Dyslipidemia and cardiovascular disease among childhood cancer survivors: a St. Jude Lifetime Cohort report

Abstract

Background

Childhood cancer survivors have increased risk of dyslipidemia and atherosclerotic cardiovascular disease (CVD). The aim of this study was to evaluate the prevalence and associated cardiovascular risks of specific lipid abnormalities among childhood cancer survivors.

Methods

Comprehensive lipid panel measurements were obtained from 4115 5-year survivors, with 3406 (mean age at evaluation = 35.2 years, SD = 10.4 years) not having previous dyslipidemia diagnosis, as well as 624 age, sex, and race and ethnicity matched community controls.

Results

Previously undiagnosed dyslipidemia with abnormal low-density lipoprotein (LDL) cholesterol (>160 mg/dL), non–high density lipoprotein (HDL) cholesterol (>190 mg/dL), HDL cholesterol (<40 mg/dL for men, <50 mg/dL for women), and triglycerides (>150 mg/dL) were identified in 4%, 6%, 30%, and 17%, respectively. Survivors without previous dyslipidemia diagnosis had higher LDL cholesterol and non-HDL cholesterol and lower HDL cholesterol than community controls. Cranial radiotherapy (relative risk [RR] = 2.2, 95% confidence interval [CI] = 1.6 to 3.0 for non-HDL cholesterol) and total body irradiation for hematopoietic cell transplantation (RR = 6.7, 95% CI = 3.5 to 13.0 for non-HDL cholesterol; RR = 9.9, 95% CI = 6.0 to 16.3 for triglycerides) were associated with greater risk of dyslipidemia. Diagnoses of low HDL cholesterol (hazard ratio [HR] = 2.9, 95% CI = 1.8 to 4.7) and elevated triglycerides (HR = 3.1, 95% CI = 1.9 to 5.1) were associated with increased risk for myocardial infarction, and diagnoses of high LDL cholesterol (HR = 2.2, 95% CI = 1.3 to 3.7), high non-HDL cholesterol (HR = 2.2, 95% CI = 1.3 to 3.7), low HDL cholesterol (HR = 3.9, 95% CI = 2.8 to 5.4), and elevated triglycerides (HR = 3.8, 95% CI = 2.7 to 5.5) were associated with increased risk for cardiomyopathy.

Conclusions

Previously undiagnosed dyslipidemia among childhood cancer survivors was associated with increased risk for myocardial infarction and cardiomyopathy. Comprehensive dyslipidemia evaluation and treatment are needed to reduce cardiovascular morbidity in this population.

Five-year survival after diagnosis of childhood cancer exceeds 85%, attributable to improvements in cancer therapy and supportive care (1,2). Unfortunately, in addition to subsequent malignant neoplasms, cardiovascular late effects markedly contribute to nonrecurrence morbidity and mortality (1,3). Survivors are 7 times more likely than the US population to die from cardiovascular-related events (4). Additionally, survivors of childhood cancer have increased rates of both atherosclerotic cardiovascular disease (CVD) (coronary artery disease [CAD], myocardial infarction [MI], and cerebrovascular accident [CVA]) and non–atherosclerotic CVD (heart failure [HF] and valvular abnormalities), as well as cardiovascular risk factors (diabetes mellitus, hypertension, and dyslipidemia) (3,5-8). Dyslipidemia requiring lipid-lowering therapy is more common in survivors than in siblings and is undiagnosed in many survivors (3,9). St. Jude Lifetime Cohort Study (SJLIFE) investigators previously identified that 23% of survivors had been diagnosed with dyslipidemia prior to baseline evaluation, and an additional 32% of participants were identified during baseline screening (3,9). In the Childhood Cancer Survivor Study (CCSS), survivors exposed to anthracycline chemotherapy who self-reported dyslipidemia were more likely to develop HF, and those with dyslipidemia who were exposed to chest-directed radiotherapy were more likely to develop HF, MI, valvular disease, and arrhythmia (3). However, these previous evaluations of dyslipidemia and adverse cardiovascular outcomes among survivors have been limited by imprecise definitions of dyslipidemia, heterogeneous ascertainment methods including self-report, and failure to assess the individual components of dyslipidemia (3,5,9,10). Moreover, past approaches do not reflect current practice of lipid evaluation among adults in the general population and thus may not adequately assess the risk attributable to dyslipidemia among childhood cancer survivors.

In the general population, evaluation and treatment of dyslipidemia have changed within the past decade. Beginning with the 2013 American College of Cardiology and American Heart Association (ACC-AHA) guidelines, the treatment threshold for lipid-lowering therapy in primary CAD prevention has shifted to a holistic approach of assessing 10-year cardiovascular risk utilizing individual lipid values (low-density lipoprotein [LDL] cholesterol, high-density lipoprotein [HDL] cholesterol, and total cholesterol) in the setting of other cardiovascular risk factors, such as age, sex, race, presence of diabetes mellitus, current smoking, or hypertension, with the recommendation for lipid-lowering therapy for those with a 10-year atherosclerotic CVD risk of at least 7.5% (11,12). Given the limitations of existing studies of dyslipidemia among childhood cancer survivors, the prevalence and associated risk of specific lipid abnormalities known to contribute to CVD risk are not known. Therefore, we used the SJLIFE cohort that included baseline comprehensive lipid panel assessments in both survivors and community controls to ascertain the prevalence of previously undiagnosed dyslipidemia of multiple categories as well as the risk of CVD associated with these abnormalities.

Methods

Participants

Eligible participants completed a baseline on-campus evaluation designed to characterize health outcomes among survivors of childhood cancer and community controls as they age (9). Survivors eligible for this study were treated for childhood cancer at St. Jude Children’s Research Hospital between 1962 and 2012, aged 18 years and older at evaluation, and at least 5 years postdiagnosis. Females were not pregnant and at least 3 months postpartum. Participants were invited for follow-up visits approximately every 5 years. SJLIFE community controls were aged 18 years and older and frequency matched by age, sex, and race and ethnicity. Exclusion criteria for community controls included being a first-degree relative of a SJLIFE participant, having a history of childhood cancer, or being pregnant. Community controls had 1 visit. Survivors and controls were excluded if they did not have at least 1 fasting lipid panel assessment as part of SJLIFE baseline evaluation. Study documents, materials, and methods were approved by the institutional review board. Participants provided written, informed consent prior to participation.

Diagnosis, treatment, and demographics

Demographic information, cancer diagnosis, and treatment information were abstracted from medical records including cumulative doses of specific chemotherapy agents and radiotherapy. Total anthracycline dose (mg/m²) was the sum of doxorubicin, daunorubicin, epirubicin, idarubicin, and mitoxantrone in doxorubicin equivalent doses (13). Radiotherapy records were centrally reviewed, and maximum body region dose estimated. Anthropometrics and the presence of diabetes mellitus and hypertension were diagnostically assessed; self-report was used to obtain family and substance use history as well as physical activity and eating index data, as shown in Table 1.

Dyslipidemia

At baseline, participants completed a multi-item questionnaire that included age at onset of any previously diagnosed cardiac events or dyslipidemia, which were then validated by medical record review. Fasting lipid panel analysis from SJLIFE evaluations included total cholesterol, LDL cholesterol, HDL cholesterol, non-HDL cholesterol, and triglycerides, for evaluation of dyslipidemia at baseline evaluation. Total cholesterol, triglycerides, and HDL cholesterol were measured using the Roche Diagnostics fourth generation homogeneous enzymatic colorimetric test (Roche Diagnostics, Indianapolis, IN, USA). LDL cholesterol was calculated using the Friedewald equation (14). The following cutoffs were used to denote dyslipidemia in specific categories: LDL cholesterol  greater than 160 mg/dL, non-HDL cholesterol greater than 190 mg/dL, HDL cholesterol less than 40 mg/dL for men and less than 50 mg/dL for women, and triglycerides greater than 150 mg/dL. Previously undiagnosed dyslipidemia was defined as a lipid abnormality in survivors whose dyslipidemia status was determined at their in-person visit and who did not have dyslipidemia reported prior to SJLIFE evaluation. Previously diagnosed dyslipidemia was defined as a lipid abnormality at evaluation among survivors with a dyslipidemia diagnosis prior to SJLIFE evaluation.

Cardiovascular outcomes

Comprehensive clinical assessment, including echocardiography and medical record review by 2 clinicians, was used to identify cardiovascular risk factors and outcomes (both at baseline SJLIFE evaluation and on follow-up evaluations), using a modified version of the National Cancer Institute’s Common Terminology Criteria for Adverse Events (version 4.03; Supplementary Table 1, available online) (15,16). We included atherosclerotic disease (CAD/MI, CVA, and peripheral vascular disease) as well as cardiomyopathy. Common Terminology Criteria for Adverse Events were graded 1-5: mild (grade 1), moderate (grade 2), severe (grade 3), life-threatening or disabling (grade 4), and fatal (grade 5) by 2 clinicians. For CVA, peripheral vascular disease, and cardiomyopathy (documented left ventricular systolic dysfunction), we included outcomes grade 2-4. For CAD/MI we included outcomes grade 3 (abnormal cardiac enzymes with electrocardiogram evidence of infarction) and grade 4 (life-threatening MI with hemodynamic instability) to more accurately classify MI. We excluded grade 2 CAD as it describes mildly abnormal cardiac enzymes without electrocardiogram evidence of ischemia, a condition commonly found without MI.

Statistical analyses

Descriptive statistics characterized the study population and compared survivors without previous diagnosis of dyslipidemia with community controls with t tests or χ² statistics as appropriate. Logistic regression analysis was performed to evaluate the association between the risk factors (cardiovascular and cancer treatment variables) and each component of dyslipidemia. The elastic-net regularization technique with tenfold cross-validation for logistic regression was used to find a more parsimonious and interpretable model of selected predictors of dyslipidemia (GLMNET R package). Risk for cardiovascular outcomes (treating noncardiovascular death as a competing event) was obtained via Fine and Gray modeling with the proportional subdistribution hazard function, using attained age as the time and adjusting for sex, race and ethnicity, cardiovascular and health behavior risk factors, and cardiovascular events prior to baseline assessment (17,18). The proportional hazards assumption in the Fine and Gray models was tested by using the cumulative sums of residuals (crskdiag R package). A separate model was run to evaluate the potential interaction effects of radiotherapy and dyslipidemia toward risk for CVD, calculating the relative excess risk due to interaction (19). Ten-year atherosclerotic CVD risk was calculated using the ACC-AHA 2013 atherosclerotic CVD risk score (CVrisk R package) and categorized using the ACC online calculator ASCVD Risk Estimator Plus (20). SAS version 9.4 (SAS Institute, Inc, Cary, NC, USA) and R version 4.2.1 were used for statistical analyses. All tests were based on 2-sided analysis, and P values less than .05 were considered statistically significant.

Results

Among 9373 eligible survivors, 4115 completed at least 1 on-campus fasting lipid evaluation, with 3406 not having previous dyslipidemia diagnosis; among 855 community controls, 624 completed at least 1 on-campus lipid evaluation (Table 1; Supplementary Figures 1 and 2, available online). Survivors did not differ from community controls by age at evaluation (mean = 35.1 [SD = 10.2] years vs 35.0 [SD = 10.4] years; P = .87) and race and ethnicity (80.4% vs 80.3% non-Hispanic White; P = .10). Among survivors, 23.5% of those previously diagnosed with dyslipidemia had a history of hypertension compared with 16.3% of those without a previous diagnosis of dyslipidemia. Hypertension was prevalent among 13.4% of community controls. Diabetes mellitus, obesity, and sedentary lifestyle were more prevalent among survivors with previously diagnosed dyslipidemia as compared with both survivors without previously diagnosed dyslipidemia and community controls. History of cranial, chest, or total body irradiation was more prevalent among survivors with previously diagnosed dyslipidemia as compared with survivors without previously diagnosed dyslipidemia. History of previous MI, CVA, peripheral vascular disease, or cardiomyopathy was rare (<5%) among survivors and community controls.

Survivors were more likely than community controls to have abnormal non-HDL cholesterol and HDL cholesterol, both overall (Table 2) and when stratified by sex but not by race (Figure 1; Supplementary Tables 2 and 3, available online). Survivors previously diagnosed with dyslipidemia were more likely than those without previously diagnosed dyslipidemia to have abnormal lipid values in all categories both overall and when stratified by sex and race (Figure 2). Among 3406 survivors without previously diagnosed dyslipidemia, newly diagnosed abnormal LDL cholesterol, non-HDL cholesterol, HDL cholesterol, and triglycerides were identified in 4%, 6%, 30%, and 17%, respectively, at their baseline clinical evaluation. With longitudinal follow-up (range = 5.1-5.4 years), among survivors without dyslipidemia prior to or at their baseline clinical evaluation (n = 2089), new abnormalities in LDL cholesterol, non-HDL cholesterol, HDL cholesterol, and triglycerides were identified in 37 (2%), 38 (2%), 166 (8%), and 134 (6%) survivors, respectively.

Among survivors without previously diagnosed dyslipidemia, treatment with total body irradiation was associated with a sevenfold (hazard ratio [HR] = 6.7, 95% confidence interval [CI] = 3.5 to 13.0) increased risk of abnormal non-HDL cholesterol and a tenfold (HR = 9.9, 95% CI = 6.0 to 16.3) increased risk of abnormal triglycerides. Cranial radiotherapy, older age at diagnosis, hypertension, diabetes mellitus, chronic kidney disease, obesity, smoking history, and sedentary lifestyle were also significantly associated with a new diagnosis of a specific lipid abnormality (Table 3); survivors of Black race had lower incidence of dyslipidemia in multiple categories, as has been found in other studies (21,22). Similar associations were seen among all survivors with lipid abnormalities (including those with previously diagnosed dyslipidemia; Supplementary Table 4, available online).

Among survivors without previously diagnosed dyslipidemia, low HDL cholesterol (HR = 2.9, 95% CI = 1.8 to 4.7) and elevated triglycerides (HR = 3.1, 95% CI = 1.9 to 5.1) were associated with an increased risk for MI during SJLIFE follow-up (Table 4). Newly diagnosed lipid abnormalities in each category (elevated LDL cholesterol: HR = 2.2, 95% CI = 1.3 to 3.7; elevated non-HDL cholesterol: HR = 2.2, 95% CI = 1.3 to 3.7; low HDL cholesterol: HR = 3.9, 95% CI = 2.8 to 5.4; elevated triglycerides: HR = 3.8, 95% CI = 2.7 to 5.5) were associated with an increased risk for cardiomyopathy. Similar associations were seen among all survivors with lipid abnormalities (including those with previously diagnosed dyslipidemia; Supplementary Table 5, available online). Table 5 evaluated the interaction of dyslipidemia and radiotherapy and risk for development of cardiovascular outcomes, as both were independently predictive of cardiovascular outcomes. Abnormal non-HDL cholesterol potentiated risk for stroke and peripheral vascular disease among those exposed to radiotherapy (relative excess risk due to interaction = 1.1, 95% CI = 0.2 to 1.9; P = .01; and relative excess risk due to interaction = 1.2, 95% CI = 0.2 to 2.3; P = .02, respectively; Table 5).

Dyslipidemia at initial SJLIFE evaluation was associated with a 40% increase in risk of all-cause mortality (HR = 1.40, 95% CI = 1.00 to 1.97), but hazard ratios for cardiovascular mortality (HR = 1.96, 95% CI = 0.83 to 4.64) did not reach statistical significance (Supplementary Table 6, available online). When evaluating 10-year risk of atherosclerotic CVD using the 2013 ACC–AHA guidelines for the general population, intermediate (7.5%) or higher 10-year risk was found among 12.7% of survivors without previously diagnosed dyslipidemia.

Discussion

Long-term survivors of childhood cancer are at increased risk of CVD and cardiac-specific mortality, in part because of exposure to therapies with known cardiovascular risks including anthracycline chemotherapy and chest-directed radiotherapy (7,23-25). Dyslipidemia has been identified more commonly among survivors than controls and is an independent predictor of CVD among survivors. However, specific lipid abnormalities commonly used in atherosclerotic CVD risk scoring have not been evaluated as independent predictors of cardiomyopathy and MI (3,5,9,10). Our study demonstrated that, among childhood cancer survivors, 1) low HDL cholesterol and elevated triglycerides were associated with increased risk for MI; 2) abnormal LDL cholesterol, non-HDL cholesterol, HDL cholesterol, and triglycerides were associated with increased risk for cardiomyopathy; and 3) non-HDL cholesterol potentiates peripheral vascular disease and stroke in a more than additive fashion. Utilizing ACC–AHA atherosclerotic CVD risk scoring, 13% of survivors without previous dyslipidemia diagnosis had intermediate or higher 10-year risk of atherosclerotic CVD, meeting class I indication for lipid-lowering therapy. However, this is likely an underestimate of cardiovascular risk as this calculation does not consider the added risk for atherosclerotic CVD known to be associated with chest-directed radiotherapy (12). Unfortunately, risk calculators for survivors of childhood cancer that include radiotherapy and other treatment exposures do not currently accommodate specific lipid abnormalities (6,26). With incorporation of this work’s findings into atherosclerotic CVD risk prediction for childhood cancer survivors, there may be indication for earlier lipid-lowering therapy in this population, potentially decreasing the burden of cardiovascular morbidity and mortality among survivors.

The identification of specific lipoprotein and triglyceride abnormalities among survivors is novel and important, as LDL cholesterol greater than 190 alone denotes eligibility for statin therapy, and HDL cholesterol and LDL cholesterol are used to predict 10-year atherosclerotic CVD risk (12). Compared with community controls matched for age, sex, and race and ethnicity, survivors had higher LDL cholesterol and non-HDL cholesterol and lower HDL cholesterol, potentially warranting earlier lipid-lowering therapy among survivors. Previous evaluations of the SJLIFE cohort ascertained the prevalence of and risks associated with dyslipidemia overall. Hudson and colleagues (9) identified dyslipidemia in 491 of 807 at-risk survivors, with 186 having diagnosis prior to SJLIFE screening, 256 diagnosed as part of screening, and 49 diagnosed subsequent to initial SJLIFE evaluation. Within the CCSS, Armstrong and colleagues (3) identified self-reported dyslipidemia to be more prevalent among survivors than sibling controls and found that dyslipidemia imparted a 25-fold increased risk for CAD among survivors exposed to chest radiotherapy and a ninefold increased risk for HF among survivors exposed to anthracyclines (3). However, these and other publications did not report abnormalities and risks related to specific lipoprotein abnormalities (3,5,9,10). The 2018 ACC–AHA Guideline on the Management of Blood Cholesterol recommends the use of the updated ACC ASCVD Risk Estimator Plus, using HDL cholesterol and LDL cholesterol to identify 10-year atherosclerotic CVD risk and recommends lipid-lowering therapy in those with an intermediate (7.5%) or greater 10-year atherosclerotic CVD risk (12). The data reported in this study underscore the need for evaluation and treatment of dyslipidemia among at-risk survivors, so those at risk for CVD can be identified for intervention, as well as the need for a survivor-specific atherosclerotic CVD risk calculator that incorporates both treatment exposures and specific lipid measures.

Children’s Oncology Group guidelines recommend biennial fasting lipid panels for childhood cancer survivors who received total body irradiation or abdominal radiotherapy, however, no frequency is specified for other survivors, instead referring to US Preventive Services Task Force guidelines for dyslipidemia screening (27). In the general population, however, there remains a lack of consensus regarding initial age for dyslipidemia screening as well as the interval for lipid panels, especially among those aged 20-39 years (this and other survivor cohorts have mean age of approximately 35 years) (28,29). Given the known earlier cardiovascular morbidity and mortality among childhood cancer survivors and marked rates of newly diagnosed dyslipidemia in this cohort, there may be justification to provide additional guidance for earlier and more frequent lipid evaluations among survivors (3,24). Future work should include the evaluation of atherosclerotic CVD risk reduction attributable to dyslipidemia screening and therapy among survivors.

Overall poor control of previously diagnosed dyslipidemia was also identified in this analysis, with many participants having uncontrolled dyslipidemia in all categories, including 13% with uncontrolled LDL cholesterol and 18% with uncontrolled non-HDL cholesterol. The 2018 ACC –AHA Guidelines on the Management of Blood Cholesterol places emphasis on reduction of LDL cholesterol once a diagnosis of dyslipidemia is made, including reduction below LDL cholesterol targets utilized in the general population, especially for those with higher overall atherosclerotic CVD risk (12). The finding of uncontrolled LDL cholesterol and non-HDL cholesterol among survivors previously diagnosed with dyslipidemia underscores the need for continued dyslipidemia follow-up among this population.

Dyslipidemia evaluation and treatment is complementary to assessment and management of other modifiable cardiovascular risk factors that increase CVD risk, including hypertension, smoking, obesity, and diabetes. These concomitant conditions occurred more frequently among survivors with dyslipidemia as compared with community controls. A previous evaluation in the CCSS cohort found increasing risk for CVD as the number of modifiable risk factors increased (3). For example, hypertension alone was associated with a sixfold increased risk of CAD among survivors exposed to chest radiotherapy, but history of both hypertension and dyslipidemia increased this relative risk to 21 (3). The long-term care plan of childhood cancer survivors should include recommendations for reduction of these modifiable risk factors as well as cardiovascular risks associated with the specific anticancer agents received. Although we found increased dyslipidemia rates among those who received cranial irradiation and total body irradiation, we did not find increased rates of dyslipidemia among those who received anthracyclines and platinum agents, as has been found in other studies (30-33).

The study’s strengths include complete lipid panel evaluation among a large population of long-term survivors, with evaluation of traditional and cancer treatment risk factors associated with dyslipidemia diagnosis. The study evaluated the risks associated with newly diagnosed dyslipidemia as part of SJLIFE screening toward 10-year atherosclerotic CVD risk and the development of MI and cardiomyopathy. The main limitation of the study is the lack of comprehensive laboratory data from the time of dyslipidemia diagnosis among those previously diagnosed with dyslipidemia. For this reason, risk profiles for dyslipidemia and CVD development were evaluated primarily for survivors without previously diagnosed dyslipidemia (n = 3406). Future survivor studies could establish atherosclerotic CVD risk on initial dyslipidemia diagnosis and determine the risk reduction imparted by continued lipid evaluation and treatment. Additionally, the etiology of cardiomyopathy in this population (ischemic vs nonischemic) was not able to be ascertained, as SJLIFE evaluation does not include coronary assessment. Another potential limitation is that all participants were treated for their childhood cancer at a single institution. Though previous evaluation of SJLIFE patients has shown participants to have similar exposures to that of other children treated in North America during this time period, these results may not be generalizable to all survivors (34-36). A predominance of non-Hispanic White individuals is a known limitation of the dataset. Given the known effects of race on dyslipidemia and cardiovascular outcomes, future survivor datasets should be powered to evaluate these differences.

In summary, these results identify an important burden of dyslipidemia among childhood cancer survivors, associated with risk of cardiovascular morbidity. This and future similar work can help establish evidenced-based guidelines for more robust dyslipidemia evaluation for survivors as compared with nonsurvivors. These data can inform clinical trials of dyslipidemia treatment to assess risk reduction for cardiovascular morbidity and mortality among survivors diagnosed with dyslipidemia.

Data availability

Data available upon request.

Author contributions

Jason F. Goldberg, MD, MS (Data curation; Formal analysis; Investigation; Writing—original draft; Writing—review & editing), Geehong Hyun, PhD (Data curation; Formal analysis), Kirsten K. Ness, PhD (Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing—review & editing), Stephanie B. Dixon, MD, MPH (Investigation; Methodology; Writing—review & editing), Jeffrey A. Towbin, MD (Investigation; Validation; Writing—review & editing), Isaac B. Rhea, MD (Conceptualization; Investigation; Writing—review & editing), Matthew J. Ehrhardt, MD (Conceptualization; Investigation; Methodology; Writing—review & editing), Deo Kumar Srivastava, PhD (Data curation; Formal analysis; Investigation; Validation), Daniel A. Mulrooney, MD (Conceptualization; Funding acquisition; Investigation; Methodology; Project administration; Writing—review & editing), Melissa M. Hudson, MD (Conceptualization; Data curation; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Writing—review & editing), Leslie L. Robison, PhD (Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Validation; Writing—review & editing), John L. Jefferies, MD (Conceptualization; Investigation; Methodology; Supervision; Writing—review & editing), Anand Rohatgi, MD (Conceptualization; Data curation; Investigation; Methodology; Supervision; Writing—review & editing), and Gregory T. Armstrong, MD MSCE (Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing—review & editing).

Funding

Support to St. Jude Children’s Research Hospital provided by the National Cancer Institute (R01 CA157838, G. Armstrong, Principal Investigator; U01 CA195547, M. Hudson and K. Ness, Principal Investigators), the Cancer Center Support (CORE) grant (P30 CA21765, C. Roberts, Principal Investigator), and the American Lebanese-Syrian Associated Charities (ALSAC).

Conflicts of interest

GA, who is a JNCI Associate Editor and co-author on this paper, was not involved in the editorial review or decision to publish the manuscript. No conflicts of interest exist.

Acknowledgements

Funders provide support of research costs and salary. The funders had no role in the study design; collection, analysis, or interpretation of the data; or the writing of the manuscript and decision to submit it for publication.

References