Diagnostic and molecular testing patterns in patients with newly diagnosed acute myeloid leukemia in the Connect® MDS/AML Disease Registry

Abstract Diagnostic and molecular genetic testing are key in advancing the treatment of acute myeloid leukemia (AML), yet little is known about testing patterns outside of clinical trials, especially in older patients. We analyzed diagnostic and molecular testing patterns over time in 565 patients aged ≥ 55 years with newly diagnosed AML enrolled in the Connect® MDS/AML Disease Registry (NCT01688011) in the United States. Diagnostic data were recorded at enrolment and compared with published guidelines. The percentage of bone marrow blasts was reported for 82.1% of patients, and cellularity was the most commonly reported bone marrow morphological feature. Flow cytometry, karyotyping, molecular testing, and fluorescence in situ hybridization were performed in 98.8%, 95.4%, 75.9%, and 75.7% of patients, respectively. Molecular testing was done more frequently at academic than community/government sites (84.3% vs 70.2%; P < .001). Enrolment to the Registry after 2016 was significantly associated with molecular testing at academic sites (odds ratio [OR] 2.59; P = .023) and at community/government sites (OR 4.85; P < .001) in logistic regression analyses. Better understanding of practice patterns may identify unmet needs and inform institutional protocols regarding the diagnosis of patients with AML.


INTRODUCTION
Acute myeloid leukemia (AML) is the most common acute leukemia in adults [1] occurring in 4.3 in 100 000 individuals in the United States [2]. AML is a heterogeneous disease with numerous cytogenetic and molecular features influencing prognosis [3][4][5].
According to current guidelines, accurate diagnosis and initial evaluation of AML requires clinical, morphological, cytogenetic, and molecular genetic assessment [3][4][5][6]. Approximately 50% of patients with AML have an abnormal karyotype, and a median of 3 gene mutations per patient [7]. More mutations are generally associated with worse prognosis; however, prognosis depends heavily on which genes are mutated [8]. The addition of diagnostic molecular genetic testing results to risk stratification models has enhanced prognostication [8,9] and, with the advent of targeted therapies, has become essential in making treatment decisions. Since 2017, four targeted therapies have been approved for treatment of AML: the FLT3 inhibitors midostaurin [10,11] and gilteritinib [12] for newly diagnosed and relapsed/refractory patients, respectively, the IDH1 inhibitor ivosidenib [13][14][15] for newly diagnosed and relapsed/refractory patients, and the IDH2 inhibitor enasidenib [16] for relapsed/refractory patients.
Despite the critical roles that diagnostic and molecular genetic testing play in AML treatment [23], our understanding of how these tests are used, particularly in older patients who are often excluded from clinical trials, is limited. Outcomes in older patients with AML remain poor; median survival ranges from 2.5 to 24.5 months, depending on risk group [24,25]. Several factors contribute to poor outcomes, including an increased probability of drug resistance and unfavorable cytogenetics [3]. Older patients often receive less intensive therapy due to perceived frailty, which can lead to worse outcomes [25,26].
We summarize diagnostic and molecular testing patterns by site, enrolment year, and adherence to guidelines in patients with AML enrolled in the Connect R MDS/AML Disease Registry.

Participants
To minimize selection bias, consecutive patients with AML seen at par-  Figure S1) included in the study electronic database.

Bone marrow blast assessment and morphology
BM blast percentage was reported for 82.1% of patients. BM blasts were measured by a manual count from the aspirate or touch preparation in 71.7% of patients (Figure 1), as recommended by the WHO guidelines [28,32].   Figure S1; Figure 2A) Cancer Network c (NCCN) molecular testing recommendations [33] ( Figure S1).  IQR, interquartile range. * A total of 346 patients received molecular testing on bone marrow specimens; dates of reporting were missing for 1 patient.

Changes in molecular genetic testing over time
To

DISCUSSION
This analysis from the Connect R MDS/AML Disease Registry suggests that, in a clinical practice setting outside of a clinical trial, physicians largely follow testing guidelines, although there is a need for further education to increase testing rates.
In this study, blast assessment was performed manually in 71.7% of Registry patients. The WHO guidelines recommend that blast percentages should be determined manually using 200-cell leucocyte differential counts of the peripheral blood, and 500-cell differential counts of all nucleated BM cells in cellular marrow aspirate smears stained with Wright-Giemsa [32,34]. A low rate of manual differentials may indicate poor aspirates rather than a lack of awareness of the WHO guidelines. For poor quality aspirates immunohistochemistry may be useful; however, use of flow cytometry to determine blast percentage by counting CD34 + cells is discouraged because not all leukemic blasts express CD34, and hemodilution and processing artifacts can cause inaccurate estimation of blast percentage [28,35]. In a physician survey, respondents reported routinely using flow cytometry, immunohistochemistry, and karyotyping with poor quality aspirates [36].  [38], increased testing as a requirement for clinical trial enrolment at academic centers, or bias that older patients are unlikely to benefit from therapy [39].
It is increasingly acknowledged that age should not be the only factor in making treatment decisions [40], partially because novel, targeted therapies may be better tolerated than conventional therapies. This may lead to an increase in molecular testing rates in older patients.
Multivariable analyses revealed that the only overlapping predictor for molecular testing at both academic and community/government sites was enrolment after 2016. This may suggest changing diagnostic practices as AML testing guidelines were updated [4,5,32]. The observation that patients with a normal karyotype were more likely to undergo molecular testing may be due to the perception that molecular genetic testing has a higher yield in normal karyotype patients, dating from an era when guidelines only recommended molecular testing in such patients; now, with the ability to therapeutically target specific mutations, molecular testing has a larger role than simply prognostication. It may also reflect economic factors, reserving costly molecular testing for patients most likely to need accurate risk assessment. One important strength of this Registry study is that it represents current routine clinical practice for newly diagnosed patients with AML across practice settings. There were few eligibility criteria other than age (≥55 years) and no comorbidity restrictions. Therefore, the data presented here closely reflect the older patients' experience.
As with all studies using data from routine clinical practice, this study was subject to limitations concerning lack of randomization and specific protocols for patient assessment and intervention. Despite this, the Registry represents one of the largest prospective cohort studies of patients with AML, treated in geographically diverse and mainly community-based settings in the United States, and provides insights into their clinical experience. Regardless of the as-yet-unknown prognostic value of molecular testing [41,42], the advent of targeted therapies for AML underscores the importance of molecular characterization of AML in all patients.