In Myeloid Neoplasms, Next-Generation Sequencing Can Be Inappropriate—and Money-Wasting

Using more careful consideration when deciding whether or not to order next-generation sequencing (NGS) for a patient with an actual or suspected myeloid neoplasm (MN) could save institutions thousands of dollars annually without compromising care, according to a Yale School of Medicine team. They have created a set of criteria to determine the appropriateness of NGS testing for MN (MN-NGS), with the goal of maximizing actionable results. Writing in eJHaem, the team also noted that the results of NGS, when not clinically indicated but performed anyway, can foster fruitless investigative paths and amplify patient anxiety.¹

Close-up illustration of a MN | image credit: sawaratch - stock.adobe.com

They noted that, indeed, MNs often harbor pathogenic mutations that go undetected by karyotyping and fluorescence in situ hybridization, and NGS is truly necessary for diagnosis, risk stratification, and therapy.² Among the 6 situations that would, if present, warrant NGS (ie, approval criteria) would be clinical suspicion of new, relapsed, or worsening disease, and end-of-induction chemotherapy.

The 6 “cancellation criteria”—situations in which these investigators say NGS clearly should not be performed—include, first, having a suspicion of only nonmyeloid disease (ie, the diagnosis is a nonmyeloid disease, or there is no suspicion for acute myeloid leukemia [AML], myelodysplasia, myeloproliferative neoplasm [MPN], or another MN). Other situations are having no suspicion of progression of a known MN; no evidence for recurrence post-transplant; a diagnosis of chronic myeloid leukemia (CML) with no concern for AML; and cases using blood when a concurrent bone marrow NGS is being performed. The 6^th and final criterion is that none of the above cancellation criteria have been met, but no approval criteria have been met either.

The actionable results that should emerge from NGS done for the proper reason include making a new MN diagnosis, characterizing a MN with baseline mutational status for follow-up purposes, and altering a patient’s treatment plan, noted the investigators.

How they tested their hypothesis:

Over 5 months, August to December 2018, the researchers applied their criteria to all (n = 174) MN-NGS tests ordered at their institution (5 other tests were cancelled by clinicians upfront). The team’s plan was to see if any tests that met their cancellation criteria yielded actionable results.

They retrospectively categorized the tests as appropriate (group A; n = 115), inappropriate (group B; n = 29), and appropriately canceled (group C; n = 30). In fact, no one's tests in group B tests actionable results, while 75 of the 115 (65%) group A tests did (P < .0001).

In total, then, about one-third (59/174) of all MN-NGS orders could be cancelled without missing any actionable variant, the authors stated, meaning patient care would not be compromised. This would result in CMS-reimbursed savings of just over $150,000 annually, they calculated.

Top approval/cancellation criteria:

In group A, the most frequently applicable approval criteria was the existence of a new diagnosis or clinical suspicion of MDS without increased blasts after another wise negative workup. In group B, the cancellation criterion most applicable was the last on the list: that the order did not meet any approval criteria. For group C, the cancellation criterion most frequent was the diagnosis of a nonmyeloid disease or that there was no suspicion for an MN.

Among the 75 actionable results, the most common outcome by far was that a new or newly recurrent diagnosis was characterized with baseline mutational status for follow-up purposes (73/75; 97%), the team related. Other actionable outcomes included the formulation of a new diagnosis (7/75; 9.3%) and a shift in treatment (10/75; 13%).

“Considering that NGS testing and reporting is time- and resource-intensive, implementing our screening criteria for MN-NGS testing may result in more directed and cost-efficient evaluation of MNs,” the researchers emphasized, adding that larger studies should be undertaken to confirm their findings.

Gene results:

The authors also included an accounting of the gene mutations most commonly found in the patients for whom NGS was ordered.

• For those with a history, or a new diagnosis, of AML at the time of ordering (n = 52), the most commonly mutated genes were RUNX1, SRSF2, and IDH2 (each with 16 variants).
• For those with a history, or a new diagnosis, of MDS at the time of ordering (n = 35), the most commonly mutated genes were TP53, TET2, and SF3B1, with 11, 9, and 4 variants, respectively.
• Among tested patients with a history of or a newly established MPN (n = 9), the most commonly mutated genes were JAK2 (6 variants) and MPL (3 variants). In these patients, the team found no triple-negative MPNs. Seven of nine patients (77.8%) demonstrated additional nondriver variants; among the nondriver variants reported to adversely affect overall survival in MPNs, 1 variant each of SRSF2, ASXL1, EZH2, and SF3B1 were present.

References

1. Gisriel SD, Howe JG, Tormey CA, et al. Optimization criteria for ordering myeloid neoplasm next-generation sequencing. eJHaem. 2024;5(5):1-7. doi:10.1002/jha2.1036
2. Reinig E, Yang F, Traer E, et al. Targeted next-generation sequencing in myelodysplastic syndrome and chronic myelomonocytic leukemia aids diagnosis in challenging cases and identifies frequent spliceosome mutations in transformed acute myeloid leukemia. AmJ Clin Pathol. 2016;145(4):497-506.doi:10.1093/ajcp/aqw016