FIGURE 1 Flow diagram showing the selection procedure
Results
3.1 Study Characteristics A total of 2 232 unique articles were identified from the search of which 27 were included upon full text screening (Figure 1). One additional paper was included based on manual reference tracking. Cohen’s K=0.94 for interrater reliability.
The characteristics of the 28 included articles are presented in Table 1. Nineteen unique study populations were identified among all included articles. Studies were conducted in North America (USA: n=20, Canada: n=4), Europe (n=3), or Australia (n=1), and published between 2001 and 2022. Only 12 studies exclusively included patients with a PBT. Participants’ mean age at diagnosis and at study inclusion ranged from 3.5 to 10.0 years and 10.3 to 19.8 years, respectively. Most studies recruited patients who completed cancer treatment, while two studies included patients during their cancer therapy, and one study included both13–15. Furthermore, fifteen articles implemented the inclusion criterion of NCI in at least one domain16–30 .
The effectiveness of the interventions was measured in nearly all studies using objective neurocognitive assessments, or proxy-reported measurements of neurocognition (Table 2). Almost half (n=13) of the articles measured intervention efficacy immediately post-intervention (less than 3 weeks after intervention or indicated as post-intervention)19–25,30–35. The other half (n=15) also considered and measured maintenance of possible intervention effects (range 12 weeks – 2.9 years after intervention completion)13–18,26–29,36–40. For results of risk of bias, we refer to supplementary information (Supplementary Figure S1).

3.2 Types of interventions for NCI

Three categories of interventions were distinguished: 1) lifestyle interventions, 2) cognitive training, and 3) pharmacological interventions. An overview of different interventions and their results are given in table 1 and 2.
Lifestyle interventions
Four articles (3 study populations) implemented a 12-week aerobic exercise training program 13,33–35. Immediate improvement in reaction times was seen when participants trained in a group setting, and these improvements could be predicted by improved fitness33,35. Furthermore, exercise training showed a positive impact on white matter microstructure (as indicated by higher fractional anisotropy based on diffusion-weighted MRI) as well as hippocampal volume, and on MEG-derived functional connectivity. The positive effect on white matter microstructure maintained at 12 weeks after completion of the training. Additionally, the training program led to an increase in cortical thickness which was also associated to a decrease in reaction time35. One study reported that no adverse events occurred, and that drop-out rate amounted 14.3%35.
Another exercise training intervention, in which training consisted of individually tailored strength-based exercises and aerobic activities (modified every 3-4 weeks), showed no significant improvement in mental health nor in objective neurocognitive functioning. By contrast, self-reported measures of community use, home living, health and safety, leisure, and self-direction did improve. Thirty-five percent of the participants dropped out of the study and one fall without physical injury was reported40.
A last exercise intervention implemented physical activity (strength, endurance, and speed exercises) with simultaneous attentional challenges (verbal stimuli). The exercise intervention was administered as a period of physical training followed by a period of combined training or reversed. Improvements were seen in all groups for verbal long-term memory (baseline compared to post-intervention), as well as for organization, and anxiety and emotional control (post-observation compared to post-intervention). Additionally, they found no difference between patients on or off cancer treatment They did not report on adverse events and drop-out rate was 16.7%13.
Finally, eurythmy therapy is a form of mind-body therapy. One study demonstrated that all patients (n=7) completing 25hours of eurythmy therapy in a period of six-months, improved in full-scale IQ and processing speed. Working memory (WM), perceptual reasoning, verbal comprehension, and visual motor integration improved in six, five, three, and five patients, respectively. This study did not report on adverse events36.
Cognitive training
Regarding cognitive training, six articles (5 study populations) employed the adaptive computerized working memory (WM) training, Cogmed26–28,31,37,38. Participants receiving Cogmed training showed improvement on attention, processing speed, WM, symbolic WM and executive functions immediately post-intervention27,28,31. WM scores and fluent cognition were improved at 9.1 weeks post-intervention, and improvement on fluent cognition remained stable at 14.5 weeks 38. Improvement on processing speed, executive functions WM, visual-spatial WM, academic achievement, emotional and behavioral problems, and social skills and remained stable at six months post-intervention27,28,38. In contrast, scores on two attention tasks changed over time: spatial span forward (WISC-V) deteriorated and omissions on the CPT improved27,28. Finally, most stable improvement over time was observed for fluent cognition and executive functioning at 13.5 months post-intervention38.
No difference was observed between adaptive and non-adaptive Cogmed training38. Compared to a non-adaptive WM training (MegaMemo), adaptive Cogmed training did show greater improvement in parent-reported learning difficulties26. Adaptive Cogmed did not show a difference in WM improvement compared to JumpMath (a workbook program directly targeting math skills), however the improvement in mathematics (greater improvement for JumpMath group) and symbolic WM (greater improvement for Cogmed group) was different among these interventions 31. Higher pre-intervention IQ and greater number of completed sessions predicted greater post-intervention improvement26,28. Nevertheless, there was no significant difference between 25 and 35 Cogmed sessions37 .
Functional MRI (fMRI) found that Cogmed training reduced activation of lateral prefrontal, left cingulate and bilateral medial frontal regions during a WM task, however, this was not associated with WM outcome27. Adverse events concerning Cogmed were either not reported, or the study stated that no adverse events had occurred26,28,37. Drop-out rates varied between 4.8% to 25%.
Another computerized intervention, Fast ForWord, focuses on training reading skills. This training effect was investigated in medulloblastoma patients during radiotherapy. Compared to the standard of care group, no significant differences were observed immediately post-intervention. At subsequent follow-up timepoints (up to five years post-intervention), specifically greater sound awareness was encountered at 2.9 years (range 1.6 – 4.5 years) post-intervention14,15. Only 16.3% of participants completed the prescribed sessions and 7.0% dropped-out of the study14. Patients were excluded from the study if they experienced an adverse event, however, it was not specified whether patients discontinued for this reason.
A last game-like cognitive training intervention that was investigated in pediatric brain tumors, was the Captain’s Log (n=9), designed to improve multiple neurocognitive domains including memory, attention, concentration, listening skills, self-control, and processing speed. After on average 28.4 sessions (range 9-53), improvement was seen for digit-span forward (WISC-III) and parent-reported attentional problems. Results of the digit-span forward (WISC-III) task were positively correlated with pre-intervention IQ. No adverse events were reported29.
In addition to interventions focusing on improving neurocognitive functioning, other interventions were identified with the aim of teaching compensatory strategies. Two studies applied a cognitive remediation program (CRP)17,18. CRP is a therapist-delivered framework designed to reinforce multidimensional aspects of attention processes, encompassing hierarchically graded massed practice, strategy acquisition, and cognitive-behavioral interventions17,18. Immediately post-intervention, improvement on academic achievement, parent- and teacher-reported attention and hyperactivity symptoms and attention, WM, memory recall, and vigilance was observed in the CRP group17,18. fMRI showed increased regional activity in the CRP group both immediately and six months after intervention, and these patterns were more closely resembling those of typical developing children18. There were no adverse events reported17.
The cognitive rehabilitation curriculum (CRC) is an online training focusing on cognitive flexibility, attention and WM. Participants showed increased processing speed, cognitive flexibility, and verbal- and visual memory immediate post-intervention. Additionally, fMRI demonstrated increased activation in the inferior, superior, and middle frontal gyrus compared to pre-intervention activation patterns. They did not report on possible adverse events30.
Two therapist-delivered training programs were studied as well. The first one was cognitive and problems-solving training, which targeted broader metacognitive functions through educational therapy, cognitive behavioral and rehabilitation approaches. Post-treatment improvements were specifically significant for reported social skills scores and quality of written expression. Adverse effects were not reported, and drop-out rate amounted 25%25. The second one was a survivor’s Journey, targeting executive dysfunctions through behavioral problem-solving therapy and metacognitive strategies. A survivor’s Journey showed to improve self-reported emotional QoL, parent reported total QoL and physical QoL immediately after the intervention. Furthermore, their exploratory analysis showed that intervention effects varied based on age at diagnosis and pre-intervention IQ. No information was provided regarding adverse events. Only 10.5% dropped-out of the study32.
Finally, the last cognitive training is neurofeedback training. This training targets specific beta frequencies associated with arousal and attention. Bot the intervention- and placebo- group showed post-intervention improvement, without significant effect of the neurofeedback training. No serious adverse events were reported and drop-out rate amounted 11.3%16.
Pharmacological interventions
Two pharmacological compounds, methylphenidate (MPH) and donepezil, have been studied in patients with a PBT. MPH was studied in six articles. Immediate neurocognitive response of a single MPH administration showed a significant effect on selective attention, impulsivity, and cognitive flexibility21. Immediate improvement in sustained attention was assessed in two studies, but only significant in one study19,21. A 3-week crossover trial (placebo, low dose MPH, and high dose MPH), showed improvement of parent- and teacher reported scores on attention and cognitive problems, improved attention deficit/hyperactivity disorder index for low and high dose MPH20,22. Additionally, teachers reported improved social and academic competences, less hyperactivity symptoms, and less problem behavior. Though, less problem behavior was only seen in the high-dose MPH group20,22. Finally, a 12-month study with individually titrated MPH dose showed improvement in reported measures of attention, hyperactivity, and attention deficit/hyperactivity disorder in both the MPH- and the control group. However, improvement was significantly larger for measures of attention and attention deficit/hyperactivity disorder in the MPH group compared to control23,24. Sustained attention and processing speed were also significantly improved, and improvement in attention was already present after 1 month with a rebound after 3 months23,24 Modest agreement was observed between parent- and teacher- ratings, and poor agreement was observed between reported ratings and performance-based measures23,24. Studies reported various adverse events in the MPH group as well as in the control group. Reported adverse events included abdominal pain, appetite loss, wheezing, dizziness, anxious, tearful, and many more19–21. The range of reported drop-out rates was between 10.5% and 33.8%19,20,23,24.
Donepezil, an acetylcholinesterase inhibitor, was described in one study and showed significant improvement on executive functioning immediately post-intervention, and these improvements persisted after a wash-out period of 12-weeks. Additionally, moderate improvements in parent-reported planning/organizing, WM, and memory were seen. The drop-out rate was 27.3%. Several adverse events were reported, 30% of the participants reported gastro-intestinal problems and 17% reported non-specific mood changes and confusion. Nevertheless, 75.0% of patients preferred re-administration after treatment completion because of perceived benefit39.