AUTHORS
Sameeran Das, Feng Gao, MD, PhD, MPH, MS2, John F. DiPersio, MD3, Geoffrey L. Uy, MD4, Samuel Urrutia, MD, MS5, Meagan A. Jacoby, MD, PhD6, David H Spencer, MD, PhD7, Eric Duncavage, MD8, Keith Stockerl-Goldstein, MD9, Ramzi Abboud, MD10
AFFILIATIONS
1 Washington University School, Saint Louis, Missouri, United States
2 Washington University School, Saint Louis, Missouri, United States
3 Washington University School, Saint Louis, Missouri, United States
4 Washington University School, Saint Louis, Missouri, United States
5 Washington University School, Saint Louis, Missouri, United States
6 Washington University School, Saint Louis, Missouri, United States
7 Washington University School, Saint Louis, Missouri, United States
8 Washington University School, Saint Louis, Missouri, United States
9 Washington University School, Saint Louis, Missouri, United States
10 Washington University School, Saint Louis, Missouri, United States
ABSTRACT
Introduction:
Patients with TP53-mutated MDS and AML face dismal outcomes even after allogeneic hematopoietic stem cell transplant (alloHCT). The role of molecular disease burden-driven assessment remains unclear. Prior next-generation sequencing studies have been bound to a limit of detection (LOD) of 2% variant allele frequency (VAF). Myeloseq, Washington University’s clinical NGS panel, now achieves an LOD of 0.1% for previously detected variants. With this ultra-sensitive assay, we studied the prognostic value of TP53 VAF pre- and post-transplant in a single-institution cohort of MDS and AML patients.
Methods:
We retrospectively analyzed 91 adults with TP53-mutated MDS or AML who underwent alloHCT at Siteman Cancer Center between 2018 and 2026 and had a diagnostic NGS. Pre-HCT Myeloseq was available in 90 patients; post-HCT Myeloseq was available in 56 at day +30 and in 61 at day +100. VAF was modeled as a continuous variable as it satisfied linearity; hence, hazard ratios (HR) reflected the risk of an outcome per 1% change in VAF. Univariate and multivariable Cox regression models assessed overall survival (OS) and relapse-free survival (RFS). Fine-gray models treated non-relapse mortality (NRM) as a competing event to relapse. Landmark analyses for post-HCT VAF excluded patients with follow-up shorter than the landmark timepoints.
Results:
The median age of the cohort was 61 years; 65.9% had AML and 34.1% had MDS; and 52.4% had biallelic/multi-hit TP53 status. Donor types included matched unrelated (42.9%), haploidentical (40.7%), and matched related (16.5%). Most patients received myeloablative conditioning (75.8%). Median VAF at diagnosis was 41% and median pre-HCT VAF was 11.5%. Myeloseq’s LOD was 0.1% in 25.3% of patients, 0.25% in 22.0%, and 2% in 52.7%, the latter representing transplants before the utilization of the high-sensitivity Myeloseq. Median follow-up time was 248 days (range 6-2245); 57% relapsed.
Pre-HCT TP53 VAF was not prognostic as a continuous variable (OS HR 1.007, p=0.107; RFS HR 1.007, p=0.092). Diagnostic VAF was significant univariately (OS p=0.045; RFS p=0.022) but lost significance after adjustment for donor type (OS p=0.151; RFS p=0.1025), suggesting a confounding effect of transplant characteristics. Neither multi-hit status (OS p=0.810) nor conditioning intensity (OS p =0.884) was significant for outcome.
In contrast, post-HCT TP53 VAF demonstrated highly significant prognostic value at day +30 and day +100 landmarks. Day +30 VAF predicted OS (HR 1.041, p=0.0002) and RFS (HR 1.034, p=0.0006) and remained significant after adjusting for donor type (OS HR = 1.040, 95% CI 1.017-1.064, p=0.0006; RFS HR 1.032, 95% CI 1.012-1.053, p=0.0019). Day +100 VAF was the strongest predictor for OS (HR 1.039, p<0.0001) and RFS (HR 1.035, p<0.0001) and was associated with relapse in competing risks (HR 1.030, 95% CI 1.019-1.041, p<0.001), while day +30 VAF was not (HR 1.000, 95% CI 0.980-1.021, p=0.962).
Conclusions:
Post-transplant Myeloseq independently predicts outcomes in TP53-mutated MDS and AML. Quantitative, continuous VAF also outperforms binary MRD detection status, supporting the role of ultra-sensitive reporting instead of threshold-based reporting. Pre-HCT disease burden measured using a more sensitive assay than prior studies did not predict outcome, arguing against deferring transplant to achieve a deeper molecular remission. Our findings support post-transplant MRD as a guide for intervention strategies in TP53-mutated MDS and AML.
DISCLOSURES
Sameeran Das: No relevant conflicts of interest to declare.