Abstract
Objective: The aim of this analysis was to examine treatment patterns in patients with acute myeloid leukemia (AML) in routine clinical practice in the United States, including factors influencing the choice of front-line treatment intensity and the effect of age and treatment line.
Methods: We used data from the Adelphi AML Disease Specific Programme, a real-world, cross-sectional survey conducted in 2015. Physicians completed patient record forms providing patients’ demographic and clinical characteristics.
Results: In total, 61 academic, non-academic, and office-based hematologists and hematology/oncology specialists provided data on 457 patients with AML; 284 had ≥20% blasts (World Health Organization defined AML) and were included in the analysis. In the front-line setting, 60% of patients received high-intensity therapy, most commonly cytarabine plus anthracycline; the most common low-intensity treatments were hypomethylating agents. Primary drivers for selecting high-intensity versus low-intensity treatment were age, performance status and comorbidities; 67%, 64% and 61% of physicians stated they would prescribe high-intensity treatment to patients aged <65 years, with good performance status or no comorbidities, respectively. In practice, patients aged <60 years were more likely to receive high-intensity induction treatment (high vs. low intensity by age p < .0001). In a selected cohort of relapsed/refractory patients, 69% of patients received high-intensity therapy (78% of patients aged <60 years and 57% of patients aged ≥60 years).
Conclusions: Most patients in this analysis of real-world survey data received well established, front-line induction therapies. Treatment intensity was determined by age, comorbidities and performance status, as recommended by guidelines.
Introduction
Acute myeloid leukemia (AML) is the most common acute leukemia in adults. In the US, the estimated prevalence of AML was 48,615 in 2011Citation1. It has been estimated that 19,520 new cases of AML will be diagnosed in 2018 and 10,670 deaths due to AML will be recordedCitation2. The incidence of AML increases with age, from 1.0 per 100,000 in patients aged 15–19 years to 28.5 per 100,000 in those aged 80–84 yearsCitation3; one-third of new cases are diagnosed in patients older than 75 yearsCitation2.
AML is characterized by accumulation of myeloid blast cells in the bone marrow, peripheral blood, and organs such as the spleen and liver, which disrupts the production of normal blood cells leading to bone marrow failureCitation4. Treatment guidelines recommend intensive induction therapy for most younger patients with AML, with treatment aimed at achieving a complete remission. For patients aged ≥60 years, a standard-dose cytarabine-based regimen is recommended. Lower-intensity therapies may be prescribed for older patients who are candidates for intensive remission therapy but have unfavorable risk factors and for those who are not candidates for intensive treatment. Allogeneic hematopoietic stem cell transplantation is strongly recommended for patients with relapsed/refractory AML and for those with intermediate- and adverse-risk features during first complete remissionCitation5.
In order to better understand the drivers and patterns of treatment regimen selection in the real world, we performed an analysis of a contemporary dataset of adult patients with AML and their treating physicians currently in routine clinical practice in the US. We investigated the treatment regimens prescribed to patients with AML, the factors that influenced prescribing patterns among physicians and age-related differences in treatment patterns. The objectives of the study were three-fold: (1) to describe patient characteristics and real-world use of AML treatments for patients on high- and low-intensity induction therapy and understand the key physician-perceived drivers of therapy selection; (2) to examine the characteristics of patients aged <60 and ≥60 years at the point of AML diagnosis and evaluate the subsequent induction treatments they receive; and (3) to examine characteristics and treatment patterns of patients considered to have relapsed/refractory disease after initial AML treatment.
Methods
Study design and data sources
This was an analysis of data drawn from the Adelphi AML Disease Specific Programme (DSP) conducted in 2015 in the US. DSPs are large, multinational surveys conducted in clinical practice with the aim of describing current disease management, disease-burden impact and associated treatment effects (clinical and physician perceived). The DSP is a point-in-time survey of physicians and their patients presenting in a real-world clinical setting.
The survey methodology was designed to facilitate collection of up-to-date data from three complementary physician-completed information sources: physician online surveys, physician-reported workload questionnaires and patient record forms. The methodology also included patient self-reported data. A complete description of the survey methods has been previously publishedCitation6. Patients and physicians included in the DSPs cannot be identified as the data they provide is disaggregated and anonymized before receipt.
DSP data is collected in accordance with Adelphi Real World procedures, which are compliant with the Health Information Technology for Economic and Clinical Health (HITECH) ActCitation7 and the Health Insurance Portability and Accountability Act (HIPAA)Citation8. The validity and representativeness of the DSP methodology have been assessedCitation9,Citation10. Measures, including logic checks, are also adopted to ensure that numeric values are appropriate and that there is no missing data. DSP data is collected by local fieldwork partners and fully anonymized. Physicians are reimbursed for their participation in the study by the local fieldwork partners at fair market rates, and the fieldwork teams adhere to national data-collection regulations. Patients take part in the survey on a voluntary basis; the authors were not included in the data collection and only had access to the raw and de-identified datasets.
Patients and physicians
During the initial data collection phase of the AML DSP, physicians were invited to participate based on their medical specialty, and were then approached by field-based interviewers who recruited the survey sample based on their eligibility according to the following criteria: qualified between 1978 and 2011, seeing a minimum of two patients with AML in a typical week and personally responsible for prescribing decisions for patients with AML.
Patients were recruited by their consulting physicians. Physicians were asked to recruit the next 6–8 patients they saw who they considered to have AML. Inclusion criteria for these patients were minimal: patients were required to have a diagnosis of AML, to be receiving or to have received active drug therapy for their AML, and to be over the age of 18 years at the time of their AML diagnosis. Patients participating in clinical trials were excluded from the survey because the treatment of those patients is often rigorously controlled by protocol and therefore not a reflection of a general practice setting. Physicians completed record forms for the patients they recruited, based on the patient’s condition at the time of presentation and other historical data. As patients were enrolled into the DSP on a consecutive basis in order of consultation, they could be included at any point during their disease: newly diagnosed, receiving induction or consolidation therapy, undergoing transplant, in follow-up, or with relapsed or refractory disease. Therefore, the DSP dataset represents multiple subpopulations of patients with AML.
Physicians provided details regarding patient demographics, clinical characteristics at the time of diagnosis and at the time of most recent presentation, current treatment, and prior lines of treatment in patients with relapsed/refractory AML. Treatment was physician selected, either by the physician participating in the survey and currently managing the patient, or another physician who had previously treated the patient. Details of treatments were included, with physicians being asked to state whether they considered induction therapy to be high or low intensity. According to this response, physicians provided details of the drug or combination of drugs that the patient received as part of their induction regimen. A list of drugs considered to form part of either a high- or low-intensity regimen is shown in Supplementary Table S1.
Table 1. Patient clinical and demographic characteristics according to treatment intensity.
Physicians were also asked to provide a blast count for the patient at the time of diagnosis. For the present analysis, only patients meeting the World Health Organization defined AML criteriaCitation11 were included in the analysis. Therefore, for inclusion in the present analysis, patients were required to have a physician diagnosis of AML and ≥20% blasts at diagnosis. Although a diagnosis of AML can be made in patients with <20% blasts, e.g. in those with recurring cytogenetic abnormalities or a classic molecular finding of AML, the current “guideline-defined” AML population was limited to adult patients with ≥20% blasts. Patients with <20% blasts at diagnosis were excluded from the analysis as baseline cytogenetic risk stratification was not available for this dataset.
Assessments
The physician survey recorded their opinions regarding the management of patients with AML and current treatment, including drivers of therapy intensity choice. Information recorded in the patient record form included patient demographics, clinical characteristics at the time of diagnosis and at the time of most recent presentation, current treatment (which was determined by the physician), and prior lines of treatment in patients with relapsed/refractory AML. High- and low-intensity treatments were determined by initially asking physicians if induction therapy was high or low intensity for each patient.
Outcomes
The primary endpoint of the survey was to describe treatment patterns for adult patients with AML in the real-life clinical setting in the US. Secondary endpoints included assessment of the influence of age on chosen treatments, physician-reported factors influencing assignment of treatment to patients, and a description of patients who had relapsed or were refractory to treatment.
Statistical analysis
Descriptive statistics were generated for all outcome variables of interest: number of patients, median, minimum, maximum and interquartile ranges (IQRs) for continuous variables, and number of patients, frequency and percentage of responses for categorical and ordinal variables.
Hypothesis tests were used to assess associations between clinical and demographic characteristic outcomes, treatment intensity (high vs. low) and patient age (<60 vs. ≥60 years). Mann–Whitney U (non-parametric) tests were used to compare numeric outcome variables, Fisher’s exact tests were used to compare binary categorical outcome variables and χ2 tests were used to compare categorical outcome variables with more than two levels of response.
All analyses were performed in Stata (Version 15.0; StataCorp LLC, College Station, TX, USA).
Results
Physicians and patients
Sixty-one physicians participated in this survey, most of whom were male (n = 45; 74%). Four physicians (7%) practiced only in a hospital-based practice, 24 (39%) worked in a comprehensive cancer center, 5 (8%) were office-based only and the remaining 28 (46%) worked in combined hospital- and office-based practices. Among the 56 hospital-based physicians, most worked in major academic institutions, universities and teaching hospitals (25 of 56 physicians; 45%), followed by regional/community hospitals affiliated with major academic institutions (n = 13; 23%) and the remainder (32%) worked in regional/community hospitals that were not affiliated with major academic institutions. A total of 457 patients with AML were included in this survey; 193 (42%) were treated at a comprehensive cancer center, 98 (21%) were treated at a regional center affiliated with a major academic institution, 137 (30%) were treated at a regional center with no major academic affiliation and the remaining 29 patients (6%) were treated by physicians who were based only at an office.
Among the 457 patients who participated in the AML DSP, 284 (62%) had a blast count of ≥20% and a confirmed diagnosis of AML and were included in the current analysis; their clinical and demographic characteristics are summarized in . Clinical and demographic characteristics of the overall population and those with ≥20% blast cells at diagnosis are shown in Supplementary Tables S2 and S3, respectively. Patients had a median age of 60 years (range 19–91 years; IQR 69–50 years). The median time since diagnosis was 4 months (range 0–70.0 months; IQR 9.0–2.0 months). Mean time since diagnosis was 7.0 months (standard deviation 8.5 months); some patients were included who had been diagnosed for ≥5 years, which skewed the distribution. Twenty of the 457 patients were classified as having acute promyelocytic leukemia (APL) and 14 patients were classified as having therapy related AML/myelodysplastic syndromes (MDS) at diagnosis of AML.
Table 2. Treatment regimens prescribed in the front-line induction setting according to patient age.
Table 3. Patient characteristics considered by physicians when choosing high- or low-intensity treatment in AML (n = 61 physicians).
Treatment
For patients treated in the front-line, induction-therapy setting, all patients received antileukemic therapy in which physicians reported that 170 patients (60%) received a high-intensity regimen, the most common being cytarabine plus an anthracycline, with or without a third agent (n = 120; 71%). The remainder of the patients (n = 114; 40%) received low-intensity induction therapy, the most common being a hypomethylating agent (n = 54; 47%). Front-line treatment regimens are summarized in . Additional analysis has been provided (Supplementary Table S4) to explore any differences in clinical approach between the practice settings. There is little difference in clinical approach between practice settings for the prescription of high-intensity treatment regimens. Notable differences are observed in the clinical approach between practice settings when prescribing low-intensity treatment regimens, with low dose cytarabine being more commonly prescribed in academic institutions and azacitidine, decitabine and anthracycline monotherapies more commonly prescribed in the non-academic setting.
Table 4. Clinical and demographic characteristics according to patient age.
Patients receiving high-intensity treatment were significantly younger (p < .0001), had a lower Eastern Cooperative Oncology Group (ECOG) performance status at diagnosis (p < .0001), and were more likely to have de novo AML (p = .0095) and have presented with a normal white blood cell count at diagnosis (p = .0228) than those receiving low-intensity treatment (). Among the comorbidities reported by patients at the time of AML diagnosis, patients receiving high-intensity treatment were significantly more likely to have depression than those receiving low-intensity treatment (p = .0149; ). Conversely, patients who received low-intensity chemotherapy were more likely to have a history of cerebrovascular disease, chronic kidney disease, congestive heart failure, peripheral vascular disease, or a history of stroke or transient ischemic attack (all p < .05 vs. high-intensity treatment).
Physician-perceived drivers of treatment choice
To further investigate the factors associated with reported treatment patterns, the 61 physicians were asked, “To which of the following patient types or patient characteristics would you generally consider prescribing low- or high-intensity chemotherapy?”. Responses are summarized in . The most commonly cited factors for selecting low-intensity treatment were age ≥65 years (41 physicians; 67%), ECOG performance status ≥2 (39 physicians; 64%) and presence of concomitant comorbidities (38 physicians; 62%). Patient characteristics most commonly cited as reasons for prescribing high-intensity treatment were age <65 years (41 physicians; 67%), ECOG performance status 0 or 1 (39 physicians; 64%) and absence of comorbidities (37 physicians; 61%). The missing responses seen are from the physicians not selecting that attribute, i.e. by not checking the box, the physicians did not view that criterion as being important in their treatment decisions.
Effect of age on treatment patterns
Characteristics of patients according to age ≥60 years and <60 years are shown in . Patients aged <60 years were significantly more likely than those aged ≥60 years to have de novo AML (p = .0002), a lower ECOG performance status (p < .0001) and a more favorable physician-determined risk category at diagnosis (p < .0001) (). Younger patients were more likely than older patients to undergo biochemical profiling (p = .018), cytogenetic analysis (p = .0427) and computed tomography (p = .0023) at diagnosis of AML.
Following their initial diagnosis, younger patients were more likely to be prescribed high-intensity (101 of 139 patients [73%]) versus low-intensity (38 of 139 patients [27%]) induction treatment (). The most commonly prescribed regimens for younger patients were based on cytarabine plus an anthracycline (70 of 139 patients; 50%), high-dose cytarabine monotherapy (9 of 139 patients; 6%) and low-dose cytarabine monotherapy (8 of 139 patients; 6%).
In contrast, the cohort of older patients was equally likely to be initiated on low- or high-intensity induction treatment (76 of 145 patients [52%] vs. 69 of 145 patients [48%]; ). The most commonly prescribed regimens for older patients were cytarabine plus an anthracycline (47 of 145 patients; 32%), low-dose cytarabine (27 of 145 patients; 19%) and one of the approved hypomethylating agents (39 of 145 patients; 26%).
Supplementary Table S5 shows the medical coverage according to age; as expected patients who are <60 years are more likely to have private insurance in place and those ≥60 years public insurance (Medicare).
Patients with relapsed or refractory acute myeloid leukemia
Among the 284 patients included in this analysis, 32 were reported to have relapsed or to be refractory to previous treatment. These patients had a median age of 56 years (range 28–91 years; IQR 64.5–51 years) (). At the time of the survey, 22 of these patients (69%) were undergoing treatment with high-intensity regimens and 10 (31%) were being treated with low-intensity treatment regimens (). Differences between the two groups were not statistically significant, although these results need to be interpreted with caution given the limited sample size of this cohort.
Table 5. Clinical and demographic characteristics of patients who relapsed after, or were considered refractory to, previous therapy according to current treatment intensity.
Table 6. Current treatment regimens for patients with disease considered relapsed or refractory to prior AML drug therapy.
Discussion
In this real-world survey of US patients with guideline-defined AML, namely patients with a physician diagnosis of AML and a blast count ≥20% at diagnosis, 60% of patients received high-intensity induction treatment regimens and the remaining 40% received low-intensity therapy. Approximately 59% of AML patients younger than 60 years compared to 41% of those older than 60 years received induction therapy with high-intensity regimens. Cytarabine/anthracycline combinations were the most common high-intensity regimens, whereas azacitidine, decitabine and low-dose cytarabine were the most common low-intensity treatment regimens. The patient’s insurance status and access restrictions may be a factor to take into consideration when interpreting these results. For example, when assessing younger patients who were not treated with intensive chemotherapy, there were no patients in this cohort who did not meet the criteria for a treatment option the physician considered most appropriate. However, in five cases the hospital formulary did not include the option that the physician considered more appropriate and in a further two cases the patient was unable/unwilling to pay the cost share required by their plan for the physician’s preferred option.
Further examination revealed significant differences between patients receiving high- and low-intensity induction regimens, including age at diagnosis, performance status and comorbidity profiles. Younger patients were more likely to receive the more aggressive, high-intensity induction regimens than older patients (73% vs. 48%, respectively), which is consistent with results from a 2015 study by Meyers and colleagues, in which younger patients with fewer comorbidities were more likely to receive chemotherapy than older patientsCitation12. Although most patients were therefore treated according to guideline recommendationsCitation5, a considerable proportion (27%) of patients aged <60 years received a low-intensity regimen. Reasons for this are unclear but could include physician/patient choice of a lower-intensity therapy because of poor performance status, comorbid conditions, a desire to receive a lower-intensity regimen or the setting in which the patient was treated.
When asked what factors influenced their prescribing decisions, physicians in this survey reported that age, performance status, and comorbidities were considerations in the decision to prescribe low-intensity treatment. Although past research has suggested that patients up to age 80 years may be considered for intensive therapiesCitation13,Citation14, lower-intensity treatment regimens are routinely chosen for this cohort of patients with AML, in part due to the poor outcomes observed in older patients treated with standard cytarabine–anthracycline induction regimensCitation5,Citation15. In practice, available real-world evidence suggests that a sizeable proportion of older patients (aged ≥60 years) often receive no antileukemic therapyCitation16 and that treatment with intensive induction strategies is uncommon in community oncology practiceCitation17. Treatment decisions for older patients are complex and should reflect the overall goals of care, which may include achieving remission, prolonging survival or preserving quality of lifeCitation18,Citation19.
The way in which the data was collected in the AML DSP enabled refinement of the original survey sample to generate a population meeting current World Health Organization definitions, i.e. a diagnosis of AML and a blast count at diagnosis of ≥20%. The “guideline-defined” patient sample was largely similar to the overall population, except that the overall patient population appeared to have better performance status, was more likely to be asymptomatic at diagnosis and was less likely to be receiving high-intensity treatment compared with the sample of patients with ≥20% blasts. The DSP study methodology therefore allowed us to examine a typical physician’s AML patient population by use of unrestrictive screening criteria (i.e. physician diagnosis of AML and treatment for AML at some point in the patient’s history), whereas the patient’s blast count at diagnosis of AML could be used to generate a sample of patients with a more rigorous definition of AML. Reasons for the differences seen between the two populations are not yet known but warrant further examination to establish whether factors other than those recommended in guidelines are used in the diagnosis of AML.
Some limitations of the DSP survey methodology should be considered. Although minimal inclusion criteria governed the selection of the participating physicians, participation is influenced by willingness to complete the survey. As a result, bias is possible as the physicians surveyed represent a pragmatic sample and may not be representative of the overall population of physicians treating AML in the US. The DSP survey population was not a random sample of patients; rather, the protocol required inclusion of a sequential number of patients seen (approximately 6–8) who had a diagnosis of AML to reduce selection bias. The diagnosis of AML was based on the physician’s judgment and not a formalized diagnostic checklist, as applied in the real-world setting. Patients seen frequently were more likely to be included in the sample. In addition, our original AML population included patients with <20% blasts, all of whom were excluded from the current analysis. This may have excluded some patients eligible for a diagnosis of AML despite their low blast countCitation11, such as those with recurrent genetic abnormalities, relapsed/refractory disease or core binding factor AML. Although the selection of patients with ≥20% blasts may therefore have excluded the rare patients fulfilling those criteria, this was balanced against the risk of including patients who did not have AML.
Despite these limitations, real-world studies, like ours, play an important part in highlighting areas of concern that are not addressed in clinical trials. Patients included in clinical trials may not be representative of the consulting population because of age restrictions and stringent eligibility criteria for clinical trialsCitation20. Patients included in clinical trials are likely to be more adherent to medication than those treated in the real-world settingCitation21 and treatment in a clinical trial has recently been associated with improved outcomesCitation22. As a result, data from real-world studies can complement clinical trials and provide insight into the effectiveness of interventions in patients commonly seen in clinical practice.
In conclusion, this analysis of patients with physician-defined AML in real-life clinical settings in the US has shown that this was a heterogeneous population. Analysis of adult patients with a more rigorously defined AML diagnosis revealed that most patients received traditional, well established regimens as front-line induction therapy for AML. The key perceived drivers of treatment selection between high- and low-intensity chemotherapy were age, the presence of comorbidities and performance status. Identification of drivers of therapy selection and the physician’s awareness of them are critical to ensuring that patients receive the most appropriate therapy to improve clinical outcomes and balance acceptable toxicity and quality of life. Better appreciation of these factors by treating physicians may help in navigating the complexity of decision-making in AML.
Transparency
Declaration of funding
This analysis was funded by Astellas Pharma, Inc.
Author contributions
Named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this manuscript, take responsibility for the integrity of the work as a whole, and have given final approval to the version to be published. B.C.M. contributed to the design of the research, assisted in the analysis and interpretation of data, and developed and critically edited the manuscript. B.J.P., S.W., C.M., C.N.B. and L.E.H.W. contributed to the design of the research, assisted in the analysis and interpretation of data, and drafted, reviewed and approved the manuscript. A.H., J.P. and A.R. contributed to the design of the research, project managed the collection of the data, assisted in the analysis and interpretation of data, and drafted, reviewed and approved the manuscript. S.C.F. contributed to the design of the research, assisted in the analysis and interpretation of data, and drafted, reviewed, edited and approved the manuscript.
Declaration of financial/other relationships
B.C.M. has disclosed that he has received research funds from Astellas Pharma, Inc. B.J.P., S.W., and C.M., have disclosed that they are current employees of Astellas Pharma, Inc. C.N.B., L.E.H.W., and S.C.F. have disclosed that they were employed by Astellas Pharma, Inc. at the time of the study. J.P. and A.R. have disclosed that they are employees of Adelphi Real World. A.H. has disclosed that she was employed by Adelphi Real World at the time of the study.
CMRO peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
Supplemental Material
Download MS Word (75.7 KB)Acknowledgments
Medical writing assistance was provided by Deirdre Carman PhD of Alispera Communications Ltd, and was funded by Astellas Pharma, Inc.
Data availability statement
The authors confirm that the questions relating directly to the analysis will be made available to scientific researchers (on request) by Adelphi Real World as a supplementary document. Access to anonymized individual participant level data will not be provided for this trial as it meets one or more of the exceptions described on www.clinicalstudydatarequest.com under “Sponsor Specific Details for Astellas”.
References
- Leukemia & Lymphoma Society. Acute Myeloid Leukemia [Internet]. 2017. [cited 2018 Apr 24]. Available from: https://www.lls.org/sites/default/files/file_assets/PS32_AML_BookletFINAL_with_insert.pdf
- National Cancer Institute Surveillance, Epidemiology, and End Results program. Cancer stat facts: acute myeloid leukemia (AML) (1992–2015). [Internet]. 2018. [cited 2018 Apr 24]. Available from: https://seer.cancer.gov/statfacts/html/amyl.html
- National Cancer Institute Surveillance, Epidemiology, and End Results Program. SEER Cancer Statistics Review 1975–2014 SEER incidence rates, age-adjusted and age-specific rates, by race and sex [Internet]. 2018. [cited 2018 Apr 24]. Available from: https://seer.cancer.gov/csr/1975_2014/results_single/sect_13_table.13.pdf
- Kumar CC. Genetic abnormalities and challenges in the treatment of acute myeloid leukemia. Genes Cancer. 2011;2:95–107.
- National Comprehensive Cancer Network Clinical practice guidelines in oncology. Acute myeloid leukemia. Version 1 [Internet]. 2018 [cited 2018 Apr 24]. Available from: www.nccn.org/professionals/physician_gls/pdf/aml.pdf
- Anderson P, Benford M, Harris N, et al. Real-world physician and patient behaviour across countries: disease-specific programmes – a means to understand. Curr Med Res Opin. 2008;24:3063–3072.
- Health Information Technology. Health Information Technology Act [Internet]. 2009. [cited 2018 Apr 24]. Available from: https://www.healthit.gov/sites/default/files/hitech_act_excerpt_from_arra_with_index.pdf
- US Department of Health and Human Services. Summary of the HIPAA Privacy Rule [Internet]. May 2003. [cited 2018 Apr 24]. Available from: http://www.hhs.gov/sites/default/files/privacysummary.pdf
- Higgins V, Piercy J, Roughley A, et al. Trends in medication use in patients with type 2 diabetes mellitus: a long-term view of real-world treatment between 2000 and 2015. Diabetes Metab Syndr Obes. 2016;9:371–380.
- Babineaux SM, Curtis B, Holbrook T, et al. Evidence for validity of a national physician and patient-reported, cross-sectional survey in China and UK: the Disease Specific Programme. BMJ Open. 2016;6:e010352.
- Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127:2391–2405.
- Meyers J, Yu Y, Kaye JA, et al. Medicare fee-for-service enrollees with primary acute myeloid leukemia: an analysis of treatment patterns, survival, and healthcare resource utilization and costs. Appl Health Econ Health Policy. 2013;11:275–286.
- Juliusson G, Antunovic P, Derolf A, et al. Age and acute myeloid leukemia: real world data on decision to treat and outcomes from the Swedish Acute Leukemia Registry. Blood. 2009;113:4179–4187.
- Becker H, Marcucci G, Maharry K, et al. Favorable prognostic impact of NPM1 mutations in older patients with cytogenetically normal de novo acute myeloid leukemia and associated gene- and microRNA-expression signatures: a Cancer and Leukemia Group B study. JCO. 2010;28:596–604.
- Döhner H, Estey EH, Amadori S, et al. European LeukemiaNet. Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood. 2010;115:453–474.
- Medeiros BC, Satram-Hoang S, Hurst D, et al. Big data analysis of treatment patterns and outcomes among elderly acute myeloid leukemia patients in the United States. Ann Hematol. 2015;94:1127–1138.
- Ma E, Bonthapally V, Chawla A, et al. An evaluation of treatment patterns and outcomes in elderly patients newly diagnosed with acute myeloid leukemia: a retrospective analysis of electronic medical records from US community oncology practices. Clin Lymphoma Myeloma Leuk. 2016;16:625–636.e3.
- Ossenkoppele G, Löwenberg B. How I treat the older patient with acute myeloid leukemia. Blood. 2015;125:767–774.
- Erba P. Finding the optimal combination therapy for the treatment of newly diagnosed AML in older patients unfit for intensive therapy. Leuk Res. 2015;39:183–191.
- Dechartres A, Chevret S, Lambert J, et al. Inclusion of patients with acute leukemia in clinical trials: a prospective multicenter survey of 1066 cases. Ann Oncol. 2011;22:224–233.
- Hubbard TE, Paradis R. Real-world evidence: a new era for health care innovation. The Network for Excellence in Health Innovation [Internet]. 2015. [cited 2018 Apr 24]. Available from: http://www.nehi.net/writable/publication_files/file/rwe_issue_brief_final.pdf
- Østgård LS, Nørgaard M, Sengeløv H, et al. Improved outcome in acute myeloid leukemia patients enrolled in clinical trials: a national population-based cohort study of Danish intensive chemotherapy patients. Oncotarget. 2016;7:72044–72056.