Impact of a 40-Gene Targeted Panel Test on Physician Decision Making for Patients With Acute Myeloid Leukemia

Abstract

Purpose: Physicians treating hematologic malignancies increasingly order targeted sequencing panels to interrogate recurrently mutated genes. The precise impact of these panels on clinical decision making is not well understood.

Methods: Here, we report our institutional experience with a targeted 40-gene panel (MyeloSeq) that is used to generate a report for both genetic variants and variant allele frequencies for the treating physician (the limit of mutation detection is approximately one AML cell in 50).

Results: In total, 346 sequencing reports were generated for 325 patients with suspected hematologic malignancies over an 8-month period (August 2018 to April 2019). To determine the influence of genomic data on clinical care for patients with acute myeloid leukemia (AML), we analyzed 122 consecutive reports from 109 patients diagnosed with AML and surveyed the treating physicians with a standardized questionnaire. The panel was ordered most commonly at diagnosis (61.5%), but was also used to assess response to therapy (22.9%) and to detect suspected relapse (15.6%). The panel was ordered at multiple timepoints during the disease course for 11% of patients. Physicians self-reported that 50 of 114 sequencing reports (44%) influenced clinical care decisions in 44 individual patients. Influences were often nuanced and extended beyond identifying actionable genetic variants with US Food and Drug Administration-approved drugs.

Conclusion: This study provides insights into how physicians are currently using multigene panels capable of detecting relatively rare AML cells. The most influential way to integrate these tools into clinical practice will be to perform prospective clinical trials that assess patient outcomes in response to genomically driven interventions.

FIG 1.

Distribution of patient diagnoses. (A) Distribution and (B) relative frequency of hematologicb disorders for the 346 samples evaluated with the sequencing panel. Diagnosis is based on the clinical history ascertained by pathologists signing out the report. Acute myeloid leukemia (AML) with acute promyelocytic leukemia (APL) subtype—designated by a light blue bar under the AML distribution bar chart—were not eligible for analysis with the sequencing panel. BALL, B-cell acute lymphoblastic leukemia; BPDCN, blastic plasmacytoid dendritic cell neoplasm; CCUS, clonal cytopenia of undetermined significance; CML, chronic myeloid leukemia; CMML, chronic myelomonocytic leukemia; CLL, chronic lymphoblastic leukemia; DLBCL, diffuse large B-cell lymphoma; ET, essential thrombocytopenia; HCL, hairy cell leukemia; ICUS, idiopathic cytopenias of undetermined significance; MCC, Merkel cell carcinoma; MDS, myelodysplastic syndrome; MM, multiple myeloma; MPN, myeloproliferative neoplasm, not otherwise specified; PMF, primary myelofibrosis; PNH, paroxysmal nocturnal hemoglobinuria; PV, polycythemia vera; TALL, T-cell acute lymphoblastic leukemia; WM, Waldenström macroglobulinemia.

FIG 2.

Distribution of variants in acute myeloid leukemia (AML) cases. (A) Heatmap of the distribution of variants in all AML cases reported. Each row represents a single gene and each column represents a sequencing report (N = 122). Colored squares denote that a variant was observed in the designated gene. Colors indicate the variant type. If there was more than one variant observed per gene within the same case, the most deleterious variant on the basis of the variant effect prediction was listed. The bottom bar indicates the FLT3 status for all 122 patients. The bar color indicates the type of FLT3 variant (internal tandem duplication v tyrosine kinase domain). (B) Percentage of cases with a variant in a given gene. Each row represents a single gene and the color indicates the variant type.

FIG 3.

Physicians ordered sequencing panels at various times throughout the acute myeloid leukemia (AML) course. (A) Distribution of timepoints. (B) Distribution of therapeutics for which the sequencing report was ordered as part of response assessment. 7 + 3, 7 days cytarabine and 3 days anthracycline; alloHCT, allogeneic stem-cell transplantation; CPX-351, liposomal cytarabine and daunorubicin; gem/ozo, gemtuzumab ozogamicin; HiDAC, high-dose Ara-C; HMA, hypomethylating agent.

FIG 4.

Physician-reported influences from sequencing reports issued for patients with acute myeloid leukemia (AML). (A) In total, 346 consecutive cases—that is, reports—were analyzed and 124 had a diagnosis of AML. Of these 124 cases, four were excluded (acute promyelocytic leukemia [APL] subtype [n = 2]; no survey returned [n = 2]). Six surveys were returned but ineligible for additional analysis. Physicians reported that the report influenced clinical decision making in 50 cases (44%). (B) Physician-reported influences. Variants associated with predictive influences are labeled. In 10 cases, variants observed on sequencing reports were used to stratify relapse risk for patients in first clinical remission (CR1). In eight cases, variants observed were used to assess for persistent molecular disease (PMD). In 64 cases, physicians reported that the results did not inform decision making. (C) There were 14 physicians who contributed at least one survey to this study.

FIG 5.

Incorporation of panel testing in disease monitoring for patients with acute myeloid leukemia (AML). Each panel represents a single patient where multiple sequencing reports were obtained. The plot indicates any variants observed with associated variant allele frequencies (VAFs) at the timepoints labeled on the x-axis. These plots also show FLT3 variant status (presence tyrosine kinase domain or internal tandem duplication [ITD]) and approximate duration of treatment. (A) Sequencing panels revealed persistent molecular disease after induction. Based on the persistent IDH2 variant, the physician initiated targeted therapy with enasidenib. (B) Physician reported that the observation of persistent molecular disease at 475 days (D) post–stem-cell transplantation (SCT) indicated the need for a second allogeneic stem-cell transplantation despite clinical remission (0% blasts on bone marrow biopsy [BMBx]). (C) Sequencing panels were used to track VAFs to evaluate the efficacy of a donor lymphocyte infusion (DLI) and gilteritinib over time. (D) Extramedullary relapse was confirmed by comparing sequencing results of the original tumor and the periesophageal lesion.