson's Chi-squared, Fisher's exact, and Wilcoxon rank-sum tests.

Results: Of 210 patients, 42 (20%) had at least one CHIP+ mutⁿ. Median age was higher in CHIP+ vs. CHIP– patients (66 vs. 59 years, p<0.001). CHIP positivity was not significantly associated with sex, race, family history, smoking, or alcohol use. No significant differences were observed in prior cancer types (e.g., breast, head and neck) or treatment (chemotherapy, radiation, surgery). Medical histories of cancer, autoimmune disease, and baseline blood counts were also similar. Out of 487 samples, 56 had detectable somatic mutⁿs. The most frequently mutⁿ were DNMT3A (n=27), PPM1D (n=10), and TET2 (n=9). By expected growth kinetics, 37 (66%) mutⁿ were slow-growing, 12 (21.4%) intermediate, 2 (3.5%) fast-growing, and 4 (7.14%) of unknown kinetic category.

Only one patient showed clonal progression from their first sample after adjusting for age-related VAF growth, involving DNMT3A. Beyond VAF-kinetics–based progression, 11 patients who had no detectable mutⁿ initially, developed new mutⁿ in subsequent years with DNMT3A (n=7) being the most common acquisition followed by PPM1D (n=2). Some of these later became undetectable, suggesting transient clones. In total, 12 patients (1 by VAF kinetics, 11 by new mutⁿ acquisition) clonally progressed over 5 years of this study (5.7%).

Six patients regressed by year 2, with gene mutⁿ VAFs becoming undetectable. These included mutⁿ in PPM1D (n=2), CBL (n=2), TP53 (n=1), and DNMT3A (n=1). Three more patients regressed in year 3 (JAK2, BCORL1, PPM1D), and 1 in year 4 (DNMT3A). In all, 10 patients showed regression over 5 years (17.8%). Notably, 5 of 10 were referred to preventive cardiology and received interventions for cardiac comorbidities. Overall, ~94% of the cohort demonstrated clonal stability or regression.

In our cohort, annual VAF growth rates varied from those reported in the literature. Slow-growing mutⁿs such as DNMT3A and TET2 had median rates of 0.83% and 2.11%, aligning with historical rates ≤5%. Intermediate mutⁿs like TP53 and PPM1D had median rates of 1.64% and –0.625%, below the expected 5–10% annual rise. JAK2, a fast-growing mutⁿ, also showed a lower rate of 0.19%. Ongoing follow-up may reveal future changes in VAF kinetics.

Conclusion: Despite being a high-risk population due to prior solid malignancy and exposure to chemotherapy or radiation, most patients in this five-year study showed clonal stability. Even among those who acquired new mutⁿs and “progressed”, VAF kinetics remained stable post-detection. CHIP+ clones were transient in some cases, reverting to CHIP- status over time. CHIP's inflammatory link to cardiovascular disease suggests preventive cardiology may impact clonal behavior, though further study is needed. As CHIP detection can elicit significant anxiety for cancer pts, our findings provide reassurance that clonal stability is the predominant pattern in longitudinal CHIP tracking. This cohort will be the control for future comparisons with higher-risk groups; such as those exposed to radioligand or cellular/immunotherapies and VA populations in the next phase of our multi-year CHIP protocol enrollment.