Central nervous system (CNS) tumors are the most common solid tumors and the leading cause of death in childhood cancer.1,2 Advances in diagnosis and treatment have increased survival rates.3–5 According to the report of the Registry of Childhood Tumors of the Sociedad Española de Oncohematología Pediátrica (Spanish Society of Pediatric Oncohematology), the 5-year survival after diagnosis of a CNS tumor is 71%.6

Surviving a brain tumor is associated with a high risk of physical, neuropsychiatric and social late effects (LEs), especially in patients managed with radiotherapy (RTx).7 The most frequent LEs are neurologic, sensory and endocrine disorders, in addition to an increased cardiovascular risk. These disorders tend to be chronic and develop more than 5 years after treatment.8–14

Late effects involving different systems can also contribute to neurocognitive changes ranging from mild learning difficulties to severe deficits, with a risk of progressive neurologic impairment years after treatment.15–17

As follow-up progresses, CNS tumor survivors and their families need to redefine their priorities beyond the fear of recurrence to focus on LEs and their impact on social skills, independence and vocational goals.18–21 The literature on childhood brain tumor survivors describes a higher prevalence of psychological maladjustment and lower educational attainment, employment rates and frequency of partnered/romantic relationships. Psychosocial assessment is important in this population,22–24 as one of the main goals of follow-up in survivors should be to minimize the risk of exclusion.25–27

Health-related quality of life (HRQoL) questionnaires are widely used in the follow-up of CNS tumor survivors, and it is surprising that, despite the multiple LEs that affect this group of patients, there are studies that do not find differences in HRQoL when comparing them to healthy controls.12 On the other hand, a study found that childhood CNS tumor survivors overestimated their level of leadership-popularity and underestimated their isolation behaviors and victimization by peers.28 This evinces the need to explore other aspects that are not addressed in these instruments, such as adaptive behavior, including daily performance, independence and cultural expectations,29 as well as other psychosocial variables, since they are a group of survivors with specific risks and whose social functioning could be impaired.27,29,30

The aim of our study was to assess morbidity in a group of Spanish childhood brain tumor survivors beyond their HRQoL to cover aspects that have not been considered previously, such as adaptive behavior, self-esteem, physical activity as well as other social indicators such as independent living, education and relationship status, which may be important in the assessment of their functioning. We also sought to estimate the probable impact of radiotherapy on the physical and psychosocial areas under study.

We conducted a cross-sectional observational study in CNS survivors included in the childhood oncology database of our hospital.

The inclusion criteria were having undergone treatment between1990 and 2010, age less than 18 years at diagnosis and active user status in the public health care system. We excluded patients currently in treatment for recurrence.

In the survivor registry, we identified 264 patients with a diagnosis of CNS tumor. Of this total, 221 were active in the health care system and we were able to locate 195, of who 150 met the inclusion criteria and 111 agreed to participate in the study.

The study was approved by the ethics committee of the hospital and we obtained written informed consent from all the patients or, if they were under 18 years of age, from their parents.

All CNS tumor survivors were active patients at the long-term follow-up unit. We reviewed their health records to collect data on clinical variables including age, sex, tumor type and location, age at diagnosis, predisposing genetic syndromes, treatment received, medical specialties involved in their follow-up, reported LEs, and duration of follow-up.

We analyzed LEs in these patients based on the information documented in medical reports and their detection in visits to the long-term follow-up unit. When a new LE was detected, survivors were referred to the appropriate specialist.

We classified LEs based on the involved organ or system applying the Common Terminology Criteria for Adverse Events (CTCAE) as modified in the St. Jude Lifetime Cohort Study,31 summarized in Table 1.

All study participants completed the following questionnaires:

• 1

  36-Item Short Form Health Survey (SF-36) for assessment of HRQoL, which includes 8 subscales (physical functioning, role limitations [due to physical problems], bodily pain, general health, energy/fatigue, social functioning, role limitations [due to emotional problems] and mental health), with each item rated on a scale from 0 to 100.28

• 2

  Rosenberg self-esteem scale, which classifies self-esteem based on the score as low (0–15), normal (15–25) or high (25–40).32

• 3

  Lansky scale (in patients aged < 16 years) or Karnofsky scale (older patients) to assess functional/performance status.

• 4

  Vineland Adaptive Behavior Scales, Second Edition (Vineland-II), administered through a personal interview with the survivor, also including a parent/legal guardian in patients aged less than 18 years and adult patients who authorized their presence. It yields scores in four domains (communication, daily living skills, socialization and motor skills) plus an adaptive behavior composite score. Each skill is rated on a scale from 0 to 2, all scores are age-standardized (100 ± 15) and, depending on the results, adaptive behavior is classified as high (>130), moderately high (115–129), adequate (86–114), moderately low (71–85) or low (<70).29

• 5

  Physical Activity Questionnaire for Adolescents (PAQ-A) or International Physical Activity Questionnaire-Short Form (IPAQ-SF), depending on the age of the participant. The first yields a score calculated as the mean score for its 9 items, each of which is rated on a 5-point scale, and the second assesses walking, moderate-intensity activity and vigorous-intensity activity, calculating the time spent on each activity (minutes) and its frequency (days) weekly and categorizes the level of physical activity as high, moderate or low.33

• 6

  Sociodemographic data form, which includes variables such as marital status, partnered/romantic relationship status, educational attainment, independent living and having an active driver’s license.

• 7

  In addition, a neurocognitive evaluation was only performed in underage patients who visited the pediatric neuropsychiatry clinic of the hospital with their parents.

The study included 111 childhood CNS tumor survivors (70 male). The median current age was 20 years (range, 12–34), the median time elapsed from diagnosis was 13 years (range, 10–26) and the median age at diagnosis was 6 years (range, 1–18). Table 2 summarizes the clinical characteristics of the CNS tumor survivors, including the diagnosis, tumor location and received treatment.

Self-esteem

The score on the Rosenberg scale was 30.2 (SD, 2.1) in the RTx group compared to 31.3 (SD, 3.3) in survivors who did not receive RTx (P = .61).

The scores were below normal in 5.6% of patients and in the high range in 43%. We found no significant differences when we compared the scores in the two treatment groups.

We found a positive and moderately strong correlation between the total scores in the quality of life questionnaire and the self-esteem questionnaire (Spearman rho = 0.462; P = .001) (Fig. 1).

Figure 1.

Association between scores in the self-esteem scale and the health-related quality of life survey.

We found a positive and moderate-to-strong correlation between the self-esteem and quality of life scores (rho = 0.462; P = .001).

Rho, Spearman correlation coefficient.

(0.06MB).

Adaptive behavior

In the comparison of adaptive behavior scores (for each domain and overall composite score) in CNS tumor survivors based on the treatment history, we found that those who received radiotherapy scored significantly lower in communication, daily living skills and socialization (Table 4).

Table 4.

Comparison of scores on the Vineland-II scales (domain and composite scores) in CNS tumor survivors who received radiotherapy versus those who did not.

Vineland-II scores	Overall CNS tumor survivors, mean ± SD (range)	RTx group mean ± SD	No RTx group mean ± SD	P
Comunication	98.7 ± 18.4 (35−120)	93.1 ± 22.1	106.52 ± 8.2	.004
Daily living skills	96.8 ± 21.3 (30−130)	89.8 ± 21.4	107.73 ± 16.7	<.001
Socialization	96.7 ± 17.2 (32−130)	93.5 ± 21.3	101.71 ± 14.6	.024
Motor skills	96.9 ± 18.4 (58−130)	95.2 ± 18.7	100.48 ± 17.8	.102
Composite score	97.9 ± 18.4 (36−118)	92.8 ± 18.9	104.2 ± 12.5	.001

CNS, central nervous system; RTx, radiotherapy; SD, standard deviation.

We used multiple linear regression to assess the predictive value of the independent variables (age at diagnosis, radiotherapy treatment and number of LEs) in relation to the Vineland-II composite score. The general linear model explained 52.2% (R2 = 0.522) of the variance in the composite score (P < .01; F3,44 = 16; β = 2.63). Each additional year of age at diagnosis was associated with an increase in the dependent variable of the Vineland composite score (β = 1.00), in contrast to increases in the dose of radiation (β = −0.26) or in the number of LEs (β = −5.35), which predicted poorer scores in the Vineland-II (P < .01).

Based on the composite adaptive behavior score, we classified survivors into 2 groups: “adequate to moderately high” and “moderately low to low”. Patients with poorer adaptive behavior scores had a significantly longer duration of follow-up in years (P = .04) (Fig. 2).

Figure 2.

Comparison of CNS tumor survivors with moderately low to low versus adequate to high composite adaptive behavior scores based on duration of follow-up.

The Wilcoxon rank test found a significant difference in the duration of follow-up between the two groups.

P = .04.

(0.06MB).

The mean score in the Lansky or Karnofsky scale used for assessment of functional status was 80.8 (SD, 12.1) in the RTx group compared to 89.6 (SD, 8.5) in survivors who did not receive RTx (P = .01).

There was a strong positive correlation between functional status and adaptive behavior scores (Spearman rho = 0.796) (Fig. 3).

Figure 3.

Association between the scores in the Vineland-II adaptive behavior scales and the Lansky/Karnofsky performance scales.

We found a strong positive correlation between the functional status and adaptive behavior scores (rho = 0.796; P = .001).

Rho: Spearman correlation coefficient.

(0.06MB).

The distribution of the diagnoses and locations of CNS tumors in this sample of survivors was consistent with the previous literature.4–6

For the purpose of analysis, the divided the sample of survivors based on whether or not they had been irradiated, since radiotherapy, in addition to being a known predictor of poorer outcomes and LEs is strongly associated with the diagnosis of more aggressive and less frequent tumors, which we grouped together in order to obtain a group size comparable to the size of the group of patients who did not receive radiotherapy, who predominantly had low-grade tumors managed with surgery alone. This stratification allowed us to analyze differences in questionnaire scores to assess the potential impact of radiotherapy.

The most frequent LEs were neurologic and cognitive, followed by endocrine and neurosensory LEs. We found greater morbidity in patients managed with radiotherapy, as described in previous studies.34–37 However, we did not find significant differences in HRQoL and self-esteem scores when we compared the two groups.

Although HRQoL is important in the cancer survivor population, most questionnaires assessing this aspect do not cover aspects such as cognitive skills, family dynamics, social interactions, partnered/romantic relationships, educational attainment or employment.38 Similarly, the use of instruments for assessment of self-esteem has been questioned due to known limitations, especially in groups with potential cognitive impairment,32 and another factor that could be at play is the underdiagnosis of LEs in this group of patients.

The observed positive and moderately strong correlation between HRQoL scores and self-esteem scores in our study suggests that these indicators may be more variable and subjective compared to others, like adaptive behavior and functional status, in CNS tumor survivors.

On the other hand, we found significant differences in adaptive behavior and functional status scores when we compared CNS tumor survivors based on whether or not they had been managed with radiotherapy. In the bivariate analysis, we also found a strong correlation between the scores in these two scales.

Adaptive behavior is a relevant indicator in this subset of survivors. A study in patients treated with conformal radiation therapy found significantly lower adaptive functioning scores compared to controls, and this treatment was also correlated to poorer neurocognitive scores.29

In the multiple linear regression analysis, the adaptive behavior scores worsened with decreasing age at treatment, especially below age 6 years, with increasing radiation dose and with increasing duration of follow-up. Furthermore, when we compared the groups with “adequate to moderately high” versus “moderately low to low” adaptive behavior scores, we found a significantly longer median duration (in years) of follow-up in the latter group. A possible explanation is that survivors in the group with poorer scores may not be able to keep pace with their same-age peers so that the developmental gap between them would widen over time, chiefly due to progressive neurologic deterioration.15,16

With respect to social variables, the percentage of adult survivors who had enrolled in university or were employed was remarkably low. In the group of CNS tumor survivors who had received radiotherapy, there was a higher percentage who had dropped out after completing primary education and a higher percentage who had never had a romantic partner.

Promoting physical activity is important in CNS tumor survivors, who are at increased risk in several areas in association with LEs. In addition, there is evidence that exercise is associated with better neurocognitive outcomes and better social integration in CNS tumor survivors.39,40 Eighty-five percent of survivors in our study reported low levels of physical activity in the questionnaires, compared to the results found in the Spanish validation study of the international physical activity questionnaire.33

Among the limitations of our study, we ought to mention that, since our hospital only offers neurocognitive evaluation in the pediatric age group, the size of the group that had undergone this assessment was insufficient for analysis of this aspect. The design and sample of the study did not allow us to make comparative analyses based on the type or location of the tumor, and, since the duration of follow-up was longer than 10 years, all survivors treated with radiotherapy had received conventional irradiation, with larger doses and fields than those used at present, and new treatment modalities such as proton therapy were not represented.

We believe that a more holistic approach to the follow-up of CNS tumor survivors, involving them actively along with their families in the detection and prevention of LEs, in addition to a more thorough assessment of functional performance and psychosocial functioning, could be a good first step to improve their transition to adult care under different specialists.

The prevalence of LEs in pediatric CNS tumor survivors is high, and even higher in those managed with radiotherapy. The history of radiotherapy was not associated with the self-reported HRQoL or self-esteem, but it was associated with adaptive behavior, the level of functioning and certain social variables, such as educational attainment and partnered/romantic relationship status.

Additional studies in larger samples and with more thorough assessments are required to improve our knowledge of long-term outcomes and follow-up in pediatric CNS tumor survivors.

The study received a doctoral grant from the Fundación Oncohematología Infantil (FOI, Pediatric Oncology and Hematology Foundation) and the Fundación para la Investigación Biomédica (Biomedical Research Foundation) of the Hospital Universitario Niño Jesús (FIBHNJ).