The Burden of Late Effects and Related Risk Factors in Adolescent and Young Adult Cancer Survivors: A Scoping Review

Simple Summary It is unclear what the risk of negative health outcomes is after cancer during adolescence and young adulthood. We conducted a review to understand the risk of second cancers, chronic conditions, and death in adolescent and young adult (AYA) cancer survivors and found factors that increase the risk. In total, 652 studies were identified, of which 106 were included in the review: 23 for second cancers, 34 for chronic conditions, and 54 for deaths. The number of included studies increased over time, from four studies in 2010 to 17 in 2020. The studies found that AYA cancer survivors are at an increased risk of second cancers, chronic conditions, and deaths. In particular, the following factors increased risk: radiation exposure for second cancers; younger attained age and earlier calendar period of diagnosis for chronic conditions; and non-Hispanic Black or Hispanic, low socioeconomic status, and earlier calendar period of diagnosis for deaths. Abstract Risk factors associated with late effects in survivors of adolescent and young adult (AYA) cancer are poorly understood. We conducted a systematic scoping review to identify cohort studies published in English from 2010–2020 that included: (1) cancer survivors who were AYAs (age 15–39 years) at diagnosis and (2) outcomes of subsequent malignant neoplasms (SMNs), chronic conditions, and/or late mortality (>5 years postdiagnosis). There were 652 abstracts identified and, ultimately, 106 unique studies were included, of which 23, 34, and 54 studies related to the risk of SMNs, chronic conditions, and mortality, respectively. Studies investigating late effects among survivors of any primary cancer reported that AYA cancer survivors were at higher risk of SMN, chronic conditions, and all-cause mortality compared to controls. There was an indication that the following factors increased risk: radiation exposure (n = 3) for SMNs; younger attained age (n = 4) and earlier calendar period of diagnosis (n = 3) for chronic conditions; and non-Hispanic Black or Hispanic (n = 5), low socioeconomic status (n = 3), and earlier calendar period of diagnosis (n = 4) for late mortality. More studies including the full AYA age spectrum, treatment data, and results stratified by age, sex, and cancer type are needed to advance knowledge about late effects in AYA cancer survivors.


Introduction
An estimated 1.2 million adolescents and young adults (AYAs), aged 15-39 years, were diagnosed with cancer in 2020 globally [1,2]. Although the incidence of AYA cancer is increasing or stable in many countries [3], screening, diagnosis, and treatment continue to improve, leading to an overall decrease in all-cause and cancer-related mortality [4]. As a result, there is a growing population of AYA cancer survivors who will spend the majority of their lives at risk of late effects due to their cancer and its treatment. However, the burden of late effects and their associated risk factors in AYA cancer survivors are not well understood.
Although several systematic reviews investigating late effects in survivors of childhood [5][6][7][8] and adult [9][10][11] cancers have been published recently, to our knowledge, only one systematic literature review investigating late effects specifically among AYA cancer survivors has been published in the last decade, and that focused on the 16-29 year age range [12]. Reviews that examine late effects in both childhood and AYA cancers [13][14][15] discuss a striking lack of studies specifically investigating late effects in AYA cancer survivors relative to a robust body of literature for childhood cancer survivors [14]. Given important social and biological differences of AYA cancer, a review of the current evidence on the relationship between tumor-, treatment-, and patient-related risk factors and common late effects in AYA cancer survivors is warranted.
We thus sought to conduct a systematic scoping review to examine the current state of the literature studying late effects among AYA cancer survivors. Specifically, we aimed to describe the burden of subsequent malignant neoplasms (SMNs), chronic conditions (including hospitalizations and medication prescriptions as surrogates), and late mortality-key late effects, which lead to substantial years of life lost or living with a disability-in AYA cancer survivors and identify risk factors for each late effect. Finally, by summarizing the available literature, we aimed to identify knowledge gaps that will inform future research.

Eligibility Criteria
The three outcomes of interest in this study were (1) SMNs; (2) chronic conditions, which were physical or psychological in nature and diagnosed by a medical professional or for which a hospitalization or pharmacological prescription acted as a surrogate for a diagnosis (e.g., diagnosis of, hospitalization from, or prescription for a disease); and (3) late mortality, defined as a death occurring more than 5 years after the original cancer diagnosis. Studies that presented results for SMNs or chronic conditions were eligible if these conditions developed at any point after diagnosis. Studies were deemed eligible for inclusion in the scoping review if they met the following criteria defined a priori: (1) The study participants included individuals with a history of cancer who were AYAs (age 15-39 years) at the time of diagnosis, including studies not exclusively focused on AYAs (e.g., 0-19 year olds) if age-stratified results, which captured the AYA subgroup, were reported; (2) the study included one or more of the three outcomes defined above; (3) the research was a prospective or retrospective cohort study to allow the examination of the relationship between the AYA cancer exposure and the late effect; and (4) the study was original research published in English in the years 2010-2020. Conference abstracts, theses, reviews, and sources of grey literature were excluded. Case-control studies were not included because they are less suitable for producing evidence of causality and are more prone to bias than cohort studies.

Study Selection
We searched PubMed using three separate searches that aligned with each of the three outcomes of interest (Supplementary Table S1). Identified titles and abstracts published between 1 January 2010 and 31 December 2020 were screened independently by two reviewers (CRB and RLD). Disagreements were resolved by discussion until consensus

Study Selection
We searched PubMed using three separate searches that aligned with each of the three outcomes of interest (Supplementary Table S1). Identified titles and abstracts published between 1 January 2010 and 31 December 2020 were screened independently by two reviewers (CRB and RLD). Disagreements were resolved by discussion until consensus was reached. If consensus could not be reached, a third expert reviewer (MMFB) was consulted. This process was also used to screen the full text of articles. The reference lists of all included articles were then examined independently by CRB and RLD, and additional articles were included in the scoping review until a consensus among the research team was reached. The following information was then abstracted for each included study: primary author, date of publication, country of origin, purpose of the study, data source, outcome ascertainment (medical records, registries, self-report, etc.), study design, type(s) of cancer studied, diagnosis period, age at diagnosis, survival entry date, overall study size or AYA cancer population size within a larger cohort, length of follow-up, key findings, and identified risk factors. Patient-(e.g., demographic, social, and lifestyle), treatment-(e.g., type and year of treatment), and tumor-related (e.g., type and histology) risk factors were all eligible based on their potential to influence the occurrence of late effects. As per the methodology of scoping reviews, we did not appraise the quality or risk of bias of the included studies [16]. Where possible, we abstracted the most-adjusted risk estimates from the results.

Results
The three unique searches in PubMed yielded a total of 652 records (SMNs: 199, chronic conditions: 187, late mortality: 266) ( Figure 1). Hand-searching of included studies identified an additional 46 eligible studies. Following title and abstract screening, 225 records were deemed potentially relevant to the scoping review and eligible for full-text screening. After full-text screening, a total of 111 studies (SMNs: 23, chronic conditions: 34, late mortality: 54), of which 106 were unique.  All studies in this review had retrospective cohort designs. The number of included studies increased over time, with 44% being published from 2018-2020 ( Figure 2). The studies were conducted predominantly in North America (63%) and Europe (29%). Only 45% captured the entire 15-39 years age range, while a further 17% focused on a childhood cohort, which overlapped with the youngest AYAs. Half of the studies made withingroup comparisons, while the remainder used an external comparison group, primarily the general population (30%), childhood cancer survivors (12%), or siblings (5%). Sixty (54%) studies included participants from a mixed-cancer population, while the remaining studies examined specific tumors or tumor groups. In equal proportions, participants entered the study at the time of diagnosis (38%) or as 5-year survivors (38%). Finally, the median sample size across the three outcomes was 3053 AYA participants, ranging from 7 to 401,287.
All studies in this review had retrospective cohort designs. The number of included studies increased over time, with 44% being published from 2018-2020 ( Figure 2). The studies were conducted predominantly in North America (63%) and Europe (29%). Only 45% captured the entire 15-39 years age range, while a further 17% focused on a childhood cohort, which overlapped with the youngest AYAs. Half of the studies made withingroup comparisons, while the remainder used an external comparison group, primarily the general population (30%), childhood cancer survivors (12%), or siblings (5%). Sixty (54%) studies included participants from a mixed-cancer population, while the remaining studies examined specific tumors or tumor groups. In equal proportions, participants entered the study at the time of diagnosis (38%) or as 5-year survivors (38%). Finally, the median sample size across the three outcomes was 3053 AYA participants, ranging from 7 to 401,287.
. Figure 2. The number of studies included in the scoping review for each outcome of subsequent malignant neoplasms, chronic conditions, and mortality according to year of publication. Duplicates (studies containing information on more than one outcome) were removed from the line indicating the total number of studies.

Subsequent Malignant Neoplasms
Studies investigating the development of any kind of SMN among a mixed-cancer population and reporting a single overall estimate for the standardized incidence ratio (SIR) found that AYA cancer survivors were 1.6 (95% confidence interval [95% CI]:1.6-1.6) to 4.3 (95% CI:3.6-5.1) times more likely to experience an SMN relative to that of a primary malignant neoplasm expected in the general population (Table 1) [17][18][19][20]. Absolute excess risk (AER) for developing an SMN in these same cohorts ranged between 15.9 (95% CI:12.1-19.8) and 25.9 (95% CI not reported) per 10,000 person-years at risk. The risk of developing specific types of SMNs ranged widely, with SIRs ranging from 1.1 (95% CI:1.1-1.2) for breast cancer to 2.0 (95% CI:1.9-2.0) for lung cancer and 3.0 (95% CI:2.6-3.5) for meningeal cancer.   For the development of any SMN after a specific first primary cancer diagnosis, SIRs ranged from 1.2 (95% CI:1.1-1.4) after a bladder cancer diagnosis in men [24] to 9.0 (95% CI:3.4-16.5) after a diagnosis of a bone cancer [24,26]. In particular, Hodgkin lymphoma (HL) survivors experienced the highest risk of SMN, which was 3 to 8 times that of the general population, [17,24] with an excess of 0.1 (95% CI not reported) to 111 (95% CI not reported) SMNs per 10,000 person-years [30,32].

Risk Factors Associated with Subsequent Malignant Neoplasms
The influence of sex on SMN risk was inconsistent (Figure 3; Supplementary  [21]. Investigations of specific primary tumor groups suggested that female sex was a risk factor for SMN in HL survivors (HR:1.8, 95% CI: 1.0-1.3), [32] but protective in melanoma (HR:0.7, 95% CI:0.6-0.8) and thyroid (OR:0.6, 95% CI:0.5-0.7) cancer survivor populations [34,38]. Similarly, older age at diagnosis was found to increase the risk of developing an SMN overall and for melanoma and thyroid cancer survivors, [25,34,38] while other studies have associated younger age at diagnosis or treatment with increased risk of SMNs in HL survivors [24,29,30].
Cancers 2021, 13, x 8 of 28 Figure 3. Descriptive representation of the direction of effect for patient-, tumor-, or treatmentrelated factors described in three or more studies on the risk for subsequent malignant neoplasms, chronic conditions and hospitalizations, and late mortality among AYA cancer survivors in cohorts with combined cancer types. Deeper colors represent more studies (n = 5) and white cells represent that no study investigated that risk factor (n = 0). Reference categories appear in parentheses. Abbreviations: AAPI, Asian-American and Pacific Islander; NS, non-significant; SES, socioeconomic status; SMN, subsequent malignant neoplasm.
Possible treatment-related risk factors were reported in eight studies. Four studies reported that having radiotherapy elevated the risk of SMNs up to 2.7 times (95% CI:1.0-7.7) that expected, among AYA cancer survivors overall and among specific tumor groups [17,18,25,32], while two studies reported no significant effect of radiotherapy on the risk of SMNs in HL survivors [25,31]. Conversely, a study noted more than a 4-fold (95% CI:3.3-5.4) increase in the risk of SMNs in survivors of HL who received radiotherapy in addition to chemotherapy, relative to those who received chemotherapy only [28]. Treatment with chemotherapy for any primary malignancy increased the risk of SMN (SIR:6.3, 95% CI:4.1-9.4) [18]. Finally, the cumulative incidence of developing an SMN was higher for survivors of acute lymphoblastic leukemia (ALL) or non-HL (NHL) who received hematopoietic transplant compared to those who did not [37,39].

Chronic Conditions
Compared to control groups, AYAs with any primary malignancy had a 38-90% increased risk of first hospitalization (Table 2) [40,41]. In particular, AYA survivors of leukemia, NHL, HL, and central nervous system (CNS), head and neck, and bone tumors were about 4 times (95% CI:3.7-4.8) [42] more likely to develop chronic conditions, as well as more severe chronic conditions that led to hospitalization, relative to the general population or a matched population [40,41,[43][44][45][46][47][48][49]. Similarly, a cohort of AYA survivors of leukemia, CNS malignancy, HL, NHL, Wilms tumor, neuroblastoma, soft-tissue sarcoma, and bone cancer had a 4.2-fold (95% CI:3.7-4.8) increased risk for developing one or more severe or disabling health conditions compared to siblings [42].  Possible treatment-related risk factors were reported in eight studies. Four studies reported that having radiotherapy elevated the risk of SMNs up to 2.7 times (95% CI:1.0-7.7) that expected, among AYA cancer survivors overall and among specific tumor groups [17,18,25,32], while two studies reported no significant effect of radiotherapy on the risk of SMNs in HL survivors [25,31]. Conversely, a study noted more than a 4-fold (95% CI:3.3-5.4) increase in the risk of SMNs in survivors of HL who received radiotherapy in addition to chemotherapy, relative to those who received chemotherapy only [28]. Treatment with chemotherapy for any primary malignancy increased the risk of SMN (SIR:6.3, 95% CI:4.1-9.4) [18]. Finally, the cumulative incidence of developing an SMN was higher for survivors of acute lymphoblastic leukemia (ALL) or non-HL (NHL) who received hematopoietic transplant compared to those who did not [37,39].
Various cancer therapies, including chemotherapy, radiotherapy, and stem cell transplants, were investigated for their role in the development of chronic conditions. In a mixed-cancer cohort, the combination of chemotherapy, radiotherapy, and surgery was associated with an 80% increase in chronic conditions compared to chemotherapy alone (95% CI:1.1-3.1), but no other treatment combinations were associated with significantly more conditions [47]. When assessed individually, receipt of radiotherapy increased the risk of stroke (IRR:3.5, 95% CI:5.9-37.2), diabetes (IRR:1.9, 95% CI:1.3-2.9), and thyroid disorders (IRR:3.1, 95% CI:2.2-4.4) for AYA survivors of any type of cancer [64], as well as diseases of the nervous system for survivors of brain tumors (HR:3.3, 95% CI:1.8-6.2) [68]. Radiotherapy was linked to a 3.4-fold risk of endocrine (total body or chest vs. none) and cardiac conditions (≥15 Gy chest radiation vs. none) for survivors of bone cancers, leukemia, CNS malignancies, HL, NHL, Wilms tumor, neuroblastoma, and soft tissue sarcoma [42]. However, radiotherapy was not associated with the overall risk of health conditions among ALL survivors [39]. Exposure to cytotoxic drugs was associated with an increased risk of CVD among 2-year survivors of any type of cancer (3.5% vs. 2.0% for radiotherapy only) [46] but was not associated with an overall increase in hospital-related morbidity among survivors of HL [32]. Finally, among survivors of NHL, HL, and ALL, stem cell transplants were also associated with an increased overall risk of more than 20 conditions of various organ systems [37,39,67].

Reference Cancer Type Number of Aya Participants Outcome Ascertainment Results
Mixed-cancer cohort Armenian S, 2020 [81] Any primary malignancy 10,574 SEER Survival rate of AYA cancer survivors was 78.5% at 25 years after diagnosis, but was at a 10.4-fold increased risk of death compared to noncancer controls (IRR = 10.4, 95% CI 9.7-11.2). Absolute excess risk for death from any cause was 12.7 per 1000 person-years (95% CI, 11.9-13.4 per 1000 person-years). Fifteen years post-diagnosis, incidence of second cancer mortality exceeded the rate of recurrence-related mortality. Lowest long-term survival in breast cancer survivors (25 years: 59.8%) and the highest long-term survival in thyroid cancer survivors (25 years: 95.3%).

201
Electronic medical records Ten-year overall survival for AYAs was about 78%.

Lymphoma cohort
Anton-Culver, 2010 [83] NHL 3489 Death certificates Overall, 1081 of 3489 people died in the study cohort. The most common causes of death were due to lymphoma-related causes and human immunodeficiency virus.
Castellino, 2011 [84] HL 1273 United States National Death Index The HR for risk of death from any cause for the 15-21 age group was 1.1 (95% CI 0.6-2.0), relative to the <10 age group. Eight percent of adolescent ALL survivors in the cohort died between 5-9.99 years after the start of the study. Less than one percent died at 10 or more years. The seven-year overall and leukemia-specific survival for cord blood transplant patients was 55% and 48%, respectively, compared to sibling-allo-HCT, which was 63% and 61%, respectively. Ten-year cumulative testicular cancer-specific mortality rate for seminoma and nonseminoma was 1.4% (95% CI 1.2% to 1.7%) and 6.1% (95% CI 5.7% to 6.7%), respectively. Significantly decreased mortality was observed for participants aged 40 at diagnosis for seminoma HR: 2.0 (95% CI 1.5 to 2.6) and nonseminoma HR: 2.1 (95% CI, 1.7 to 2.6).

Risk Factors Associated with Late Mortality
Among AYAs diagnosed with any type of primary cancer, females experienced significantly less excess mortality compared to males when adjusted for years of follow-up, calendar period, and comorbidities (HR:0.7; 95% CI:0.6-0.7) (Figure 3; Supplementary  Table S4) [70]. One study provided evidence that the conditional survival advantages, specifically for circulatory and respiratory disease deaths, experienced by females for many cancers were most pronounced in the years closer to diagnosis [76]. Females were also reported to have an overall survival advantage relative to males in the following tumorspecific populations: NHL (HR:0.6, 95% CI:0.6-0.8) [83], HL (HR:0.8, 95% CI:0.7-0.9) [31,86], melanoma (HR:0.7, 95% CI:0.6-0.8) [101], AML (HR:0.8, 95% CI:0.7-0.9) [85] and thyroid cancer (HR:0.4, 95% CI:0.3-0.5) [105]. When specific causes of death were assessed, two studies found that males surviving any type of cancer were more at risk of dying from respiratory conditions [77], whereas female survivors of HL and NHL were more at risk of dying from CVD [74]. There was a lack of consensus as to how age at diagnosis affects the risk of late mortality (Figure 3). One study reported significantly more excess mortality  [72]. When assessed for specific tumor types, AYAs diagnosed at an older age compared to a younger age also had poorer survival outcomes after NHL [83], HL [86], testicular cancer [106], Ewing sarcoma [108], thyroid cancer [105], and melanoma [101].
Regarding tumor-related risk factors, having more extensive disease or a higher stage of cancer at diagnosis [81,83,102,107,110] was associated with increased risks in late mortality for tumor-specific and mixed-cancer cohorts. Excluding one study [18], studies reported that excess mortality from AYA cancer decreased over calendar time, with people diagnosed, treated, or enrolled in studies in later years experiencing better survival outcomes [31,32,70,76,80,90,100,106]. For treatment-related risk factors, the evidence associating radiotherapy, chemotherapy, and surgery with late mortality was mixed among the included studies. Two studies conducted in a combined cancer population associated radiotherapy with a 2.0-(95% CI:1.3-3.1) and 1.5-(95% CI:1.0-2.1) fold increase in death from any cause and SMNs, respectively [18,81], while tumor type-specific studies reported non-significant associations [105,110] or decreased risks [31]. Conversely, among survivors of any type of AYA cancer, the risk of late mortality from any cause was not statistically different based on receipt of chemotherapy [18], though specific tumor types, such as bone and soft tissue sarcomas (SMR:3.2, 95% CI:2.3-4.3) [104], were found to have increased risk of late mortality if exposed to chemotherapy. Finally, not having surgery was associated with an increase in cancer-specific death among survivors of head and neck squamous cell carcinoma (HR:1.6; 95% CI:1.2-2.1) [110] but not in combined cancer [18] or bone and soft tissue sarcoma cohorts [104].

Discussion
For AYA cancer survivors, the development of late effects after cancer can exacerbate the challenges of being a young person, such as balancing social relationships, new careers, and education with limited practical knowledge and financial resources [113,114]. Recognizing the poor understanding of late effects in AYA cancer survivors, this scoping review was conducted to identify published peer-reviewed research that reported on SMNs, chronic conditions, and late mortality risks in this under-researched population. Our findings emphasize the high burden of negative health outcomes experienced by AYA cancer survivors later in life. What remains less clear, based on the evidence from this review, is which factors contribute most to the increased risk of developing late effects, in part due to the lack of detailed treatment exposure investigations and the high level of heterogeneity among the studies. Genetic and lifestyle factors may also modify the mechanisms of established relationships between cancer therapies and late effects [115,116], yet references to factors like obesity, smoking, and physical activity were scarce within the reviewed literature [46].
It is understood generally that cancer treatments, namely, radiotherapy and chemotherapy, are associated with morbidity and premature mortality, and multimodal therapies intensify these risks [47,[117][118][119]. Studies included in this review reported that AYAs in mixed-cancer cohorts experience an estimated 2.6-, 1.9-, and 10.4-fold increase in the risk for SMNs, hospitalization from any health condition, and 25-year all-cause mortality, respectively, compared to control groups [25,81,120]. Studies of childhood cancer survivors have shown a dose-response relationship between radiotherapy and the development of secondary sarcomas and breast, thyroid, CNS, and gastrointestinal cancers [22,[121][122][123]. Similarly, specific chemotherapeutic drugs have been identified as risk factors for the late effects investigated in this review, with specific thresholds (e.g., high-dose intravenous methotrexate defined as any single dose ≥1000 mg/m 2 ) noted by the Children's Oncology Group Long-Term Follow-Up Guidelines for Survivors of Childhood, Adolescent and Young Adult Cancers [124]. Based on the literature identified in this review, detailed treatment-related risk factors cannot be identified, as few studies include the necessary treatment exposure data to make such assertions. While entities such as the International Late Effects of Childhood Cancer Guideline Harmonization Group are developing guidelines for long-term follow-up for AYA cancer survivors, the evidence used is drawn primarily from studies of childhood cancer survivors [125].
Treatment type is an important predictor of late morbidity, but unmodifiable risk factors may also play a role [117]. Sex-related differences in survival may be related to biological factors such as sex hormones and immune response, behavioral factors such as females' increased self-awareness of their bodies [126] or tendency towards health-seeking behaviors compared to men [127], and clinical factors such as differences in screening practices (e.g., colorectal cancer) [70]. Age-related differences in the development of late effects may be fundamentally different between younger and older AYAs, as 15-20 year-olds may still be undergoing growth associated with puberty and the rapid proliferation of tissues brought on by sex hormones [117]. These tissues are particularly vulnerable to damage caused by radiation, potentially affecting the maturation of organs and systems associated with teenaged growth spurts (e.g., gonads and musculoskeletal system) [117,128]. In contrast, AYAs aged 20-39 may be exposed to a different set of carcinogenic exposures at work (e.g., asbestos) or in life (e.g., alcohol and smoking) that might alter their risk profile for late effects.
Variable background risks for mortality and chronic conditions also contextualize the findings in this review. Among studies that examined the risk of chronic conditions by attained age, most found that compared to their counterparts in the general population, AYAs of a younger attained age had a heightened risk of developing these late effects than AYAs of an older attained age. This finding may be explained by the background risk of chronic conditions increasing with age, leading to less excess morbidity among older AYA cancer survivors.

Opportunities for Future Research
The identification of effective cancer treatments with fewer short-and long-term side effects is a research priority for AYA cancer survivors [129]. Before interventions can be improved for long-term safety, the risks associated with current treatments must be better understood. Through this scoping review, we can propose several avenues for future research to help elucidate existing knowledge gaps. Broadly, there is a paucity of evidence generated from large, well-characterized cohort studies with lengthy follow-up, which include persons aged 15-39 and report results according to tumor type, sex, and age at diagnosis [130]. The existence of a body of such evidence would allow health professionals to stratify the risk of their patients for late effects and refer high-risk patients for suitable interventions [119]. Detailed treatment information, such as chemotherapy and radiation dosing, is also needed. Unlike childhood cancer survivors, for whom clinical cohorts such as the North American Childhood Cancer Survivor Study [131] and the St. Jude Lifetime Cohort Study [132] exist, few cohorts of AYA cancer survivors include this level of granularity due to the substantially greater number of AYA cancer survivors and the fact that many cohorts are established using population-based cancer registries in which only crude treatment information is recorded [49].
Another area of future investigation relates to SMNs, as it was the least investigated late effect that we explored. SMNs are a well-studied outcome after childhood cancer and a study included in this scoping review found that AYAs surviving any type of cancer have a higher absolute risk of developing an SMN compared to children or older adult survivors [17]. Lung cancer is reported to account for a large proportion of these excess cancers in AYA cancer survivors, highlighting the need to consider the effect of modifiable lifestyle factors, like smoking, as risk factors for late effects [24]. Some estimates suggest that 35% of second cancers developed by adult cancer survivors can be attributed to the adverse effects of alcohol and smoking [133].
Indeed, we identified only one study [46] that studied lifestyle factors such as physical activity, overweight and obesity, alcohol use, or smoking as possible risk factors for late effects in AYA cancer survivors, despite the potential for poor health behaviors to modify the risk of therapy-related complications, with or without interaction with genetics [115,134]. This area is one of active research within adult oncology that has led to changes in clinical practice [115,135] as positive health behaviors have shown some promise for shielding cancer survivors from SMNs, chronic conditions, and premature mortality, both directly and indirectly. Given that AYA cancer survivors have a significantly higher prevalence of unhealthy lifestyle behaviors relative to individuals with no history of cancer [136][137][138], this knowledge gap presents another important opportunity for future work [118].
Finally, there is a dearth of evidence generated from populations in low-and middleincome countries (LMICs), despite these countries bearing the vast majority of cancer deaths [139]. Compared to high-income countries, survival rates in LMICs are estimated to be 11% to 81% lower, meaning that many people will not survive their cancer long enough to develop late effects [140]. Underdiagnoses, misdiagnoses, delayed presentation, and unavailability of treatment or abandonment of therapy contribute to lower rates of survival [141]. Some of these challenges may be corrected by rectifying the general shortage of healthcare professionals who manage patient loads that are often significantly higher than in most high-income countries [141]. Resource constraints further limit access to cancer drugs deemed essential by the World Health Organization [142,143] as well as the capacity for population-based cancer registries [144]. Without adequate cancer-related surveillance, it is difficult to develop effective policy priorities for improving cancer outcomes and health equity in LMICs. AYA oncology should be a priority matter due to the significant social and economic implications of loss of life due to cancer in this age group [145], which accounts for 40% of the world's population [114].

Limitations
Despite the comprehensive nature of our scoping review, we must recognize several limitations. First, we found substantial heterogeneity across the included studies in terms of cancer types, subgroups, data sources, reference groups, methodology, and analysis. Additionally, we have not examined the quality of the evidence generated from each of the studies because grading of evidence was beyond the scope of this review. As a result, the findings of this review should be interpreted with caution as the included studies have not been assessed according to their individual clinical and methodological contexts. Another limitation of our review is that the search strategy was limited to one language, one database, and 11 years of evidence, which may contribute to selection bias and the lack of studies conducted in LMICs. Generally, English-language restrictions have not been shown to impact the conclusions of systematic reviews [146]. Additionally, we decided a priori to not include "childhood" or "adult" in our search terms, even though the AYA age group can overlap with each of these more general definitions. However, we are confident that through hand-searching of reference lists, we identified most or all childhood and adult survivor cohorts with results that are relevant to this review.

Conclusions
AYA cancer survivors experience a high level of morbidity many years into their survival. In this comprehensive scoping review, we summarized peer-reviewed literature published since 2010, describing the late-effect burden and examining patient-, tumor-, and treatment-related risk factors associated with late effects after AYA cancer. In doing so, this review highlighted substantial gaps in knowledge about the experience of AYA cancer survivors compared to what is known about children and middle-aged and older adults. Large, observational cohort studies including the full AYA age spectrum, long follow-up, detailed treatment exposure data, and results stratified by tumor type, age at diagnosis, and sex are still needed to help reduce the disease burden and increase quality of life in this unique cancer survivor population.