The Value of Next-Generation Sequencing in the Screening and Evaluation of Hematologic Neoplasms in Clinical Practice

Victoria Northrup, MSc; Allison Maybank, MSc; Nancy Carson, PhD; Tarek Rahmeh, MD


Am J Clin Pathol. 2020;153(5):639-645. 

In This Article


This retrospective review shows that the overall likelihood of obtaining positive NGS in suspected myeloid disorders was much higher when testing was performed on bone marrow aspiration than on peripheral blood (72% vs 47%). This difference, however, might be at least partially attributed to the variation in clinical indications between the different types of samples. Our frequency of identified variants in bone marrow was similar to a previously reported frequency.[5] The frequency in peripheral blood has not been reported previously. The review also shows that NGS is a significantly more sensitive test for the screening and evaluation of hematologic neoplasms than conventional CG.

In addition, the likelihood of NGS yielding positive results was strongly dependent on the associated peripheral blood presentation, ranging between 0% in patients with thrombocytopenia only (0/6) and 100% of patients with circulating blasts in the peripheral blood (26/26).

There was a strong association between the clinical indication and the likelihood of NGS yielding positive results. The highest yield was in de novo AML (100%), followed by AML follow-up (78%). Patients with anemia, cytopenias, leukocytosis, polycythemia, and mixed presentations also had a significant rate of positivity, ranging between 20% and 69%. The lowest yield was among cases of thrombocytopenia and neutropenia (0% and 14%, respectively).

Of note is that the clinical indication did not always mirror the exact peripheral blood picture at the time of testing for various reasons, such as fluctuations in blood counts, whether there was a recent blood transfusion, and compounding factors such as recent infection.

Molecular genetics have become an integral component of the classification schemes for AML, MPNs, and MDS/MPN. Furthermore, certain gene variants provide significant prognostic information for these disorders and aid in directing the type of therapy needed for patients. Therefore, testing for specific gene variants has become part of the standard of care in these patients.[1,12] While in theory it is still technically possible to perform standalone polymerase chain reaction–based molecular tests for virtually all genes implicated in myeloid neoplasms, such an approach would be too cost prohibitive and labor intensive to be implemented in clinical practice, hence the adoption of NGS in our laboratory. Although there are myeloid panels available with more genes than the AmpliSeq and Oncomine myeloid panels,[3,6] these panels provide information on clinically valuable variants in a cost-effective manner, making them ideal for a clinical setting.

The value of molecular genetics continues to evolve in MDS. They have, however, fallen short of playing a significant diagnostic role, which still relies heavily on bone marrow morphology and conventional CG. This is mainly due to the limited specificity of MDS-associated gene variants, which can be encountered in other myeloid neoplasms, as well as their occasional detection in otherwise healthy elderly individuals.[13]

Notwithstanding this limitation, NGS in our institution was performed in several patients as a screening tool for myelodysplasia on peripheral blood samples to distinguish benign and reactive conditions from potentially neoplastic ones, thus preventing unnecessary invasive procedures to many patients while providing prognostic information for patient management if a myelodysplasia is confirmed. While this approach appears to be promising in certain clinical presentations, such as anemia, in which 55% of cases tested were positive, in other types of presentations, such as thrombocytopenia and neutropenia, where the yield is much lower, NGS does not appear to be a cost-effective first-line approach. The number of cases in this group of patients, however, is too small to draw definite conclusions, and evaluation of larger groups of patients will be required.

In several other patients in whom a full hematologic workup was undertaken and a diagnosis of MDS was confirmed, NGS in our experience had provided valuable prognostic information regarding risk of transformation to AML and overall survival.

In 29% of positive NGS cases, variants detected had not been reported previously, and 28% of those were categorized as VUS at the time of the study. When taking into account previously reported variants, 34% of all positive NGS studies revealed at least one VUS. Considering the relative novelty of NGS in clinical hematology practice, these findings are not unexpected. We predict that the clinical value of NGS will exponentially increase over the next few years as more and more laboratories implement NGS in their practices and more of these variants are reported in the medical literature, along with their clinical context, or shared in national and international databases.

Previous studies have assessed the value of targeted NGS in the diagnostic workup of myeloid neoplasms, and their results have confirmed the ability of this relatively novel tool to detect clinically actionable genetic lesions in a significant proportion of patients.[8–10] While these studies were essentially restricted to confirmed cases of hematologic malignancies, our cohort of patients encompasses a wide range of clinical presentations, such as isolated anemia or neutrophilia, in which MDS or MPN was only part of a broad differential diagnosis. We believe that our approach will contribute to the evaluation of NGS as a first-line screening tool for suspected hematologic neoplasia.

Although this study provides evidence of the importance of NGS in clinical diagnostics of myeloid neoplasms, it is not without its limitations. In 22 cases, clinical information was not available to the researchers and therefore could not be used to determine sensitivity or specificity. In addition, the calculations on the sensitivity of CG vs NGS were limited to the samples that had both performed. Since many of the peripheral blood samples did not have accompanying bone marrow specimens, these cases could not be used to compare the two assays. In addition, there were clinical presentations for which we had only a handful of cases, and therefore drawing conclusions regarding the utility of NGS for these presentations will require further follow-up in a larger cohort of patients.

In conclusion, this study summarizes our experience in applying NGS in clinical hematology practice and evaluates its role as a tool for the detection of MDS, confirmation of MPNs, and the diagnostic assessment, prognostic stratification, and posttreatment follow-up of AML and other myeloid neoplasms. The role of NGS is expected to expand as more gene variants are reported in clinical practice and further information regarding their pathogenicity becomes available.