1638 / 1708
pts). Cerebrospinal fluid shunt: EIQ was higher at baseline
(12.58 pts) in patients who did not have a CSF shunt
(
P
Z
.0478), and those without CSF shunts had a lower rate
of decline in math (
P
Z
.0025, 4.79 pts/y) and reading scores
(
P
Z
.0319, 2.32 pts/y). Extent of resection: Baseline math
scores were higher in patients who underwent
<
GTR
(
P
Z
.0091, 9.97 pts). Gross total resection was associated
with a slower rate of decline in reading scores (
P
Z
.0269,
2.25 pts/y) than in those who underwent
<
GTR. Age at RT:
With the exception of math and reading scores, age (time of
diagnosis or irradiation) had a highly significant impact on
the rate of decline in all test scores (EIQ,
P
Z
.0141; Math,
P
Z
.1832; Reading,
P
Z
.0688; Spelling,
P
Z
.0424).
Impact of mean radiation dose on longitudinal
trends in cognitive scores
The longitudinal trends in cognitive scores were modeled
by time since irradiation and mean dose. Increasing mean
dose to all volumes had a statistically significant negative
impact on EIQ. Increasing mean dose to all normal tissue
volumes except for the infratentorial brain and hippocampi
had a statistically significant negative impact on math
scores. Increasing mean dose to the right temporal lobe had
a statistically significant negative impact on reading scores.
The impact of increasing mean dose to the right hippo-
campus was borderline significant. When age was included,
Table 3
Effect of increasing mean dose on cognitive test scores by brain volume
Normal tissue volume
Estimated IQ
WIAT math
WIAT reading
WIAT spelling
þ
þ
þ
þ
Entire brain
.0121
<
.0001
.0096
.0007
n.s.
<
.0001
n.s.
.0002
Supratentorial brain
.0161
<
.0001
.0251
.0009
n.s.
<
.0001
n.s.
.0002
Temporal lobe (left)
.0032
<
.0001
.0184
.0013
n.s.
<
.0001
n.s.
.0002
Temporal lobe (right)
.0005
<
.0001
.0053
.0009
.0184
<
.0001
n.s.
<
.0001
Hippocampus (left)
.0751
<
.0001
n.s.
.0025
n.s.
<
.0001
n.s.
.0001
Hippocampus (right)
.0130
<
.0001
n.s.
.0016
n.s.
<
.0001
n.s.
<
.0001
Infratentorial brain
.0002
<
.0001
n.s.
.0034
n.s.
<
.0001
n.s.
.0001
Abbreviation:
n.s.
Z
not significant. Other abbreviations as in
Table 2 .P
values are grouped in columns according to the inclusion (
þ
) or exclusion ( ) of age in the model.
70
60
50
40
30
20
10
0
3 yrs 5 yrs 8 yrs 10 yrs 12 yrs 15 yrs
3 yrs 5 yrs 8 yrs 10 yrs 12 yrs 15 yrs
70
60
50
40
30
20
10
0
Intratent
Hippo L
Hippo R
Temporal L
Temporal R
Brain
Supratent
Intratent
Hippo L
Hippo R
Temporal L
Temporal R
Brain
Supratent
70
60
50
40
30
20
10
0
3 yrs 5 yrs 8 yrs 10 yrs 12 yrs 15 yrs
Intratent
Hippo L
Hippo R
Temporal L
Temporal R
Brain
Supratent
Intratent
Hippo L
Hippo R
Temporal L
Temporal R
Brain
Supratent
3 yrs 5 yrs 8 yrs 10 yrs 12 yrs 15 yrs
70
60
50
40
30
20
10
0
Reading
Spelling
Math
Estimated IQ
Dose (Gy)
Dose (Gy)
Dose (Gy)
Dose (Gy)
Fig. 1.
Estimated iso-effect curves of mean dose by brain volume and age at the time of irradiation representing a
probability of below-average IQ or academic achievement 5 years after treatment. Each graph represents a different cognitive
test, and each curve represents a different normal tissue volume. Missing estimates indicate that the model calculated a dose
that was outside the range of dose used to generate the models. Brain
Z
entire brain volume; EIQ
Z
estimated intelligence
quotient; Hippo L
Z
left hippocampus; Hippo R
Z
right hippocampus; Infratent
Z
infratentorial brain; Math
Z
Wechsler
Individual Achievement Test (WIAT) math scores; Reading
Z
WIAT reading scores; Spelling
Z
WIAT spelling scores;
Supratent
Z
supratentorial brain; Temporal L
Z
left temporal lobe; Temporal R
Z
right temporal lobe.
Volume 90 Number 3 2014
Radiation effects in medulloblastoma
557