Radiation therapy increases long-term gonadotropin deficiency, hypothalamic-pituitary dysfunction, and brainstem necrosis in pediatric patients with brain tumor

Article information

Radiat Oncol J. 2026;44(2):126-136

Publication date (electronic) : 2026 June 18

doi : https://doi.org/10.3857/roj.2025.00885

Feng-Chin Lee^1,2,3, Chi-Hsiang Chung^4,5,6, Sheng-Feng Lin^1,7,8, Wei-Fong Kao^1,9, Sen-Kuang Hou^1,9, Hsi Chang^10,11, Peter C. Hou¹², Wu-Chien Chien^,4,5,13, Chun-Chieh Chao^,1,9,14,15

¹Department of Emergency Medicine, Taipei Medical University Hospital, Taipei, Taiwan

²Department of Emergency Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan

³Division of Pediatric Neurology, Department of Pediatrics, Taipei Medical University Hospital, Taipei Medical University, Taipei, Taiwan

⁴Department of Medical Research, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan

⁵School of Public Health, National Defense Medical Center, Taipei, Taiwan

⁶Taiwanese Injury Prevention and Safety Promotion Association, Taipei, Taiwan

⁷Department of Public Health, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan

⁸School of Public Health, College of Public Health, Taipei Medical University, Taipei, Taiwan

⁹Department of Emergency Medicine, School of Medicine, College of Medicine, Taipei Medical University Hospital, Taipei, Taiwan

¹⁰Department of Pediatrics, School of Medicine, Taipei Medical University, Taipei, Taiwan

¹¹Department of Pediatrics, Taipei Medical University Hospital, Taipei, Taiwan

¹²Division of Emergency Critical Care Medicine, Department of Emergency Medicine, Brigham and Women’s Hospital, Mass General Brigham, Harvard Medical School, MA, USA

¹³Graduate Institute of Life Sciences, National Defense Medical Center, Taipei, Taiwan

¹⁴Graduate Institute of Clinical Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan

¹⁵Graduate Institute of Injury Prevention and Control, College of Public Health, Taipei Medical University, Taipei, Taiwan

Correspondence: Wu-Chien Chien Department of Medical Research, Tri-Service General Hospital, National Defense Medical Center, Taipei 114, Taiwan Tel: +886-2-27372181 (ext. 8101) E-mail: chienwu@ndmctsgh.edu.tw

Chun-Chieh Chao Department of Emergency Medicine, Taipei Medical University Hospital, No. 252, Wuxing St, Xinyi District, Taipei 110, Taiwan Tel: + 886-2-87923311 (ext. 19189) E-mail: chaosees@gmail.com

Received 2025 December 15; Revised 2026 April 17; Accepted 2026 May 8.

Abstract

Introduction

Radiotherapy (RT) is widely used to effectively treat primary and metastatic brain tumors in adult and pediatric patients [1]. The clinical differences between brain tumors in adults and children are striking. Compared with brain tumors in adults, pediatric tumor types (mostly glial and neuronal) are more sensitive to RT [2]. RT is the most effective nonsurgical treatment for pediatric gliomas, resulting in 5-year progression-free survival estimates exceeding 80% [3].

To optimize the therapeutic efficacy, RT strategies have been modified dramatically in recent decades, developing various RTs, such as intensity modulated RT, volumetric modulated arc therapy, and proton beam RT. Modified RTs were reported to improve long-term survival of brain tumor patients [4]; however, RT-related late side effects, including neurocognitive, neuroendocrine dysfunction, cerebrovascular complications, and hearing deficits were often observed in pediatric patients with glioma [5-8].

Late-delayed treatment increases the likelihood of developing brain complications and permanent neurologic disability [9]. Consequently, alteration or even discontinuation of the antineoplastic treatment plan may be required to prevent further neurologic injury. As a result, early identification of pediatric patients at high risk for developing late brain injury is essential. In addition, the development of neuroprotective strategies that modify existing antineoplastic treatment regimens is clinically significant. However, no nationwide population-based cohort study has been performed to fully assess the extent to which RT affects overall late side effects in pediatric patients with brain tumor in a nationwide population, including endocrinopathies (e.g., hypothalamic pituitary dysfunction, gonadotropin deficiency, and hypothyroidism), late-delayed complications (e.g., post-RT brain stem necrosis, sensorineural hearing loss, and RT-induced non-senile cataracts), and late cerebrovascular complications (e.g., cavernous/arteriovenous malformation, transient cerebral ischemia, and stroke). Therefore, the present study aimed to fully evaluate associations between RT and late side effects in pediatric patients with brain tumors in the nationwide population of Taiwan.

Materials and Methods

2. Study design and sample

Fig. 1 depicts the study population selection for this population-based, retrospective cohort study. A total of 34,092 patients with brain cancer were initially selected from the NHIRD (2000–2015). Patients aged > 18 years; those with a brain tumor diagnosed before the index date; those with late effects before tracking; patients without tracking and those with unknown sex were excluded. After exclusions, a cohort of pediatric patients with brain tumor (n = 3,292) was established, identifying patients by using ICD-9-CM 191 (malignant neoplasm of brain), 192.0 (malignant cranial nerve tumor), 192.1 (malignant cerebral meninges tumor), 194 (malignant neoplasm of other endocrine glands and related structures), or 239.6 (neoplasm of unspecified nature of brain). Among 3,292 pediatric patients with brain tumor, 815 cases received RT (ICD-9-CM V58.0) but 2,477 patients did not undergo RT. Then, two-fold matching by age, sex, and the year of the index date was performed to select 1,630 matched cases without RT, selecting 815 cases with RT and 1,630 matched cases without RT who represented the final analytic sample. For all hospitals and medical centers in Taiwan, the brain tumor was diagnosed according to guidelines of the Taiwan Society of Neuro-Oncology, while the RT exposure procedure, exposure dose and exposure location were based on guidelines of the Taiwan Society of Therapeutic Radiology and Oncology.

Fig. 1.

Flowchart of study population selection. RT, radiotherapy.

Results

1. Patients’ characteristics

A total of 2,445 pediatric patients with brain tumors were included in this study, consisting of 815 patients with RT and 1,630 patients without RT (Table 1, Fig. 1). No significant differences in age, sex, salary-based insured premium, and location of residence were found between the two groups at baseline and endpoint (all p > 0.05); however, CCI_R, prior chemotherapy, prior brain surgery, urbanization level at baseline and the hospital level of medical care were significantly different between the two groups (all p < 0.05) (Table 1).

Table 1.

Characteristics of study population at endpoint

2. Associations between patients’ characteristics and incidence of RT-related late side effects

The incidence of late side effects increased significantly in the RT group compared with that in the non-RT group (21.96% vs. 16.32%, p < 0.05) (Table 1). After adjusting for possible confounding variables, multivariable Cox regression analysis revealed that the risk of late side effects in the RT group was approximately 1.173-fold (adjusted HR [aHR], 1.173; 95% CI, 1.063 to 1.430; p < 0.05) (Table 2) higher than that in the non-RT group. In addition, high CCI_R (aHR, 1.032; 95% CI, 1.004 to 1.060), prior chemotherapy (aHR, 1.679; 95% CI, 1.299 to 2.170), prior brain surgery (aHR, 1.856; 95% CI, 1.466 to 2.350), and high level of care were also associated with an elevated risk of late side effects, while older age (≥1 year) was associated with a decreased risk of late side effects (all p < 0.05) (Table 2). After stratifying by the variables listed in Table 2, multivariable Cox regression analysis adjusted for these variables disclosed that each subgroup among patients with RT had a higher risk of late side effects than did subgroups among patients without RT (all p < 0.05) (Supplementary Table S1), except for the older age subgroup, suggesting associations between age and RT-related late side effects.

Table 2.

Cox-regression analysis of factors associated with RT-related late side effects

Kaplan-Meier plot and log-rank tests showed that the RT group also had a high risk of late side effects compared with that of the non-RT group across ≥7 years of tracking (all p < 0.05) (Fig. 2A). Furthermore, high insurance premium, high urbanization level, and high hospital level of care were also associated with incident RT-related late side effects (all p < 0.05) (Table 2).

Fig. 2.

Kaplan-Meier plot for cumulative incidence of late effects in patients with (solid line) and without radiotherapy (RT) (dotted line). (A) Overall late effects. (B) Long-term endocrinopathies. (C) Late-delayed complications. (D) Late cerebrovascular complications. ^*p < 0.05.

3. Associations between RT and specific late side effects

Compared to the non-RT group, the RT group had higher proportion of long-term endocrinopathies (13.3% vs. 9.8%, p < 0.05) and late-delayed complications (8.1% vs. 4.5%, p < 0.05) and a slightly lower proportion of late cerebrovascular complications (2.2% vs. 2.6%, p > 0.05) (Supplementary Table S2) than the non-RT group. For long-term endocrinopathies, the proportion of hypothalamic pituitary dysfunction (10.1% vs. 7.5%, p < 0.05) and gonadotropin deficiency (1.7% vs. 0.8%, p < 0.05) was also higher in the RT group than in the non-RT group. For late-delayed complications, the proportion of stroke (and other and ill-defined cerebrovascular disease) (1.1% vs. 0.4%, p < 0.05) and post-RT necrosis of the brain stem (4.8% vs. 2.0%, p < 0.05) was also higher in the RT group than in the non-RT group. For late cerebrovascular complications, the proportion of cavernous/arteriovenous malformation (0.3% vs. 1.0%, p < 0.05) was lower in the RT group than that in the non-RT group.

Among the results of multivariate Cox regression analysis after adjusting for the variables listed in Table 2, the RT group exhibited 1.172-fold (95% CI, 1.020 to 1.398) and 1.796-fold (95% CI, 1.250 to 2.578) higher risk of long-term endocrinopathies and late-delayed complications, respectively, than did the non-RT group (all p < 0.05) (Table 3). However, no significant differences were found in late cerebrovascular complications between the two groups. For long-term endocrinopathies, the RT group had 1.602-fold (1.010–3.718) and 1.120-fold (1.005–1.548) higher risk for gonadotropin deficiency and hypothalamic pituitary dysfunction, respectively, than did the non-RT group (all p < 0.05). For late-delayed complications, the RT group had 3.208 (1.953–5.297)-fold higher risk for post-RT necrosis of the brain stem than did the non-RT group (p < 0.05).

Table 3.

Multivariable Cox-regression analysis of RT-related late side effects subgroup

Kaplan-Meier plot and log-rank tests revealed that the RT group also had a higher risk of long-term endocrinopathies and late-delayed complications, respectively, than did the non-RT group across ≥8 years of tracking (all p < 0.05) (Fig. 2B, 2C, Supplementary Tables S3, S4). No significant differences were found in the risk of late cerebrovascular complications between the two groups across all years of tracking (all p > 0.05) (Fig. 2D, Supplementary Table S5).

Discussion and Conclusion

Results of the present study demonstrated associations between RT and the profile of overall late side effects in pediatric patients with brain tumors in the national population of Taiwan. The key finding was that the RT group had higher risk of late side effects compared to the non-RT group, except for those with older age and prior surgery or chemotherapy. In these late side effects, the risk of long-term endocrinopathies and late-delayed complications were higher in the RT group than in the non-RT group, however, no significant differences were found in the risk of late cerebrovascular complications between the two groups. Associations were also noted between age and the presence of RT-related late side effects. For long-term endocrinopathies, the RT group had higher risk of gonadotropin deficiency and hypothalamic pituitary dysfunction than did the non-RT group. For late-delayed complications, the RT group had higher risk for post-RT necrosis of brain stem than did the non-RT group.

Treatment strategies for pediatric brain tumor often include a combination of surgery, RT, and chemotherapy. Nevertheless, many pediatric brain tumor survivors are at significant risk for late side effects, which can be attributed to direct neurologic damage to the developing brain caused by surgery on the central nervous system [12]. Chemotherapy or RT may lead to late side effects, including possible neurocognitive and endocrine effects [13,14]. Similarly, the results of the present study exhibited that not only RT (aHR, 1.173) but also prior surgery (aHR, 1.856) or chemotherapy (aHR, 1.679) were independent predictors for increased risk of late side effects, suggesting that late side effects are not solely attributable to RT, and that prior chemotherapy and surgical interventions also may constitute significant risk factors, potentially exceeding the contribution of RT. In addition, the contemporary precision RT techniques (e.g., intensity-modulated RT, volumetric modulated arc therapy, or proton therapy) have recently become more widely adopted in pediatric brain tumor [15], enabling substantial sparing of normal organs, hence the actual incidence of late side effects in clinical practice attributable to RT may be considerably lower than that observed in the present cohort.

Higher radiation doses (36 Gy) correlate with greater late cognitive decline [14]. Furthermore, previous studies have reported that younger children are more sensitive to RT, which may complicate their risk for late side effects such as neurocognitive decline [7,16]. The impact of age also exceeded that of the radiation dose; patients younger than 5 years experienced the greatest decline in cognition [7]. Shalitin et al. [8] also reported that pediatric patients with brain tumors undergoing RT were at increased risk of late endocrine side effects, particularly those treated with cranial radiation and diagnosed at a younger age. Avadiappan et al. [17] also observed that younger patients receiving RT were at a greater risk of developing cerebrovascular complications. Compared to those studies, findings of the present study have added further evidence that, among pediatric patients, patients with young age (<1 year) and undergoing RT were at an increased risk of overall late side effects when compared to patients with older age (≥1), supporting the association between pediatric age and RT-related overall late side effects, with a cut-off value of age 1-year. On the basis of these findings, we recommend that pediatric patients with brain tumors and undergoing RT require regular monitoring and frequent neurologic evaluations to ensure normal development. Moreover, additional educational support is essential for patients’ recovery.

Cerebral radiation necrosis is a well-characterized toxicity associated with RT, as first described in adult patients [18]. However, little data is available regarding the incidence and risk of radiation necrosis in pediatric patients. Plimpton et al. [19] observed that pediatric patients undergoing RT for brain tumors may have an increased likelihood of developing cerebral radiation necrosis compared to adult counterparts. Furthermore, recent studies have revealed that brainstem injury is a rare complication of photon therapy [20,21]. Results of the present study demonstrate further that pediatric patients with brain tumors and undergoing RT had a higher (3.208-fold) risk of post-RT necrosis of the brain stem than did pediatric patients not receiving RT, which may worsen cerebral cellular injury and vascular damage leading to tissue hypoxia. Radiation necrosis has been reported to begin as acute cellular injury, in which endothelial cell death causes platelet aggregation and thrombus formation, ultimately leading to occlusion of microvessels [18,22]. Therefore, we hypothesized that the aforementioned mechanism may possibly be responsible for RT-related brainstem necrosis in pediatric patients with brain tumors. Those cellular and physiological alterations may further lead to cerebral atrophy, causing mass lesions, resulting in confusion, seizures, and other significant neurologic effects [23]. Helton et al. [24] indicated that late-delayed cerebral radiation necrosis occurs 6-12 months to years after RT. Therefore, our recommendation is to routinely perform MRI monitoring for post-RT brain- stem injury in pediatric patients within this time period. Various approaches such as corticosteroids [25], bevacizumab [26], hyperbaric oxygen therapy [27], anticoagulation [28], and oral vitamin E [29] may be recommended for treatment of cerebral radiation necrosis.

Late endocrine side effects also occur frequently and are often severe in pediatric patients with brain tumors [8]. The most common late endocrine side effects is GH deficiency, which may occur as soon as the first year after the initiation of RT [5]. It occurs in nearly 100% of patients who receive a radiation dose of 36 Gy or more to the pituitary region, and is less common with radiation doses of 18–24 Gy, but may not become evident until 10 years after treatment [12]. Laughton et al. [30] reported that around 1.8 years after RT, 84% of pediatric patients with embryonal brain tumors were diagnosed with GH deficiency, and 10% of patients developed thyroid-stimulating hormone (TSH) deficiency. Furthermore, the 4-year cumulative incidence rates for GH and TSH deficiency were 93% and 23%, respectively [30]. Bahl et al. [31] showed that the 5-year cumulative incidence rate was 68% for GH deficiency and 52% for hypothyroidism in pediatric patients with medulloblastomas (MB) who underwent RT. The present study further demonstrated that, in up to 16 years of tracking, the majority of pediatric patients with brain tumors and RT had 1.120-fold higher risk for hypothalamic pituitary dysfunction than did their counterparts without RT. In response to this finding, we suggest that the goal here is to prevent radiation-induced hyposuppression of the hypothalamic-pituitary-adrenal (HPA) axis, but not to reduce the radiation dose, since radiation-induced toxicity in the HPA axis was not reduced by lower RT doses in low-risk MB patients [32].

Vatner et al. [33] reported that the 4-year actuarial rate for gonadotropin deficiencies was 4.1% in pediatric and young adult patients with brain tumor after RT. In a large cohort of 748 pediatric patients with brain tumor and RT, the estimated 40-year cumulative incidence of gonadotropin deficiency was 24.4% [34]. In particular, the present study further pointed that pediatric patients with RT had a 1.602-fold higher risk for gonadotropin deficiency than those without RT, implying that the risk of gonadotropin deficiencies is increasing steadily over time even long after the completion of RT. Thus, despite improved long-term survival of pediatric patients with brain tumor and RT, the RT-related gonadotropin deficiency may substantially affect their quality of life and medical expenditures later [35].

Both short and long-term cerebrovascular complications may arise following RT. Cerebral microbleeds can be detected as early as eight months after RT, while cavernous malformations and atherosclerosis can present decades later [6,36]. However, in the present study, after 16-year tracking, no significant differences were found in late cerebrovascular complications, including cavernous/arteriovenous malformation, stroke, cerebral atrophy, and moyamoya disease, between RT and non-RT groups. Cerebral microbleeds are RT-induced small vessel vasculopathies that are associated with poorer executive function, poor working memory, cognitive decline, and stroke in pediatric patients with brain tumors [37,38]. Hence, we still recommend that clinicians must monitor these potential late cerebrovascular complications continuously after using RT to treat pediatric patients with brain tumor.

Beyond the survivorship challenges commonly encountered within the broader pediatric oncology population, pediatric brain tumor survivors face a pronounced risk of enduring endocrinopathies, particularly GH deficiency. Fischer et al. [12] demonstrated that GH deficiency was present in nearly all patients who received radiation doses of ≥ 36 Gy targeting the pituitary region. Nevertheless, GH deficiency can be ameliorated through growth hormone replacement therapy (GHRT) [39]. However, given the established role of GH in tumorigenesis [40], concerns have been raised regarding the potential causal relationship between GHRT and tumor recurrence or the emergence of secondary malignancies. Evidence to date reassures, however, that GHRT does not elevate the risk of tumor recurrence or secondary malignancies in pediatric brain tumor survivors [41]. Although such patients exhibit an increased incidence of meningiomas, the increase is likely attributable to prior cranial irradiation rather than to GHRT administration. Further investigation is needed to determine the risk of tumor occurrence or secondary malignancies in pediatric patients with brain tumor after RT.

The present study was strengthened by the analysis of data from the longitudinal NHIRD, which represents a nationwide population in Taiwan. Nevertheless, this study has several limitations, including, first, the data were extracted from an administrative database, so the possibility of selection bias cannot be ruled out and the observational study design does not allow inferences of causation. Also, due to the insurance claim limit, patients’ follow-up time is necessarily restricted and since no appropriate ICD-9-CM codes for the classification of neurocognitive decline are available in the administrative database, we were not able to assess it in the present study, and because most patients with pediatric brain tumor receive combined therapy (RT, chemotherapy, and surgery) rather than RT alone, the late side effects are likely to be caused by combined therapy, even though multivariate Cox regression analysis was adjusted by these variables. The NHIRD also does not provide detailed information on possible brain tumor- or RT-related risk factors such as smoking, diet, genetic variants, structural birth defects, types of tumors, location of tumors, brain surgery, radiation dose, treatment volume, fraction size, and image guidance, which could be potential confounding factors and these factors could not be evaluated, necessitating prospective multicenter studies to confirm results of the present study.

RT, in addition to prior surgery or chemotherapy, is an independent predictor for increased risk of late side effects in pediatric patients with brain tumors, particularly in patients who are younger at tumor diagnosis. RT-related late side effects, including gonadotropin deficiency, hypothalamic pituitary dysfunction, and brainstem necrosis increase over time. Lifelong medical surveillance and close follow-up for potential RT-related late side effects are essential in young pediatric patients, thereby reducing morbidity and mortality.

Supplementary Materials

Article information Continued