This session discusses the increasingly common use of NGS panels in the evaluation of patients with peripheral blood cytopenias and the utility of such panels in the diagnosis of clonal hematopoiesis of indeterminant potential (CHIP), clonal cytopenias of uncertiain significance (CCUS) and myelodysplastic syndromes (MDS).
Below is a list of questions received from the live webinar attendees. Click on any question to reveal the speaker's response.
We classify all mutations following the AMP/CAP/ASCO guidelines for somatic variants in cancer (J Mol Diagn. 2017;19:4-23). Broadly speaking, missense mutations that are outside of known hotspots are often classified as being of unknown clinical significance. Similarly, for frameshift mutations where truncating variants are a known mechanism of disease, most examples will be classified as being of strong or potential clinical significance.
In the current WHO classification, establishing a definitive diagnosis of MDS requires evidence of morphologic dysplasia or the presence of an MDS-defining cytogenetic abnormality. Somatic mutations in the absence of these findings may represent CHIP or CCUS rather than an overt MDS. Normal results using a well-designed NGS panel have a high negative predictive value for MDS, such that the absence of MDS-associated clonal mutations suggests cytopenias are more likely to be non-neoplastic.
While DNMT3A, TET2 and ASXL1 are the most frequent CHIP mutations, many other genes may also be mutated in CHIP (including JAK2 and TP53).
In the absence of morphologic dysplasia or an MDS-defining cytogenetic abnormality, cases of cytopenia with somatic mutations would be classified as clonal cytopenia of uncertain significance (CCUS).
CEBPA is difficult to sequence as it has many very GC-rich areas.
CCUS and MDS cannot be distinguished based on molecular data alone. MDS is distinguished from CCUS by the presence of morphologic dysplasia in the peripheral blood and/or bone marrow or the presence of an MDS-defining cytogenetic abnormality.
In the current WHO classification, establishing a definitive diagnosis of MDS requires evidence of morphologic dysplasia or the presence of an MDS-defining cytogenetic abnormality. It remains an open question whether data such as numbers of mutations or VAF should be added to our criteria to establish a diagnosis of MDS in the future.
In general, peripheral blood and bone marrow demonstrate similar genetic findings in MDS. In our experience, it is rare to have cytopenias so profound that adequate DNA may not be obtained. In the presence of marked neutropenia, bone marrow may be the preferred sample.
In chronic myeloid neoplasms, we see a median number of 2 Tier 1/Tier 2 mutations per case. VUSs are generally even fewer. When we first began NGS testing, it would frequently take us an hour or more per case to investigate and annotate our variants. As we have improved our informatics pipeline, reduced artifacts, built up a database of variant interpretations, and generally gained experience, many cases can be annotated in perhaps 15 minutes.
With the exception of selected codons (such as JAK2 V617F), our informatics pipeline filters out all variants present below 4% VAF. All variant calls are manually reviewed to eliminate artifacts.
Many genes that are somatically mutated in MDS may also be mutated in the germline (e.g. RUNX1, TP53, ETV6, GATA2, CEBPA). Whether additional germline predisposition genes should also be included is a difficult question that should be addressed through detailed discussions amongst your hematopathologists, molecular pathologists and clinicians.
As the number of genes associated with germline predisposition to myeloid malignancy is quite large, specific panels for such disorders can be clinically useful. In our experience, the observed VAFs for somatic mutations are quite similar in buffy coat versus straight bone marrow aspirate material.
VAF data alone is generally insufficient to distinguish between somatic and germline status with certainty when sequencing a neoplastic sample. When we identify a variant that suggests the possibility of a germline condition, we will often suggest repeat testing on skin fibroblasts or other germline sample.
We classify all mutations following the AMP/CAP/ASCO guidelines for somatic variants in cancer (J Mol Diagn. 2017;19:4-23).
Within our hematopathology group, our practice is to order an NGS panel on all new cases of AML and, when required to reach a diagnosis, MPN. In patients with cytopenias and/or suspected MDS, NGS panels are typically ordered by our clinical teams when wishing to assess prognosis, screen patients for clinical trials, or select a potential targeted therapy for treatment. Any testing strategy implemented within your pathology practice should also be discussed with your clinical teams.
Interpreting splice site alterations from DNA sequencing alone can be challenging. In general, if we see an alteration that eliminates a canonical splice site in a gene where splice variants are a known mechanism of disease, we will typically classify such changes are being of clinical significance. Other types of splicing variants may be interpreted as a VUS.
Our assay has a limit of detection of 5% VAF at most loci. JAK2 V617F has been validated down to 1% VAF.
We will perform testing on either peripheral blood or bone marrow. If both are available, bone marrow is generally preferred.
We do not routinely perform flow cytometry for suspected MDS in our hematopathology group.
I anticipate that metaphase karyotyping will remain a critical part of bone marrow evaluation for many years. When we implement NGS-based fusion detection, we will likely reduce but not eliminate the need for FISH testing.
Establishing a diagnosis of MDS, including hypocellular MDS, requires morphologic dysplasia and/or an MDS-defining abnormality by metaphase karyotyping. Mutational data alone does not distinguish between hypoplastic MDS and aplastic anemia. With that said, the spectrum of mutations observed in aplastic anemia is somewhat different than those found in MDS and so mutational data may offer a clue in challenging cases even if it does not allow for a definitive diagnosis.
Our clinical teams will generally follow patients with ICUS or CCUS relatively closely (for example, every 3-6 months).
In MDS, we see similar findings in peripheral blood and bone marrow samples.
For myeloid testing, the core set of 34 genes suggested by the AMP chronic myeloid neoplasm working group is a good starting point (J Mol Diagn. 2018;20:717–737).
We do not currently examine gene expression levels by real-time PCR for MDS in our clinical laboratory.