How does an expanded targeted gene panel influence clinical decision making for patients with AML?

Background

AML remains a challenging hematologic malignancy to treat, with many patients exhibiting a poor prognosis and poor clinical outcomes. At all stages of AML treatment, including during first-line treatment, ongoing response reassessment, and both refractory and relapse disease, genetic mutations inform prognosis and treatment decisions. Some of these decisions are clear; for example, specific targeted therapies have been approved for AML with FLT3 mutations. However, other genetic mutations have a less clear role; they may help determine the prognosis and the presence of high-risk disease, but not they do not directly influence treatment choices.

The genetic landscape of AML is complex, with over 250 genetic aberrations described,¹ and there is a high degree of heterogeneity between patients, as well as clonal variation throughout each patient’s treatment journey. Therefore, in AML, genetic information can influence prognosis, risk stratification, and treatment in many ways. However, using this information can be challenging due to difficulty interpreting the clinical significance of genomic terminology, a lack of report standardization, and a lack of established and clinical trial data to reference.

With a view to exploring this issue, Erica Barnell and colleagues presented results from a survey of clinicians responsible for managing patients with AML, exploring how the results of a 40-gene targeted panel test influenced their treatment decisions.¹ These results were published in Precision Medicine in January 2021, and here we summarize the findings.

Study design

This study was a single-center study performed at Washington University School of Medicine over an 8-month period between August 2018 and April 2019. During this time, all consecutive reports ordered at any stage of treatment were analyzed. The panel analyzed 40 genes or genetic hotspots (including gene variation/variant allelic frequency). A survey to evaluate the use and interpretation of the targeted genetic panel in clinical decision making was sent to all physicians treating a patient with a primary diagnosis of AML. Patients with acute promyelocytic anaemia (M3 AML) were excluded.

Results

In total, 109 patients were identified during the study period (58 male [53.2%]; 51 female [46.8%]), and these patients were managed by a total of 14 physicians. Following exclusion, 122 gene panel reports were analyzed, which had a median turnaround of 12 days (range, 4–36 days). In 11% of patients, the report was requested on more than one occasion.

Survey responses were provided for 120 of 122 reports issued (98% response rate). Of these 120 reports:

• 114 reports were taken forward for analysis;

• two patients were lost to follow-up; and
• four patients declined treatment.

Of the 122 panels, only 13 samples had no observed variants; eight of these samples were from patients in clinical remission. The median number of failed genes across all reports was two (range, 0–24). These genes included WT1 (n = 211), CUX1 (n = 197), and CEBPA (n = 111). Furthermore, only two genes on the panel were not identified in any report.

Influence of reports on clinical decision making

Of the 114 sequencing reports analyzed, physicians reported that 50 (44%) influenced clinical decision making. These 50 reports affected 56 decisions, in the following ways:

• 38 influenced treatment choice;
• ten influenced risk stratification in first clinical remission; and
• eight influenced the ongoing measurement of persistent molecular disease.

Treatment decisions

In total, 35 of the 50 reports that influenced clinical decision making did so by influencing treatment decisions, and three patients were influenced more than once. This included:

• 14 patients who were recommended a hypomethylating agent due to a TP53/TET2 variant;
• 12 patients with a FLT3 mutation who were recommended midostaurin/gilteritinib;
• eight patients who were recommended IDH1/IDH2 inhibitors;
• one patient with a KIT variant who was recommended midostaurin/gilteritinib;
• and three patients who were recommended for a clinical trial due to the presence of unusual variants or persistent disease.

Decisions about transplant in first remission

For ten patients, the panel results influenced a decision about progression to allogeneic hematopoietic stem cell transplantation (allo-HSCT). Allo-HSCT was not recommended in six patients in full clinical remission based on:

• biallelic CEBPA variants in two patients;
• KIT variant status in one patient;
• lack of high-risk variants in one patient; and
• disappearance of co-variants present at diagnosis in one patient.

Allo-HSCT was recommended for four patients in full clinical remission based on the panel results.

Conclusion

The results of this study indicate that physicians responsible for making clinical decisions in the management of patients with AML are increasingly utilizing genetic information identified through genetic panels, including co‑occurring variants and established genetic aberrations. Furthermore, decisions are being influenced at every stage, from diagnosis through to complete remission. There is also a clear role for targeted-genetic panels alongside standard of care analyses with regards to monitoring for residual molecular disease.

The study itself describes several limitations. It was a single-center study, involving physicians with a high degree of genomic literacy and experience. Furthermore, among these physicians, there was a high degree of variance in the number of patients being managed by each physician (range, 1–18) and there was a large range in the number of cases reported to be influential (range, 0–90%). These limitations suggest there may be difficulty translating the findings to other hematologic centers, particularly as physicians will have differing levels of experience with interpreting genomic information.

Further work is required to bring expanded targeted genetic panels into clinical practice alongside the current standards of care and staging criteria, including larger multicenter studies to truly assess the impact and utility of genetic panels in clinical decision making.