Use of cancer-directed therapy at the end of life among adolescents and young adults

Abstract

Background

Adolescents and young adults frequently receive chemotherapy near death. We know less about the use of targeted agents and immunotherapy or trends over time.

Methods

We conducted a retrospective cohort study of 1836 adolescents and young adults with cancer who died between 2009 and 2019 after receiving care at 1 of 3 sites (Dana-Farber Cancer Institute, Kaiser Permanente Northern California, and Kaiser Permanente Southern California). We reviewed electronic health data and medical records to examine use of cancer-directed therapy in the last 90 days of life, including chemotherapy, targeted therapy, immunotherapy, and investigational drugs.

Results

Over the study period, 35% of adolescents and young adults received chemotherapy in the last 90 days of life; 24% received targeted therapy, 7% immunotherapy, and 5% investigational drugs. Additionally, 56% received at least 1 form of systemic cancer-directed therapy in the last 90 days of life. After adjustment for patient sex, race, ethnicity, age, site of care, diagnosis, and years from diagnosis to death, the proportion of adolescents and young adults receiving targeted therapy (odds ratio [OR] = 1.05 per year of death, 95% confidence interval [CI] = 1.02 to 1.10; P = .006), immunotherapy (OR = 1.27, 95% CI = 1.18 to 1.38; P < .0001), and any cancer-directed therapy (OR = 1.04, 95% CI = 1.01 to 1.08; P = .01) in the last 90 days of life increased over time.

Conclusions

More than half of adolescents and young adults receive cancer therapy in the last 90 days of life, and use of novel agents such as targeted therapy and immunotherapy is increasing over time. Although some adolescents and young adults may wish to continue cancer therapy while living with advanced disease, efforts are needed to ensure that use of cancer-directed therapy meets preferences of adolescents and young adults approaching death.

Decisions about cancer-directed therapy in patients with advanced disease are often complex. Conversations about prognosis, inherent prognostic uncertainty, patients’ ability to tolerate treatment, and patient preferences guide decisions about continued efforts to treat the disease. Yet, late life chemotherapy is not without its costs; among older adults, palliative chemotherapy is not associated with improvements in patient quality of life and can come at the expense of increased suffering (1-3). As a result, organizations such as the American Society of Clinical Oncology have cautioned against late life chemotherapy for patients with advanced disease (4).

The use of cancer-directed therapy at the end of life has received limited attention in adolescents and young adults, for whom cancer is the leading disease-related cause of death (5). Previous work has identified high rates of medically intensive measures at the end of life among adolescents and young adults, including high rates of chemotherapy utilization (6-8). However, prior work in adolescents and young adults has not examined use of newer agents such as targeted therapies or immunotherapy, even though consideration of such agents may better reflect a changing landscape of cancer care delivery in the current era.

In addition, prior work in adolescents and young adults has focused on chemotherapy use in the last 2 weeks of life (6-8), consistent with existing quality measures for older adults (9). However, cancer treatment in the last 3 months of life has also been used in older adults to offer a more complete picture of care for patients living with advanced cancer (10). During this time, many adolescents and young adults have their first conversations about prognosis, palliative care, and hospice, and many begin to reframe their care around quality of life rather than cure (11,12). Although cancer treatment during this phase may have greater prospect of benefit than at the very end of life, it also requires consideration of patient goals of care and any potential impact, positive or negative, on quality of life. In addition, many early phase trials exclude patients with a life expectancy of less than 3 months, underscoring concerns about use of early phase therapy during this time (13).

To more fully understand adolescents and young adults’ use of systemic cancer treatment at the end of life, we examined trends in use of cancer-directed therapy in the last 3 months of life among adolescents and young adults with cancer, including chemotherapy, targeted therapy, immunotherapy, and investigational agents over time at 3 large centers from 2009 to 2019. We also examined use of chemotherapy in the last 2 weeks of life, which is an existing quality measure, as well as newly initiated cancer therapy in the last 14 and 30 days of life.

Methods

We included adolescents and young adult cancer patients who died between January 2009 and December 2019 after receiving care at 1 of 3 sites: Dana-Farber Cancer Institute (DFCI), Boston, Massachusetts; Kaiser Permanente Northern California (KPNC); and Kaiser Permanente Southern California (KPSC). DFCI is a comprehensive cancer center that provides specialized cancer care to children, adolescents, and adults. KPNC and KPSC are integrated health plans and care delivery systems that provide comprehensive health services to members throughout California.

Each site used electronic health record (DFCI) or cancer registry (KPNC, KPSC) data to identify eligible adolescents and young adults aged 12-39 years at death, an age range chosen to encompass existing definitions of adolescents and young adult (14-16). We included patients enrolled in the KPSC or KPNC insurance plans or receiving care at DFCI more than 90 days prior to death. Eligible patients were diagnosed with stage IV cancer or stage I-III cancer or leukemia with either a new metastasis by International Statistical Classification of Diseases version 9 codes or more than 1 chemotherapy regimen with more than 90 days between episodes of administration, indicating likely recurrence (17,18). Manual chart review was used to confirm recurrence. Site investigational review boards approved the study.

Data were obtained using a combination of electronic health data and medical records from each site. At DFCI, eligible patients were identified using an electronic clinical database, the Oncology Data Retrieval System, which was also used to ascertain sociodemographic information, treatment information, and outpatient appointments. Medical records were used to obtain additional data elements such as inpatient and outpatient care delivered.

KPNC and KPSC maintain internal cancer registries that report cancers diagnosed or treated at health system–affiliated medical centers to the National Cancer Institute Surveillance, Epidemiology and End Results Program. Each site also maintains research and clinical databases with complete information on membership, diagnosis, procedures, pharmacy and infusions, health-care utilization, and outside (non-network) claims. Medical records were used for additional inpatient and outpatient information when necessary.

Whenever possible, we used electronic clinical databases for measures used in this study. Trained medical record abstractors at each site also conducted manual medical record review to abstract data elements that were not available in databases. Records were first abstracted by 2 reviewers, and conference calls were conducted weekly across all 3 sites to ensure consistency. We assessed agreement, and once agreement reached at least 95%, reviewers began coding independently, with second abstraction in 10% of records. We continued monthly conference calls to maintain consistency and address any discrepancies.

Data collected included patient sex; dates of birth, diagnosis, and death; cancer type and stage; race; ethnicity; and treatment received. We evaluated use of systemic cancer-directed therapy in the last 90 days of life, including chemotherapy, targeted therapy, immunotherapy, and investigational drugs. We also evaluated use of chemotherapy in the last 14 days of life, which is an existing quality measure for older adults (9), as well as initiation of new cancer treatment in the last 30 and 14 days of life. At DFCI, treatment information was abstracted manually from medical records. At KPNC and KPSC, treatment information was available via infusion databases and pharmacy records, including medication names and administration or prescription dates and duration of supply.

Chemotherapy was defined as traditional cytotoxic chemotherapy. Targeted therapy was defined as medications that target proteins that control cell growth, division, and spread, including small molecule inhibitors and monoclonal antibodies targeting cell surface markers. Immunotherapy was defined as agents that use the immune system to target cancer, including checkpoint inhibitors, T-cell therapies, and vaccines. Investigational drugs were defined as medications that were being tested for safety or efficacy, either on clinical trials or via compassionate use mechanisms; of note, investigational status was not readily obtained via pharmacy records, which may have resulted in limited ascertainment of investigational drugs at Kaiser sites. A data dictionary including all medications defined in KPNC and KPSC pharmacy records as cancer-directed therapy, with further manual review by the study team to ensure that the list was comprehensive, was used to categorize each agent, ensuring consistency in coding. Each medication was classified in only 1 category; for example, investigational agents were always classified as investigational drugs even if they were also targeted agents or immunotherapy, and monoclonal antibodies were always classified as targeted agents.

Statistical methods

Descriptive statistics were used to summarize patient characteristics of the analytic cohort. The utilization of cancer-directed treatment at the end of life was summarized using proportions and corresponding 95% confidence intervals (CIs). Multivariable logistic regression models were used to assess associations between each modality of cancer-directed therapy and patient characteristics, including year of death, patient sex (male vs female), race (White, Black, Asian, and Other categories), ethnicity (Hispanic or Latino, non-Hispanic or Latino, and not documented), age at death (12-24 years, 25-39 years), cancer type (hematologic malignancy, solid tumor, brain tumor), and years from diagnosis to death, adjusted for site of care (DFCI, KPNC, and KPSC). Site adjustments were made to account for potential differences in data ascertainment as well as differences in unmeasured patient characteristics at each site. Odds ratios (ORs) for site of care are not shown. Year of death and years from diagnosis to death were included as continuous variables in the multivariable logistic regression models, without applying any transformation. The degree of association was summarized by odds ratio per unit increase for these continuous variables. A 2-sided P value of .05 was used for statistical significance. Use of investigational drugs was predominantly identified at 1 site (DFCI), which may be related to true differences in use or to differences in data ascertainment; we therefore limited multivariable models for investigational drugs to DFCI patients only.

Results

Patient characteristics

Just over half of included adolescents and young adults were female (55%; Table 1). A majority of adolescents and young adults were documented in records as White (61%), 8% were Black, and 12% Asian; 28% were documented as Hispanic or Latino. Race and ethnicity varied by site of care, with more White patients at DFCI (77%) than KPNC (43%) and KPSC (58%, P < .001; Supplementary Table 1, available online). Half of KPSC patients were documented as Hispanic compared with 22% at KPNC and 8% at DFCI (P < .001). Median age at death was 33 years.

Table 1.

Patient characteristics^a

Characteristics	Overall No. (%) (n = 1836)
Age at death
Mean (SD)	30 (7)
Median (IQR)	33 (25-37)
Age at diagnosis, y
Mean (SD)	28 (8)
Median (IQR)	30 (23-35)
Sex
Female	1006 (55)
Male	825 (45)
Not documented	5 (<1)
Race
American Indian or Alaska Native	5 (<1)
Asian	220 (12)
Black or African American	146 (8)
More than 1 race	10 (1)
Native Hawaiian or Other Pacific Islander	14 (1)
Other	38 (2)
White	1120 (61)
Not documented	283 (15)
Ethnicity
Hispanic or Latino	506 (28)
Not Hispanic or Latino	720 (39)
Not documented	610 (33)
Primary cancer site
Bone or soft tissue	251 (14)
Brain	161 (9)
Breast	251 (14)
Gastrointestinal	354 (19)
Genitourinary	206 (11)
Head, neck, thyroid	70 (4)
Leukemia	154 (8)
Lung	83 (5)
Lymphoma	108 (6)
Skin	61 (3)
Other	134 (7)
Not documented	3 (<1)
Stage at diagnosis
I-III	1045 (57)
IV	520 (28)
Not documented	271 (15)

a

IQR = interquartile range.

Cancer-directed therapy in the last 90 days of life

Over the 11-year study period, 35% (95% CI = 33% to 38%) of patients received chemotherapy in the last 90 days of life (Table 2); 24% (95% CI = 22% to 26%) received targeted therapy, 7% (95% CI = 6% to 8%) immunotherapy, and 5% (95% CI = 4% to 6%) investigational drugs. Just over half (56%; 95% CI = 53% to 58%) received at least 1 form of systemic cancer-directed therapy in the last 90 days of life. Of adolescents and young adults, 42% received only 1 form of systemic cancer therapy in the last 90 days of life, 13% received 2, and less than 1% received 3 (Supplementary Table 2, available online).

Table 2.

Cancer-directed therapy at the end of life^a

Characteristic	No. (%)	95% CI
Chemotherapy in the last 90 days of life	650 (35)	33% to 38%
Targeted therapy in the last 90 days of life	432 (24)	22% to 26%
Immunotherapy in the last 90 days of life	123 (7)	6% to 8%
Investigational drug in the last 90 days of life	89 (5)	4% to 6%
Any cancer-directed therapy in the last 90 days of life	1021 (56)	53% to 58%
Chemotherapy in the last 14 days of life	242 (13)	12% to 15%

a

CI = confidence interval.

Figure 1 graphically depicts changes in use of systemic therapy in the last 90 days of life over the study period. In multivariable models adjusted for patient age at death, sex, race, ethnicity, diagnosis, years between diagnosis and death, and site of care, the proportion of adolescents and young adults receiving targeted therapy (OR = 1.05 per year of death, 95% CI = 1.02 to 1.10; P = .006; Table 3) increased over the study period. Similarly, the adjusted odds of immunotherapy in the last 90 days of life increased over the study period (OR = 1.27 per year of death, 95% CI = 1.18 to 1.38; P < .001) as did use of any cancer-directed therapy (OR = 1.04, 95% CI = 1.01 to 1.08; P = .01). Use of chemotherapy and investigational drugs in the last 90 days of life did not change statistically significantly over the study period.

Figure 1.

Trends in the use of cancer-directed therapy in the last 90 days of life, 2009-2019.

Table 3.

Odds of cancer-directed therapy at the end of life.^a

	Chemotherapy in last 90 days of life		Targeted therapy in last 90 days of life		Immunotherapy in last 90 days of life		Investigational drug last 90 days of lifeb		Any cancer therapy in last 90 days of life		Chemotherapy in last 14 days of life
Characteristics	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Year of death	1.00 (0.96 to 1.03)	.8	1.05 (1.02 to 1.10)	.006	1.27 (1.18 to 1.38)	<.001	1.01 (0.90 to 1.12)	>.9	1.04 (1.01 to 1.08)	.01	1.00 (0.95 to 1.05)	>.9
Age at death, y
12-24	—		—		—		—		—		—
25-39	1.46 (1.13 to 1.90)	.004	1.03 (0.77 to 1.38)	.8	0.80 (0.51 to 1.28)	.3	0.49 (0.30 to 0.82)	.006	1.13 (0.88 to 1.44)	.3	1.16 (0.81 to 1.68)	.4
Sex
Female	—		—		—		—		—		—
Male	0.98 (0.79 to 1.21)	.8	1.39 (1.09 to 1.76)	.007	1.18 (0.79 to 1.77)	.4	0.81 (0.49 to 1.32)	.4	1.01 (0.83 to 1.24)	.9	0.85 (0.62 to 1.15)	.3
Race
Asian	1.20 (0.85 to 1.67)	.3	1.36 (0.94 to 1.94)	.1	0.67 (0.32 to 1.30)	.3	1.98 (0.75 to 4.69)	.14	1.23 (0.89 to 1.71)	.2	0.70 (0.38 to 1.21)	.2
Black	1.26 (0.85 to 1.86)	.3	1.09 (0.69 to 1.68)	.7	0.16 (0.03 to 0.52)	.01	0.93 (0.31 to 2.31)	.9	1.07 (0.73 to 1.57)	.7	1.25 (0.72 to 2.08)	.4
Other race or not documented	1.02 (0.74 to 1.37)	>.9	0.72 (0.50 to 1.02)	.07	0.67 (0.35 to 1.21)	.2	1.50 (0.65 to 3.25)	.3	0.82 (0.61 to 1.09)	.2	1.05 (0.67 to 1.62)	.8
White	—		—		—		—		—		—
Ethnicity
Hispanic or Latino	—		—		—		—		—		—
Not documented	1.05 (0.73 to 1.50)	.8	1.02 (0.67 to 1.56)	>.9	1.34 (0.65 to 2.80)	.4	3.67 (1.10 to 17.1)	.056	1.06 (0.75 to 1.49)	.8	0.77 (0.46 to 1.30)	.3
Not Hispanic or Latino	0.92 (0.67 to 1.25)	.6	1.05 (0.75 to 1.47)	.8	1.17 (0.64 to 2.17)	.6	3.74 (1.17 to 17.0)	.047	1.00 (0.75 to 1.34)	>.9	1.04 (0.68 to 1.61)	.9
Cancer type
Brain tumor	—		—		—		—		—		—
Hematologic cancer	2.73 (1.74 to 4.37)	<.001	0.52 (0.33 to 0.82)	.005	2.04 (0.96 to 4.64)	.07	0.83 (0.38 to 1.82)	.6	1.30 (0.86 to 1.96)	.2	3.43 (1.82 to 6.97)	<.001
Solid tumor	1.76 (1.18 to 2.69)	.007	0.48 (0.32 to 0.71)	<.001	1.16 (0.59 to 2.52)	.7	0.73 (0.38 to 1.46)	.4	1.04 (0.73 to 1.49)	.8	1.38 (0.76 to 2.73)	.3
Years from diagnosis to death	0.96 (0.93 to 1.00)	.05	0.97 (0.93 to 1.01)	.2	1.00 (0.93 to 1.06)	>.9	0.91 (0.82 to 1.0)	.06	0.97 (0.94 to 1.00)	.05	0.93 (0.87 to 0.98)	.02

a

Multivariable logistic regression models by increasing year of death, adjusted for site, patient race, ethnicity, cancer type, time from diagnosis to death, and age at death. Bold values are those in which P < .05. CI = confidence interval; OR = odds ratio.

b

Limited to adolescents and young adults from Dana-Farber Cancer Institute only given the small number at other sites who received investigational agents.

Additional factors were associated with cancer-directed therapy use in adjusted models. Chemotherapy use in the last 90 days of life was higher among older patients aged 25-39 years at death (OR = 1.46, 95% CI = 1.13 to 1.90; P = .004) relative to patients aged 12-24 years and among patients with hematologic malignancies (OR = 2.73, 95% CI = 1.74 to 4.37; P < .001) and solid tumors (OR = 1.76, 95% CI = 1.18 to 2.69; P = .007) relative to brain tumors. Black adolescents and young adults were less likely than White adolescents and young adults to receive immunotherapy (OR = 0.16, 95% CI = 0.03 to 0.52; P = .01), whereas Hispanic patients were less likely to receive investigational agents (OR for use of investigational agents among non-Hispanic patients = 3.74, 95% CI = 1.17 to 17.0; P = .047) relative to Hispanic patients. Older patients aged 25-39 years at death also had lower odds of receiving investigational agents (OR = 0.49, 95% CI = 0.30 to 0.82; P = .006) relative to patients aged 12-24 years. All models were adjusted for site of care, with the exception of the model for investigational drugs, performed for DFCI patients only.

Chemotherapy in the last 14 days of life

In the last 14 days of life, 13% of adolescents and young adults received chemotherapy (95% CI = 12% to 15%; Table 2). Use of chemotherapy in the last 14 days of life did not change over the study period (OR = 1.00, 95% CI = 0.95 to 1.05; P > .9). Patients with hematologic cancers had higher odds of chemotherapy use in the last 14 days of life (OR = 3.43, 95% CI = 1.82 to 6.97; P < .001) relative to brain tumor patients, and adolescents and young adults who lived a longer time between diagnosis and death had lower odds of receiving chemotherapy in this time frame (OR = 0.93, 95% CI = 0.87 to 0.98; P = .02). Figure 2 depicts use of chemotherapy in the last 14 days of life over the study period.

Figure 2.

Trends in the use of chemotherapy in the last 14 days of life, 2009-2019.

Newly initiated cancer-directed therapy in the last 14 and 30 days of life

Supplementary Table 3 and Supplementary Figure 1 (available online) show new therapies initiated in the last 14 and 30 days of life over the study period. Overall, 13% of adolescents and young adults started a new therapy in the last month of life; most included chemotherapy (9%), while 4% started a new therapy in the last 14 days of life, including chemotherapy for 2.4% of adolescents and young adults.

Discussion

In this study of more than 1800 adolescents and young adults, more than half received cancer-directed therapy in the last 3 months of life, including 35% who received chemotherapy. Furthermore, 13% of adolescents and young adults received chemotherapy in the last 2 weeks of life, when benefit is expected to be very limited. The rates of chemotherapy use that we found exceed rates seen in older adults; in one study of patients on Medicare, for example, 25% received chemotherapy in the last 3 months of life and 5% in their last 2 weeks (10). Although there is no established optimal rate of chemotherapy for adolescents and young adults approaching death, these findings are consistent with prior work showing heightened use of medically intensive care at the end of life among adolescents and young adults (6-8).

Adolescents and young adults may use cancer therapy at the end of life for a number of reasons. Work in pediatrics demonstrates that relatively few parents prioritize avoidance of chemotherapy at the end of life (19,20), and adolescents and young adults may feel the same way. Some adolescents and young adults may also want to do everything possible to live as long as possible, and young people may tolerate therapy better than older adults. In addition, oncologists and family caregivers may be reluctant to discontinue treatment in young people who should in other circumstances have their lives ahead of them. As a result, use of cancer therapy may be consistent with goals of care for many adolescents and young adults and their families.

However, we also know that conversations about prognosis and goals of care generally occur late in life among adolescents and young adults (11), raising the question of whether some adolescents and young adults might choose to forgo therapy if they recognized that benefits may be limited. Not all cancer therapy will offer greater length of life for adolescents and young adults, and it may come at cost of quality of life, as has been found in older adults (3). Late life treatment may also deprive adolescents and young adults of other aspects of supportive care, including hospice care, which often precludes use of cancer therapy. Strategies are needed to ensure that adolescents and young adults make decisions about cancer-directed therapy that reflect realistic expectations of risk and benefit and incorporate personal values. Previous work has shown that when goals of care discussions start earlier, before the last month of life, adolescents and young adults are less likely to use chemotherapy at the end of life (11). Ensuring that conversations about personal goals take place also meets preferences of adolescents and young adults, many of whom want to engage in conversations about their wishes for care and have their values heard and understood by care teams (21,22).

During the study period, use of targeted agents and immunotherapy increased. These agents may be better tolerated than conventional chemotherapy. However, that chemotherapy use remained constant as use of these agents rose is notable. Newer agents do not appear to be used in place of chemotherapy but rather in addition to it. This raises concern that overall care intensity for adolescents and young adults at the end of life may increase as new options for treatment emerge.

We also found that some subgroups of adolescents and young adults were more likely to receive cancer therapy at the end of life. Older adolescents and young adults and those with hematologic cancers tended to receive more chemotherapy, whereas younger patients tended to receive more investigational drugs. Patients with hematologic cancers such as leukemias are known to experience more intensive measures at the end of life, and the progression from recurrence to death can be foreshortened in many. In addition, early phase trials are more frequently available in pediatric settings, which may have influenced this finding; however, because most investigational agent use was identified at a single site in our study, this finding may not be generalizable to all centers. Finally, Black and Hispanic patients had lower use of immunotherapy and investigational drugs, respectively. Although patient race and ethnicity differed by site, findings were present after adjustment for site or, in the case of Hispanic patients receiving investigational drugs, were also present within the DFCI population, suggesting that these results are not solely secondary to site differences. Whether racial and ethnic differences represent disparities in access to newer agents or, alternatively, a greater preference to forgo these agents near the end of life is not known.

This study is limited by its focus on 3 large centers; data may not be generalizable to a wider population. The 3 centers include 2 health systems and 1 academic cancer center, which may have different approaches to use of cancer therapy. We also used different data ascertainment methods at the different sites. Use of pharmacy records resulted, for example, in low rates of identified investigational drugs at Kaiser sites, which may reflect the data curation process as well as possible true differences. Thus, although some variation by site was identified in multivariable models, our study was not designed to assess site-of-care differences as a primary predictor of differences in care received.

Also of note, although we examined use of cancer therapy in the last 3 months of life, we do not know how cancer therapy impacted quality or length of life. Newer agents in particular could have offered patients the opportunity to continue cancer-directed therapy after progression while also minimizing side effects of conventional chemotherapy. Further work is needed to better understand how type of cancer therapy impacts patient experiences during this phase of life. We also do not have information on use of all types of cancer-directed therapy in the last 14 days of life. Strengths include the large and racially and ethnically diverse population across an 11-year period and inclusion of newer agents.

Cancer-directed treatment at the end of life remains prevalent, with stable chemotherapy use, while other therapies are on the rise. Adolescents and young adults may use cancer treatment near the end of life for a range of reasons, in some cases reflecting personal goals to pursue cure or longer life. However, late life cancer treatment could also reflect unrealistic hopes for benefit and may come at a cost of quality of life. Continued efforts are needed to ensure that use reflects patient values and realistic expectations for outcomes.

Data availability

Underlying primary data will be made available upon request in compliance with university and institutional review board regulations. Materials may be transferred to others under the terms of a material transfer agreement. Access to data generated under the project will be available for research purposes. When data are shared, all personal identifiers will be removed at the time databases are provided, and variables that could lead to deductive disclosure of the identity of individual research patients will be redacted.

Author contributions

Jennifer Mack, MD, MPH (Conceptualization; Data curation; Funding acquisition; Investigation; Writing—original draft), Colin Cernik, MS (Data curation; Formal analysis; Writing—review & editing), Lanfang Xu, MS (Data curation; Formal analysis; Writing—review & editing), Cecile A. Laurent, MS (Data curation; Formal analysis; Writing—review & editing), Lauren Fisher, MPH (Data curation; Methodology; Project administration; Supervision; Writing—review & editing), Nancy Cannizzaro, BA (Project administration; Writing—review & editing), Julie Munneke, BA (Project administration; Writing—review & editing), Robert M. Cooper, MD (Conceptualization; Investigation; Writing—review & editing), Joshua R. Lakin, MD (Conceptualization; Investigation; Writing—review & editing), Corey M. Schwartz, MD (Conceptualization; Investigation; Writing—review & editing), Mallory Casperson, MS (Conceptualization; Investigation; Writing—review & editing), Andrea Altschuler, PhD (Conceptualization; Investigation; Writing—review & editing), Lori Wiener, PhD (Conceptualization; Investigation; Writing—review & editing), Lawrence Kushi, ScD (Conceptualization; Investigation; Writing—review & editing), Chun R. Chao, PhD (Conceptualization; Investigation; Writing—review & editing), and Hajime Uno, PhD (Conceptualization; Data curation; Formal analysis; Methodology; Writing—review & editing).

Funding

This study was supported by the National Cancer Institute (U01 CA218651 to JWM) and, in part, by the intramural program of the National Cancer Institute (LW).

Conflicts of interest

The authors have no conflicts of interest to disclose.

Acknowledgements

The National Cancer Institute was not involved in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.