Received: 5 November 2016 Revised: 1 April 2017 Accepted: 3 April 2017

DOI: 10.1002/pbc.26619

Pediatric
RESEARCH ARTICLE Blood &
The American Society of
Cancer Pediatric Hematology/Oncology

Impact of induction chemotherapy, hyperfractionated

accelerated radiotherapy and high-dose thiotepa on brain
volume loss and functional status of children with primitive
neuroectodermal tumour

Elwira Szychot1,2 Kiran Seunarine3 Kshitij Mankad4 Steffi Thust5

Chris Clark3 Mark N. Gaze6 Antony Michalski2

1 Department of Clinical Studies, The Institute of

Cancer Research, London, United Kingdom Abstract

2 Department of Paediatric Oncology, Great Background: The introduction of aggressive chemo-radiotherapy regimens has improved over-
Ormond Street Hospital for Children NHS all survival in children with primitive neuroectodermal tumours (PNET). However, these combi-
Foundation Trust, London, United Kingdom nations may result in neurotoxicity. Previously reported magnetic resonance imaging abnormali-
3 Developmental Imaging and Biophysics Section,
ties in children receiving intensive sequential chemotherapy, hyperfractionated accelerated radio-
Institute of Child Health, University College
London, London, United Kingdom
therapy (HART) and high-dose thiotepa prompted us to investigate the degree of brain volume loss
4 Department of Radiology, Great Ormond and patients’ functional status after therapy.
Street Hospital for Children NHS Foundation
Methods: We retrospectively reviewed clinico-radiological data of children with PNET treated in
Trust, London, United Kingdom
5 Department of Diagnostic Neuroradiol- this way at our centre.
ogy, National Hospital for Neurology and
Results: We studied 14 children treated between December 2009 and April 2013. Data were not
Neurosurgery, University College London
Hospitals NHS Foundation Trust, London, United complete for one child. Performance status was severely restricted in four children, and mildly to
Kingdom moderately impaired in 7 of the 13 children. Eleven of 13 children showed mild-to-severe gener-
6 Department of Oncology, University Col-
alised neuroparenchymal atrophy, in 7 of whom neuroparenchymal volume loss was moderate to
lege London Hospitals NHS Foundation Trust, severe. Of these seven, six had received high-dose thiotepa. There was no correlation between
London, United Kingdom
brain volume loss and Lansky performance status. However, unexpected neurotoxicities, such as
Correspondence
Elwira Szychot, Department of Clinical Studies,
symptoms of transverse myelitis, were observed.
The Institute of Cancer Research, 15 Cotswold
Conclusion: Measurement of brain volume loss in patients treated with HART and high-dose
Road, Sutton SM2 5NG, London, United
Kingdom. thiotepa may not be sufficient to predict function. However, correlation of brain volume loss due to
Email: Elwira.Szychot@icr.ac.uk late neurotoxicity with performance decline may be more obvious over longer period of follow-up.
Grant sponsor: NIHR University College London The combination of HART and myeloablative courses of thiotepa is associated with severe neuro-
Hospitals Biomedical Research Centre.
toxicity and subsequent decline in performance status in a significant proportion of patients.

KEYWORDS
brain volume, chemotherapy, Lansky, medulloblastoma, radiotherapy

1 INTRODUCTION

Primitive embryonal tumours are the commonest malignant tumours

Abbreviations: CCLG, Children’s Cancer and Leukaemia Group; CNS, central nervous system;
CSF, cerebrospinal fluid; CT, computerised tomography scan; GOSH, Great Ormond Street
Hospital for Children NHS Foundation Trust; HART, hyperfractionated accelerated
radiotherapy; MRI, magnetic resonance imaging; PNET, primitive neuroectodermal tumours;
ROI, region of interest; SIOP, International Society of Paediatric Oncology; sPNET,
supratentorial primitive neuroectodermal tumours; UCLH, University College London
Hospitals NHS Foundation Trust; WHO, World Health Organization
of the central nervous system (CNS) in childhood with an annual inci-
dence of 3.56/100,000 person-years.1 The 2016 World Health Organi-
zation (WHO) updated classification of the embryonal tumours of the
CNS incorporates molecular parameters in addition to histology.2 In
the previous WHO classification, primitive neuroectodermal tumours
(PNET) were classified according to their microscopic similarities and

Pediatr Blood Cancer. 2017;e26619. wileyonlinelibrary.com/journal/pbc 

c 2017 Wiley Periodicals, Inc. 1 of 8
https://doi.org/10.1002/pbc.26619
2 of 8
encompassed medulloblastoma (posterior fossa PNET), pineoblastoma
and supratentorial PNET (sPNET).3 Prior to the 2016 WHO classifica-
tion of CNS tumours, treatment protocols were based on the entities
described in the 2007 classification.
Although the overall outcome for children with CNS tumours has
improved significantly over the last 20 years, the survival of patients
with metastatic disease or sPNET remains poor.4 Aggressive chemo-
radiation treatment regimens have been introduced, attempting to
improve the prognosis of high-risk PNET.5–10 These new treatment
protocols improved the overall 5-year survival of children with high-
risk disease from 50% to 70% over the last 10 years.4,11,12
As survival rates have increased, it has become clear that improve-
ment has come at a price.4,13 Physical deficits, neuro-psychological
impairment and decreased quality of survival have been reported.13–15
There is evidence that high-dose chemotherapy, combined with
radiotherapy for childhood brain tumours, may result in a compound
injury to the normal parenchyma of the developing brain.16 Parenchy-
mal volume loss, generalised white matter signal changes and focal
enhancing white matter lesions have recently been reported in chil-
dren treated with a protocol which utilised high-dose sequential
chemotherapy, hyperfractionated accelerated radiotherapy (HART)
and myeloablative therapy with autologous stem cell rescue.5,17,18 In
the HART schedule, radiotherapy is administered in a larger number
of smaller fractions than in standard practice (typically 1.3 Gy rather
than 1.8 Gy; hyperfractionation) within a shortened time period (e.g.,
4.5 vs. 6 weeks; acceleration) by giving more than one fraction per day
(typically two, with an inter-fraction interval of at least 6 hr for nor-
mal tissue recovery).5 The biological effect of the radiation dose to the
fast-growing malignancy is greater both as a result of the higher dose
and the reduced chance of repopulation in the shortened overall time,
while normal tissues are relatively spared.
There is increasing concern that combination therapies may result
in neurotoxicity, but the clinical impact of these findings remains to
be determined.18 Detection of multiple anomalies on magnetic res-
onance imaging (MRI) scans in children treated for high-risk PNET
at Great Ormond Street Hospital for Children NHS Foundation Trust
(GOSH) and University College London Hospitals NHS Foundation
Trust (UCLH), London, UK, along with the reports from other institu-
tions, prompted us to investigate whether the degree of brain volume
loss correlates with the functional status of patients after the com-
pletion of therapy.17,19 We reviewed the MRI changes and functional
outcomes of consecutive patients treated for high-risk PNET with com-
bination chemotherapy and radiotherapy at our hospitals over a 5-year
period.
2 METHODS
2.1 Patients
This study includes all children <16 years of age with high-risk medul-
loblastoma and sPNET, including metastatic pineoblastoma, who were
treated at GOSH and UCLH between December 2009 and April 2013
according to U.K. national guidelines (CCLG Guidelines, where CCLG
SZYCHOT ET AL .
is Children’s Cancer and Leukaemia Group) which were based on the
Milan Cancer Centre protocol published by Gandola et al.5 The cohort
consisted of 11 males and 3 females with a median age of 7.6 years
(range, 3–13) at the time of diagnosis. Nine children had medulloblas-
toma, four with sPNET and one child had metastatic pineoblastoma.
Eight patients were Caucasian, three children came from the Indian
subcontinent and three from the Middle East. Metastatic staging at
diagnosis showed that three patients had Mo disease, one had M1 , two
had M2 and eight had M3 disease. Nine children had residual disease
≥1.5 cm2 after surgery. The clinical characteristics of these patients
are summarised in Table 1. Institutional approval was obtained for ret-
rospective data extraction from clinical notes, hospital databases and
serial imaging studies.
2.2 Therapy methods
Maximal safe tumour resection was attempted where possible. After
surgery, patients received four cycles of induction chemotherapy con-
sisting of intravenous methotrexate 8 g/m2 , plus vincristine 1.5 mg/m2
(days 1 and 8), etoposide 2.4 g/m2 (week 2), cyclophosphamide 4 g/m2
plus vincristine 1.5 mg/m2 (week 5) and carboplatin 800 mg/m2 .
Disease response was evaluated before radiotherapy planning to
guide the radiation dose. Children received HART to the whole CNS
and a boost, as described below. After radiotherapy, children in com-
plete remission pre-radiotherapy received maintenance chemother-
apy over 54 weeks: intravenous vincristine 1.5 mg/m2 as a bolus every
3 weeks (a total of 18 doses) and a total of six courses of oral lomus-
tine 80 mg/m2 every 9 weeks. In children with persistent disseminated
disease prior to HART or those with sPNET, the maintenance therapy
after HART was replaced with two myeloablative courses of thiotepa
(900 mg/m2 /course) followed by autologous stem cell rescue with a
minimum dose of 3.0 × 106 /kg of CD34+ . The autologous stem cells
had been collected during induction therapy. Of the 11 patients treated
with myeloablative courses of thiotepa followed by autologous stem
cell rescue, 8 received two courses while the remaining 3 children
received one course.
2.3 Radiotherapy technique
After induction chemotherapy, patients were stratified by their
response on the reassessment MRI scan, and also by age. Those
younger than 10 years with complete or partial response received
lower dose craniospinal radiotherapy with a higher dose posterior
fossa boost (to keep the dose to the primary site the same regard-
less of response). In phase one of the whole neuraxis, all patients were
irradiated with two daily 1.3 Gy fractions at least 6 hr apart, 5 days a
week using 6 MV photons.5 Typically, one direct posterior field cov-
ered the spine; two were used in larger patients. The head was treated
with two lateral opposed fields with protection of the facial structures.
Three junction positions were used to ensure that there was no over-
dosing or underdosing at field junctions. Children aged 10 and above
and younger poor responders received a total dose of 39 Gy in 30 frac-
tions of 1.3 Gy over 21 days to the neuraxis. Those under the age of 10,
SZYCHOT ET AL . 3 of 8

TA B L E 1 Patient clinical data and MRI findings

Lansky Lansky Brain volume Follow-up

No. Age Dgn Thiotepa score 1 score 2 Neurology loss (outcome) Radiotherapy
1. 10 MBL 0 40 60 Ataxia, PN MD 1 year 8 months CNS—39 Gy/30F
(Alive) PS – 21Gy/14F

2. 10 MBL 2 70 70 Ataxia, DDK SV 4 years 5 months CNS—31.2 Gy/24F

(alive) PS - 28.5 Gy/19F

3. 5 MBL 0 20 60 Ataxia, hemipharesis, ML 1 year 2 months CNS—39 Gy/30F

nystagmus (died) PS—21 Gy/14F
Met—9 Gy/6F

4. 7 MBL 2 70 70 Ataxia MD 5 years 7 months CNS—39 Gy/30F

(alive) PS—21 Gy/14F
Met—9 Gy/6F

5. 3 MBL 2 70 70 TM ML 1 year 3 months CNS—39 Gy/30F

(died) PS—20.2 Gy/14F

6. 6 MBL 2 70 70 Ataxia, DDK, MD 1 year 1 months CNS—39 Gy/30F

hemiparesis (died) PS and Met—9 Gy/6F
PS —12 Gy/8F

7. 10 MBL 2 10 10 Hemiparesis, FNP, NO 8 months (died CNS—39 Gy/13F

squint, nystagmus, C1) PS—21 Gy/14F
LM
8. 9 MBL 1 70 70 TM ML 1 year 2 months CNS—39 Gy/13F
(died) PS—21 Gy/14F

9. 13 MBL 1 – – – MD 2 years (died) No data

10. 5 sPNET 2 80 80 Hemiparesis SV 5 years 4 months CNS—31.2 Gy/14F
(alive) PS—28.5 Gy/19F

11. 11 sPNET 2 70 70 PN, weakness, LM – 5 years 4 months CNS—39 Gy/30F

(alive) PS—21 Gy/14F

12. 3 sPNET 0 70 70 PN, seizures, enuresis, NO 5 years 9 months CNS—31.2 Gy/14F

bowel incontinence (alive) PS—28.5 Gy/19F

13. 8 sPNET 2 80 70 Ataxia ML 11 months (died CNS—39 Gy/13F

C2) PS—21 Gy/14F

14. 7 PBL 1 10 10 Spastic quadriparesis, SV 1 year 11 months CNS—39 Gy/30F

global developmental (alive) PS—21 Gy/14F
delay Met—7.8 Gy/6F

Age, patient’s age in years; Alive, patient alive in December 2016; Brain volume, neuroparenchymal volume loss; CNS, whole central nervous system radio-
therapy; DDK, dysdiadochokinesia; Dgn, histopathological diagnosis; Died, patient died of disease progression (relapse); Died C1, patient died from complica-
tions of treatment—acute respiratory distress syndrome; Died C2, patient died from complications of treatment—gastrointestinal bleeding; FNP, facial nerve
palsy; Follow-up, duration of follow-up (years and months); Lansky score 1, minimal score recorded during follow-up appointments after the end of therapy;
Lansky score 2, score recorded at the last follow-up; LM, limited mobilisation; LPMD, leptomeningeal disease; MBL, medulloblastoma; MD, moderate; Met,
metastasis radiotherapy, in addition to CNS radiotherapy; ML, mild; Neurology, neurological findings post therapy; NO, no change; PBL, pineoblastoma; PN,
peripheral neuropathy; PS, primary site radiotherapy (either posterior fossa or supratentorial), in addition to CNS radiotherapy; Radiotherapy: Site, dose
(Gy)/number of fractions (F); sPNET, supratentorial primitive neuroectodermal tumour; SV, severe; Thiotepa, the number of courses of high-dose thiotepa;
TM, symptoms of transverse myelitis; VT, third ventricle.

who had achieved either complete or partial remission, received 31.2
Gy in 24 fractions of 1.3 Gy over 16 days.
During phase two, children received irradiation with intensity mod-
ulated arc therapy either to the whole posterior fossa (in children with
medulloblastoma) or to the tumour bed (those patients with sPNET).
In children who had received a total dose of 39 Gy to the whole
neuraxis, 21 Gy in 14 fractions of 1.5 Gy was administered over 9
days. This took the total primary site dose, in those receiving the
higher craniospinal dose, to 60 Gy in 44 fractions over 30 days. In
those who had received only 31.2 Gy, the radiation dose was 28.5 Gy
in 19 fractions of 1.5 Gy over 14 days. This took the total primary
tumour dose, in those receiving the lower craniospinal dose, to 59.7
4 of 8
F I G U R E 1 MRI images showing decline in corpus callosum volume following treatment. Image on the left (baseline): segmented corpus callosum
(red) at diagnosis. Image on the right: 33% decline in corpus callosum volume (red) following treatment.
Gy in 43 fractions over 30 days (assuming no unscheduled interrup-
tions). Composite plans of the whole CNS and primary site boost were
reviewed carefully to ensure that doses to the brain stem and cervi-
cal cord did not exceed recognised radiation tolerance. Children with
persisting localised small volume metastatic lesions following induc-
tion chemotherapy could be given a boost up to 9 Gy in six, twice
daily fractions of 1.5 Gy. No patients in this series received additional
radiotherapy to the posterior fossa for persistent primary site disease.
Patients with diffuse leptomeningeal disease did not receive a further
boost.
2.4 Performance status
The Lansky Scale is designed for patients under the age of 16 and was
used to determine the functional status of children after completion
of treatment. The performance scale between 10 and 100 was used to
define the score that best represented the child’s activity status.
A score of 100 indicated that the child was able to carry out
fully normal activity. A score between 50 and 60 denoted moderate
activity restrictions, while scores of 40 and below indicated severe
restrictions.20 Performance status ratings were obtained on children
following completion of chemotherapy. The children scored retrospec-
tively based on the information from consultations recorded in medical
and nursing notes, computer data and clinic letters.
2.5 MRI evaluation
Radiological assessment of disease was performed preoperatively,
within 72 hr post surgery and before HART. After completing treat-
ment, children underwent MRI studies at 3 monthly intervals during
the first year and 6 monthly thereafter. T2-weighted images, pre-
and post-contrast T1-weighted sequences as well as fluid-attenuated
inversion recovery and diffusion weighted imaging/apparent dif-
fusion coefficient were obtained in each patient. Serial MRI scans
from presentation until the last follow-up were retrospectively and
quantitatively analysed by two specialist paediatric neuro-radiologists
assessing the trajectory of corpus callosum volume loss with a
consensus reading blinded to the original report.
SZYCHOT ET AL .
2.6 Brain volume analysis
Quantitative brain volume analysis was undertaken using measure-
ment of corpus callosum volume. Specifically, the corpus callosum was
manually segmented by drawing a region of interest (ROI) on the mid-
sagittal slice of structural images taken after surgery and at follow-
up (Fig. 1). Images were selected such that they were as similar as
possible (similar contrasts/resolutions) at both time points. The vol-
ume of the ROI was then calculated using Jenkinson FMRIB Software
Library’s stats command and normalised by slice thickness to account
for differences in acquisition.21 Finally, the absolute change in vol-
ume was compared between neuroparenchymal atrophy groups (mild,
moderate or severe) using an analysis of covariance, implemented in
statistical software- the CAR package in R.22
3 RESULTS
3.1 Neurological outcome and Lansky performance
status
We recorded the maximum decline in Lansky performance status as
well as the score acquired at last follow-up (Table 1). We were not able
to determine performance status in 1 of 14 patients as care was trans-
ferred to another centre. Performance status was severely restricted
in 4 of 13 children (Lansky score 10–40). Mild to moderate impairment
was noted in 7 of 13 children (Lansky score 50–70). Two of 4 patients
with severely restricted performance status showed either severe
or moderate neuroparenchymal volume loss. One of those patients
experienced a profound clinical deterioration after the first course of
myeloablative thiotepa. A computerised tomography (CT) brain scan
revealed parenchymal changes with extensive swelling of brain stem
structures. Subsequently, the patient required decompression of pos-
terior fossa. A brain biopsy showed nonspecific features. No pathogens
were identified in the brain tissue or cerebrospinal fluid (CSF). His neu-
rological function was much worse than expected from the changes on
his brain MRI scan.
Two children treated with HART and high-dose thiotepa suffered
from transverse myelitis. One patient developed a significant lethargy
with marked ataxia. An MRI scan showed extension of a previously
SZYCHOT ET AL .
FIGURE 2 Boxplots showing change in corpus callosum volume for
patients with mild (left), moderate (middle) and severe (right) neuro-
parenchymal volume losses
noted enhancing signal abnormality within the upper cervical cord with
some swelling of the cervical cord. Similar changes were noted on a
scan post high-dose treatment and were considered to be part of a
post-therapy process. The second child developed lethargy and sub-
sequently left-sided hemiparesis. A CT scan showed a moderate-sized
chronic subdural haematoma causing mass effect. No pathogen was
identified in CSF or blood cultures. His hemiparesis never resolved.
An MRI scan demonstrated patchy enhancement affecting the cervi-
cal cord at C1–C2 and probably also the upper thoracic cord. However,
subsequent scans showed disease progression. Neurological findings
of these patients post therapy are summarised in Table 1.
3.2 Brain volume
Data were obtained in 13 of 14 patients. Adequate analysis of the cor-
pus callosum was not possible in one patient due to the poor quality
of available images. Eleven of the 13 children showed mild-to-severe
generalised neuroparenchymal atrophy. Moderate to severe neuro-
parenchymal volume loss was observed in 7 of 13 patients. A review
of the clinical data did not identify any alternative cause of atrophy in
these children. Figure 1 shows an example of corpus callosum segmen-
tations at two time points in a patient with severe generalised neuro-
parenchymal atrophy. Figure 2 shows a box plot of corpus callosum vol-
ume loss in each generalised neuroparenchymal atrophy group. Corpus
callosum volume loss was significantly higher in the group of severely
affected children compared to the mildly affected group, while the dif-
ference between the moderate and severe groups was close to signifi-
cance (Fig. 2). Eleven children were treated with myeloablative courses
of thiotepa. Six of these children developed moderate to severe neu-
roparenchymal atrophy, three developed mild neuroparenchymal vol-
ume loss. It is possible that further progression to moderate or severe
neuroparenchymal atrophy could have subsequently occurred in some
5 of 8
F I G U R E 3 Scatter plot showing change in corpus callosum volume
and Lansky score
of the less severely affected children, as three of them died early and
one died of treatment complications. The three patients who did not
receive myeloablative therapy with thiotepa showed mild, moderate or
no neuroparenchymal volume loss.
The median age at diagnosis for those children who showed severe
and moderate neuroparenchymal volume loss was 7.1 years. The
median age at diagnosis of patients with mild or no neuroparenchymal
atrophy was 8.1 years.
3.3 Brain volume versus Lansky score
There was no correlation between brain volume loss and Lansky
performance status (Fig. 3).
3.4 Survival
Six children of the 14 were alive at a median follow-up of 1.8 years
(range, from 8 months to 5.7 years). Five of those patients currently
have no evidence of disease. Six patients died of disease progression
and two of treatment complications.
4 DISCUSSION
The fact that brain tumour survivors can develop changes in the struc-
ture of white matter, leading to neurocognitive and sensory deficits,
has been well documented.23 In 2006, Shan et al. for the first time
quantitatively evaluated morphologic white matter damage in the
developing brains of children treated for medulloblastoma with radi-
ation therapy. The results revealed significant deficits in the integrity
of white matter and these white matter density changes may have
explained the volume loss they reported.24
In 2014, Thust et al. reported multiple anomalies including brain vol-
ume loss on MRI scans in children treated for high-risk PNET accord-
ing to the CCLG Guidelines at GOSH and UCLH. Brain volume loss
in this cohort was associated with leukodystrophy of cerebellar and
cerebral white matter. While Shan et al. applied fractal geometry to
quantify the morphologic effect of neurotoxicity, Thust’s group used a
6 of 8
subjective visual grading system to estimate neuroparenchymal vol-
ume loss.17,24 In our study, we have optimised a quantitative brain
volume analysis using corpus callosum segmentation and measure-
ments based on previous reports indicating that cranial irradiation
would adversely affect the development of the corpus callosum, which
is the largest commissure of the brain’s white matter.25 This tech-
nique enabled us to exclude the impact of surgical resection on neuro-
parenchymal volume loss. Otherwise it would be difficult to differenti-
ate between chemo-radiation neurotoxicity and effects of the surgery
itself. We observed mild to severe degree of neuroparenchymal volume
loss in 11 of 14 patients, which is consistent with the previous report
from our centre.17 In 2000, Reddick et al. also reported a significant
volume loss of normal appearing white matter in children treated with
craniospinal irradiation for medulloblastoma.26 There were no signif-
icant differences in the incidence of white matter loss when patients
were stratified by age at the time of craniospinal therapy.26 This is con-
sistent with our findings, as in this study the median age of children
with severe and moderate neuroparenchymal volume loss was simi-
lar to those with mild or no brain atrophy. Interestingly, in our study,
patients diagnosed with medulloblastoma had greater neuroparenchy-
mal volume loss compared to those with sPNET. In contrast, no imaging
abnormalities were described in the U.K. study evaluating treatment
with HART in patients with medulloblastoma in a protocol that did not
use high-dose chemotherapy with stem cell rescue.9
Debilitating cognitive declines have also previously been reported
in children treated with craniospinal radiotherapy.24 The initial tox-
icity review of an intensive sequential high-dose thiotepa therapy
given after HART regimen according to the ‘Milan’ protocol showed
no severe neurological sequelae of treatment, with a median follow-up
of 82 months.5 In the Gandola study, 2 of 32 children who completed
HART developed self-resolving, unspecific encephalitis during the first
myeolablative course and the second course was omitted, no long-term
neurotoxicity was reported.5
In contrast, the U.K. audit presented at International Society of Pae-
diatric Oncology (SIOP) 2015 revealed severe, unexpected neurotoxi-
city in 8 of 19 patients treated with intensive chemotherapy, HART and
high-dose thiotepa. Seven of these patients developed a rapid onset
of global neurological deterioration with bulbar dysfunction. All these
children had no neurological impairment prior to the treatment with
thiotepa. As in Thust’s report, leukoencephalopathic changes were
noted on MRI scans of these patients.27
In our cohort one of our 14 patients suffered a clinical deterioration
with morphological features resembling encephalitis following the first
course of myeloablative thiotepa. Diffuse signal abnormality within
cerebrum, cerebellum, brain stem and deep grey matter subsequently
evolved into severe neuroparenchymal atrophy and severe restriction
of performance status.
As no other causative factor was identified, these features may
have resulted from treatment neurotoxicity. Neurological sequelae
of treatment were also observed in three other patients with neu-
roparenchymal volume loss; two patients developed symptoms of
transverse myelitis and one ataxia and cerebellar symptoms. Review
of the radiotherapy plans and clinical charts of patients with myelitis
excluded a possibility of the cervical spine inclusion in the posterior
SZYCHOT ET AL .
fossa boost volumes, a possible association reported by the SIOP
Brain Working Group.28 Radiological and clinical findings were also
associated with an impairment of performance status. The SIOP Brain
Working Group consisting of European Radiation Oncologists and
Physicists held a meeting in 2014 to address the issues of ‘unexpected’
neurotoxicities reported in different countries following treatment of
240 children with high-risk PNET. Significant neurotoxicities, such as
brainstem/cerebellum radionecrosis, and three cases of myelitis were
reported.28
In contrast to our results, in 2004 Fouladi et al. reported localised
atrophy only in one of 134 children treated with high-dose chemother-
apy, craniospinal radiotherapy and topotecan for medulloblastoma or
PNET. Twenty of 134 patients developed white matter lesions and
only a minority of children suffered from permanent neurologic deficits
and morphological changes suggesting that the risk of neurological
deficit may be protocol specific. However, the presence of white matter
changes was associated with a significant neurocognitive decline.29
The conclusions from our report are limited by the small number
of patients and retrospective assessment of their performance sta-
tus. Based on the clinical data available, we were unable to define
risk factors for neuroparenchymal volume loss. However, substantial
brain volume loss as well as unexpected neurotoxicity were evident
in our cohort of children following treatment. Our observations are
consistent with those published last year by Vivekanandan et al. who
reported that a number of U.K. centres were experiencing problems
with severe neurotoxicity in children treated according to the CCLG
Guidelines.13 The reported neurotoxicity included symptoms related to
damage of both brain stem and spinal cord (myelitis).
Although neurotoxicity is a well-recognised side effect of the whole
CNS irradiation, these observations appear to show that the toxicity
following treatment according to the CCLG Guidelines is more severe
than expected.
When we started this study, we hypothesised that there would be
a correlation between brain volume loss and Lansky performance sta-
tus in our patients following treatment. Our data do not support this
hypothesis, although they are limited by a small sample size. Another
limitation of our study is the retrospective estimation of the perfor-
mance status based on the clinical description of patients in clinic. We
accept that this is potentially inaccurate, but the scores were not rou-
tinely assessed in clinics. Thus, this was the best quality of data we
could extract retrospectively. We required a numerical score to evalu-
ate the correlation with brain volume numerical data. It should also be
noted that there are possible limitations of the Lansky scoring method
because of its comprehensiveness. Different domains of disability are
presented as a single sum score so it is difficult to identify changes
within a specific domain.
Another important limitation of our study is that the image parame-
ters varied, in particular the in-place resolution, between subjects and
time points. Thus, an accurate manual outlining of the callosum and its
quantification was difficult. Further difficulties may have also resulted
from the impact of irradiation therapy on the variability of the corpus
callosum’s size and shape with respect to other brain structures.30–32
Our results suggest that the measurement of brain volume loss
in patients treated according to the CCLG Guidelines may not be an
SZYCHOT ET AL .
accurate predictor of function. However, the correlation of brain vol-
ume loss due to late neurotoxicity with performance decline may be
more obvious over a longer period of follow-up than examined here.
The outcome of this study validates previous results of neuro-
parenchymal volume loss published by Thust et al., using a different
method of estimating brain volume. Our results also confirm previous
suggestion that the combination of HART and myeloablative courses
of thiotepa is associated with severe neurotoxicity and subsequent
decline in performance status in a subset of patients. Further clinical
research should focus on investigating the correlation between neu-
roparenchymal volume loss, neurocognitive sequelae and performance
status in a larger cohort of patients.
ACKNOWLEDGEMENTS
We acknowledge NIHR support of Dr. Elwira Szychot. Dr. Mark N. Gaze
is supported, in part, by the NIHR University College London Hospitals
Biomedical Research Centre.
CONFLICT OF INTEREST
The authors declare that there is no conflict of interest.
REFERENCES
1. Arora RS, Alston RD, Eden TOB, Estlin EJ, Moran A, Birch JM. Age—
incidence patterns of primary CNS tumors in children, adolescents,
and adults in England. Neuro Oncol. 2009;11(4):403–413.
2. Louis DN, Perry A, Reifenberger G, et al. The 2016 World Health Orga-
nization classification of tumors of the central nervous system: a sum-
mary. Acta Neuropathol. 2016;131(6):803–820.
3. Louis DN, Ohgaki H, Wiestler OD, et al. The 2007 WHO classifi-
cation of tumours of the central nervous system. Acta Neuropathol.
2007;114(2):97–109.
4. Jakacki RI, Burger PC, Zhou T, et al. Outcome of children with
metastatic medulloblastoma treated with carboplatin during cran-
iospinal radiotherapy: a Children’s Oncology Group Phase I/II study.
J Clin Oncol. 2012;30(21):2648–2653.
5. Gandola L, Massimino M, Cefalo G, et al. Hyperfractionated acceler-
ated radiotherapy in the Milan strategy for metastatic medulloblas-
toma. J Clin Oncol. 2009;27(4):566–571.
6. Pizer BL, Clifford SC. The potential impact of tumour biology on
improved clinical practice for medulloblastoma: progress towards
biologically driven clinical trials. Br J Neurosurg. 2009;23(4):364–
375.
7. Yuh GE, Loredo LN, Yonemoto LT, et al. Reducing toxicity from cran-
iospinal irradiation: using proton beams to treat medulloblastoma in
young children. Cancer J. 2004;10(6):386–390.
8. Taylor RE, Bailey CC, Robinson KJ, et al. Outcome for
patients with metastatic (M2-3) medulloblastoma treated with
SIOP/UKCCSG PNET-3 chemotherapy. Eur J Cancer. 2005;41(5):727–
734.
9. Taylor RE, Howman AJ, Wheatley K, et al. Hyperfractionated accel-
erated radiotherapy (HART) with maintenance chemotherapy for
metastatic (M1-3) medulloblastoma—a safety/feasibility study. Radio-
ther Oncol. 2014;111(1):41–46.
10. von Hoff K, Hinkes B, Gerber NU, et al. Long-term outcome and clin-
ical prognostic factors in children with medulloblastoma treated in
the prospective randomised multicentre trial HIT’91. Eur J Cancer.
2009;45(7):1209–1217.
7 of 8
11. Chintagumpala M, Hassall T, Palmer S, et al. A pilot study of risk-
adapted radiotherapy and chemotherapy in patients with supratento-
rial PNET. Neuro Oncol. 2009;11(1):33–40.
12. Gajjar A, Chintagumpala M, Ashley D, et al. Risk-adapted craniospinal
radiotherapy followed by high-dose chemotherapy and stem-cell res-
cue in children with newly diagnosed medulloblastoma (St Jude
Medulloblastoma-96): long-term results from a prospective, multicen-
tre trial. Lancet Oncol. 2006;7(10):813–820.
13. Vivekanandan S, Breene R, Ramanujachar R, et al. The UK experience
of a treatment strategy for pediatric metastatic medulloblastoma com-
prising intensive induction chemotherapy, hyperfractionated accel-
erated radiotherapy and response directed high dose myeloablative
chemotherapy or maintenance chemotherapy (Milan strategy). Pediatr
Blood Cancer. 2015;62(12):2132–2139.
14. Liptak C, Brinkman T, Bronson A, et al. A social program for adoles-
cent and young adult survivors of pediatric brain tumors: the power of
a shared medical experience. J Psychosoc Oncol. 2016;34(6):493–511.
15. Dietrich U, Wanke I, Mueller T, et al. White matter disease in children
treated for malignant brain tumors. Childs Nerv Syst. 2001;17(12):731–
738.
16. Ruben JD, Dally M, Bailey M, Smith R, McLean CA, Fedele P. Cerebral
radiation necrosis: incidence, outcomes, and risk factors with emphasis
on radiation parameters and chemotherapy. Int J Radiat Oncol Biol Phys.
2006;65(2):499–508.
17. Thust SC, Blanco E, Michalski AJ, et al. MRI abnormalities in chil-
dren following sequential chemotherapy, hyperfractionated acceler-
ated radiotherapy and high-dose thiotepa for high-risk primitive neu-
roectodermal tumours of the central nervous system. J Med Imaging
Radiat Oncol. 2014;58(6):683–690.
18. Kennedy C, Bull K, Chevignard M, et al. PNET4 study group of the
Brain Tumour Group of the European branch of the International Soci-
ety of Paediatric Oncology (SIOP-E). Quality of survival and growth in
children and young adults in the PNET4 European controlled trial of
hyperfractionated versus conventional radiation therapy for standard-
risk medulloblastoma. Int J Radiat Oncol Biol Phys. 2014;88(2):292–
300.
19. Spreafico F, Gandola L, Marchianò A, et al. Brain magnetic resonance
imaging after high-dose chemotherapy and radiotherapy for child-
hood brain tumors. Int J Radiat Oncol Biol Phys. 2008;70(4):1011–
1019.
20. Lansky SB, List MA, Lansky LL, Ritter-Sterr C, Miller DR. The
measurement of performance in childhood cancer patients. Cancer.
1987;60(7):1651–1656.
21. Jenkinson M, Beckmann CF, Behrens TE, Woolrich MW, Smith SM.
FSL. Neuroimage. 2012;62(2):782–790.
22. Fox J, Weisberg HS. An R Companion to Applied Regression. 2nd ed.
London: SAGE Publications Ltd; 2011.
23. Filley CM, Kleinschmidt-DeMasters BK. Toxic leukoencephalopathy.
N Engl J Med. 2001;345(6):425–432.
24. Shan ZY, Liu JZ, Glass JO, Gajjar A, Li CS, Reddick WE. Quan-
titative morphologic evaluation of white matter in survivors of
childhood medulloblastoma. Magn Reson Imaging. 2006;24(8):1015–
1022.
25. Palmer SL, Reddick WE, Glass JO, Gajjar A, Goloubeva O, Mulhern
RK. Decline in corpus callosum volume among pediatric patients with
medulloblastoma: longitudinal MR imaging study. AJNR Am J Neurora-
diol. 2002;23(7):1088–1094.
26. Reddick WE, Russell JM, Glass JO, et al. Subtle white matter vol-
ume differences in children treated for medulloblastoma with conven-
tional or reduced dose craniospinal irradiation. Magn Reson Imaging.
2000;18(7):787–793.
8 of 8
27. Sastry J, Pizer B, Thorp N, et al. UK national audit reveals unex-
pected neurotoxicity of children with STPNET treated with intensive
chemotherapy, hyperfractionated radiotherapy (HART) and tandem
high dose thiotepa consolidation. Pediatr Blood Cancer. 2015; 62(Suppl
4):S216.
28. Gandola L, Horan G, Meroni S, et al. Intensive treatment of high-risk
medulloblastoma (HR-MB): how to learn from toxicities in an European
setting of radiotherapists and physicists of the SIOP Brain Tumour
Working Group? Neuro Oncol. 2016; 18(Suppl. 3):iii108.
29. Fouladi M, Chintagumpala M, Laningham FH, et al. White matter
lesions detected by magnetic resonance imaging after radiotherapy
and high-dose chemotherapy in children with medulloblastoma or
primitive neuroectodermal tumor. J Clin Oncol. 2004;22(22):4551–
4560.
30. Wang D, Shi L, Chu WC, Paus T, Cheng JC, Heng PA. A comparison
of morphometric techniques for studying the shape of the corpus cal-
losum in adolescent idiopathic scoliosis. Neuroimage. 2009;45(3):738–
748.
SZYCHOT ET AL .
31. Collinson SL, Gan SC, Woon PS, et al. Corpus callosum morphology
in first-episode and chronic schizophrenia: combined magnetic res-
onance and diffusion tensor imaging study of Chinese Singaporean
patients. Br J Psychiatry. 2014;204(1):55–60.
32. Pekala JS, Mamourian AC, Wishart HA, Hickey WF, Raque JD. Focal
lesion in the splenium of the corpus callosum on FLAIR MR images: a
common finding with aging and after brain radiation therapy. AJNR Am
J Neuroradiol. 2003 May;24(5):855–861.
How to cite this article: Szychot E, Seunarine K, Mankad
K, et al. Impact of induction chemotherapy, hyperfractionated
accelerated radiotherapy and high-dose thiotepa on brain vol-
ume loss and functional status of children with primitive neu-
roectodermal tumour. Pediatr Blood Cancer. 2017;00:e26619.
https://doi.org/10.1002/pbc.26619