MDS without clonal marker: When depth outperforms breadth Free

Background:

Myelodysplastic neoplasms (MDS) are a heterogeneous group of clonal myeloid neoplasms. The interplay of mutations shapes disease phenotype, progression, and response to therapy, highlighting clonality as a central feature of MDS biology. However, in around 10-20% of MDS patients no clonal aberrations are found by standard diagnostic work up. The question remains: expanding gene panels or focusing on clinically relevant markers?

Aim:

To comprehensively characterize MDS cases with normal karyotypes and without detectable mutations utilizing deep sequencing and extended gene panels.

Methods:

We analyzed 636 samples of non-therapy related MDS patients (WHO-HAEM5, median age: 73 years [23-93]; female: 42%) by standard diagnostics (cytomorphology, cytogenetics, molecular genetics) and whole genome sequencing (WGS; median coverage: 100x). Mutation status of 57 myeloid malignancy-associated genes ('myeloid gene panel') were derived from WGS data with a 10% variant allele frequency (VAF) detection limit. WGS data was further evaluated for small nucleotide variants in 723 cosmic cancer genes ('cgc genes') and exome-wide. Protein-altering variants were classified using a four-tier system based on ClinVar, Cosmic, functional impact predictors and gnomAD population frequency. (Likely) pathogenic variants were considered as somatic mutation, while the remaining variants were classified as variants of unknown significance (VUS) or benign polymorphisms. Cases lacking somatic mutations in the myeloid gene panel in WGS analysis were additionally analyzed by targeted deep next-generation sequencing (NGS; Watchmaker DNA Library Prep Kit, median coverage with UMIs: 1,100x; detection limit: 0.5% VAF).

Results:

56/636 cases were detected without any genetic aberrations (MDS-NEG), representing 9% of all cases and 16% of morphologically-defined MDS. MDS-NEG were significantly younger, more female and had a better median overall survival (OS, median age: 66 vs. 73 years, p<0.001; female: 63% vs. 41%, p=0.002; median OS: 7.4 vs. 4.9 years, p=0.001), with the majority of cases being diagnosed as MDS with low blasts (71% [40/56]) and very low or low IPSS-M risk category (91% [51/56]). Targeted deep sequencing led to the identification of 15 mutations in 14/56 (25%) patients (MDS-CHIP) with a median VAF of 1.8% (range: 1.0-3.4%) and the expected enrichment of DNMT3A (n=3) and TET2 (n=2). Interestingly, MDS-CHIP showed detrimental OS compared to other MDS-NEG cases (5.3 vs. 12.6 years, p=0.018). No difference in age or blast counts were detected between the two groups.

Expanding the mutation analysis to the cgc gene panel, 30 (likely) pathogenic variants in 30 genes and 95 VUS in 81 genes were identified affecting 45% (25/56) and 79% (44/56) of MDS-NEG patients, respectively. Notably, the detection of additional variants in cgc genes did not impact OS within these patients. Further extending the analysis to the entire coding exome revealed 677 additional (likely) pathogenic variants and 1,983 additional VUS in overall 2,347 genes affecting all patients (56/56). In the MDS-NEG cohort the median number of mutations per patient increased from 0 [0-2] (likely) pathogenic variants and 0 [0-2] VUS by targeted deep sequencing to 0 [0-2] and 1 [0-5] by cgc gene analyses to 12 [4-23] and 36 [20-53] by exome analyses, indicating a large amount of information within the exome with so far unknown significance. Further, WGS analysis revealed chromosomal abnormalities in 4 patients (4q and 6p copy-neutral loss of heterozygosity (CN-LOH); 3 Mb deletion on 11q; 1.3 Mb gain on 3q22), indicating potential evidence for clonality not detectable by standard cytogenetics due to CN-LOH detection and higher resolution of WGS analysis (although a germline alteration cannot be definitively excluded). However, cytogenetic abnormalities detected by WGS did not impact on OS (median 6.2 vs. 7.4 years; p=0.322). Further, there was no overlap with MDS-CHIP cases, ending up in 18/56 cases with additional clonal markers detected by either targeted deep sequencing of 57 myeloid genes or WGS with higher resolution for structural/numerical aberrations.

Conclusions:

Broadening the analysis spectrum offered limited additional information, but in-depth analysis identified prognostically relevant subclones (VAF <5%) and small-scale chromosomal aberrations. Thus, a focused, in-depth approach may be more clinically efficient and relevant than a broader investigation.