Performance and clinical utility of Spanish germ line genetic testing criteria for hematologic neoplasms predisposition

TO THE EDITOR:

Recent studies indicate that germ line pathogenic or likely pathogenic (P/LP) variants in genes associated with hereditary hematopoietic malignancies (HHMs) may be present in up to ∼19% of myeloid neoplasms in adults.1, 2, 3, 4 Early recognition is pivotal for clinical decision-making, including allogeneic hematopoietic stem cell transplantation donor selection, extrahematologic surveillance, genetic counseling, and cascade testing.^5,6 Nevertheless, HHMs remain underdiagnosed due to incomplete penetrance, variable expressivity, and heterogeneous testing. Although several initiatives have proposed clinical criteria for germ line testing, no universal consensus exists.^7,8

To address this, a multidisciplinary team in Catalonia, Spain, developed standardized germ line testing criteria. Initially implemented within the Catalan public health system,^9,10 these criteria were later adopted nationally through a 2024 consensus led by the Spanish Society of Hematology and Hemotherapy (https://www.sehh.es/publicaciones/libros-sehh). As detailed in Table 1, the framework distinguishes “patient criteria,” which encompass individuals with hematologic malignancies (HM) or aplastic anemia (AA); and “unaffected individual criteria” for those without HM/AA but with other indications warranting testing, such as extrahematologic manifestations, persistent cytopenia, or a relevant family history.

Table 1.

Germ line testing criteria

Main group	Subcategory	Detailed criteria
Patient criteria	Personal history	HM, primarily myeloid neoplasm (MDS or AML), or AA diagnosed at age ≤50 years. Therapy-related myeloid neoplasm at any age if associated with core TP53 tumors∗ and with TP53 mutation, or diagnosed at age ≤50 years. Hypoplastic MDS, regardless of age. Features suggestive of hereditary predisposition or congenital BM failure (eg, thrombocytopenia, dysmorphic features, idiopathic pulmonary fibrosis, hearing loss, etc). Eligible for alloHSCT from a related donor†
	Family history	≥1 first- or second-degree relative with a hematologic disorder compatible with hereditary predisposition, with at least 1 diagnosed at age ≤50 years. ≥1 first- or second-degree relative with a tumor associated with core TP53 mutations,∗ with a TP53 mutation, diagnosed at age ≤50 years. ≥1 first- or second-degree relative with features of hereditary predisposition or congenital BM failure.
	Genetic molecular features	P/LP variants associated with germ line hematologic disorders, with VAF of >30% for single-nucleotide variants or >20% for insertions/deletions. For TP53 alterations, testing is recommended only in therapy-related myeloid neoplasms associated with core TP53 tumors∗diagnosed at age ≤50 years Acute lymphoblastic leukemia with molecular alterations (ETV6, IKZF1, PAX5, and RUNX1). Hypodiploid acute lymphoblastic leukemia with TP53 alterations.
Unaffected individuals criteria	Personal history	Features suggestive of hereditary predisposition syndrome or congenital BM failure. Poor mobilization of hematopoietic progenitor cells (CD34+ cell count <10 per μL on the fourth day of mobilization).
	Family history	≥2 relatives with a hematologic disorder compatible with hereditary predisposition to myeloid neoplasms, with at least 1 diagnosed at age ≤50 years, if the affected relative is unavailable for testing. ≥2 relatives with features of hereditary predisposition or congenital BM failure.

alloHSCT, allogeneic hematopoietic stem cell transplantation; AML, acute myeloid leukemia.

^∗TP53 core tumors: soft-tissue sarcomas, osteosarcoma, breast cancer (especially premenopausal), brain tumors (eg, astrocytoma, glioblastoma, medulloblastoma), and adrenocortical carcinoma.

^†This category was added in the 2024 criteria review; it was not considered in this study because it was newly introduced and not used consistently across the entire study cohort.

In this study, we evaluated the clinical criteria that prompted germ line testing for HHMs and assessed their diagnostic yield and clinical utility.

In 2022, we initiated a prospective, multicenter registry to collect data from individuals undergoing germ line testing according to the Spanish criteria as part of routine clinical care (supplemental Methods).

Between October 2022 and June 2025, 198 unrelated individuals meeting ≥1 criterion, were enrolled. Most participants were male (114/198 [58%]) and of European ancestry (166/198 [84%]). Hematologic abnormalities were present in 96% of individuals (190/198), including HM in 77.8% (154/198), AA in 8.6% (17/198), and clonal cytopenia of undetermined significance or unexplained cytopenia in 9.6% (19/198). The median age at diagnosis for HM was 52.5 years (interquartile range, 41.3-63.8). Across the cohort, acute myeloid leukemia and myelodysplastic neoplasm (MDS) were the most frequent diagnoses (70/198 [35%] and 58/198 [29%], respectively; supplemental Table 2).

Germ line DNA was obtained from several tissues: hair follicles in 45.5% (90/198), buccal swabs in 26.3% (52/198), T-lymphocytes in 6.1% (12/198), cultured skin fibroblasts in 5.0% (10/198), bone marrow (BM) stromal cells in 4.5% (9/198), and blood in 12.6% (25/198) of individuals, restricted exclusively to individuals without HM without clonal events in the somatic next-generation sequencing panel and/or cytogenetic study.

Testing was performed in site-designated laboratories using either whole-exome sequencing or a targeted next-generation sequencing panel that included a standardized HHM gene set (supplemental Table 1). When tumor-only sequencing identified P/LP variants in a HHMs predisposition gene with variant allele frequency (VAF) suggestive of germ line origin (>30% for single-nucleotide variants or >20% for insertions/deletions, consistent with current European Society For Medical Oncology cutoffs),^11,12 confirmatory Sanger sequencing was performed on DNA from nonhematopoietic tissues to establish germ line status.

Most individuals (147/198 [74.2%]) met a single criterion; 23.7% (47/198) met 2, 1.5% (3/198) met 3, and 0.5% (1/198) met 4. The most frequent triggers were HM/AA at age ≤50 years (44.4% [88/198]), and germ line–suspicious VAF on tumor sequencing (43.9% [87/198]), followed by family history in affected individuals (10.6% [21/198]) and HHM-suggestive features in affected (9.1% [18/198]) or unaffected (8.1% [16/198]) individuals (Figure 1B).

Figure 1.

Genes affected and diagnostic yield per criterion. (A) Distribution of individuals with P/LP variants in predisposition genes within our cohort. (B) Individuals meeting each criterion and P/LP variants identified per criterion. The number at the top of each bar indicates the total number of individuals meeting that criterion. Dark purple bars represent the number of patients in whom a P/LP variant was identified. FH, family history; TRMN, therapy-related myeloid neoplasms.

Germ line variants were identified in 39.9% (79/198) of tested individuals, including P/LP variants in 30.3% (60/198). Among P/LP variants, the most frequently affected gene was DDX41 (43.3% [26/60]), followed by RUNX1 (13.3% [8/60]), RTEL1 (8.3% [5/60]; 1 individual with a biallelic variant), ETV6 (6.7% [4/60]), GATA2 (6.7% [4/60]), ELANE (6.7% [4/60]), NAF1 (3.3% [2/60]), and single individuals (1.7% [1/60] each) of ERCC6L2 (biallelic, confirmed in trans), CSF3R, TP53, TINF2, CHEK2, ATM, and TERT (Figure 1A; supplemental Table 3). Detailed cytogenetic and molecular characterization of the patients with DDX41, GATA2, RUNX1, and ERCC6L2 variants, highlighting specific genotype-phenotype correlations (eg, DDX41 with normal karyotype, GATA2 with chromosome 7 and 8 abnormalities), is provided in the supplemental Results.

Importantly, diagnostic yield varied markedly by testing trigger. It was highest among HHM-suggestive features (61.1% [11/18] in affected and 43.8% [7/16] in unaffected individuals), followed by germ line–suspicious VAF (41.4% [36/87]), hypoplastic MDS (22.2% [2/9]), and HM/AA at age ≤50 years (21.6% [19/88]). A family history–based approach identified a causative germ line variant in 36.4% (4/11) of unaffected individuals and 19.1% (4/21) of affected individuals (Figure 1B).

Beyond diagnostic yield, clinical utility was immediate. Allogeneic hematopoietic stem cell transplantation donor selection was tailored in 15 patients (19%: 10 DDX41, 5 RUNX1). Medical management was influenced by genetic findings in 2 individuals (2.5%): 1 patient with DDX41-associated MDS received venetoclax plus hypomethylating agents, guided by the reported sensitivity of germ line DDX41–related disease.13, 14, 15 Another patient with chronic neutrophilic leukemia carrying a CSF3R-p.T618I germ line variant was managed conservatively with close clinical and molecular monitoring because of stable neutrophil counts for >10 years.¹⁶ Cascade testing was offered to at-risk relatives, leading to testing in 153 family members from 60 families; 43.8% (67/153) were carriers of the familial P/LP variant and enrolled in early-detection follow-up programs.

This study represents the first Spanish multicenter and the largest prospective clinical cohort to undergo germ line genetic testing according to standardized national criteria. Using these criteria, we identified P/LP variants in approximately one-third of individuals tested, exceeding the ∼15% typically reported for patients with solid tumors,^17,18 and aligning closely with the recent prospective validation of the Nordic guidelines. Tesi et al reported a P/LP yield of 35% (30/85) and a substantially higher yield in the CytoMol-trigger group (defined by germ line–suspicious VAF or −7/del(7q) in individuals aged <50 years) of 52% (17/33). These results mirror our findings, with a 41.4% yield (36/87) among VAF-triggered cases and DDX41 representing the most frequently mutated gene in both cohorts.¹⁹

This evidence supports that VAF-triggered testing is a highly efficient entry point, even in the absence of a striking family history. We defined our VAF cutoffs according to European Society For Medical Oncology recommendations^11,12; however, in our cohort, all confirmed germ line variants initially detected through tumor testing had VAFs ≥35%, with the lowest confirmed single-nucleotide variants at 39%. These findings suggest that, in HM, applying a pragmatic VAF threshold ≥35% may improve the prioritization of patients for germ line confirmation. Notably, other groups have also proposed similar cutoffs, typically ∼40%, further supporting this approach.^2,19

Interestingly, our highest diagnostic yield was observed among HHM-suggestive features, both in affected and unaffected individuals. Our findings indicate that phenotype-driven testing can uncover predisposition to HHMs regardless of whether an individual has already developed cancer or BM failure. Individuals with mild or moderate clinical presentations (eg, idiopathic cytopenia, abnormal skin pigmentation, or lymphedema) should also be considered for testing.

One major strength of this study lies in the direct clinical implications of germ line findings, both in guiding patient management and enabling cascade testing. In 1 illustrative family, we identified a germ line NAF1 LP (p.(V228Nfs∗3)) variant through the HHM-suggestive features criterion (pulmonary interstitial disease, premature graying, long-standing cytopenia, and liver damage). Cascade testing of the proband’s brother confirmed the same germ line variant, prompting baseline BM evaluation and pulmonary screening, which revealed MDS with ring sideroblasts and early interstitial lung disease.

Our study demonstrates the feasibility of applying standardized national criteria within a public health care network, encompassing both affected and unaffected individuals. Strengths include its prospective, multicenter design, and counseling-first workflow. Limitations include potential ascertainment bias, as clinician-triggered referrals likely favored readily recognizable indicators (eg, high VAF or younger age), possibly undercapturing subtler HHM phenotypes or family histories, and platform heterogeneity between targeted panels and whole-exome sequencing. Through the implementation of these criteria, we strengthened our recommendation for the use of hair bulb DNA as the alternative to the gold standard of cultured fibroblasts. Because hair follicles are a nonhematopoietic tissue, false negatives caused by somatic reversion are averted, and our previous results show no relevant contamination.²⁰ However, as a limitation, T-lymphocytes and blood (only in individuals without HM) were used in a limited number of analyses. In both tissues, the presence of clonal hematopoiesis may be a confounding factor. To mitigate this, variants identified as germ line in T cells and blood were validated via family segregation analysis, alternative tissues, or phenotypic correlation (supplemental Table 3). Regarding somatic reversion mechanisms, although no highly suspected patients (eg, monosomy 7) were tested using T cells, we acknowledge the possibility of false negatives.

In conclusion, real-world application of standardized Spanish criteria yielded P/LP variants in ∼30% of tested individuals leading to clinically actionable interventions, with highest yields in HHM-suggestive phenotypes and tumor-informed triggers.

The study was approved by the Vall d'Hebron University Hospital Ethics Committee (PR(AG) 129/2023) and by institutional review boards of all participating centers and was conducted in accordance with the Declaration of Helsinki.

Conflict-of-interest disclosure: The authors declare no competing financial interests.