Paper Details

ABSTRACT

Background: Most MDS patients harbor detectable mutations with prognostic and therapeutic implications. However, a subset lacks any detectable mutation by standard-of-care testing, representing a diagnostically challenging and understudied population. Here, we characterize the clinical features and outcomes of patients with MDS and no detectable mutations by panel-based NGS, and apply whole-genome sequencing (WGS) or longitudinal testing subsets to uncover cryptic alterations with potential implications for diagnosis and treatment.

Methods: We retrospectively analyzed data from patients diagnosed with MDS between 2017 and 2022 who were treated at The University of Texas MD Anderson Cancer Center (MDACC) or Institut de Recerca contra la Leucèmia Josep Carreras. All patients underwent sequencing at the time of diagnosis using targeted, panel-based next generation sequencing (NGS), assessing mutations in either 81 or 50 genes. A subset of MDACC samples without detectable mutations via panel-based sequencing further underwent WGS (Tempus xH) or longitudinal testing with an NGS panel. 

Results: Among 1178 patients with MDS, 7% (n=89) did not have any mutations detected by targeted-panel testing, hereafter referred to as the panel-negative (PN) cohort. Among these, 46% (n=41) had normal karyotype. Patients in the PN cohort were younger (median age 67 vs 70 years old) and showed less cytopenia compared to those with detectable mutations by panel testing. In the PN cohort, 19 patients (21%) had MDS with increased blasts. Megakaryocytic dysplasia was the most frequent finding in the bone marrow (91%), and 55% of patients had trilineage dysplasia. Abnormal flow cytometry findings (such as altered immunophenotype in myeloblasts and myelomonocytic cells) were found in 82% of patients with available flow cytometry data.

According to the IPSS-R score, 62% of PN patients were very low or low risk, 16% intermediate and 21% high or very-high risk. When the IPSS-Molecular (IPSS-M) score was applied, risk levels were downstaged in 58% of patients. With a median follow-up time of 42 months (95% CI 41-52 months), the median OS of patients in the PN cohort was 86 months, with 2- and 5-year OS rates of 77.7% and 61.3%, respectively. These outcomes were significantly better than patients who had detectable mutations by panel testing (2- and 5-year OS: 44.6% and 24.9%, respectively; p<0.001).

Twenty-nine samples in the PN cohort underwent WGS (16 with a diploid karyotype), which uncovered driver alterations in 66% (n=19). Notably, three patients harbored novel cryptic MECOM rearrangements with partner genes CD164, CDK6 and SPHK1. Additionally, one patient with normal karyotype was found to have a cryptic NUP98::PRRX2 rearrangement. Four patients had mutations in the UBA1 gene (responsible for VEXAS syndrome), one patient had a mutation in the canonical hotspot location p.M41L and three patients presented mutations in non-canonical but previously described UBA1 codons. Two patients with a prior diagnosis of MDS with ring sideroblasts were found to have ALAS2 mutations p.R191Ter and p.R191Q, possibly due to diagnosis of congenital sideroblastic anemia. In 10 patients, no clear detectable driver genetic abnormality was found. All these patients had variants in genes potentially involved in leukemogenesis with unproven roles in myeloid disease, including: epigenetic regulators (BRD3, BRD4, CHD4), DNA damage response and repair (POLQ, PRKDC, FANCI), or tumor suppressor genes (VHL, EXT1).

Among all 89 patients in the PN cohort, 18 patients had available bone marrow NGS data from at least 2 timepoints. Ten (56%) of these patients persisted without detectable mutations by panel NGS at any of the timepoints. The remaining eight patients (44%) in the PN cohort with longitudinal data exhibited an emergent, detectable mutation at a later timepoint. The most common mutations detected at a second timepoint were TP53 (n=4 patients), RUNX1 (n=2), and IDH1 (n=2).

Conclusions: We found that a substantial fraction of patients with MDS (7%) lacked any driver alterations detected by targeted, panel-based NGS. However, performing WGS on 29 of these patients, revealed 66% (n=19) had clinical mutations with either diagnostic or prognostic significance. WGS and longitudinal testing have the potential to uncover actionable alterations that might otherwise be missed by targeted panel-based sequencing assays, especially in patients with dysplasia and no detectable clonality.