Abstract
Background
Health-related quality of life (HRQoL) was recently demonstrated to improve throughout therapy for high-risk pediatric Hodgkin lymphoma (HL); however, average scores may not reflect individual differences. This study aimed to identify subgroups of patients with similar HRQoL trajectories from pre- to post-therapy.
Methods
AHOD1331 trial participants aged 11-20 (n = 268; mean [SD] age = 15.6 [1.9]; 48% male) completed the Child Health Ratings Inventories–Global scale (HRQoL) prior to treatment, after cycle 2, after cycle 5, and the end of treatment. Group-based trajectory models (GBTMs) identified latent clusters of individuals with similar HRQoL patterns over time. Multivariable multinomial logistic regression estimated the association between a priori defined characteristics and membership in trajectory-based groups. Log-rank tests examined differences in post-T4 progression-free survival (PFS) by trajectory groups.
Results
GBTM identified 3 HRQoL groups: Group 1 (consistently unfavorable [25.7%]), Group 2 (moderate-and-increasing [44.8%]), and Group 3 (consistently favorable [29.5%]). Older age (odds ratio = 1.24, 95% confidence interval = 1.03 to 1.50; P = .022), female sex (2.48, 1.23 to 4.99; P = .011), and Hispanic ethnicity (2.31, 0.97 to 5.50; P = .059) were associated with increased odds of membership in Group 1 vs Group 3. Older age (1.18, 1.00 to 1.39; P = .038) and B-symptoms (2.18, 1.09 to 4.33; P = .027) were associated with increased odds of Group 2 membership vs Group 3. Group membership was not associated with post-T4 PFS.
Conclusions
A subgroup of high-risk pediatric HL patients experience persistently poor HRQoL, starting at diagnosis and continuing through therapy. Age, female sex, Hispanic ethnicity, and B-symptoms were linked to worse HRQoL. These findings can help identify patients at higher risk for poor HRQoL and guide intervention.
ClinicalTrials.gov
Introduction
Each year approximately 1200 children and adolescents are diagnosed with classic Hodgkin lymphoma (HL) in the United States.1 Although overall survival is greater than 95% in North America, patients endure intensive treatments associated with significant acute side effects and late effects.2,3 Therefore, contemporary trials in high-risk and advanced stage disease have focused on evaluating novel therapies that maintain disease control but minimize acute and long-term burden. Health-related quality of life (HRQoL) is a multidimensional concept that incorporates a variety of domains including physical, emotional, and psychological well-being.4 Measuring HRQoL as reported by the patient is important to understand the impact of illness and identify when and how to intervene to mitigate acute and long-term effects of therapy. Additionally, HRQoL has been associated with progression-free survival (PFS) and overall survival (OS) among patients with cancer.5-9
Until recently, little was known about HRQoL among pediatric patients with classical HL at diagnosis and throughout therapy. We recently reported that, on average, self-reported HRQoL improved during treatment among patients with high-risk HL, treated on the Children’s Oncology Group (COG) AHOD1331 trial.10 Patients on this trial were randomized to receive either Brentuximab vedotin (BV) with a multiagent chemotherapy backbone of AVE-PC (doxorubicin, vincristine, etoposide, prednisone, and cyclophosphamide, hereafter “BV arm”) or ABVE-PC (doxorubicin, bleomycin, vincristine, etoposide, prednisone, and cyclophosphamide, hereafter “standard arm”). The BV arm had superior therapeutic efficacy compared with the standard arm (cumulative incidence of relapse 7.5% vs 17.1%).11 Additionally, patients on the BV arm experienced a faster and more significant improvement in HRQoL.10 These data indicated that despite comparable rates of disease burden and toxicity between study arms at early imaging assessment, treatment with BV-AVEPC was associated with a more rapid and sustained improvement in/HRQOL. However, it is possible that there is significant heterogeneity in individual patient’s experience in HRQoL that is masked by reporting overall group means.
Therefore, in this secondary analysis of longitudinal self-reported HRQoL data in the patient-reported outcomes (PROs) cohort (N = 309) from AHOD1331, we aimed to identify groups of patients who experienced similar trajectories of HRQoL using group-based trajectory modeling.12,13 We also aimed to identify patient and disease factors associated with these trajectories and examine how the trajectories predict PFS. Understanding if there is a subgroup or groups of patients who did not experience improvement in HRQoL is critical to our understanding of the patient experience in pediatric and adolescent cHL. Furthermore, understanding risk factors for poor HRQoL will help us identify individuals at high risk and design tailored interventions.
Methods
Participants
COG study AHOD1331 (NCT02166463) enrolled patients with newly diagnosed high-risk classical HL (stage IIB with bulk, IIIB, or IVA and IVB) aged 2-21 years.11 Patients were ineligible if they had nodular lymphocyte predominant HL, were pregnant, had a known immunodeficiency, or received systemic corticosteroids within 28 days of enrollment. The first half of the planned enrollment sample was included in a required, preplanned PRO sub-study (n = 309). Demographic and clinical characteristics did not differ between the PRO sub-study participants and the subsequent trial enrollees.14 Participants 11 to 20 years old (n = 280) self-reported their HRQoL and were eligible for these analyses.
Participants self-reported demographic characteristics at the time of study enrollment and clinical characteristics were abstracted from electronic health records. Participants were asked to self-report race, which was further classified into Black or African American, White, Other, or not reported; the other group included those who reported they were American Indian or Alaskan Native, Asian, or Native Hawaiian or other Pacific Islander. For this analysis, patients were followed for disease progression and survival through March 31, 2024. Disease progression was defined by modified Lugano criteria.11,15 The protocol was reviewed and approved by the Pediatric Central Institutional Review Board, and the local IRBs of participating sites. Written informed consent from patients 18 years of age or older and parents/guardians and child assent for those younger than age 18 was obtained.
Treatment protocol
Patients were randomly assigned to receive 5 cycles of chemotherapy in 1 of 2 study arms: ABVE-PC (standard arm) or BV-AVE-PC (BV arm), as previously described.11 Patients with large mediastinal adenopathy (LMA) and/or those with slow-responding lesions based on interim PET scans (at second cycle) received involved site radiation therapy (ISRT) after the fifth cycle of chemotherapy.
HRQoL
The Child Health Ratings Inventories (CHRIs)–Global scale is a 7-item unidimensional global HRQoL measure that has been validated in pediatric cancer populations and is available in English and Spanish.4,10 The scale’s reference period was either “in general” (eg, “doing overall”) or regarding the previous 7 days. Responses were captured by a 5-point Likert scale. Scores are scaled from 0 to 100 where higher scores indicate better functioning. Participants completed the CHRIs global scale prior to the initiation of protocol therapy (T1), day 8 of cycles 2 (T2) and 5 (T3), and 6-8 weeks after the end of all planned therapy, inclusive of radiation (T4).
Statistical analysis
To identify trajectories of HRQoL, we used group-based trajectory models (GBTMs) which are a class of discrete mixture models. This method assumes that the population is composed of distinct groups with distinct trajectories. It is an extension of mixture modeling that uses maximum likelihood methods to estimate membership probabilities for multiple trajectories.13 The SAS PROC TRAJ macro package was used for this analysis (http://www.andrew.cmu.edu/user/bjones/).12 Scores from all 4 HRQoL assessment time points were used to generate trajectory groups. Several criteria12,13,16 were used to determine the ideal number of groups: (1) improvement in Bayes factor (approximated as 2 times the changes in Bayesian Information Criterion [BIC]) in the saturated model (greater number of groups) relative to the null model (fewer groups), (2) minimum group size (at least 5% of sample), (3) average posterior probability of group assignment (at least 70%), and (4) clinical interpretability. Whereas complete data at all times points is not a requirement of GBTM, we conducted sensitivity analyses that repeated the GBTM among those with complete data at all time points to ensure our results were robust.
Once the HRQoL trajectories were identified, mean HRQoL scores for each time point within each trajectory group were generated using linear mixed models to account for missingness. The fixed effects were categorical time (T1, T2, T3, and T4), trajectory group, and a time by trajectory group interaction. We identified a priori pretreatment covariates that may be associated with group membership, including age, sex, race, ethnicity, parent employment, household income, parent education, insurance, stage, histology, large mediastinal mass (LMA), and any B-symptoms (fever, unintentional weight loss, or significant fatigue). Because GBTM uses data from all time points, including prior to treatment, all covariates examined must have been measured prior to treatment initiation. Multinomial logistic regression was used to estimate the association between each a priori identified covariate and group membership. Any covariate with a Type III effect P value less than .10 was subsequently included in a multivariable multinomial logistic regression model. Kaplan-Meier curves and log-rank tests were used to compare post-T4 PFS (based on a March 31, 2024, data freeze) by trajectory group membership, where post-T4 PFS was defined as the time from completion of the T4 assessment to the earliest of progression or death, whichever occurred first. PFS was chosen as the outcome based on previously reported associations between PFS and HRQoL.5-7 Cox proportional hazards models were used to quantify between-trajectory group effects and test for a study arm by group membership interaction, as the primary trial results yielded a statistically significant difference in the cumulative incidence of relapse between study arms.11 All statistical testing was 2-sided and considered statistically significant at P < .05. All analyses were completed using SAS 9.4 (SAS Institute, Cary, NC, United States).
Results
In total, 268 participants completed the CHRIs-Global at one or more time points and are included in these analyses (Figure S1). Demographic and clinical characteristics were well balanced by study arm. Overall, participants were 52% female and 57% Non-Hispanic White; 60% had LMA, and 58% had stage IV disease (Table 1).
Table 1.
Participant characteristics.
| ABVE-PCa | Bv-AVEPCb | Overall | |
|---|---|---|---|
| (n = 137) | (n = 131) | (n = 268) | |
| Age, mean (SD) | 15.8 (1.9) | 15.5 (1.9) | 15.6 (1.9) |
| Sex | |||
| Female | 74 (54.0%) | 65 (49.6%) | 139 (51.9%) |
| Male | 63 (46.0%) | 66 (50.4%) | 129 (48.1%) |
| Race | |||
| Black or African American | 15 (10.9%) | 15 (11.5%) | 30 (11.2%) |
| Otherc | 5 (3.6%) | 5 (3.8%) | 10 (3.7%) |
| Unknown or not reported | 14 (10.2%) | 8 (6.1%) | 22 (8.2%) |
| White | 103 (75.2%) | 103 (78.6%) | 206 (76.9%) |
| Ethnicity | |||
| Hispanic or Latino | 23 (16.8%) | 23 (17.6%) | 46 (17.2%) |
| Not Hispanic or Latino | 108 (78.8%) | 101 (77.1%) | 209 (78.0%) |
| Unknown or not reported | 6 (4.4%) | 7 (5.3%) | 13 (4.9%) |
| Parent employment | |||
| Full time | 69 (50.4%) | 66 (50.4%) | 135 (50.4%) |
| Part-time | 14 (10.2%) | 14 (10.7%) | 28 (10.4%) |
| Full-time homemaking | 24 (17.5%) | 23 (17.6%) | 47 (17.5%) |
| Other | 16 (11.7%) | 14 (10.7%) | 30 (11.2%) |
| Unknown | 14 (10.2%) | 14 (10.7%) | 28 (10.4%) |
| Household income | |||
| <$20 000 | 29 (21.2%) | 15 (11.5%) | 44 (16.4%) |
| $20 000-$39 999 | 21 (15.3%) | 24 (18.3%) | 45 (16.8%) |
| $40 000-$59 999 | 22 (16.1%) | 15 (11.5%) | 37 (13.8%) |
| $60 000-$79 999 | 12 (8.8%) | 15 (11.5%) | 27 (10.1%) |
| ≥$80 000 | 36 (26.3%) | 42 (32.1%) | 78 (29.1%) |
| Unknown | 17 (12.4%) | 20 (15.3%) | 37 (13.8%) |
| Parent education | |||
| <High school | 18 (13.1%) | 11 (8.4%) | 29 (10.8%) |
| High school | 41 (29.9%) | 41 (31.3%) | 82 (30.6%) |
| Some college | 46 (33.6%) | 43 (32.8%) | 89 (33.2%) |
| ≥College degree | 17 (12.4%) | 18 (13.7%) | 35 (13.1%) |
| Unknown | 15 (10.9%) | 18 (13.7%) | 33 (12.3%) |
| Insurance | |||
| Private | 74 (54.0%) | 65 (49.6%) | 139 (51.9%) |
| Public | 48 (35.0%) | 48 (36.6%) | 96 (35.8%) |
| Private and public | 5 (3.6%) | 3 (2.3%) | 8 (3.0%) |
| None | 2 (1.5%) | 5 (3.8%) | 7 (2.6%) |
| Unknown | 8 (5.8%) | 10 (7.6%) | 18 (6.7%) |
| Stage | |||
| IIB | 32 (23.4%) | 26 (19.8%) | 58 (21.6%) |
| III | 26 (19.0%) | 28 (21.4%) | 54 (20.1%) |
| IVA | 32 (23.4%) | 33 (25.2%) | 65 (24.3%) |
| IVB | 47 (34.3%) | 44 (33.6%) | 91 (34.0%) |
| Histology | |||
| Nodular sclerosis | 113 (82.5%) | 100 (76.3%) | 213 (79.5%) |
| Hodgkin lymphoma, NOS | 20 (14.6%) | 23 (17.6%) | 43 (16.0%) |
| Other | 3 (2.2%) | 8 (6.1%) | 11 (4.1%) |
| Unknown | 1 (0.7%) | 0 | 1 (0.4%) |
| LMA | |||
| Yes | 84 (61.3%) | 78 (59.5%) | 162 (60.4%) |
| No | 52 (38.0%) | 52 (39.7%) | 104 (38.8%) |
| Unknown | 1 (0.7%) | 1 (0.8%) | 2 (0.7%) |
| B symptoms | |||
| Yes | 104 (75.9%) | 97 (74.0%) | 201 (75.0%) |
| No | 31 (22.6%) | 33 (25.2%) | 64 (23.9%) |
| Unknown | 2 (1.5%) | 1 (0.8%) | 3 (1.1%) |
Abbreviations: B symptoms = fever, night sweats, weight loss; LMA = large mediastinal adenopathy; NOS = not otherwise specified.
ABVE-PC regimen includes doxorubicin, bleomycin, vincristine, etoposide, prednisone, and cyclophosphamide.
BV-AVE-PC regimen includes Brentuximab vedotin doxorubicin, vincristine, etoposide, prednisone, and cyclophosphamide, 3 LMA defined as transverse tumor diameter >1/3 the thoracic diameter at the dome of the diaphragm on a 6-foot posterior-anterior upright chest radiograph.
The other race group included American Indian or Alaskan Native, Asian, and Native Hawaiian or other Pacific Islander.
Identification of HRQoL trajectory groups
According to our selection criteria (Table 2), a model with 3 groups had the best fit and is shown in Figure 1. Three HRQoL trajectory groups were identified: consistently unfavorable HRQoL (Group 1 = 25.7%), moderate-and-increasing HRQoL (Group 2 = 44.8%), and consistently favorable HRQoL (Group 3 = 29.5%). Group 1 had a consistently unfavorable trajectory, which started with low HRQoL prior to therapy, was sustained until the end of therapy, and experienced only a slight improvement after the end of the therapy. Group 2 was considered to have moderate HRQoL prior to therapy, which improved over the course of therapy. Last, patients in Group 3 started with favorable HRQoL scores that were maintained throughout therapy. Within groups, mean HRQoL did not differ by study arm (Figure 2). In sensitivity analyses restricted to the 188 subjects who provided HRQoL data at all time points, trajectory groups were similar, with a slightly higher proportion of subjects in the consistently favorable group (Figure S2).
Table 2.
Group-based trajectory model selection.
| Number of groups | Group sizes | Probability of group assignment | BIC | 2(ΔBIC)a |
|---|---|---|---|---|
| 1 | 268 | 100 | −3950.41 | – |
| 2 | 131 (48.9) | 91.5 | −3843.31 | 107.1 |
| 137 (51.1) | 92.6 | |||
| 3 | 69 (25.7) | 86.9 | −3828.41 | 14.9 |
| 120 (44.8) | 83.8 | |||
| 79 (29.5) | 86.9 | |||
| 4 | 10 (3.7) | 87.2 | −3829.41 | −1 |
| 84 (31.3) | 85.2 | |||
| 102 (38.1) | 81.9 | |||
| 72 (26.9) | 85.5 |
Abbreviation: BIC = Bayesian Information Criterion.
Level of strength is based on Bayes Factor (tested model vs reference [null] model), where 2ln(Bayes factor) ≈ 2(ΔBIC). If 2ln(Bayes factor) ≤2, no evidence for tested model; 2 < 2ln(Bayes factor) ≤6, weak evidence for tested model; 6 < 2ln(Bayes factor) ≤10, Strong evidence for tested model; 2ln(Bayes factor)>10, very strong evidence for tested model (adapted from Jones et al.13).
Figure 1.
Mean health-related quality of life scores by trajectory group. Abbreviations: CHRIs = Child Health Ratings Inventories; CI = confidence interval.
Figure 2.
Mean Child Health Ratings Inventories (CHRIs) score by trajectory group and treatment arm. The means and 95% confidence interval (CI) for the ABVE-PC arm are presented with a hashed line the means and 95% CI for the BV-AVEPC arm are presented with a solid line.
Risk factors for group membership
Results from univariable analysis for pretreatment risk factors associated with group membership are presented in Table S1. Although treatment with BV was associated with a lower odds of membership in Group 1 (consistently unfavorable group) in univariate models (odds ratio [OR] = 0.50, 95% confidence interval [CI] = 0.26 to 0.97; P = .040; Table S1), the overall effect was not statistically significant (P = .121), and therefore, study arm was not included in multivariable models. In multivariable models (Table 3), older age (1.24, 1.03 to 1.50; P = .022), female sex (2.48, 1.23 to 4.99; P = .011), and Hispanic ethnicity (2.31, 0.97 to 5.50; P = .059) were associated with increased odds of membership in Group 1 (consistently unfavorable group) vs Group 3 (the consistently favorable group). Older age (1.18, 1.00 to 1.39; P = .038) and presence of any B-symptoms (2.18, 1.09 to 4.33; P = .027) were associated with increased odds of membership in the Group 2 (moderate-and-increasing trajectory group) compared with Group 3 (consistently favorable).
Table 3.
Odds of group membership associated with baseline demographic and clinical factors.
| Group 2: Moderate-and-increasing (vs Group 3: Consistently favorable) |
Group 1: Consistently unfavorable (vs Group 3: Consistently favorable) |
||||
|---|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | P a | |
| Age | 1.18 (1.00 to 1.39) | .038 | 1.24 (1.03 to 1.50) | .022 | .043 |
| Sex (ref = male) | |||||
| Female | 1.79 (0.98 to 3.29) | .060 | 2.48 (1.23 to 4.99) | .011 | .032 |
| Ethnicity (ref = not Hispanic or Latino) | |||||
| Hispanic or Latino | 1.07 (0.46 to 2.47) | .884 | 2.31 (0.97 to 5.50) | .059 | .074 |
| B symptoms (ref = no) | |||||
| Yes | 2.18 (1.09 to 4.33) | .027 | 1.79 (0.82 to 3.91) | .147 | .079 |
Treatment arm, race, parent employment, household income, parent education, insurance, stage, histology, and large mediastinal mass were not associated with group membership (type III P > .10; Table S1).
P value for type III effect.
Association with post-T4 PFS
The median follow-up for post-T4 PFS was 6 years (minimum = 0.00, maximum = 8.23), with 35 participants experiencing disease progression after T4. There was no association between trajectory group and post T-4 PFS (P = .711). Based on published difference in the cumulative incidence of relapse by study arm9 and a statistically significant interaction between study arm and group membership in Cox proportional hazards models for post-T4 PFS (P = .088), comparisons of post-T4 PFS by trajectory groups were conducted separately within study arm. Trajectory group membership was not associated with PFS in either study arm (BV arm, P = .115; or standard arm P = .265, Figure 3).
Figure 3.
Kaplan-Meier plots of post-T4 PFS by group-based HRQoL trajectory group, stratified by treatment group, along with log-rank tests for each. Panel A represents the post-T4 survival among those in the BV-AVE-PC arm and panel B represents post-T4 survival among the ABVE-PC arm. Group 1 (consistently unfavorable) is represented with a solid line, Group 2 (moderate and increasing) is represented with a dotted line, and Group 3 (consistently favorable) is represented by a dashed line.
Discussion
These data demonstrate that 25% of pediatric patients with high-risk HL experience poor HRQoL at diagnosis that persists throughout treatment, despite similar disease burden, high-risk features, and highly efficacious therapy.11 These data add important nuance to the previously reported results of an overall improvement of HRQoL based on group means over time.10 These previously reported group means are highly influenced by the majority and are masking the poor HRQoL experience of approximately 25% of patients. These new data highlight the interindividual heterogeneity of HRQoL and the importance of examining the variability between participants. Strikingly, HRQoL scores prior to treatment may be indicative of the HRQoL trajectory throughout therapy. Furthermore, we report that B-symptoms, female sex, older age, and Hispanic ethnicity in our sample were associated with membership in worse HRQoL trajectory groups. Identifying high-risk factors may help develop and apply potential targets for additional supportive care interventions.
Although we previously demonstrated overall improvement throughout therapy for pediatric high-risk HL, there appears to be a pretreatment separation of HRQoL that carries forward throughout treatment. Our data suggest that if patients have very low HRQoL prior to therapy, they are unlikely to see significant improvements throughout their course of therapy. This is the first time we have identified such patterns in pediatric HL. The presence of B-symptoms was associated with an increased odds of membership in the moderate-and-increasing HRQoL group (Group 2, P = .027), and to a lesser degree the consistently unfavorable group (Group 1, P = .079) relative to the odds of membership in the consistently favorable HRQoL group (Group 3), suggesting that systemic symptoms may be drivers of HRQoL in pediatric high-risk HL. Although patients on AHOD1331 experienced excellent disease control overall, trajectory groups were not associated with study arm and did not predict post-T4 PFS in our analysis.11 Therefore, factors other than disease burden must be contributing the persistent low HRQoL in a subgroup of HL. These may include anxiety, depression, cognitive reserve, resilience, fatigue, and social determinants of health, all of which were not measured in the present study.17 Future research is needed to understand the multifactorial biological and social drivers of initial and persistent low HRQoL in pediatric HL.
Nonetheless, HRQoL at diagnosis could serve as an effective screener for patients who may require additional supportive care resources throughout therapy and beyond. Consistent with previous studies of HRQoL in pediatric cancer,10,17,18 older adolescents, females, and Hispanic patients were more likely to belong to a worse HRQoL trajectory group and should also be targeted for symptom monitoring, need for psychosocial support, and other supportive care interventions that may improve HRQoL. Distressing physical and emotional symptoms have been previously associated with worse HRQoL, suggesting that effective symptom and toxicity management may improve HRQoL.19,20 However, although symptom monitoring was recently demonstrated to improve patient-reported symptom burden in a randomized trial of pediatric patients with cancer, it did not improve HRQoL.21 Patients experience significant stressors beyond disease burden and treatment toxicity due to disruption in roles and routines.22,23 These stressors may be ameliorated by existing interventions that demonstrate promising improvements in HRQoL including problem-solving skills training,24 yoga or physical activity,25,26 and art therapy.27,28 Future research is needed to explore the potential of combining symptom monitoring and management with these interventions to improve HRQoL in those most at risk for persistently low HRQoL.
HRQoL improved for approximately 75% of patients by the end of initial treatment. However, data from long-term survivors of pediatric HL demonstrate impairments in HRQoL across multiple domains, specifically in emotional and social functioning, and persistent fatigue.29-31 Furthermore, in our analysis, a subgroup of patients do not experience a resolution of their poor HRQoL by the end of therapy. Therefore, it is possible that poor HRQoL may persist from therapy into long-term survivorship in a subgroup of patients, rather than developing years later due to late effects. Previous work has demonstrated that coping skills are associated with HRQoL during the first 5 years after treatment for childhood cancer.32 It may also be related to survivors’ self-management and self-activation, decreasing their engagement in survivorship care to screen for and manage late effects that may impact long-term HRQoL. Additional longitudinal research that spans from therapy to long-term survivorship is needed to examine if poor HRQoL persists from treatment and identify the risk factors and drivers of this persistent trajectory to inform the timing and mode of intervention. This is of critical importance in survivors of HL who have been at an exceptionally high risk of late effects.33-35
We were unable to detect an association between HRQoL trajectory groups and post-T4 PFS. This is in contrast with several studies from adult patients with solid tumors and lymphoma demonstrating the predictive power of pretreatment HRQoL on survival outcomes.5,8,9,36 Given the small number of events (n = 35), it is possible that we were underpowered to detect such an association. It is also possible that the global measure of HRQoL used here is not related to survival in the same ways the multidimensional measures of HRQoL used in the adult setting are. Additional evidence is needed to confirm that HRQoL is not predictive of survival in pediatric high-risk HL and examine if age is modifying these associations. It may be that younger patients are better able to tolerate and adhere to therapy, having fewer competing comorbidities, and better HRQoL, diminishing HRQoL’s association with survival. Regardless, HRQoL may not be affecting the efficacy of treatment, but is impacting the patient’s overall well-being.
This analysis has several strengths, including the use of group-based trajectory modeling to identify previously unknown subgroups of patients whose HRQoL trajectory differs from the overall sample means. Although completion rates of HRQoL measures on treatment were very high (>90%), completion rates by the end of therapy (T4) were appreciably lower than during treatment. We previously published that those who were less likely to complete T4 were from historically underrepresented racial and ethnic backgrounds.10 Previous literature suggests these groups experience disproportionately lower HRQoL, which helps to explain why the proportion of patients in the consistently favorable group increased when we restricted the GBTM to a sample with complete data.37,38 Efforts are underway to tailor future studies to identify and overcome the barriers associated with low rates of PRO completion at the end of and after therapy. We also were unable to account for the impact of acute treatment-related toxicity on HRQoL trajectories. Because GBTM analyses included pretreatment time points, all risk factors for trajectory group membership had to be measured prior to treatment, excluding treatment factors such as acute toxicity and the use of radiation therapy. Although, per protocol, all participants with an LMA would be prescribed RT,11 we did not see an association between LMA and trajectory group membership. Additional measures, methods, and studies are needed to understand how acute toxicity during therapy, additional lines of therapy, blood-based biomarkers such as erythrocyte sedimentation rate, and social determinants of health may influence someone’s HRQoL trajectory. The small number of events limited our ability to detect associations with post-T4 PFS. Future pooled analyses may be better powered to detect such an effect.3 Last, the CHRIs is a global measure of HRQoL and we were unable to evaluate any specific HRQoL domains to examine potential aspects of persistent poor HRQoL. Building on the recent Adolescent and Young Adult PRO Initiative, led by Drs. Parsons and Roth,39 ongoing HL trials such as the AHOD2131 trial in early-stage HL (NCT05675410) have since included multi-domain PRO measures of HRQoL (eg, PROMIS 25 or 29) that will help to fill this gap.
In summary, a subgroup of patients with pediatric HL experience persistently poor HRQoL, whereas others experience improvements or sustained favorable HRQoL, from pre- to posttreatment. High-risk HL patients who are older adolescent age, female sex, and Hispanic ethnicity, or have B-symptoms need to be screened for targeted interventions to address the risk of persistently poor HRQoL.
Supplementary Material
Acknowledgments
We would like to thank all the participants of AHOD1331.
Role of the funders: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of data; or preparation of the manuscript for publication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Prior presentations: These data were presented, in part, at the International Society for Childhood Adolescent and Young Adult Hodgkin Lymphoma, Memphis, TN, in October 2023 and the International Symposium on Hodgkin Lymphoma in Cologne, Germany, in October 2024.
Contributor Information
AnnaLynn M Williams, Division of Supportive Care in Cancer, Department of Surgery, University of Rochester Medical Center, Rochester, NY, United States.
Angie Mae Rodday, Institute for Clinical Research and Health Policy Studies, Tufts Medical Center, Boston, MA, United States.
Lindsay A Renfro, Division of Biostatistics, Department of Population and Public Health Sciences, University of Southern California and Children’s Oncology Group, Los Angeles, CA, United States.
Yue Wu, Department of Biostatistics, University of Florida, Gainesville, FL, United States.
Tara O Henderson, Department of Pediatrics, University of Chicago Pritzker School of Medicine and Comer Children’s Hospital, Chicago, IL, United States.
Frank G Keller, Department of Pediatrics, Emory University School of Medicine, Atlanta, GA, United States.
Angela Punnett, Division of Hematology-Oncology, Department of Paediatrics, Hospital for Sick Children and University of Toronto, Toronto, ON, Canada.
David Hodgson, Department Radiation Oncology, University of Toronto, Toronto, ON, Canada.
Kara M Kelly, Department of Pediatrics, Roswell Park Comprehensive Cancer Center, University at Buffalo Jacobs School of Medicine and Biomedical Sciences, Buffalo, NY, United States.
Sharon M Castellino, Department of Pediatrics, Emory University School of Medicine, Atlanta, GA, United States.
Susan K Parsons, Institute for Clinical Research and Health Policy Studies, Tufts Medical Center, Boston, MA, United States.
Author contributions
AnnaLynn M. Williams (Conceptualization, Data curation, Formal analysis, Methodology, Resources, Software, Visualization, Writing—original draft, Writing—review & editing), Angie Mae Rodday (Conceptualization, Data curation, Formal analysis, Methodology, Visualization, Writing—original draft, Writing—review & editing), Lindsay A. Renfro (Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing), Yue Wu (Formal analysis, Methodology, Validation, Visualization, Writing—original draft, Writing—review & editing), Tara O. Henderson (Conceptualization, Funding acquisition, Investigation, Methodology, Writing—original draft, Writing—review & editing), Frank G. Keller (Conceptualization, Investigation, Writing—original draft, Writing—review & editing), Angela Punnett (Conceptualization, Writing—original draft, Writing—review & editing), David Hodgson (Conceptualization, Writing—original draft, Writing—review & editing), Kara M. Kelly (Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Resources, Writing—original draft, Writing—review & editing), Sharon M. Castellino (Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Resources, Writing—original draft, Writing—review & editing), and Susan K. Parsons (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing)
Supplementary material
Supplementary material is available at JNCI: Journal of the National Cancer Institute online.
Funding
This work was supported by the Children’s Oncology group and the National Cancer Institute at the National Institutes of Health to the Children’s Oncology Group (U10CA098543), NCTN Statistics and Data Center Grant (U10CA180899), NCTN Operations Center Grant (U10CA180886), the NCORP Grant (UG1CA189955), and an R00CA256356 award to A.M.W. This research was also supported by a research grant from the Leukemia and Lymphoma Society (PI S.K.P.), the St Baldrick’s Foundation (PI S.K.P.), and Seagen Inc (PI T.O.H.). The funding bodies did not have a formal role in the development or interpretation of the study.
Conflicts of interest
The authors report no conflicts of interest. T.H., who is a JNCI Associate Editor and coauthor on this paper, was not involved in the editorial review or decision to publish the manuscript.
Data availability
An individual-level de-identified dataset containing the variables analyzed in the primary results paper can be expected to be available upon request. Requests for access to COG protocol research data should be sent to: datarequest@childrensoncologygroup.org. Data are available to researchers whose proposed analysis is found by COG to be feasible and of scientific merit and who agree to the terms and conditions of use.
References
- 1. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74:12-49. [DOI] [PubMed] [Google Scholar]
- 2.Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 18 Regs Research Data + Hurricane Katrina Impacted Louisiana Cases, November 2018. Sub (1975-2016 Varying)—Linked to County Attributes—Total U.S., 1969-2017 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, released April 2019, based on the November 2018 submission. Accessed May 2022.
- 3. Castellino SM, Giulino-Roth L, Harker-Murray P. , et al. Children’s Oncology Group’s 2023 blueprint for research: Hodgkin lymphoma. Pediatr Blood Cancer 2023;70 Suppl 6:E30580. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Parsons SK, Shih MC, Mayer DK, et al. Preliminary psychometric evaluation of the Child Health Ratings Inventory (CHRIs) and Disease-Specific Impairment Inventory-Hematopoietic Stem Cell Transplantation (DSII-HSCT) in parents and children. Qual Life Res. 2005;14:1613-1625. [DOI] [PubMed] [Google Scholar]
- 5. Gotay CC, Kawamoto CT, Bottomley A, et al. The prognostic significance of patient-reported outcomes in cancer clinical trials. J Clin Oncol. 2008;26:1355-1363. [DOI] [PubMed] [Google Scholar]
- 6. Coates AS, Hürny C, Peterson HF, et al. Quality-of-life scores predict outcome in metastatic but not early breast cancer. International Breast Cancer Study Group. J Clin Oncol. 2000;18:3768-3774. [DOI] [PubMed] [Google Scholar]
- 7. Sloan JA, Zhao X, Novotny PJ, et al. Relationship between deficits in overall quality of life and non-small-cell lung cancer survival. J Clin Oncol. 2012;30:1498-1504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Thompson CA, Yost KJ, Maurer MJ, et al. Quality of life at diagnosis predicts overall survival in patients with aggressive lymphoma. Hematol Oncol. 2018;36:749-756. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Jung HA, Park S, Cho JH, et al. Prognostic relevance of pretreatment quality of life in diffuse large B-cell lymphoma patients treated with rituximab-CHOP: results from a prospective cohort study. Ann Hematol. 2012;91:1747-1756. [DOI] [PubMed] [Google Scholar]
- 10. Williams AM, Rodday AM, Pei Q, et al. Longitudinal health-related quality of life among patients with high-risk pediatric Hodgkin lymphoma treated on the Children’s Oncology Group AHOD 1331 Study. J Clin Oncol. 2024;42:3330-3338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Castellino SM, Pei Q, Parsons SK, et al. Brentuximab vedotin with chemotherapy in pediatric high-risk Hodgkin’s lymphoma. N Engl J Med. 2022;387:1649-1660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Jones BL, Nagin DS. Advances in group-based trajectory modeling and an SAS procedure for estimating them. Sociol Methods Res. 2007;35:542-571. [Google Scholar]
- 13. Jones BL, Nagin DS, Roeder K. A SAS procedure based on mixture models for estimating developmental trajectories. Sociol Methods Res. 2001;29:374-393. [Google Scholar]
- 14. Parsons SK, Rodday AM, Pei Q, et al. Performance of the FACT-GOG-Ntx to assess chemotherapy-induced peripheral neuropathy (CIPN) in pediatric high risk Hodgkin lymphoma: report from the Children’s Oncology Group AHOD 1331 study. J Patient Rep Outcomes. 2023;7:113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Cheson BD, Fisher RI, Barrington SF, et al. ; United Kingdom National Cancer Research Institute. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32:3059-3068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kim SY. Determining the number of latent classes in single- and multi-phase growth mixture models. Struct Equ Modeling. 2014;21:263-279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Bakker A, Streefkerk N, Bakker A, et al. A systematic review of health-related quality of life in children and adolescents during treatment for cancer. EJC Paediatr Oncol. 2023;2:100134. [Google Scholar]
- 18. Rodday AM, Terrin N, Parsons SK, HSCT-CHESS™ Study. Measuring global health-related quality of life in children undergoing hematopoietic stem cell transplant: a longitudinal study. Health Qual Life Outcomes. 2013;11:26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Rosenberg AR, Orellana L, Ullrich C, et al. Quality of life in children with advanced cancer: a report from the PediQUEST Study. J Pain Symptom Manage. 2016;52:243-253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Smith AW, Bellizzi KM, Keegan TH, et al. Health-related quality of life of adolescent and young adult patients with cancer in the United States: the adolescent and young adult health outcomes and patient experience study. J Clin Oncol. 2013;31:2136-2145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Dupuis LL, Vettese E, Grimes AC, et al. symptom screening linked to care pathways for pediatric patients with cancer: a randomized clinical trial. JAMA. 2024;332:1981-1991. 10.1001/jama.2024.19585 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Patterson V, Olsavsky A, Garcia D, et al. Impact of sociodemographic factors, stress, and communication on health-related quality of life in survivors of pediatric cancer. Pediatr Blood Cancer. 2024;71:e31001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Rodriguez EM, Dunn MJ, Zuckerman T, et al. Cancer-related sources of stress for children with cancer and their parents. J Pediatr Psychol. 2012;37:185-197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Viola AS, Kwok G, Levonyan-Radloff K, et al. Feasibility and acceptability of Bright IDEAS-Young Adults: a problem-solving skills training intervention. Cancers (Basel). 2022;14:3124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Orsey AD, Park CL, Pulaski R, et al. Results of a pilot yoga intervention to improve pediatric cancer patients’ quality of life and physical activity and parents’ well-being. Rehabil Oncol. 2017;35:15-23. [Google Scholar]
- 26. Wurz A, Chamorro-Vina C, Guilcher GM, et al. The feasibility and benefits of a 12-week yoga intervention for pediatric cancer out-patients. Pediatr Blood Cancer. 2014;61:1828-1834. [DOI] [PubMed] [Google Scholar]
- 27. Abdulah DM, Abdulla BMO. Effectiveness of group art therapy on quality of life in paediatric patients with cancer: a randomized controlled trial. Complement Ther Med. 2018;41:180-185. [DOI] [PubMed] [Google Scholar]
- 28. Cullen KR, DiMaggio-Potter ME, Klimes-Dougan B, et al. The impact of a creativity camp intervention on depression and well-being in adolescents. Child Psychiatry Hum Dev. 2024. 10.1007/s10578-024-01766-3 [DOI] [Google Scholar]
- 29. Calaminus G, Dorffel W, Baust K, et al. Quality of life in long-term survivors following treatment for Hodgkin’s disease during childhood and adolescence in the German multicentre studies between 1978 and 2002. Support Care Cancer. 2014;22:1519-1529. [DOI] [PubMed] [Google Scholar]
- 30. Zeltzer LK, Recklitis C, Buchbinder D, et al. Psychological status in childhood cancer survivors: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2009;27:2396-2404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Williams AM, Mirzaei Salehabadi S, Xing M, et al. Modifiable risk factors for neurocognitive and psychosocial problems after Hodgkin lymphoma. Blood. 2022;139:3073-3086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Fisher RS, Sharp KMH, Prussien KV, et al. coping trajectories and the health-related quality of life of childhood cancer survivors. J Pediatr Psychol. 2021;46:960-969. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Bhakta N, Liu Q, Yeo F, et al. Cumulative burden of cardiovascular morbidity in paediatric, adolescent, and young adult survivors of Hodgkin’s lymphoma: an analysis from the St Jude Lifetime Cohort Study. Lancet Oncol. 2016;17:1325-1334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Castellino SM, Geiger AM, Mertens AC, et al. Morbidity and mortality in long-term survivors of Hodgkin lymphoma: a report from the Childhood Cancer Survivor Study. Blood. 2011;117:1806-1816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Oeffinger KC, Stratton KL, Hudson MM, et al. Impact of risk-adapted therapy for pediatric Hodgkin lymphoma on risk of long-term morbidity: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2021;39:2266-2275. 10.1200/JCO.20.01186:JCO2001186 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Fukushima T, Suzuki K, Tanaka T, et al. Global quality of life and mortality risk in patients with cancer: a systematic review and meta-analysis. Qual Life Res. 2024;33:2631-2643. 10.1007/s11136-024-03691-3 [DOI] [PubMed] [Google Scholar]
- 37. Kirchhoff AC, Spraker-Perlman HL, McFadden M, et al. Sociodemographic disparities in quality of life for survivors of adolescent and young adult cancers in the behavioral risk factor surveillance system. J Adolesc Young Adult Oncol. 2014;3:66-74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Reeves TJ, Mathis TJ, Bauer HE, et al. Racial and ethnic disparities in health outcomes among long-term survivors of childhood cancer: a scoping review. Front Public Health. 2021;9:741334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Roth ME, Parsons SK, Ganz PA, et al. Inclusion of a core patient-reported outcomes battery in adolescent and young adult cancer clinical trials. J Natl Cancer Inst. 2023;115:21-28. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
An individual-level de-identified dataset containing the variables analyzed in the primary results paper can be expected to be available upon request. Requests for access to COG protocol research data should be sent to: datarequest@childrensoncologygroup.org. Data are available to researchers whose proposed analysis is found by COG to be feasible and of scientific merit and who agree to the terms and conditions of use.