вход по аккаунту


Effect of therapeutic ionizing radiation on the human brain.

код для вставкиСкачать
Effect of Therapeutic Ionizing Radiation on
the Human Brain
R. Grant Steen, PhD,1– 4 David Spence, BS,1 Shengjie Wu, MS,5 Xiaoping Xiong, PhD,5 Larry E. Kun, MD,6
and Thomas E. Merchant, DO, PhD4,6
We test a hypothesis that fractionated radiation therapy within a therapeutic dose range is associated with a dose-related
change in normal brain, detectable by quantitative magnetic resonance imaging. A total of 33 patients were examined by
quantitative magnetic resonance imaging to measure brain tissue spin-lattice relaxation time (T1) before treatment, and
at various times during and after radiation therapy. A T1 map was generated at each time point, and radiation therapy
isodose contours were superimposed on the corresponding segmented T1 map. Changes in white matter and gray matter
T1 were analyzed as a function of radiation therapy dose and time since treatment, controlling for patient age and tumor
site. In white matter, a dose level of more than 20Gy was associated with a dose-dependent decrease in T1 over time,
which became significant 6 months after treatment. There was no significant change in T1 of gray matter over time, at
radiation therapy doses of less than 60Gy. However, GM in close proximity to the tumor had a lower T1 before therapy.
Our results represent the first radiation dose-response data derived from pediatric brain in vivo. These findings confirm
that white matter is more vulnerable to radiation-induced change than is gray matter, and suggest that T1 mapping is
sensitive to radiation-related changes over a broad dose range (20 to 60Gy). Human white matter T1 is not sensitive to
radiation therapy of less than 20Gy, and gray matter T1 is unchanged over the dose range used to treat human brain
tumor. The reduction of gray matter T1 near the tumor could result from compression of cortical parenchyma near the
growing tumor mass, or from tumor cell invasion directly into the parenchyma. If brain T1 is a surrogate for radiation
effect, reducing the volume of normal white matter receiving more than 20Gy could be an important treatment planning
Ann Neurol 2001;50:787–795
Radiation therapy (RT) is the single most effective
nonsurgical treatment for brain tumor, and radiation
doses of less than 60Gy yield dose-related increases in
patient survival in many brain tumor types.1 However,
ionizing radiation is associated with neurological and
neurocognitive alterations after treatment of brain tumors, suggesting that RT can damage the brain.2 The
potential for RT-induced damage of normal brain thus
limits the dose of radiation that can be safely used
when treating brain tumor.3
The physiology of radiation-induced damage to
brain is poorly understood.3 Pathological changes at
autopsy are generally limited to white matter (WM)
and can include demyelination, coagulative necrosis, or
focal mineralization.4,5 Pathology suggests that the primary locus of RT injury is either the oligodendrocytes
with resultant axonal demyelination,6,7 or the vascular
endothelial cells, with resultant effects on integrity of
the blood–brain barrier (BBB).8 –10 However, autopsy
studies are limited by the fact that changes are usually
not observed until months after therapy, and autopsied
patients are generally those with the most extensive and
heavily treated tumor.11 A methodology is needed to
monitor normal brain for signs of radiation response,
especially if there is sufficient sensitivity to characterize
acute or diffuse changes. Such a method would be useful in the RT planning process, and might help to
identify those patients most at risk of tumor progression or treatment-related sequelae.
Up to 50% of brain tumor patients show a change
in WM by conventional magnetic resonance imaging
(cMRI) after radiation,12–16 so cMRI provides some
insight into the response of human brain to therapeutic
RT.17,18 However, the appearance of WM damage by
cMRI often correlates poorly with the severity of neurocognitive outcomes,16,19 –21 arguing that cMRI is not
sensitive to diffuse changes in the brain.
Preliminary work from our laboratory suggests that
From the Departments of 1Diagnostic Imaging, 5Epidemiology and
Biostatistics, and 6Radiation Oncology, St Jude Children’s Research
Hospital, Memphis; and Departments of 2Pediatrics, 3Radiology,
and 4Biomedical Engineering, University of Tennessee School of
Medicine, Memphis, TN.
Published online Nov 1, 2001; DOI: 10/1002/ana.10029
Address correspondence to Dr Steen, Department of Diagnostic Imaging, St Jude Children’s Research Hospital, 332 North Lauderdale,
Memphis, TN 38105-2794. E-mail:
Received May 4, 2001, and in revised form Jul 26, 2001. Accepted
for publication Aug 9, 2001.
© 2001 Wiley-Liss, Inc.
Table 1. Summary Data for Study Patients
Age at RT (yr)
Prescribed RT (Gy)
Follow-up time (wk)
No. of MRIs evaluated per patient
Tumor size at diagnosis (cm)
Summary data for 33 study patients, including the mean, standard deviation, and range of reported values. Among study patients, all diagnoses
were biopsy-proven and included ependymoma (13 patients); juvenile pilocytic astrocytoma (9 patients); craniopharyngioma (5 patients);
low-grade astrocytoma (1 patient); anaplastic astrocytoma (1 patient); ganglioglioma (1 patient); astroblastoma (1 patient); pleomorphic xanthoastrocytoma (1 patient); and glioblastoma multiforme (1 patient).
RT ⫽ radiation therapy.
quantitative MRI (qMRI) is more sensitive than cMRI
to diffuse brain changes.18,22 In a preliminary study,
we found that radiation administered to normal brain
during treatment of focal brain tumor was associated
with qMRI abnormalities in WM, but not in gray matter (GM). In WM, radiation doses within the therapeutic range were associated with a dose-dependent decrease in tissue spin-lattice relaxation time (T1).
However, this preliminary work reported a small sample of patients, which limited our ability to detect effects at low RT doses.18 We report a prospective study
of 33 patients followed for less than 1 year after RT.
We show that qMRI provides a sensitive, quantitative,
and objective measure of radiation effect in the brain,
which is quantitatively related to radiation dose over
the therapeutic range.
Patients and Methods
Protocol Overview and Study Design
A prospective phase II study of image-guided conformal RT
for pediatric brain tumor was opened to accrual at St Jude
Children’s Research Hospital in 1997. Eligibility criteria included the following:
• Patient age: 1.5 to 21 years
• Histologically confirmed primary brain tumor
• Focal tumor (no dissemination of tumor within or beyond the brain)
• Histological type requiring only focal irradiation
• No prior RT
• No ongoing chemotherapy (corticosteroids excluded)
• Adequate performance status (ECOG 0 –3)
Parents or guardians of children signed an informed
consent after a detailed description of the protocol. Patients received an examination including qMRI before
the initiation of RT. During the qMRI examination,
images were acquired enabling measurement of T1 in
WM and cortical GM. The examination was repeated
at weeks 3 and 5 of RT, and every 3 months after the
start of RT.
Annals of Neurology
Vol 50
No 6
December 2001
We report on 33 consecutive patients who completed their
planned course of RT (Table 1). Each patient was treated
using conformal techniques and the total dose prescribed was
either 54.0 or 59.4Gy, using conventional fractionation
(1.8Gy/day). Most patients (n ⫽ 18) received a minimum of
5 qMRI examinations; additional qMRI examinations were
obtained at week 39 (n ⫽ 15) and week 52 (n ⫽ 11) for
some patients.
Conformal Radiation Therapy
Conformal three-dimensional (3D) treatment plans employing 4 to 25 beams were developed using the “PLan University of North Carolina” treatment planning system.23 Most
treatments were delivered on MLC-equipped linear accelerators (Primus and Primart; Siemens Medical Systems, Concord, NJ). Patients were immobilized with a stereotactic head
frame, a thermoplastic face mask, or a vacuum bag molded
to the patient; general anesthesia was used when necessary.
Treatment guidelines specified a 10mm anatomically defined
clinical target volume (CTV) for low-grade tumors, whereas
patients with high-grade tumors had a 20mm anatomically
defined CTV. The geometric margin used to define the
planning target volume was 5mm for all patients. Targeting
followed ICRU guidelines,23 and tissue outside the CTV is
regarded as normal for this analysis. Isodose contours were
generated for each patient in a plane corresponding to the
qMRI image slice. Because the tumor was not always in or
near the qMRI slice, the T1 map often did not include tissue
that received the highest RT dose.
Conventional MRI
All MRI was performed on a 1.5-T imager (Siemens Medical
Systems), using a standard quadrature head coil. Conventional MRI was used to evaluate response to treatment and
to identify patients with progressive disease. Patients with
progressive disease were eliminated from analysis, because an
acute change in T1 in these patients could be attributable to
radiation or to the effect of tumor. In patients free of progressive disease, a T1 change is assumed to be due to radiation alone.
Quantitative MRI
Quantitative imaging was done with a precise and accurate
inversion–recovery (TurboPAIR) method optimized and validated previously in our laboratory,18,22,24 –26 which requires
about 4 minutes of imaging time.18 A single transverse slice
through the basal ganglia was selected, to show the same
structures for all patients and controls.
Measurement of T1
All qMRI images were analyzed on an SGI Indy workstation.
Pixels identified as noise were excluded by statistical criteria27
and remaining pixels were submitted to a curve-fitting algorithm.24,27,28 A T1 equation was solved for ␣ (spin-density),
k (cosine of the effective flip angle of the inversion pulse),
and T1 in each pixel. The T1 solution was used to produce
a parametric map, wherein pixel grayscale value is equivalent
to the T1 relaxation time.
Segmentation of qMRI Images
The qMRI images were analyzed using a fully automated
neural network algorithm,29 adapted specifically for segmenting TurboPAIR images.30 This method segments and classifies tissue as GM, WM, cerebrospinal fluid (CSF), partial
volume of GM⫹WM, partial volume of GM⫹CSF, or
background. The non-normalized signal intensity from each
pixel in the base images was used as input to a 3 ⫻ 3 singlelayer Kohonen self-organizing map. After segmentation, each
of the nine levels in the segmented image was classified by a
fully automated technique, and a pseudo-color image of the
brain was created.29
The pseudo-color images were imported into PhotoShop v
4.0.1 (Adobe Systems, San Jose, CA), and extrameningial tissues were erased (Fig 1A). Yellow pixels correspond to GM
and dark green pixels correspond to WM. Pale green pixels
correspond to volumes that contain a partial volume of
GM⫹WM, and such pixels were not further analyzed. Central WM and GM structures were erased (see Fig 1B) so that
these pixels were also excluded from analysis. This was done
for several reasons; primarily because we lack adequate control data from which to determine expected WM values in
central structures of the brain,24 and also because dosimetric
lines were often too densely placed in central structures to
assess T1 as a function of RT dose.
Analyzing T1 as a Function of Radiation Dose
To determine the relationship between RT and changes in
T1, isodose contours for each patient were superimposed on
the corresponding segmented T1 map (see Fig 1C and D).
T1 was then calculated as a function of RT dose, for WM
and GM, and T1 was recorded as a function of time since
RT. The completed data set encodes the following: tissue
type (WM and GM); radiation classified into 8 dose ranges
(⬍5Gy; 5 to ⬍10Gy; 10 to ⬍20Gy; 20 to ⬍30Gy; 30 to
⬍40Gy; 40 to ⬍50Gy; 50 to ⬍54Gy; and ⱖ54Gy); and
time since RT (pretreatment, and weeks 3, 5, 13, 26, 40,
and 53 after RT).
Statistical Tests
Our analysis was designed to estimate longitudinal trends in
T1 of WM and GM as a function of RT dose. Because each
patient had multiple qMRI examinations, T1 times are intercorrelated. We used a mixed linear model to analyze
data31,32; thus, T1 is the response variable, each patient is
treated as a data cluster, the day from initiation of RT is the
longitudinal variable, and RT dosage to tissue is the primary
covariate. Because patients were of different ages, patient age
at RT was included in the model. Initially, we used a longitudinal model, freeing the linear trend in T1 in each brain
tissue from the influence of RT dosage. This was done by
using RT as a categorical variable, rather than as an ordinal
variable. This model suggested that the estimated intercepts
and slopes of the trends in T1 changed in accordance with
RT dosage. Therefore, we fitted a surface model, in which
the longitudinal trends in T1 were linearly related to RT
dosage. The T1 trends predicted by the surface model
matched well to those in the simple longitudinal model. We
believe the surface model gives a better estimation by eliminating a layer of T1 variability due to error.18
Patient Summary
The mean follow-up of patients was 34 weeks, and an
average of 5 examinations per patient were evaluated.
Only 18% of patients had disease progression at 1 year,
and these patients were excluded from analysis. Because
the qMRI evaluations of patients were completed at
close to the scheduled examination dates,18 patient examinations are pooled by time since RT.
Brain T1 as a Function of Radiation Dose
Brain tissue T1 was measured in 1,692 regions of interest among the 159 examinations evaluated (data not
shown). An equivalent number of regions were evaluated in WM and GM, so the sensitivity of the method
is comparable in both tissues. About 70% of the data
were acquired within 3 months of RT, and only 8% of
data were acquired at the 1-year examination, so our
analysis is more sensitive at short follow-up intervals.
Mean WM T1 at each dose level is plotted (Fig 2A).
This suggests that WM exposed to more than 20Gy
shows a reduction in T1 by week 26, which persists
through week 53. There may be a transient elevation
in T1 at the highest dose of RT at week 13, but this
elevation was not significant. Because of the apparent
difference between WM exposed to less than 20Gy and
WM exposed to more than 20Gy, a 20Gy cutpoint is
used in some analyses to follow. Individual doseresponse curves for mean WM T1 at each time interval
are plotted (see Fig 2B). There is a generalized downward trend in T1 at follow-up intervals of longer than
6 months (filled symbols). There is also a suggestion
that T1 is elevated near the tumor at the end of RT
(week 13), and that T1 immediately adjacent to the
tumor is high at most time points compared with
WM, which received 40 to 50Gy (see Fig 2B).
Mean GM T1 at each dose level is also plotted (Fig
3A). This data suggests that GM exposed to more than
Steen et al: Radiation Effects on Human Brain
Fig 1. Illustration of analysis of brain T1 as a function of radiation dose. (A) Segmented T1 map, showing white matter (WM) in
dark green and cortical gray matter (GM) in yellow. Tissue that is a partial volume of WM and GM is shown in pale green, and
such tissue was not analyzed. (B) A “cored” T1 map, showing how central structures are erased before further analysis. (C) Radiation isodose lines shown superimposed on a computed tomography image at the same slice level as the T1 map. (D) Isodose lines
have been transferred to the T1 map, and all tissue which received ⱕ5Gy or ⬎0Gy has been erased. The remaining tissue is WM
(dark green) or GM (yellow) that received between 5 to 10Gy, from which average T1 values were calculated.
Annals of Neurology
Vol 50
No 6
December 2001
RT was always negative, meaning that WM T1 declines after RT. WM exposed to less than 20Gy
showed no significant change over time, but WM exposed to more than 20Gy showed a significant decrease
in T1 with time. The significance level was relatively
low for RT under 30Gy ( p ⫽ 0.05), but significance
was high for dose levels of more than 30Gy ( p ⱕ
0.01). These results show that radiation over 20Gy has
a significant effect on WM T1 (see Fig 2).
Changes in mean GM T1 over time were also analyzed as a function of RT, controlling for patient age
(Table 3). This analysis shows no significant change in
Fig 3. Gray matter T1 and dose-response curve. (A) Mean T1
of GM as a function of radiation dose in 33 patients. Hollow
symbols are radiation dose levels that are significantly different
from the filled symbols at every time point. 〫, RT ⬍5Gy; ‚,
RT 5–10Gy; ƒ, RT 10 –20Gy; , RT 20 –30Gy; €, RT
30 – 40Gy; 䊐, RT 40 –50Gy; ⽧, RT 50 –54Gy; Œ, RT
⬎54Gy. (B) Mean GM T1 as a function of RT dose at various time points after RT. 〫, week 0; ‚, week 3; ƒ, week 5;
“, week 13; ⽧, week 26; Œ, week 40; ‹, week 53.
Fig 2. White matter T1 and dose-response curves. (A) Mean
T1 of WM as a function of radiation dose in 33 patients.
Hollow symbols are radiation dose levels that are not associated with a significant change in T1 over time, whereas filled
symbols show a progressive decline in T1 with time. 〫, RT
⬍5Gy; ‚, RT 5–10Gy; ƒ, RT 10 –20Gy; Š, RT 20 –30Gy;
⽧, RT 30 – 40Gy; Œ, RT 40 –50Gy; , RT 50 –54Gy; ‹,
RT ⬎54Gy. (B) Mean WM T1 as a function of radiation
therapy (RT) dose at various time points after RT. 〫, week
0; ‚, week 3; ƒ, week 5; “, week 13; ⽧, week 26; Œ,
week 40; ‹, week 53.
50Gy has a lower T1 at most time points, compared
with GM exposed to less than 50Gy, so a 50Gy cutpoint is used in some analyses. Individual doseresponse curves for mean GM T1 at each time interval
are also plotted (see Fig 3B). There is again a downward trend in T1 except at week 26, when T1 is generally elevated.
Mean WM T1 was analyzed as a function of RT
dose, controlling for patient age (Table 2). The slope
of the dose-response relationship between WM T1 and
Steen et al: Radiation Effects on Human Brain
Table 2. White Matter T1 as a Function of Radiation Dosea
Radiation Dose (Gy)
Intercept (⫹SD)
Slope (⫹SD)
713 ⫾ 9
707 ⫾ 10
704 ⫾ 9
698 ⫾ 9
703 ⫾ 10
699 ⫾ 12
705 ⫾ 14
738 ⫾ 25
⫺0.06 (⫾0.03)
⫺0.04 (⫾0.03)
⫺0.04 (⫾0.03)
⫺0.08 (⫾0.03)
⫺0.13 (⫾0.03)
⫺0.14 (⫾0.03)
⫺0.19 (⫾0.03)
⫺0.17 (⫾0.05)
5 to ⬍10c
10 to ⬍20c
20 to ⬍30
30 to ⬍40
40 to ⬍50
50 to ⬍54
54 to 59
Change in white matter (WM) T1 as a function of RT dose according to a random coefficient mixed model using all WM T1 data.
Tests for the significance of the change in T1 over time, at each of the various radiation dose levels; a low p-value indicates that there is a
significant effect of radiation.
No significant radiation-related change in WM T1 was seen at a dose level of ⬍20 Gy.
NS ⫽ not significant; WM ⫽ white matter.
GM T1 as a function of RT, suggesting that GM is
not sensitive to radiation ⬍60Gy within 1 year. However, the plot of GM T1 over time (Fig 3) suggests that
T1 was low before RT and remained low over the
course of follow-up evaluation.
Tissue exposed to high doses of RT is different from
tissue exposed to lower RT doses (Table 4). In WM,
this trend did not reach significance until 6 months
after RT, whereas a significant T1 decrement was seen
in GM exposed to high-dose radiation before radiation
was given. Because GM close to tumor would receive a
higher RT dose by prescription, Table 4 implies that
tumor proximity has an effect on GM T1 at the earliest time interval.
We sought to determine whether therapeutic radiation
has a measurable effect on brain T1 in children, at
dose levels lower than what is known to cause cMRIvisible brain changes. Because T1 is a property of the
water molecule, with a strong effect on the ability to
visualize brain structures by cMRI, qMRI is sensitive
to changes below the resolution of cMRI methods.22,26
The results reported here confirm that T1 mapping is
sensitive to radiation-related changes in human brain.
In WM, a radiation dose greater than 20Gy was associated with a decrease in T1 (see Table 2), which became significant 6 months after treatment (see Table
4). In GM, there was no significant change in T1 over
time as a function of RT less than 60Gy (see Table 3),
although T1 was lower in GM close to the tumor even
before RT (see Table 4). This may reveal an effect of
tumor on adjacent GM. Because no significant changes
in GM T1 were seen even at the highest RT doses (see
Table 3), GM appears to tolerate the effects of therapeutic RT.
Radiation damage to the brain is common among
patients receiving RT for brain tumor. Up to 37% of
patients show evidence of radiation injury by 10 years
after treatment, and 70% of patients who develop radiation damage do so within 2 years.33 Past cMRI
studies suggest that the primary effect of RT on human
brain is an increase in signal of WM, consistent with
edema.12–16 Edema would produce an increase in measured T1 of WM,16 as water itself has a long T1
time.25 Conversely, we report that RT is associated
Table 3. Gray Matter T1 as a Function of Radiation Dosea
Radiation Doseb (Gy)
5 to ⬍10
10 to ⬍20
20 to ⬍30
30 to ⬍40
40 to ⬍50
50 to ⬍54
54 to 59
Intercept (SD)
Slope (SD)
1095 (⫾ 17)
1088 (⫾ 16)
1090 (⫾ 16)
1089 (⫾ 18)
1067 (⫾ 23)
1053 (⫾ 21)
1014 (⫾ 24)
1008 (⫾ 20)
⫺0.09 (⫾0.06)
⫺0.10 (⫾0.06)
⫺0.10 (⫾0.06)
⫺0.05 (⫾0.07)
⫺0.03 (⫾0.08)
⫺0.01 (⫾0.06)
⫺0.00 (⫾0.10)
⫺0.15 (⫾0.06)
Change in gray matter T1 as a function of RT dose according to a random coefficient mixed model using all GM T1 data.
No significant radiation-related change in gray matter T1 was seen at any dose level.
Tests for the significance of the change in T1 over time, at each of the various radiation dose levels; a low p-value indicates that there is a
significant effect of radiation.
GM ⫽ gray matter; NS ⫽ not significant.
Annals of Neurology
Vol 50
No 6
December 2001
Table 4. White Matter and Gray Matter T1 as a Function of RT Dose Categorya
⫾ SD
White Matter T1
Gray Matter T1
Pre-RT T1
710 ⫾ 13
705 ⫾ 13
1093 ⫾ 20
1059 ⫾ 21
Week 3
701 ⫾ 8
694 ⫾ 8
1077 ⫾ 20
1052 ⫾ 21
Week 5
713 ⫾ 9
704 ⫾ 9
1103 ⫾ 17
1064 ⫾ 18
Week 13
709 ⫾ 9
705 ⫾ 9
1067 ⫾ 15
1028 ⫾ 16
Week 26
708 ⫾ 9
689 ⫾ 8
1081 ⫾ 19
1057 ⫾ 19
Week 40
694 ⫾ 13
674 ⫾ 13
1073 ⫾ 17
1037 ⫾ 18
Week 53
702 ⫾ 13
680 ⫾ 13
1043 ⫾ 19
1020 ⫾ 21
Comparison of white matter T1 and gray matter T1 as a function of RT dose category at various times after treatment.
Results from 1 patient at week 26 were excluded because this patient was a dramatic outlier, with T1 ⬎1,450msec at 30, 40, 50, and
ⱖ 54Gy.
with a progressive decrease in T1 of WM, which appears to contradict these earlier findings. However, the
segmentation method used to identify WM and GM
would have classified edematous WM as a partial volume of WM and GM.29 Pediatric radiologists who
read the cMRI films did not note gross edema in any
patient at the slice level analyzed. Because edema visible by cMRI typically occurs at longer time intervals
and after higher RT doses that we report, or occurs in
the context of tumor recurrence, we believe that our
results do not contradict earlier reports.12–16
Our results suggest that WM is more vulnerable
than GM to radiation damage. Histological studies
have reported extensive radiation-related damage to
WM, but not to GM,5 and radiation necrosis is largely
restricted to WM, even in primates subjected to much
higher (and biologically more damaging) single-fraction
doses than are used clinically.4 Recent evidence suggests that acute radiation-induced dementia is associated with diffuse demyelination, astrocytic gliosis, and
necrosis of WM, without apparent damage to GM.34
Both x-ray tomography16,35 and cMRI6,21 demonstrate
that atrophy of WM is a common occurrence after
high-dose RT. In children irradiated for medulloblastoma, RT was associated with a decrease in WM volume, but there was no significant change in GM volume.29 The volume of normal-appearing WM
decreased after 23.4 to 36.0 Gy, and patients with
WM loss showed cognitive impairment.36
The most striking finding we report is that WM exposed to doses greater than 20Gy differs from WM
exposed to less than 20Gy by week 26 (see Table 4).
This finding is not explained by edema, as edema
would elevate T1 rather than reducing it. The RTrelated reduction in T1 is marginally significant at less
than 30Gy (see Table 2), but T1 reduction is compelling at RT doses of more than 30Gy. However, there is
a generalized increase in WM T1 near tumor at every
time interval except week 5 (see Fig 2B). This suggests
that the BBB in WM may open after RT, or that tumor has an impact on integrity of the BBB. Both of
these tentative conclusions are consistent with an extensive history of research.8,10,33,37,38 In any case, we
do not have direct evidence that T1 change is indicative of pathology, nor do we yet know whether the T1
changes we describe are irreversible. However, the fact
that T1 changes in a dose-dependent manner in WM
argues that we are characterizing a biological response
of the human brain at a radiation dose substantially
lower than has been reported previously.
We have hypothesized that T1 is a measure of tissue
density, with reduced T1 associated with increased tissue density (or decreased tissue water) in WM and
GM.18 Thus RT, which is known to induce gliosis in
WM,5,6 could produce a dose-related decrease in T1
because of a dose-related increase in gliosis. In GM, the
pretreatment trend in T1 could be due to compression
of cortical parenchyma near a growing tumor mass, or
to tumor invasion directly into cortical parenchyma.18
We believe that qMRI may offer sufficient sensitivity
to detect pathological changes in cell density that could
potentially result from RT. Quantitative computed tomography (CT) is able to detect a 0.2% change in tissue density, which would correspond to a loss in cell
density of about 2% in WM.39 Such a small change in
cell density cannot be detected by standard histological
methods and would require quantitative tissue analysis.
If qMRI is comparable to quantitative CT for characterizing radiation damage in the brain, then qMRI may
offer sensitivity comparable to quantitative analysis of
biopsy material. Since biopsy of normal brain is undesirable, qMRI may be the most sensitive method available to assess brain response to RT, without exposing
the patient to additional radiation.18
Steen et al: Radiation Effects on Human Brain
Radiation-induced changes in WM are usually not
seen by cMRI at RT doses of less than 50Gy in
adults.3 Yet evidence has grown that RT doses of less
than 45Gy can damage the brain, especially in children. The most sensitive indication of RT damage to
brain comes from studies of cognitive impairment (CI)
among patients who received low-dose RT. Many neurological, neurocognitive, or behavioral effects are observed at RT doses of below 45Gy,12,13,19,29,35–36,40 –52
especially in children. There is a dose–response relationship between radiation and CI, as RT doses of less
than 18Gy apparently do not produce CI,52 but higher
RT doses are associated with CI. In young children
exposed to only 18 to 25Gy, average intelligence quotient (IQ) is greater than 10 points below normal,19,44,47– 49 and there is a significant IQ difference
between children who received more than 25Gy and
those who received approximately 35Gy.44,47,48 However, patients who receive RT often receive concurrent
chemotherapy, so it is usually not possible to attribute
CI to RT unequivocally.3
Our results demonstrate that WM exposed to less
than 20Gy radiation does not undergo a significant
change in T1, even at 6 to 12 months after radiation
exposure. We also find no evidence that T1 changes in
GM in a dose-dependent manner at less than 60Gy. If
qMRI is sensitive to pathological brain changes, our
results suggest that there should be no damage to WM
exposed to less than 20Gy and no damage to GM exposed to less than 60Gy. Because conformal RT can be
used to minimize the volume of brain exposed to more
than 20Gy RT,18 our results suggest that conformal
RT may offer a substantial benefit to brain tumor patients. Our results also imply that T1 may provide an
objective surrogate marker for the sequelae of RT.
This work was made possible by support from the American Cancer
Society (RPG CDE-98803 to T.M.), the National Heart, Lung, and
Blood Institute (RO1 HL60022 to R.G.S.), the National Cancer
Institute/Cancer Center Support Grant (P30-CA21765), and the
American Lebanese Syrian Associated Charities.
We thank Mary Fitchpatrick, Mary Freeman, Crystal Manchester,
and Mark Summers, who performed all the MRI studies, as well as
the patients and families of St Jude, who dedicated their time and
1. Shrieve D, Tofilon PJ, Coleman CN, et al. Radiation biology.
In: Louis DN, Posner JB, eds. Brain tumor progress review
group. Bethesda, MD: National Cancer Institute, 2000. http://
2. DeAngelis LM. Brain tumors. N Engl J Med 2001;344:
114 –123.
3. Schultheiss TE, Kun LE, Ang KK, Stephens LC. Radiation response of the central nervous system. Int J Radiat Oncol Biol
Phys 1995;31:1093–1112.
Annals of Neurology
Vol 50
No 6
December 2001
4. Caveness WF. Pathology of radiation damage to the normal
brain of the monkey. Natl Cancer Inst Monog 1977;46:57–76.
5. Burger PC, Mahaley MS, Dudka L, Vogel FS. The morphologic effects of radiation administered therapeutically for intracranial gliomas: a postmortem study of 25 cases. Cancer 1979;
44:1256 –1272.
6. Asai A, Matsutani M, Kohno T, et al. Subacute brain atrophy
after radiation therapy for malignant brain tumor. Cancer
7. Peterson K, Rosenblum MK, Powers JM, et al. Effect of brain
irradiation on demyelinating lesions. Neurology 1993;43:
8. Rubin P, Gash DM, Hansen JT, et al. Disruption of the blood–brain barrier as the primary effect of CNS irradiation. Radiother Oncol 1994;31:51– 60.
9. Pena LA, Fuks Z, Kolesnick RN. Radiation-induced apoptosis
of endothelial cells in the murine central nervous system: protection by fibroblast growth factor and sphingomyelinase deficiency. Cancer Res 2000;60:321–327.
10. O’Connor MM, Mayberg MR. Effects of radiation on cerebral
vasculature: a review. Neurosurgery 2000;46:138 –151.
11. Emami B, Lyman J, Brown A, et al. Tolerance of normal tissue
to therapeutic irradiation. Int J Radiat Oncol Biol Phys 1991;
21:109 –122.
12. Packer RJ, Zimmerman RA, Bilaniuk LT. Magnetic resonance
imaging in the evaluation of treatment-related central nervous
system damage. Cancer 1986;58:635– 640.
13. Curnes JT, Laster DW, Ball MR, et al. Magnetic resonance
imaging of radiation injury to the brain. AJR 1986;7:389 –394.
14. Curran WJ, Hecht-Leavitt C, Shut L, et al. Magnetic resonance
imaging of cranial radiation lesions. Int J Radiat Oncol Biol
Phys 1987;13:1093–1098.
15. Tsuruda JS, Kortman KE, Bradley WG, et al. Radiation effects
on cerebral white matter: MR evaluation. Am J Neuroradiol
1987;8:431– 437.
16. Constine LS, Konski A, Ekholm S, et al. Adverse effects of
brain irradiation correlated with MR and CT imaging. Int J
Radiat Oncol Biol Phys 1988;15:319 –330.
17. Boesiger P, Greiner R, Schoepflin RE, et al. Tissue characterization of brain tumors during and after pion radiation therapy.
Magn Reson Imaging 1990;8:491– 497.
18. Steen RG, Koury M, Granja CI, et al. Effect of ionizing radiation on the human brain: white matter and gray matter T1 in
pediatric brain tumor patients treated with conformal radiation
therapy. Int J Radiat Oncol Biol Phys 2001;49:79 –91.
19. Kramer JH, Norman D, Brant-Zawadzki M, et al. Absence of
white matter changes on magnetic resonance imaging in children treated with CNS prophylaxis therapy for leukemia. Cancer 1988;61:928 –930.
20. Wilson DA, Nitschke R, Bowman ME, et al. Transient white
matter changes on MR images in children undergoing chemotherapy for acute lymphoblastic leukemia: correlation with neuropsychologic deficiencies. Radiology 1991;180:205–209.
21. Ball WS, Prenger EC, Ballard ET. Neurotoxicity of radio/
chemotherapy in children: pathologic and MR correlation.
Am J Neuroradiol 1992;13:761–776.
22. Steen RG, Reddick WE, Mulhern R, et al. Quantitative MRI
of the brain in children with sickle cell disease reveals abnormalities unseen by conventional MRI. J Magn Reson Imaging
23. Zhu Y, Stovall J, Butler L, et al. Comparison of two immobilization techniques using portal film and digitally reconstructed
radiographs for pediatric patients with brain tumors. Int J Radiat Oncol Biol Phys 2000;48:1233–1240.
24. Cho S, Jones D, Reddick WE, et al. Establishing norms for
age-related changes in proton T1 of human brain tissue in vivo.
Magn Reson Imaging 1997;15:1133–1143.
25. Steen RG, Gronemeyer SA, Kingsley PB, et al. Precise and accurate measurement of proton T1 in human brain in vivo: validation and preliminary clinical application. J Magn Reson Imaging 1994;4:681– 691.
26. Steen RG, Ogg R, Reddick WE, Kingsley PB. Age-related
changes in the pediatric brain: quantitative magnetic resonance
(qMRI) provides evidence of maturational changes during adolescence. Am J Neuroradiol 1997;18:819 – 828.
27. Reddick WE, Ogg RJ, Steen RG, Taylor JS. Statistical error
mapping for reliable quantitative T1 imaging. J Magn Reson
Imaging 1996;6:244 –249.
28. Kingsley PB, Ogg RJ, Reddick WE, Steen RG. Correction of
errors caused by imperfect inversion pulses in MR imaging
measurement of T1 relaxation times. Magn Reson Imaging
1998;16:1049 –1055.
29. Reddick WE, Mulhern RK, Elkin TD, et al. A hybrid neural
network analysis of subtle brain volume differences in children
surviving brain tumors. Magn Reson Imaging 1998;16:
413– 421.
30. Glass JO, Reddick WE, Goloubeva O, et al. Hybrid artificial
neural network segmentation of precise and accurate inversion
recovery (PAIR) images from normal human brain. Magn Reson Imaging 2000;18:1245–1253.
31. Rutter CM, Elashoff RM. Analysis of longitudinal data: random coefficient regression modelling. Stat Med 1994;13:
32. Little RC, Milliken GA, Stroup WS, Russell DW. SAS system
for mixed models. Cary, NC: SAS Institute, 1996.
33. Langleben DD, Swgall GM. PET in differentiation of recurrent
brain tumor from radiation injury. J Nucl Med 2000;41:
34. Vigliani MC, Duyckaerts C, Hauw JJ, et al. Dementia following treatment of brain tumors with radiotherapy administered
alone or in combination with nitrosourea-based chemotherapy:
a clinical and pathological study. J Neuro-oncol 1999;41:
35. Peylan-Ramu N, Poplack DG, Pizzo PA, et al. Abnormal CT
scans of the brain in asymptomatic children with acute lymphocytic leukemia after prophylactic treatment of the central nervous system with radiation and intrathecal chemotherapy.
N Engl J Med 1978;298:815– 818.
36. Mulhern RK, Reddick WE, Palmer SL, et al. Neurocognitive
deficits in medulloblastoma survivors and white matter loss.
Ann Neurol 1999;46:834 – 841.
37. Sage MR, Wilson AJ. The blood–brain barrier: an important
concept in neuroimaging. Am J Neuroradiol 1994; 15:
601– 622.
38. Shin RK, Stern JW, Janss AJ, et al. Reversible posterior leukoencephalopathy during the treatment of acute lymphoblastic
leukemia. Neurology 2001;56:388 –391.
39. Fike JR, Cann CE, Turowski K, et al. Radiation dose response
of normal brain. Int J Radiat Biol Phys 1988;14:63–70.
40. Meadows AT, Gordon J, Massari DJ, et al. Declines in IQ
scores and cognitive dysfunctions in children with acute lymphocytic leukemia treated with cranial irradiation. Lancet 1981;
41. Kun L, Mulhern RK, Crisco JJ. Quality of life in children
treated for brain tumors: intellectual, emotional, and academic
function. J Neurosurg 1983;58:1– 6.
42. Jannoun L, Bloom HJG. Long-term psychological effects in
children treated for intracranial tumors. Int J Radiat Oncol Biol
Phys 1990;18:747–753.
43. Mulhern RK, Fairclough D, Ochs J. A prospective comparison
of neuropsychological performance of children surviving leukemia who received 18 Gy, 24 Gy, or no cranial irradiation.
J Clin Oncol 1991;9:1348 –1356.
44. Silber JH, Radcliffe J, Peckham V, et al. Whole brain irradiation and decline in intelligence: the influence of dose and age
on IQ score. J Clin Oncol 1992;10:1390 –1396.
45. Archibald YM, Lunn D, Ruttan LA, et al. Cognitive functioning in long-term survivors of high-grade glioma. J Neurosurg
46. Jankovic M, Brouwers P, Valsecchi MG, et al. Association of
1800 cGy cranial irradiation with intellectual function in children with acute lymphoblastic leukemia. Lancet 1994;344:
224 –227.
47. Mulhern RK, Kepner JL, Thomas PR, et al. Neuropsychological functioning of survivors of childhood medulloblastoma randomized to receive conventional or reduced-dose CSI: a Pediatric Oncology Group study. J Clin Oncol 1998;16:
48. Grill J, Renaux VK, Bulteau C, et al. Long-term intellectual
outcome in children with posterior fossa tumors according to
radiation doses and volumes. Int J Radiat Oncol Biol Phys
49. Kieffer-Renaux V, Bulteau C, Grill J, et al. Patterns of neuropsychological deficits in children with medulloblastoma according to craniospinal irradiation doses. Dev Med Child Neurol
50. Peper M, Steinvorth S, Schraube P, et al. Neurobehavioral toxicity of total body irradiation: a follow-up in long-term survivors. Int J Radiat Oncol Biol Phys 2000;46:303–311.
51. Mulhern RK, Palmer SL, Reddick WE, et al. Risks of young
age for selected neurocognitive deficits in medulloblastoma are
associated with white matter loss. J Clin Oncol 2001;19:
472– 479.
52. Mulhern RK, Wasserman AL, Fairclough D, Ochs J. Memory
function in disease-free survivors of childhood acute lymphocytic leukemia given CNS prophylaxis with or without 1,800
cGy cranial irradiation. J Clin Oncol 1988;6:315–320.
Steen et al: Radiation Effects on Human Brain
Без категории
Размер файла
761 Кб
effect, radiation, ionizing, therapeutic, brain, human
Пожаловаться на содержимое документа