While certain tumors, including many low-grade gliomas, require only neurosurgical intervention, this treatment is not necessarily neuropsychologically “benign.” Mild intellectual and attention deficits are measurable after neurosurgical intervention but prior to adjuvant therapies [13]. Children with low-grade cerebellar astrocytomas treated only with neurosurgical resection displayed below normative-level performance on measures of general intelligence and academic and adaptive functioning at a mean of 103 days following surgery [14]. Similar results were reported for children who had surgical resections of midline or hemispheric low-grade gliomas; they displayed mild cognitive dysfunction relative to population norms when assessed an average of 111 days after surgery [15]. Cognitive dysfunction was greater in patients with left-hemisphere tumors relative to those with right-hemisphere tumors. Longitudinal studies are needed, however, to determine if there is natural recovery of function for PBTS treated with neurosurgical interventions alone or whether the deficits observed in the initial months following surgery persist.

Perioperative complications confer additional risks to neuropsychological functioning and tend to be more common in very young patients who are already more likely to experience long-term deficits. For example, cerebellar mutism syndrome (CMS) is a postsurgical complication that is seen in approximately 25 % of patients with posterior fossa tumors. The syndrome is characterized by diminished or absent speech, dysarthria, and linguistic difficulties. CMS is often accompanied by ataxia, hypotonia, and emotional lability. Children who develop CMS are more likely to be left handed and have unilateral, localized damage within the cerebello-thalamo-cerebral pathway at the level of the right cerebellum [16]. Disruption of this pathway is thought to interfere with communication between the posterior fossa and left frontal cortical regions that underlie language production [16]. CMS is also associated with increased long-term cognitive deficits compared to those without this complication. Children who develop CMS demonstrate significantly lower performance on measures of processing speed, attention, working memory, executive processes, cognitive efficiency, and academic performance in reading, spelling, and mathematics at one year post diagnosis relative to patients without CMS [17]. Other postoperative complications and infections, difficult-to-control seizures, and antiseizure medication also contribute to neuropsychological morbidity in PBTS.

Radiation therapy is generally considered the most significant risk factor for long-term neuropsychological deficits; higher dose and volume of radiation are associated with poorer outcomes, particularly in young children. Children with posterior fossa tumors have been shown to display a 2- to 4-point decline per year in intelligence which is evident at 1 year posttreatment with rapid decline initially following treatment and more gradual decline thereafter [18]. Although there is individual variability in intellectual decline, losses of 25–30 full-scale IQ points are not uncommon and the majority of children treated with cranial radiation prior to 7 years of age require special education. However, much of the research on long-term radiation effects was done with children who received higher doses of radiation than are currently used (5,400 or 3,600 vs. 2,400 cGy craniospinal), and current radiation doses and delivery techniques may be less neurotoxic. Intellectual outcome has been reported to be less impacted in those patients treated with 2,400 cGy dose to the brain and spine [19]; however, this effect has not been observed consistently, leading to questions concerning other patient factors that may influence sensitivity to radiation.

Stereotactic conformal radiotherapy administered directly to the tumor and a margin of surrounding brain protects normal brain tissue and appears to reduce overall cognitive morbidity. Children with low-grade glioma or craniopharyngioma who were treated with conformal radiation therapy showed no decline in learning performance when assessed at a mean of 4–5 years following treatment [20]. Jalali and colleagues showed that conformal radiation administered to pediatric brain tumor patients was associated with clinically significant intelligence declines in one third of the overall sample [13]. In their sample, clinically significant decline in intelligence was associated with younger age at time of treatment and higher radiation dose to the left temporal lobe. Thus, relative to conventional radiation therapy, conformal radiation appears to result in cognitive sparing. However, there continues to be a need for long-term data on specific radiation doses to various brain regions and their associated cognitive outcomes.

Understanding the contribution of chemotherapy agents to long-term neuropsychological outcomes in PBTS is complicated as single chemotherapy agents are rarely given in isolation. There is vast literature detailing the neuropsychological effects of methotrexate, much of it in pediatric acute lymphoblastic leukemia (ALL) patients. Methotrexate depletes folate within the central nervous system (CNS) and neuropsychological impairments are related to both individual and cumulative methotrexate doses. For medulloblastoma patients, intrathecal methotrexate, in addition to a standard radiation therapy and chemotherapy protocol, was associated with increased neuropsychological deficits, particularly in those under 10 years of age [21]. Radiation therapy and intrathecal methotrexate seem to have a synergistic effect when combined, and it has been hypothesized that radiation therapy increases the permeability of the blood–brain barrier to neurotoxic chemotherapy.

More recently, methotrexate has been used with brain tumor patients under 3 years of age in an effort to avoid or delay radiation, with positive neuropsychological outcomes. For example, Rutkowski and colleagues reported on intellectual outcomes for 14 children with medulloblastoma diagnosed prior to age 3 years who were treated with chemotherapy-only protocol including intravenous chemotherapy and intraventricular methotrexate [22]. Average IQ at a mean of 5 years post diagnosis was significantly lower than healthy age-matched controls, but higher than historical medulloblastoma controls who received radiation therapy.

The Head Start II protocol utilized postsurgical high-dose chemotherapy followed by autologous hematopoietic cell transplantation to avoid or delay craniospinal radiation, and radiotherapy was required for a minority of patients (33 %). Patients less than 10 years of age at the time of diagnosis demonstrated low average intelligence and visual-motor integration at baseline (following completion of induction therapy, but prior transplant or radiation) [23]. Follow-up testing at a mean of 2.8 years later showed stable cognitive performance with average social-emotional and behavioral functioning. Increased time from diagnosis was associated with lower intelligence, poorer reading, and delayed verbal memory compared to patients more recently diagnosed. Because the full neuropsychological effects of this treatment may not have fully manifested, continued monitoring of this cohort is warranted. Nevertheless, chemotherapy-based protocols that attempt to avoid or delay radiation, particularly in young children, appear to be beneficial to preserving cognitive development.

Glucocorticoid steroids (e.g., prednisone and dexamethasone) are directly associated with acute neurobehavioral difficulties in children and longer-term memory difficulties. Other chemotherapy agents indirectly impact cognitive development and academic performance. Platinum-based chemotherapeutic agents such as cisplatin are ototoxic at higher doses and may induce irreversible, bilateral hearing loss. Neuropsychological deficits, specifically in the language domains, are thought to be at least partially associated with hearing loss at an early age. Another example is vincristine-related neuropathy which may impact writing and other fine motor functions.

Radiation therapy disrupts the protracted process of myelination within the brain, which normally begins prenatally and continues through childhood and adolescence. The resulting white-matter damage is directly related to the nature and severity of neuropsychological impairment. White matter is composed of glial cells that provide structural and physiological support within the CNS and insulate axons (myelin). The functional integrity of white matter is essential for cognitive efficiency as it facilitates the transmission of electrical signals along axons. Thus, glial cells play an important role in neural transmission and are also involved in neurogenesis. These cells are sensitive to treatment-related damage because they are rapidly proliferating.

Observed changes in brain tissue following radiation therapy involve glial atrophy, demyelination, white-matter-specific necrosis, and decline in normal-appearing white-matter volume [24]. Neuroimaging measures of white-matter tissue damage, particularly in prefrontal and frontal regions, predict poor attention [25] and intellectual functioning [26] following cranial radiation. Cranial radiation also has a detrimental impact on the growth of new neurons in the hippocampus [27] with decreased hippocampal volume related to deficits in declarative memory [28].

Regardless of the treatment used, neuropsychological effects appear to be cumulative. The cognitive effects of radiation therapy are measureable as early as one year after treatment and there is continual decline until late adolescence, at which time anticipated age-related gains in cognition begin to plateau. While previously acquired skills and knowledge are retained, there is a slowing in the rate of new skill acquisition, resulting in a steadily increasing gap in performance between PBTS and age-matched peers. Based on a meta-analysis of 29 studies of intellectual and attentional functioning of PBTS across a range of tumor types, locations, and treatments, De Ruiter and colleagues reported that longer time since diagnosis was more highly predictive of overall IQ than either radiation or chemotherapy, accounting for 41 % of the total variance [29].