After the Scan Goes Clear: What Nobody Tells Parents About Brain Tumor Survivorship
Key Findings
Across 274 pediatric brain tumor survivors (PBTS) in the Hong Kong territory-wide registry, ADHD showed up in 10.58% (vs. 4.86% in the general pediatric population, adjusted RR 1.66, p = .009). Autism showed up in 6.93% (vs. 2.27%, adjusted RR 3.95, p < .001).
Six kids (2.2%) had both diagnoses. The gap is real, but the comparator has issues worth knowing about. More on that in the Fine Print. [T1]
Every additional year older at tumor diagnosis cut ADHD/autism risk by about 10% (RR 0.90, p < .001). The association held up even after adjusting for tumor type, surgery, seizure history, endocrine dysfunction, and hydrocephalus (RR 0.89, p < .001).
Translation: the developing brain is more vulnerable. A tumor or its treatment at age 3 lands differently than the same tumor at age 13. If your kid was diagnosed young, long-term neurobehavioral surveillance isn't optional. [T1]
Survivors with a seizure history had intellectual disability at 16.28%, vs. 2.16% of those without (p < .01). That's an eight-fold gap, and it's one of the strongest associations in the paper.
Tumor location also mattered. Supratentorial tumors (the ones in the upper brain regions that govern attention and executive function) predicted higher hyperactivity scores on the SDQ. Emotional problems showed up hardest in the meningioma subgroup (42.86%) and the medulloblastoma subgroup (19.44%). [T1]
The researchers built a sleep score from two things: overall quality (parent/self-rated) and whether the kid was hitting NIH age-adjusted duration targets. Better sleep predicted:
Fewer ADHD symptoms (β = -0.30, FDR q = .006). Fewer emotional symptoms (β = -0.57, FDR q = .015). Also lower anxiety, depression, and stress scores, all surviving correction for multiple comparisons.
This is the only finding in the paper that parents can actually change tomorrow. And it survived the statistical stress-test that a lot of other findings in this paper didn't. [T1]
Why It Matters
Pediatric brain tumors used to be a death sentence. They're not anymore. Five-year relative survival for pediatric CNS tumors is around 97-98% for non-malignant tumors and 75-80% for malignant tumors, per CBTRUS. [T1] That's a real shift, especially on the non-malignant side. It's also a whole new problem.
More kids surviving means a bigger population of families learning what happens after the MRI comes back clear. Your kid beat the tumor. But the developing brain took a hit. Sometimes from the tumor itself, sometimes from surgery, sometimes from radiation or chemo. And the neurobehavioral echoes can show up years later as ADHD, autism features, emotional dysregulation, or intellectual disability.
If your family was already navigating ADHD or autism before a brain tumor entered the chat, this data hits twice. You now have a kid with compounding risk factors, and an oncology team that likely isn't screening for neurobehavioral issues unless you push for it.
Here's the part that matters: neurobehavioral challenges in survivors aren't destiny. The age-at-diagnosis finding is informational (you don't get to pick when a tumor shows up). But the sleep finding is a lever you can actually pull.
The Fine Print
The headline numbers (4.86% ADHD and 2.27% autism in Hong Kong kids) come from the authors' own pull of Hong Kong's Clinical Data Analysis and Reporting System (CDARS), an electronic medical records database. That's not an independent epidemiological estimate. It's a healthcare-system contact rate.
Why that matters: PBTS have way more healthcare contact than the average kid. Mandatory oncology follow-up, developmental surveillance, specialist visits. More chances for a provider to flag ADHD or autism and put a code in a chart. The comparison is inflated by design.
An independent 2025 Hong Kong community study (Wong et al.) estimates community autism prevalence at 2.57%, close to the authors' 2.27% figure. That partially de-risks the autism number. The ADHD number has no such independent cross-check.
The paper reads as if ADHD and autism developed after the brain tumor. Dig into the methods and the picture gets muddier. Of 29 kids with ADHD, 7 had pre-existing diagnoses. Of 19 with autism, 9 were pre-existing. So roughly a quarter of ADHD cases and half of autism cases were already on the books before the oncologist showed up.
Some of these kids were neurodivergent first. For intracranial germ cell tumors specifically, there's prior literature (Liu 2021) suggesting shared genetic susceptibility between autism and germinomas. A pre-existing autism diagnosis in a kid who later develops a germ cell tumor isn't a treatment sequela. It's the same underlying biology pointing at both.
The abstract implies chemo and radiation are associated with emotional problems and lower quality of life. In the simpler model (adjusting only for follow-up time and sex), yes. In the fully adjusted model (adding tumor type, surgery, seizures, endocrine dysfunction, hydrocephalus), those treatment effects disappeared.
The radiotherapy → quality of life finding (β = -0.20, p = .041) also didn't survive FDR correction for multiple comparisons (q = .123). That's a real finding demotion.
Disease severity is the likely confound. Kids with nastier tumors get more aggressive treatment AND have worse outcomes. Treatment and severity ride together. The paper is honest about this in the discussion. The abstract, less so.
Single territory (generalizability to Western populations isn't established). Survey response rate 107 of 162 contactable families (66%). Responders skewed younger and closer to diagnosis than non-responders, so older long-term survivors are underrepresented.
The sleep score is pragmatic (two items plus a PCA composite) but not a validated sleep instrument like the CSHQ. It does the job for a signal-detection study. It's not ready for clinical use.
Yes, with asterisks. The elevated neurobehavioral burden in pediatric brain tumor survivors replicates across multiple independent studies (Shabason 2019, Wang 2022 meta-analysis, Padovani 2012 for the radiation mechanism). The age-at-diagnosis finding replicates too.
The sleep finding is the least vulnerable to the critiques above. It survives FDR correction, it's biologically plausible, it aligns with prior work from van Kooten and van Litsenburg, and it's actually actionable. That's the takeaway worth keeping.
What To Do With This
Ask your care team for routine neurobehavioral screening. Not just MRI scans. ADHD, autism, emotional symptoms, intellectual disability screening should be standard follow-up, not something you have to fight for. If your oncology team doesn't do it, ask for a referral to pediatric neuropsychology.
Track sleep seriously. Total hours (compare to NIH age-adjusted recommendations) and quality. Write it down. Log bedtime, wake time, nighttime wakeups, and how they feel in the morning. If sleep is bad, screen for sleep-disordered breathing, which is common after hypothalamic or brainstem injury.
If your kid was diagnosed young (under 7), push harder for surveillance. The age-at-diagnosis finding is one of the most robust in the paper. Younger kids at diagnosis carry higher neurobehavioral risk, even when they look "fine" in the short term.
If there was a seizure history, ask for formal IDD screening. 16.28% of survivors with seizures in this cohort had intellectual disability. That's not a rounding error. WISC testing and adaptive functioning assessment belong in the follow-up plan.
The sleep finding generalizes beyond brain tumor survivors. Better sleep quality + hitting age-appropriate duration targets correlates with fewer ADHD symptoms and less emotional distress. That's consistent with the broader sleep literature in ND kids.
Practical moves: a consistent wind-down routine, screens off 30-60 minutes before bed, a dark cool room, and paying attention to the stuff ADHD meds can do to sleep onset (talk to your prescriber about timing).
Tools like Brainloot can help you log sleep alongside symptoms so you can spot the patterns your care team can act on. The paper showed a statistical correlation. Your logs show what it looks like in your kid.
Routine neurobehavioral monitoring belongs in standard survivorship care. DSM-based screening with ADOS or ADI-R for autism suspicion, SWAN for ADHD, WISC-IV plus adaptive functioning for IDD suspicion. Build it into the follow-up protocol.
Age at diagnosis is a stratification variable. Kids diagnosed under 7 get closer surveillance, full stop. Kids with seizure history get IDD screening. Kids with supratentorial tumors get attention/EF assessment early and often.
Sleep assessment belongs in every follow-up visit. Even a validated two-item screen is better than nothing. If sleep is poor, screen for sleep-disordered breathing and offer behavioral sleep interventions before reaching for medication.
Kids who survive brain tumors aren't just cancer patients anymore. They're survivors with developing brains that took a hit. Some of those kids will have ADHD or autism that looks indistinguishable from the kind that shows up in families who never saw an oncologist.
The honest read on this paper: neurobehavioral issues are more common in these kids than in the general population, younger kids are more vulnerable, and the one thing parents can actually move is sleep. The comparator stats are inflated and the treatment-effect story collapses under adjustment, but the core pattern replicates in other cohorts.
Ask for the screening. Track the sleep. Advocate harder if your kid was diagnosed young or had seizures. That's the whole playbook.