Pathogenic Germline Variants in the DNA Damage Repair Pathway Are Enriched in Patients with Hematologic Malignancies with Prior Exposure to Cytotoxic Therapies Compared to Patients with Multiple Cancers without Prior Exposure

Background: Germline predisposition is increasingly recognized in myeloid neoplasms (MN), with a higher frequency in patients with therapy-related MN (Singhal et al., Leukemia 2021). An unequivocal diagnosis of germline predisposition carries actionable considerations for patient management, including the selection of donor stem cell source, adjustment of the conditioning regimen, and management of complications following allogeneic stem cell transplant (Saygin et al., Blood Advances 2023). However, the prevalence of pathogenic germline variants (PGVs) in different types of MN remains unknown. This study examines the frequency and type of PGVs in patients with haematological malignancy (HM) with or without prior exposure to cytotoxic therapies.

Method: This multi-institution study assessed the prevalence of PGVs in 341 patients with HM. Genomic DNA was extracted from germline samples, including hair-root DNA, mesenchymal stromal cells, skin fibroblasts, or peripheral blood T lymphocytes.

Results: PGVs were identified in 23.5% (n=80) of 341 HM patients compared to 4% and 3% of patients with single non-hematological cancer and age-matched controls, respectively (P <0.0001). The median age at diagnosis was not different in HM patients with or without PGVs (65.5 vs. 61 years; P=0.05), and PGV frequencies were 26.9%, 27.6%, 24.4%, and 16.5% in patients aged ≤50, 51–59, 60–69, and ≥70 years (P=0.22), respectively.

Of the 83 PGVs identified, DDX41 (n=13, 15.7%), CHEK2 (n=9, 10.8%), BRCA1 (n=6, 7.2%) and GATA2 (n=5, 6%) were the most frequent. Half of all PGVs were in genes involved in DNA damage repair (DDR) pathways (n=43, 51.8%), followed by PGVs in DDX41 (n=13, 15.7%), and transcription factors (n=11, 13.3%). Interestingly, there were no DDX41 variants in patients diagnosed ≤50 years of age (n=85), however, this population had the highest frequency of PGVs in transcription factors (n=7, 30.4%; P=0.053). Patients ≥70 years of age had PGVs primarily in DDR pathways (n=7, 43.8%) and DDX41 (n=4, 25%). Although the majority of PGVs were distributed evenly across genders, the frequency of DDX41 PGVs was higher in men than women (25% vs. 3.1%; P=0.008).

We further analyzed the distribution of PGVs across different HM subtypes, including primary MN (P-MN; n=92), multiple cancers including at least one HM (n=236), and bone marrow failure (BMF; n=13). Among the patients with ≥2 cancers, with at least one HM, 55 (23.3%) did not receive cytotoxic therapy (CT) prior to HM (MC-HM), while 181 patients (19%) did receive CT prior to HM diagnosis (CT-HM).

The frequency of PGVs did not differ between patients with P-MN and those with MC-HM or CT-HM (30% vs. 19.9%, P=0.056). The relatively high frequency of P-MN in our cohort is probably due to ascertainment bias, as P-MN patients were screened if they had family history, phenotypic features, or a younger age at presentation than expected, whereas patients with ≥2 cancers including HM were screened regardless of other features.

The PGV frequency was comparable between MC-HM and CT-HM (24% vs. 19%; P=0.43). However, the PGV pattern differed between the two groups. We observed striking enrichment of PGV in the DDR pathway of patients with CT-HM compared to those with MC-HM (70.6% vs. 38.5%; P=0.04). PGV in DDX41 (84.6% vs. 41.9%; P=0.002) and transcription factors (72.2% vs. 41.9%; P=0.43) were more prevalent in MC-HM compared to cases without PGV.

PGVs in the DDR pathway were associated with enrichment of complex karyotype (44.4% vs. 19%; P=0.003) and monosomal karyotype (37.0% vs. 13.9%; P=0.002), in addition to lower frequency of somatic ASXL1 somatic mutations (3.8% vs. 20.5%; P=0.04) compared to patients without PGV. DDX41 PGVs were associated with somatic DDX41 mutations (66.7% vs. 0.9%; P <0.001), and these cases were less likely to harbor complex karyotype (0.0% vs. 19.0%; P=0.08) compared to those without PGV.

Conclusion: The frequency of PGVs was higher in MN patients compared to other cancer patients and age-matched healthy controls. PGVs in the DDR pathway were highly prevalent in HM patients with prior exposure to cytotoxic therapies and were associated with poor-risk cytogenetic features. Our observation of a high frequency of PGVs in older MN patients warrants standardization of germline testing at diagnosis to guide optimal management of patients and their families.