Identifying potential germline variants from sequencing hematopoietic malignancies

Abstract

Next-generation sequencing (NGS) of bone marrow and peripheral blood increasingly guides clinical care in hematological malignancies. NGS data may help to identify single nucleotide variants, insertions/deletions, copy number variations, and translocations at a single time point, and repeated NGS testing allows tracking of dynamic changes in variants during the course of a patient's disease. Tumor cells used for NGS may contain germline, somatic, and clonal hematopoietic DNA alterations, and distinguishing the etiology of a variant may be challenging. We describe an approach using patient history, individual variant characteristics, and sequential NGS assays to identify potential germline variants. Our current criteria for identifying an individual likely to have a deleterious germline variant include a strong family history or multiple cancers in a single patient, diagnosis of a hematopoietic malignancy at a younger age than seen in the general population, variant allele frequency > 0.3 of a deleterious allele in a known germline predisposition gene, and variant persistence identified on clinical NGS panels, despite a change in disease state. Sequential molecular testing of hematopoietic specimens may provide insight into disease pathology, impact patient and family members' care, and potentially identify new cancer-predisposing risk alleles. Ideally, individuals should give consent at the time of NGS testing to receive information about potential germline variants and to allow future contact as research advances.

Graphical abstract

Figure 1.

Illustrative case of deleterious DDX41 variants identified during clinical evaluation of a hematopoietic malignancy. (A) Family history revealed 2 cancers in the patient/proband (red circle), a blood cancer of unclear nature in the father, and a head and neck cancer in the paternal grandmother. (B) Molecular profiling via a 150-gene clinical NGS panel identified 171 total variants. After annotation, filtering for clinical relevance, and individual verification by the in-house pathologist, 7 (4%) were reported as variant of uncertain significance, and the 2 (1.1%) DDX41 variants were reported as pathogenic on a final document provided to the treatment team. (C) DDX41 allele VAF is graphed throughout the patient’s clinical course for the 2 variants identified. AML, acute myeloid leukemia.

Figure 2.

Etiology of DNA alterations in a representative sample of NGS sequencing. The box contains a population of cells (each circle) undergoing bulk NGS. Each color represents a DNA alteration of different etiology. VAF for the representative example is shown with typical ranges observed on clinical NGS panels. This example does not account for CNVs or DNA structural aberrations. *Lower detection limit depends on the depth and platform of NGS.

Figure 3.

A complex clinical case and VAF of deleterious variants seen over time. A 51-year-old white man had an 8 × 10–cm mass that was determined to be diffuse large B-cell lymphoma (DLBCL). He received 6 cycles of rituximab/cyclophosphamide/doxorubicin/vincristine/prednisone (R-CHOP), 2 cycles of etoposide/methylprednisolone/high-dose cytarabine/cisplatin (ESHAP), and radiation to the mediastinum, ultimately achieving a complete response. At age 57, a screening prostate-specific antigen (PSA) was 12.4 ng/mL. Prostate biopsy showed a 4 + 4 = 8 Gleason score adenocarcinoma, and the patient had a prostatectomy with normalization of his PSA. At age 61 years, he was diagnosed with essential thrombocytosis, with JAK2 p.Val617Phe. He eventually progressed to AML, when a detailed family history was obtained. (A) Family history revealed numerous relatives with cancer: mother, breast cancer (55 years old); father, prostate cancer (69 years old); maternal aunt, ovarian cancer (37 years old); maternal cousin, unknown type of leukemia; paternal grandmother, uterine cancer; paternal grandfather, head and neck cancer; paternal aunt, breast cancer (70 years old); paternal cousin, brain tumor (75 years old); paternal cousin, lymphoma (70 years old); paternal cousin, unknown type of leukemia (12 years old). Molecular profiling at AML diagnosis showed a complex karyotype, including deletions of the long arms of chromosomes 5 and 7. NGS of predominantly leukemia cells from a BM biopsy showed a TP53 mutation and a deletion within BRCA1. The patient underwent induction chemotherapy, and molecular profiling at clinical remission demonstrated persistence of the BRCA1 deletion and loss of the TP53 mutation. Germline genetic testing on DNA derived from the patient’s cultured skin fibroblasts confirmed a germline BRCA1 deletion. He underwent an allogeneic HSCT using an unrelated donor, given the potential risk of the familial BRCA1 deletion, which had been found in an HLA-matched sibling. (B) The VAF of DNA alterations are plotted over time and show persistence of the germline BRCA1 deletion at a relatively high VAF prior to HSCT; the acquired clonal JAK2 and TP53 variants prior to HSCT; and an acquired TSC2 variant post-HSCT of donor origin. Lessons from this case include: (1) The patient was diagnosed with 3 cancers by the time germline testing was performed: DLBCL, prostate cancer, and AML. Genetic counseling and testing were warranted at the time of his first cancer based on his extensive family cancer history. (2) BRCA1 and BRCA2 are Fanconi anemia-like genes, encoding proteins important for DNA repair pathways active in the BM. Individuals with BRCA pathway mutations are at increased risk for the development of hematopoietic malignancies. In fact, cancer predisposition syndromes generally thought of as predisposing to solid tumors also increase the risk for hematopoietic malignancies, such as Lynch and Li-Fraumeni syndromes.

Figure 4.

VAFs for TP53, BRCA1, and BRCA2 and over time in patients with hematopoietic malignancies. VAFs of DNA alterations in TP53 (A), BRCA1 (B), and BRCA2 (C) in individual patients at the University of Chicago are graphed over time. Each point indicates an individual variant identified in an in-house NGS assay, and red lines connect likely somatic variants; likely germline variants are shown in blue.

Figure 5.

Suggested algorithm for identifying patients with a deleterious germline cancer predisposition variant. When a patient is diagnosed with a hematopoietic malignancy, clinical history and tumor biopsies are performed. Personal history of prior cancer (1 other hematopoietic malignancy or solid tumor, including melanoma in an individual younger than 50 years of age), diagnosis at a younger age than seen in the general population for a given cancer, or a strong family history of cancer (relative diagnosed with cancer within 2 generations of the patient) should prompt a skin biopsy and comprehensive germline testing. If tumor-only NGS identifies a known cancer-predisposition variant and the VAF is > 0.3, germline testing of the variant should follow. As additional NGS tests are performed to monitor the patient’s clinical course, persistent deleterious variants with VAF > 0.3 should prompt consideration of germline status. This is especially warranted if the deleterious variant is present in a gene associated with cancer risk. In the future, systematic collection of a skin biopsy at the time of the initial BM biopsy and culturing of fibroblasts to obtain germline DNA may become standard (dotted line). Once a deleterious germline variant is confirmed, variant/gene-specific surveillance should be followed for the patient (including a risk assessment for cancer involving organs outside the BM), genetic counseling and germline testing should be offered to appropriate family members, and potential risks should be considered if the patient were to undergo related HSCT from a family member sharing the allele. NGS, next-generation sequencing; VAF variant allele frequency. *Comprehensive testing that includes all genes and variant types that confer cancer risk is not standardized and requires careful review of testing options.