Evaluation of patients and families with concern for predispositions to hematologic malignancies within the Hereditary Hematologic Malignancy Clinic (HHMC)

Abstract

Introduction

Although multiple predispositions to hematologic malignancies exist, evaluations for hereditary cancer syndromes (HCS) are underperformed by most hematologist/oncologists. Criteria for initiating HCS evaluation are poorly defined, and results of genetic testing for hereditary hematologic malignancies have not been systematically reported.

Patients/Methods

From April 2014 to August 2015, 67 patients were referred to the Hereditary Hematologic Malignancy Clinic (HHMC). Referral reasons included (1) bone marrow failure or myelodysplastic syndrome in patients ≤50 years, (2) evaluation for germline inheritance of identified RUNX1, GATA2, or CEBPA mutations on targeted next-generation sequencing panels, (3) strong personal and/or family history of malignancy. Cultured skin fibroblasts were utilized for germline DNA in all patients with hematologic malignancy.

Results

Eight (12%) patients were clinically diagnosed with a HCS; 4 patients with RUNX1-related familial platelet disorder (FPD)/AML, and one patient each with dyskeratosis congenita, Fanconi anemia, germline DDX41, and Li-Fraumeni Syndrome (LFS). Two patients with concern for FPD/AML and LFS, respectively, had RUNX1 and TP53 variants of unknown significance. Additionally, four patients with prior HCS diagnosis (1 LFS, 3 FPD/AML) were referred for further evaluation and surveillance.

Conclusion

In this HHMC-referred hematologic malignancy cohort, HCS was confirmed in twelve patients (18%). HCS identification provides insight for improved and individualized treatment, and screening/surveillance opportunities for family members. The HHMC has facilitated HCS diagnosis, and advocates that with increased clinical awareness of hematologic malignancy predisposition syndromes, more patients who may benefit from evaluation can be identified. Mutation panels intended for prognostication may provide increased clinical suspicion for germline testing.

Introduction

Hereditary cancer syndromes (HCS) such as Li-Fraumeni Syndrome (LFS), hereditary non-polyposis colorectal cancer (HNPCC), and hereditary breast and ovarian cancer syndrome (HBOC) are well-established familial syndromes within solid tumor oncology. The identification of heterozygous germline mutations in the well-annotated cancer susceptibility genes TP53, MLH1, MSH2, MSH6, PMS2, BRCA1 and BRCA2 impact treatment decisions of an affected patient directly, and identify individuals who may benefit from potentially life-saving screening and prevention approaches among affected family members. There are currently over 200 hereditary cancer susceptibility syndromes defined, estimated to account for 5–10% of all cancer diagnoses.(1, 2)

Within hematologic malignancies, there is increasing appreciation for the genetic basis of certain inherited bone marrow failure (BMF) syndromes such as Fanconi anemia (FA) and telomere syndromes including dyskeratosis congenita (DC).(3) BMF syndromes have been characteristically diagnosed in children, although approximately 10% of patients with FA and a majority of short telomere syndrome patients are diagnosed as adults.(4–8) Despite critical therapy implications such as avoidance of DNA crosslinking agents, high-intensity regimens, and informing appropriate sibling selection for stem cell transplantation (SCT), HCS evaluation among patients with hematologic malignancies is thought to be routinely underperformed, particularly among adults. Within the past decade, nearly a dozen adult-onset inherited leukemia and myelodysplastic syndrome (MDS) predisposition syndromes have been defined (Table 1). Although considered quite rare, the low level of clinician awareness of these syndromes, coupled with the frequently insufficient family history obtained during the routine care of patients with hematologic malignancies, suggests these syndromes are considerably under-diagnosed. Furthermore, among patients with strong family histories of hematologic malignancies who are tested for currently known hereditary predisposition syndromes, many lack an identifiable genetic cause, suggesting that additional pathogenic loci remain to be discovered.(9)

Table 1.

Inherited Syndromes Predisposing to Adult-Onset Hematologic Malignancy

Syndrome	Gene	Inheritance	Hematologic Malignancy Risk	Other hematologic abnormalities	Non-hematologic complications
Familial platelet disorder with propensity to myeloid malignancies	RUNX1	AD	MDS/AML/T-cell ALL	Thrombocytopenia, bleeding propensity, aspirin-like platelet dysfunction	Eczema
Thrombocytopenia 2	ANKRD26	AD	MDS/AML	Thrombocytopenia, bleeding propensity	None described
Familial AML with mutated CEBPA	CEBPA	AD	AML	None	None described
Familial AML with mutated DDX41	DDX41	AD	MDS/AML, CMML	None	None described
Thrombocytopenia 5	ETV6	AD	MDS/AML, CMML, B-cell ALL, multiple myeloma	Thrombocytopenia	Possible risk for solid tumors
Familial MDS/AML with mutated GATA2	GATA2	AD	MDS/AML/CMML	None or Neutropenia, monocytopenia (MonoMAC syndrome)	Lymphedema, sensorineural hearing loss, extra-genital warts, other
Telomere syndromes (dyskeratosis congenita)	TERC,TERT 1CTC1, DKC1, NHP2, NOP10, RTEL1, TINF2, WRAP53, ACD, PARN	AD, AR	MDS/AML	Macrocytosis, aplastic anemia	Pulmonary, hepatic fibrosis; head, neck, ano-genital SCC
Familial aplastic anemia with SRP72 mutation	SRP72	AD	MDS	Aplastic anemia	Hearing loss
Familial B- cell ALL with PAX5 mutation	PAX5	AD	ALL	None	None described
Germline SH2B3	SH2B3	AR	ALL	None	None described
Li-Fraumeni syndrome	TP53	AD	ALL (esp hypodiploid cytogenetics), therapy-related myeloid disorders	None	Multiple solid tumors (breast, sarcoma, brain)
Fanconi anemia2	FANCA, FANCB, FANCC, BRCA2, FANCD2, FANCE, FANCF, FANCG, FANCI, BRIP1, FANCL, FANCM, PALB2, RAD51C, SLX4	AR	MDS/AML	Aplastic anemia	Head, neck, anogenital SCC, developmental delay, skeletal and renal anomalies

AD=autosomal dominant, AR=autosomal recessive, MDS=myelodysplastic syndrome, AML=acute myeloid leukemia, ALL=acute lymphoblastic leukemia, CMML=chronic myelomonocytic leukemia, SCC=squamous cell carcinoma

¹These DC genes are typically diagnosed in infants/children with DC

²The majority of individuals with Fanconi anemia are diagnosed in childhood; however, 10% of patients lack classic birth defects and present in adulthood

Moreover, with the growing importance of next-generation sequencing (NGS) technologies to evaluate for prognostically-significant or targetable genomic alterations in patients with hematologic malignancies, detecting mutations that could be of either somatic or germline origin (e.g. RUNX1, CEBPA, GATA2) has become increasingly available and commonplace. Given the difficulty of obtaining a germline “control” in the hematologic malignancy population, determination of whether the NGS-identified mutation is somatically-acquired or inherited is not straightforward; yet accurate classification is critical for optimal patient management.

The Hereditary Hematologic Malignancy Clinic (HHMC) was established for the dedicated evaluation and monitoring of patients with hematologic malignancies and suspected germline predisposition to malignancy. We herein report our experience and initial results from the first 18 months of genetic counseling and evaluation.

Patients and Methods

Patient Selection

From April 2014 to August 2015, 67 patients between the ages of 13 to 80, from 65 unique families, were referred to the HHMC at The University of Texas M.D. Anderson Cancer Center (UTMDACC). Approximately 500 patients with AML and 350 with MDS were referred to the Department of Leukemia at UTMDACC during this same time period. For manuscript clarity given the varied reasons for HHMC referral, the primary reason for HHMC referral has been apportioned to one of the following groups/cohorts, which are detailed separately within the manuscript:

1. Diagnosis of BMF, aplastic anemia (AA) or MDS in patients ≤50 years of age, or any age with concerning clinicopathologic features of a bone marrow failure syndrome.(7, 10)

2. Evaluation for germline inheritance in patients with hematologic malignancies and ≥1 known mutation(s) of RUNX1, GATA2, or CEBPA on a targeted NGS MDS/AML panel

3. A hematologic malignancy patient with multiple primary malignancies and/or a strong family history of cancer(s) at early age

4. A hematologic malignancy patient with one or more first-degree relatives, or ≥2 second-degree relatives, with hematologic malignancies.

5. Referral for evaluation, counseling and continued follow-up for patients with a previously identified HCS with associated risk for hematologic malignancy.

Sequencing of UTMDACC Patient Samples

NGS-based CLIA-approved targeted mutation analysis of 28 genes implicated in hematologic malignancies is routinely performed on DNA from bone marrow aspirates in all MDS and AML patients referred to UTMDACC as previously described.(11, 12) The coding sequence of ABL1, ASXL1, BRAF, DNMT3A, EGFR, EZH2, FLT3, GATA1, GATA2, HRAS, IDH1, IDH2, KIT, KRAS, MDM2, IKZF2, JAK2, MLL, MPL, MYD88, NOTCH1, NPM1, NRAS, PTPN11, RUNX1, TET2, TP53 and WT1 are sequenced. CEBPA analysis is performed separately as previously described via Sanger sequencing.(13) Identified sequence variants are classified based on analytic findings and curated reference databases (e.g. COSMIC, dbSNP). Known germline polymorphisms, defined by membership in dbSNP build 138/1000 genomes dataset and/or presence in >20% of our patient population were excluded.

Genetic Evaluation

All patients referred to the HHMC underwent a standard assessment including evaluation, family history, and pedigree analysis by a certified genetic counselor (S.A.B., M.M.) and hematologist/oncologist with expertise in HCS (C.D.D.). Patients with a clinical phenotype suggesting a specific HCS underwent syndrome-specific genetic testing, such as chromosome breakage analysis for FA evaluation or lymphocyte telomere length (TL) by flow-FISH for DC. Patients referred for evaluation of a variant identified within our UTMDACC NGS cancer panel underwent gene or site-specific CLIA-approved germline testing performed on cultured skin fibroblasts. Clinical genetic testing performed was patient-specific, based on personal and family history, current medical knowledge, and available clinical testing. For patients from “Group D” with a family history of hematologic malignancies, clinical testing was provided using an Agilent Clinical Research Exome kit which has been iteratively revised to include the following genes (with the percentage of the coding region covered at >10X) by exome sequencing: ANKRD26 (100%), CEBPA (61%), DDX41 (98%), ETV6 (100%), GATA2 (100%), RUNX1 (92%), SRP72 (100%), and TP53 (100%).

In all patients with active hematologic malignancy or post-allogeneic transplantation, a 4mm skin punch biopsy was performed, and skin fibroblasts were cultured to obtain germline DNA. Blood or saliva specimens for germline DNA analysis were obtained for the evaluation of consented family members. Written informed consent was obtained following institutional guidelines in accordance with the Declaration of Helsinki.

Statistical Analysis

Personal and family histories, demographics, and clinicopathologic characteristics were obtained from the electronic health record and genetic counseling session results. Patient characteristics are summarized using median (range) for continuous variables and frequency (percentage) for categorical variables.

Results

Sixty-seven patients from 65 unique families were referred to the HHMC. Evaluated patients who declined germline genetic testing (n=2) are included. Results are detailed below by reason for patient referral.

BMF/AA/MDS at young age [Group A]

Eight patients (Table 2) were referred for a history of BMF/AA/MDS at a young age, defined here as ≤ 50 years of age at diagnosis, or any age with clinical symptoms suggestive of a BMF syndrome (i.e. short stature, skeletal abnormalities, pulmonary fibrosis, abnormal skin, dentition or hair).(7, 10) The median age was 39 (range 14–55 years), and five were female. Testing for FA was performed by diepoxybutane (DEB) induced chromosomal breakage assay, and DC by TL flow-FISH analysis and/or gene panel testing. A TL below the age-adjusted first percentile, in the setting of confirmatory clinical history, is consistent with a telomere syndrome.(14) Existing gene panel testing identifies a causative mutation in approximately 50–70% of patients with DC (e.g. TERT, TERC), while the remaining DC patients have uninformative genetic panel testing resulting from unknown genetic cause. In this patient cohort, germline RUNX1 testing was also performed in one patient with a lifelong history of thrombocytopenia and multiple additional family members with thrombocytopenia, MDS and AML suggestive of FPD/AML.

Table 2.

Patients referred to the HHMC for BMF/MDS/AA at young age

Age/Sex	Dx and Age at Dx	PMHx	Family History	Cytogenetics	Testing Performed
44/M	Bone marrow failure/hypocellular MDS	Gilbert’s Syndrome	Not significant	46,XY[20]	DEB assay negative 10% age-adjusted telomere length
27/F	Anemia age 12 PNH diagnosis age 22	History of sessile serrated adenoma	Father with CRC age 34 with history of multiple polyps Both paternal grandparents with colon cancer	46,XX[20]	DEB assay negative DC panel negative MLH1, MSH2, MSH6, PMS2, EPCAM negative
35/F	Hypocellular MDS age 32	Abnormal teeth, premature grey	Not significant	45,XX,-5,der(7;17)(p10;q10),+mar[3] 46,XX[26]	DEB assay negative DC panel negative
53/M	MDS age 50	Ankylosing spondylitis Pulmonary Fibrosis Premature gray hair	Brother with pulmonary fibrosis age 47, father age 50	46,XY,+1,der(1;21)(q10;q10)[1]/46,XY[19]	<1% telomere length Heterozygous TERT R865H mutation
14/F	Anemia since birth Bone marrow failure age 10	Congenital hip dysplasia, absent thumbs bilaterally and absent right radius	Brother with aortic stenosis at birth Consanguinity present	46,XX[20]	DEB assay positive Homozygous BRIP1 p.R798*
58/M	MDS age 55 AML age 58	Premature gray, abnormal nails and dentition, short stature	Not significant	46,XY,del(11)(q13q23)[10]/46,idem,del(20)(q11.2q13.3)[2]/46,XY[8]	DEB assay negative Telomere length normal
53/F	MDS age 48 AML age 51	Longstanding easy bruising	1 sister AML age 6, 1 sister MDS age 48, 1 brother with NHL and 1 brother AML age 31. 3 of 4 first maternal cousins with AML	46,XX,t(2;22)(p23;q13.1), del(7)(q22q32)[20]	DEB assay negative Germline RUNX1 c.719delC p.P240Hfs*14
19/F	MDS age 18 s/p SCT Relapsed MDS age 19	Developmental delay, distal symphalangism, short stature, congenital hip malformations	Not significant	40~45,XX,-3,add(5)(q13),−6, −7, −14, −15, −18, −21,+0~4mar[cp20]	DEB assay (of skin fibroblasts) negative

BMF = bone marrow failure, AA = aplastic anemia, MDS = myelodysplastic syndrome, SCT = stem cell consultation, PNH = paroxysmal nocturnal hemoglobinuria, DEB = diepoxybutane, NHL = non-hodgkin lymphoma, N/A = not applicable

Three (38%) of eight Group A patients were diagnosed with a HCS. The pedigrees of these three individuals are depicted in Figure 2(a–c). One 53-year-old male with a personal history of MDS, pulmonary fibrosis, short stature, premature grey hair, whose father and brother also have idiopathic pulmonary fibrosis, had an age-adjusted TL <1% in lymphocytes and granulocytes consistent with a telomeropathy, and a pathogenic heterozygous TERT R865H variant on gene panel testing.(15) A 14-year-old female with BMF diagnosed at age 11, also with congenital skeletal abnormalities including radial arm defects, was diagnosed with FA by DEB assay, with homozygous pathogenic BRIP1 p.R798* variants on FA gene sequencing.(16) A 54-year-old female with lifelong thrombocytopenia and easy bruising/bleeding, MDS diagnosed at age 48, and a strong family history of MDS/AML was identified to have a germline RUNX1 p.P240Hfs deletion leading to a diagnosis of FPD/AML. In these three patients, screening for family members and appropriate selection of siblings for SCT donors has been possible.

Figure 2.

Pedigrees of patients/families with a hereditary cancer syndrome identified. a) Dyskeratosis Congenita (germline TERT p.R865H), b) Fanconi anemia (homozygous BRIP1 p.R798*), c) FPD/AML (germline RUNX1 p.P240Hfs deletion), d) FPD/AML (germline RUNX1 c.1098_1103dupCGGCAT duplication), e) FPD/AML (germline RUNX1 p.K194N); f) Li-Fraumeni Syndrome (germline TP53 p.R175H), g) FPD/AML (germline RUNX1 gene deletion spanning exons 1–6), h) germline DDX41 (p.M1I)

Arrowhead indicates the index patient referred to the HHMC. NOS (green): hematologic malignancy, not otherwise specified; AML (red): acute myeloid leukemia; EOS: eosinophilia; thrombocytopenia (light blue): long-standing history of thrombocytopenia; MPN (yellow): myeloproliferative neoplasm; MDS (red): myelodysplastic syndrome; AA (red): aplastic anemia.

Testing for germline status of CEBPA, GATA2 and RUNX1 mutations identified from NGS-based prognostication panel [Group B]

Germline inheritance of CEBPA, GATA2 or RUNX1 mutations lead to well-described, autosomal dominant, hematologic malignancy syndromes (Table 1) with variable penetrance.(17–21) Seventeen patients were referred for germline investigation of CEBPA, GATA2 and/or RUNX1 mutations, after one or more of these mutations was identified from an NGS-based hematologic malignancy panel performed routinely for disease prognostication at our institution. As germline samples in patients with hematologic malignancies are not typically obtained, standard sequencing cannot distinguish between germline or somatic nature, and some mutations identified on leukemia sequencing panels may be inherited. An emphasis for referral was placed on patients in whom mutations were present near-heterozygous (40–60%) or near-homozygous (≥90%) allelic frequency, and/or with the presence of multiple mutations within of one of these genes. During this time period at our institution, 771 patients with MDS or AML had NGS panel testing and 88 CEBPA (11%), 15 GATA2 (2%) and 105 RUNX1 (14%) pathogenic variants were reported.

Two CEBPA-mutated patients (one single-mutant and one double-mutant CEBPA also with GATA2 mutation), 6 GATA2-mutated patients, and 10 RUNX1-mutated patients were referred (Table 3) for germline evaluation. Upon genetic counseling, one 36-year-old male with an NGS-identified RUNX1 p.P200fs was ascertained to have a significant family history of thrombocytopenia and leukemia. He declined germline evaluation, for fear of placing the blame of his and his sister’s leukemia onto his mother, who notably shares a history of lifetime mild thrombocytopenia. Results of germline analysis by site-specific mutation testing of cultured fibroblasts in the remaining 16 patients revealed that 0/2 CEBPA, 0/6 GATA2, and 2/9 RUNX1-mutated patients with NGS-identified mutations were of germline origin.

Table 3.

Referral for germline evaluation of presumed somatic mutations

Age/Sex	NGS Panel	Variant %	Alteration	Cytogenetics	Heme Malignancy & Significant PMHx	Family History of Malignancy	Germline Mutation?
69/F	CEBPA	3.7%	c.146del p.Pro49fs	46,XX[20]	AML dx age 69 s/p SCT	Brother with AML age 48, unknown details	N
	NPM1	2.7%	c.860_863dupTCTG p.W288fs
	TET2	29.4%	c.1793del p.N598fs
	TET2	27.4%	c.4063_4087del p.A1355fs
59/F	CEBPA	35.9%	c.946_948dupGAG p.E316dup	46,XX[20]	AML dx age 59	Father died of pancreatic cancer age 33 Paternal cousin with AML in 50s	N
	CEBPA	27.3%	c.68dupC p.H24fs
	GATA2	36.3%	c.962T>A p.L321H
	TET2	32.5%	c.4145A>G p.H1382R
55/M	GATA2	52.1%	c.1169_1177del p.K390_G392del	46, XY [20]	ITP at age 21 Burkitt Lymphoma age 43 MDS/MPN age 55 Recurrent viral infections including HSV pneumonia Monomac FLOW	None	N
	DNMT3A	46.6%	c.2645G>A p.R882H
	PTPN11	4.7%	c.182A>T p.D61V
62/M	GATA2	48.1%	c.1168_1170del p.K390del	47,XY,+1,der(1;7) (q10;p10),+8[20]	MDS dx age 62 Hospitalized for PNA at age 57 Monomac Flow	None	N
	GATA2	49.8%	c.573dupG p.S192fs*9
	IKZF2	46.5%	c.1404T>G p.I468M
	TET2	49.4%	c.5738G>A p.G1913D
78/M	GATA2	50.64	c.1172_1175del p.E391fs*85	46,XY[20]	Prostate cancer s/p XRT age 76. MDS/MPN classified as atypical CML	Brother with unknown leukemia age 77, deceased	N
	ASXL1	47.8%	c.2967dupT p.E990*
	TET2	50.2%	c.4048G>A p.E1350K
80/F	GATA2	52.1%	c.973_974insAT p.M325fs*2	46,XX[20]	Hx of rheumatoid arthritis s/p IMIDs MDS/MPN-U	Sister with post-menopausal breast cancer	N
	NRAS	45.5%	c.34G>C p.G12R
57/M	GATA2	49.36%	c.1168A>G p.K390E	47,XY,+8[20]	Thrombocytopenia for 20 yrs MDS/MPN classified as atypical CML Multiple recurrent viral infections Culture-negative necrotic skin ulcers Monomac Flow	None	N
	KIT	49.5%	c.2089C>A p.H697N
36/M	RUNX1	45.5%	c.593_596dupATGG p.P200fs*	46,XY[20]	Thrombocytopenia and abnormal platelet function tests since age 10 AML dx age 36	Sister with ALL deceased age 14 Mother with lifetime thrombocytopenia	Declined germline testing
	EGFR	10.1%	c.1231C>T p.P411S
46/M	RUNX1	50.4%	c.582A>C p.K194N	46,XY[28] 46,XY,del(11)(q1 3q23)[2]	Thrombocytopenia since age 13 Mild dysplasia but no MDS diagnosis	2 daughters with thrombocytopenia; one with MDS age 6	RUNX1 confirmed as germline
55/M	RUNX1	39.1%	c.601C>T p.R201*	46,XY,t(8;13)(p1 1.2;q12)[19] 46,XY[1] FGFR1 gene rearrangement detected by FISH	Myeloid neoplasm with eosinophilia and FGFR1 gene rearrangement	Brother poor SCT mobilizer with same G387A VUS. Another brother with sarcoma in his 30s, alive	Germline RUNX1 G387A VUS
	RUNX1	46.3%	c.1160G>C p.G387A
71/F	RUNX1	26.3%	c.593A>G p.D198G	46,XX[20]	MDS No preceding cytopenias	None	N
	RUNX1	54.6%	c.796C>T p.Q266*
	ASXL1	63.0%	c.1900_1922del p.E635fs
54/M	RUNX1	48.0%	c.602G>A p.R201Q	46,XY,t(9;22)(q3 4;q11.2)[20]	PH+ ALL	None	N
	MLL	49.8%	c.6859G>A p.E2287K
13/F	RUNX1	46.0%	c.1098_1103dupCGGCAT	46,XX,inv(9)(p12 q13)[20] *constitutional chromosomal polymorphism	Known thrombocytopenia since infancy	N (limited paternal information)	RUNX1 Confirmed as germline
	TET2	47.0%	p.I366_G367dup
			c.1598T>C p.M533T
45/M	RUNX1	46.4%	c.497G>A p.R166Q	47,XY,+8[10] 48,idem,+9[8] 46,XY[2]	Thrombocytopenia for 3 yrs prior to MDS age 45	Mother with thrombocytosis	N
	DNMT3A	38.3%	c.2644C>T p.R882C
	IDH1	33.6%	c.394C>T p.R132C
	IDH2	14.1%	c.419G>A p.R140Q
	ASXL1	48.0%	c.4493C>T p.T1498M
58/M	RUNX1	6.0%	c.602G>A p.R201Q	46,XY,dup(11)(q 13q23)[6]/46,XY[ 14]	MDS	None	N
	RUNX1	40.8%	c.820del p.Q274fs
55/M	RUNX1	5.0%	c.595G>T p.G199W	46,XY,del(20)(q1 1.2q13.3)[6]/46,XY[14]	Therapy-related MDS age 55 History of PH+ ALL age 53	None	N
	RUNX1	6.0%	c.806-2A>T
	RUNX1	22.2%	c.819_835del p.I273fs
57/M	RUNX1	41.1%	c.402dupT p.G135fs	46,XY[20]	MDS dx age 56	Brother with HCV and HCC	N
	NRAS	23.7%	c.35G>A p.G12D

Germline GATA2 mutations lead to GATA2 (MonoMAC) syndrome, characterized by profound B-lymphocyte, monocyte, NK, and dendritic cell deficiencies, frequent atypical infections and a propensity to develop MDS/AML.(22–25) Three patients referred for germline GATA2 testing had experienced recurrent and disseminated viral and fungal infections and culture-negative pneumonia. Notably, flow cytometry was consistent with the anticipated phenotype of germline GATA2 mutations, i.e. the lack of a defined monocyte population, poorly-defined lymphocyte population, and decreased frequency of NK cells with specific loss of the CD56^bright subset (26) (Supplemental Figure 1); yet skin fibroblast analysis for constitutional GATA2 mutations, including evaluation for intronic variants, was wild-type in all patients. This suggests the presence of an additional unidentified germline GATA2 mutation in these three patients, or alternatively that acquired GATA2 mutations may impart a MonoMAC-like phenotype at the time of somatic GATA2 acquisition.

Two patients referred for NGS-identified RUNX1 mutations were identified to have a germline RUNX1 mutation leading to the diagnosis of FPD/AML (Figure 2d and 1e). One patient was a 13-year-old female with lifetime mild thrombocytopenia and a concern for MDS at an outside institution. Her bone marrow biopsy at referral demonstrated abnormal megakaryocytes without morphologic criteria for MDS and a RUNX1 duplication (c.1098_1103dupCGGCAT). She was subsequently referred to the HHMC where her RUNX1 duplication was recognized to be of germline origin; she continues to be monitored with clinical exam and surveillance blood work every six months. The second patient was a 46-year-old man with a presumptive historical diagnosis of ITP status-post splenectomy, whose father was deceased from CMML transformed to AML at age 77, and whose daughter was referred to our institution at age 5 for deletion 5q- MDS. Both the daughter and father had a RUNX1 p.K194N variant detected on institutional NGS sequencing; confirmed in the father to be germline through clinical RUNX1 sequencing. The p.K194N variant is predicted to be deleterious by in-silico analysis programs (e.g. Polyphen-2 and SIFT). Functional testing of the RUNX1 p.K194N variant by luciferase reporter assay demonstrated lack of transactivation compared to wild-type RUNX1, confirming likely pathogenicity (Supplemental Figure 2).

Figure 1.

Flow diagram of patients referred to the Hereditary Hematologic Malignancy Clinic (HHMC) including primary reason for referral and brief summary of findings

BMF indicated bone marrow failure; AA: aplastic anemia; MDS: myelodysplastic syndrome; HCS: hereditary cancer syndrome; MPN: myeloproliferative neoplasm; CLL: chronic lymphocytic leukemia; NHL: non-Hodgkin lymphoma

Of additional interest, a 55-year-old male with myeloid neoplasm with eosinophilia and FGFR1 gene rearrangement had two RUNX1 mutations identified on sequencing panel, a p.R201* truncating mutation and p.G387A missense variant of uncertain significance (VUS). The p.R201* was confirmed somatic, and the p.G387A was present in his germline, and clinically classified as a VUS by the testing laboratory. His brother was his SCT donor, and he was notably characterized as a poor mobilizer (<4 million/kg CD34+ cells obtained) and shared the germline p.G387A VUS.

Multiple primary malignancies or strong family history of cancers at early ages (Group C)

Thirteen patients with hematologic malignancies were referred for evaluation due to a history of multiple primary malignancies and/or a strong family history of relatives with cancer(s) (Table 4). The median age was 52 years (range 25–73). Seven of 13 patients met Revised Chompret criteria for TP53 testing.(27) In addition, one 40-year-old female with a historical diagnosis of LFS and personal history of breast cancer and sarcoma was referred at the time of a new diagnosis of therapy-related AML (Table 5). Her germline TP53 abnormality consisted of a deletion spanning exons 10–11, which incidentally was not identified on her leukemia NGS panel-based testing, likely due to the difficulty of identifying large deletions/duplications on NGS-based sequencing platforms not specifically designed for this purpose.

Table 4.

Referral for history of multiple primary malignancies and/or family history of relatives with cancer at early ages

Age/ Sex	Cancer History	Family History	Somatic Mutation Testing	Cytogenetics	Met Revised Chompret Criteria?	Germline Mutation?
58/M	Testicular seminoma age 10 months (tx surgery only) Papillary thyroid cancer age 44 tx RAI Essential thrombocytosis/MPN age 44	Son with astrocytoma age 5 Mother died age 63 of lung cancer Maternal aunt with bone cancer age 50	JAK2-V617F	46,XY[20]	No	TP53 Negative
46/F	Rectal cancer age 32 Cervical carcinoma in situ age 34 Bronchioalveolar carcinoma age 35 CML age 45	Father lung cancer age 68 Paternal grandmother with rectal cancer age 75	N/A	46,XX,t(9;22)(q34;q11.2)[19] 46,idem,t(5;16)(q31;p13.1)[1]	Yes	TP53 Negative MMR testing also negative
48/M	Hodgkin Lymphoma age 15 s/p XRT CML age 29 s/p SCT Meningioma age 47 s/p resection Papillary and follicular thyroid cancer age 48	2 paternal uncles with unknown cancers in 50–60s	N/A	N/A	No	TP53 Negative
68/F	Colorectal cancer age 52 s/p surgery and chemoXRT Osteosarcoma age 63 s/p surgery and chemotx Acute myeloid leukemia age 68	Unremarkable	TP53 c.507_530del p.M169_P177delin sI 68.3% allelic frequency	43~46,XX,del(5)(q13q35),+6,+9,−10, der(11)dup(11)(q23q24) add(11)(q24)dup(11)(q23q24) add(11)(q24),der(11)t(5;11)(q13;p15)dup(11)(q23q24) add(11)(q24)dup(11)(q23q24)add(11)(q24),−13,der(13;17)(q10;q10), der(16)add(16)(p13.3)ins(?;11)(?;q23q24)x2add(11)(q24),+19,add(21)(p11.2),+1~2mar[cp18]	No	TP53 Negative
40/F	Astrocytoma age 12 s/p chemoXRT Neurofibroma thigh age 31 Brainstem gliosarcoma age 35 s/p chemoXRT Pre-B ALL age 37	Sister lymphoma age 32 Sister melanoma age 35	ALL sample: KRAS c.38G>A p.G13D Gliosarcoma: BRAF c.1799T>A p.V600E	46,XX,add(8)(p21),−9,i(9)(q10),t(14;19)(q32;p13),+der(?)t(9:?)(p13;?)[18]	Yes	MLH1, MSH2, MSH6, PMS2, EPCAM, NF1, TP53 all negative
25/M	T-cell ALL and synchronous low-grade astrocytoma age 24	Daughter with rhabdomyosarcoma age 2	TP53 c.524G>A p.R175H NOTCH1 c.4754T>C p.L1585P NOTCH1 c.7003dupC p.L2335fs*19	1~45,XY,−1,add(1)(p13),add(1)(p36.1),add (5)(q31),der(6)add(6)(p12)dup(6) (q23q23),+7,−12,−16,del(17)(p11.2),−19,+21,+1~2mar[cp13]/46,XY[7]	Yes	Germline TP53 p.R175H mutation *93.8% allelic frequency, presumed LOH
71/F	DCIS age 57 s/p surgery and tamoxifen CLL age 66 Bilateral papillary thyroid cancer age 69 s/p RAI Pulmonary neuroendocrine tumor age 70	Brother with bronchioalveolar carcinoma age 66 Brother with testicular cancer age 64 Father with lung cancer, nonsmoker at 69	N/A	46,XX[20]	No	TP53 and PTEN negative
45/M	Rhabdomyosarcoma age 42 s/p adriamycin/ifosfamide × 6 cycles and 50 Gy XRT and surgery Acute Myeloid Leukemia age 44	Sister died at age 3 of cancer Father died at age 43 with cancer of unknown primary	TP53 c.800G>A p.R267Q (59.2% allelic frequency) TP53 c.467G>A p.R156H (54.9% allelic frequency) TET2 c.4138C>T p.H1380Y	44,XY,del(5)(q13),add(7)(q11.2), −11,−12,−17,−17,+r,+mar[18]/44,XY,del(5)(q13q33),add(7)(q11.2),−11,−12,−17,−17,+1~2mar[cp2]/	Yes	Both TP53 mutations identified in germline and clinically classified as VUS *clinical TP53 testing in his mother initiated for additional characterization of inheritance
62/M	Liposarcoma age 24 s/p XRT and surgery Leiomyosarcoma R leg Myxoid malignant fibrous histiocytoma age 60 CLL/SLL age 60	Identical twin brother with CLL age 55 and leiomyosarcoma right leg age 58. Deceased age 61. Sister with DCIS age 49. Niece with melanoma age 12. Mother with breast cancer age 44, deceased age 46	N/A	N/A		TP53 negative
52/F	CLL age 44	Brother with sarcoma age 20 Mother with breast cancer age 36	TP53 c.743 G>A p.R248Q (17.5% allelic frequency)	46,XX[20]		TP53, BRCA1 and BRCA2 negative
28/F	Diploid AML diagnosed during pregnancy age 22	Nephew with sarcoma age 14	N/A	46,XX[20]		TP53 testing negative
73/M	Gastric cancer age 45 s/p partial gastrectomy Prostate cancer age 68 s/p surgery Hodgkin Lymphoma age 70 s/p chemotx Marginal zone lymphoma age 72 s/p Rituxan Myelodysplastic syndrome age 73	One daughter with meningioma One brother with prostate cancer age 60	TP53 c.614A>G p.Y205C (21.2% allelic frequency) TET2 c.3765C>A p.Y1255*	45,XY,add(2)(p12),−5,−7,t(11;17)(q13;p11.2),+mar[7]/46,XY[13]		Declined testing
57/M	CML diagnosed age 54	Sister deceased breast cancer age 37 Maternal aunt with breast cancer age 70	BRCA1/2 and TP53 tumor testing negative	46,XX,t(9;22)(q34;q11.2)[18]/46,XX[2]		TP53, BRCA1 and BRCA2 negative

Table 5.

HHMC referral for patient with a previously identified hereditary cancer syndrome

Age/Sex	Significant PMHx	Family History	Previously Identified Germline Mutation
57/M	Lifetime mild thrombocytopenia (70–90 × 109/L) Recurrent sinus infections	Daughter with lifetime severe thrombocytopenia	RUNX1 p.W279*
27/M	Lifetime mild thrombocytopenia (80–120 × 109/L) and elective surgeries complicated by major bleeding Chronic sinusitis and recurrent ear infections Presumptive dx of ITP	Son and daughter with severe thrombocytopenia identified at age 1 and 4	RUNX1 p.T104*
40/F	Age 30: Stage II ER/PR+ invasive ductal breast cancer s/p neoadjuvant Taxotere and Xeloda followed b FEC (5-FU, epirubicin and cyclophosphamide) followed by surgery, radiation therapy and Tamoxifen. Age 35: Fibromyxoid sarcoma of R thigh s/p neoadjuvant Adriamycin/ifosfamide and XRT. Age 37: Spindle cell sarcoma R thigh s/p gemcitabine and Taxotere Age 40: Acute myeloid leukemia with complex cytogenetics	Son deceased age 6 medulloblastoma Daughter deceased age 12 metastatic adrenal gland tumor Mother deceased age 35 gastric cancer Maternal grandmother deceased age 45 gastric cancer	TP53 deletion spanning exons 10 through 11 *not apparent on NGS-based AML prognostication panel
42/M	Lifetime mild thrombocytopenia (70–80 × 109/L)	Mother with AML age 60. Maternal grandmother with AML age 45. Sister and niece with lifetime thrombocytopenia	RUNX1 p.R166*

One 24-year-old male with a synchronous diagnosis of hypodiploid T-cell ALL and low-grade astrocytoma had a confirmed germline deleterious TP53 p.R175H mutation identified. His NGS panel notably identified the p.R175H mutation with an allelic frequency reported as >90%, presumably due to somatic loss of heterozygosity (LOH). (Figure 2f)

A 44-year-old man with rhabdomyosarcoma diagnosed at age 42 treated with surgery, chemotherapy and radiation was referred to the HHMC at the time of diagnosis of therapy-related AML with complex cytogenetic abnormalities. His sister died at age 3 of an unknown malignancy, and his father died at age 43 of cancer of unknown primary. His leukemia NGS panel identified two TP53 variants, p.R267Q and p.R156H. TP53 testing of cultured fibroblasts demonstrated both variants present in his germline. In-silico analysis of p.R267Q lists this change as damaging (SIFT) and probably damaging (Polyphen), and recent analysis has demonstrated <10% TP53 activity with this mutant.(28) Likewise, in-silico analysis lists p.R156H as damaging (SIFT) and probably damaging (Polyphen), with <20% TP53 activity. Thus, a diagnosis of LFS is considered very likely and possibly related to two pathogenic germline TP53 variants; however, this is currently unable to be clinically confirmed. One additional patient declined germline testing, and the remainder of evaluated patients tested negative for germline TP53 abnormalities.

Patients with >1 first-degree relative or >2 2^nd degree relatives with hematologic malignancies (Group D)

This diverse cohort includes patients from 11 families with MDS/AML, 2 families with JAK2 p.V617F-mutated MPNs, and 8 families with multiple affected members having B-cell malignancies including CLL and NHLs. One of the 11 families with MDS/AML was identified to have a RUNX1 heterozygous 5′-end deletion of at least exons 1–6 (Figure 2g). While this specific deletion has not been previously reported, other exon-level deletions spanning the RUNT homology domain of RUNX1 are confirmed pathogenic variants leading to FPD/AML. Once again, the NGS-based AML prognostication panel performed on the index patient in this family notably failed to identify this large RUNX1 deletion.

In the index patient from another of the 11 MDS/AML families, which includes two brothers diagnosed with diploid AML at age 64 and 67 (Figure 2h) and mother with non-Hodgkin lymphoma at age 84, a previously described pathogenic start-loss substitution in DDX41 (c.3G>A, p.M1I) was detected.(29)

Research-based and iterative clinically-actionable sequencing is ongoing for the remaining families. Notably ETV6 and DDX41 was not part of the original clinical panel evaluated in the majority of the MDS/AML families. As it is now possible to sequence these genes in a CLIA-approved manner on cultured skin fibroblasts, this is being performed in all MDS/AML families above as clinically possible.

Referral for evaluation and monitoring in patients with known/previously identified HCS Group E)

In addition to the female patient with LFS described above who was referred for additional counseling and evaluation, three adult individuals were referred for follow-up and care related to germline RUNX1 mutations leading to FPD/AML. In all three instances, either their children or a sibling’s child was first diagnosed with FPD/AML after evaluation for childhood thrombocytopenia and bleeding complications, and appropriate clinical screening of family members led to their FPD/AML diagnosis. All three have a longstanding history of mild to moderate thrombocytopenia with variable bleeding/bruising propensity. With current ages of 27, 42 and 57, they continue to be followed with laboratory work and clinical evaluation biannually.

Discussion

Over 18 months, 67 patients from 65 families with hematologic malignancies were evaluated for known or suspected hereditary cancer predisposition syndromes in a dedicated clinic, providing several important and valuable insights. First, while overall rare, eight adolescents and adults (12%) were diagnosed with a HCS by the HHMC within this time interval (Figure 2), offering evidence that with improved awareness of hereditary hematologic malignancies, more patients who may benefit from genetic counseling and testing can be referred, and careful attention to clinical and pathologic information can identify high-risk patients and families.

Second, many of the referred patients have a striking family history of hematologic malignancies, yet no known HCS was identified, supporting the notion that additional germline genetic aberrations exist which remain to be identified. Much insight can be gained regarding function and importance of involved genes and pathways through these discoveries. The 700-kb duplication leading to overexpression of ATG2B and GSKIP, increasing progenitor sensitivity to thrombopoietin (TPO) and leading to increased risk of MPN exemplifies this opportunity.(30) Available clinical-based screening must be iterative and continually updated based on available evidence to incorporate validated discoveries. Panel-based genetic testing (e.g. http://dnatesting.uchicago.edu/) is clinically available and provides a cost-effective approach incorporating known relevant genes. In patients and families for whom a predisposition syndrome is suspected but available clinical testing is negative, consideration of research-based testing and ongoing periodic surveillance is advisable.

Third, HCS identification is not only beneficial for the patient with regards to informing appropriate treatment strategies and selection of sibling transplant donors, but may also benefit family members. Screening potentially-affected individuals within confirmed HCS families can provide clarity and surveillance opportunities for affected yet asymptomatic family members. Current expert opinion guidelines for carriers recommend a bone marrow biopsy at diagnosis, biannual physical exam and CBC, and repeat bone marrow evaluation at the time of any changes from baseline.(31) Evaluation for clonal hematopoiesis in patients with predispositions to hematologic malignancies, which can for example be detected in >80% of asymptomatic germline RUNX1 carriers by age 50, may also provide utility as a mechanism of disease surveillance.(32)

Fourth, patients with presumed somatic mutations identified on NGS-based cancer panels may instead have a germline event, and evaluation of the patient must consider this possibility. Germline inheritance may be more likely in situations of biallelic mutations or near-heterozygous or homozygous allelic frequency, the latter due to somatic LOH. This is an area of ongoing research; notably the “threshold” allelic frequency to warrant germline evaluation is not standardized and further research is needed. It is essential to remember that NGS-panel testing often fails to identify large deletions/duplications, unless specific assay designs and bioinformatics pipelines are utilized. Mutation panels intended for somatic testing for prognostication purposes are not definitive for the germline analysis of patients, and at best can only be used to supplement high clinical suspicion to initiate germline analysis.

Fifth, several patients with ostensibly somatic acquisition of GATA2 mutations experienced frequent atypical infections including disseminated viral and fungal infections and flow cytometry results consistent with monoMAC, suggesting a GATA2-specific clinical phenotype, regardless of constitutional or acquired nature. To the best of our knowledge, this is the first report describing a cellular deficiency phenotype in patients without identified constitutive GATA2 mutations.

Finally, the method of obtaining germline DNA is of paramount importance for patients with hematologic malignancies. Blood is not acceptable due to the inability to distinguish germline from somatic mutations, and saliva is not recommended due to frequent tumor contamination.(33) Skin fibroblasts are the most appropriate germline controls. In our clinic to date, only two patients declined germline evaluation, neither due to the procedural requirement, suggesting this is not a significant burden for most patients. We did not experience any 4mm punch biopsies leading to insufficient fibroblast culture growth or DNA procurement, suggesting even smaller biopsies in the future may be sufficient. The time required to attain results from a skin biopsy must be considered, as it typically requires 6–8 weeks from the time of procedure to obtaining results.

In conclusion, four new patients/families were diagnosed with FPD/AML, and one patient each with DC, FA, LFS and germline DDX41-related MDS/AML. Additionally, we identified two patients with germline VUS results, one each in RUNX1 and TP53, in patients/families with strong clinical concern for these syndromes. Improved understanding of the genetic basis of hereditary cancer predisposition syndromes provides additional diagnostic information, insights into treatment strategies, and more accurate evaluation and risk assessments to family members. Surveillance of affected yet asymptomatic family members will lead to improved understanding of the mechanisms of cancer progression and should be the focus of improved surveillance and prevention techniques.

Clinical Practice Points.

• Multiple predispositions to hematologic malignancies exist, yet evaluation for hereditary cancer syndromes (HCS) in patients with hematologic malignancies are underperformed by most hematologist/oncologists.

• In patients with hematologic malignancy, reasons for referral to genetic counseling should include: bone marrow failure or myelodysplastic syndrome in patients ≤ 50 years, strong personal and/or family history of malignancy, and testing for germline status of mutations (i.e. CEBPA, GATA2, and RUNX1) identified on next-generation sequencing (NGS) panels intended for prognostication.

• Patients with presumed somatic mutations identified on NGS-based cancer panels may instead have a germline event.

• Cultured skin fibroblasts are the most appropriate germline DNA source for patients with hematologic malignancies.

• HCS Identification can inform appropriate treatment strategies and selection of sibling donors, and may also provide screening and surveillance opportunities for affected family members.

• Clinically available HCS testing is limited to the current state of knowledge, thus evaluation and genetic testing of patients with concern for a HCS must include periodic surveillance and iterative testing.

• With improved awareness of hematologic malignancy predisposition syndromes, more patients who may benefit from genetic counseling and testing can be identified; careful attention to personal and family history and pathologic data can identify high-risk patients and families.