obp.20
CL
AIL
DES
NGEL
TOM
nd De
Yorkh
hool o
nd, Un
likely
hy ra
ther w
ordia
iven d
tionat
failure occurs immediately after treatment in 97.5% of patients) decreases with increasing age at treatment. ESD
As
can
of
po
ov
rad
oo
de
cu
ye
est
Th
of
po
ma
me
wo
ma
R
F.R
pit
Int. J. Radiation Oncology Biol. Phys., Vol. 62, No. 3, pp. 738–744, 2005
Copyright © 2005 Elsevier Inc.at birth is 20.3 Gy; at 10 years 18.4 Gy, at 20 years 16.5 Gy, and at 30 years 14.3 Gy. We have calculated 95%
confidence limits for age at premature ovarian failure for estimated radiation doses to the ovary from 1 Gy to
the ESD from birth to 50 years.
Conclusions: We report the first model to reliably predict the age of ovarian failure after treatment with a known
dose of radiotherapy. Clinical application of this model will enable physicians to counsel women on their
reproductive potential following successful treatment. © 2005 Elsevier Inc.
Ovarian failure, Radiotherapy, Fertility, Oocytes.
INTRODUCTION
survival rates for children and adolescents treated for
cer continue to improve, a population of young women
reproductive age emerges for whom issues of fertile
tential are paramount. Impaired fecundity and premature
arian failure are recognized potential late sequelae of
iotherapy to the ovaries.
The human ovary contains a fixed pool of primordial
cytes, maximal at 5 months of gestational age, which
clines with increasing age in a biexponential fashion,
lminating in the menopause at an average age of 50–51
ars. For any given age, the size of the oocyte pool can be
imated based on a mathematical model of decline (1, 2).
e rate of oocyte decline represents an instantaneous rate
temporal change determined by the remaining population
ol, which increases around age 37 years when approxi-
tely 25,000 primordial oocytes remain, and precedes the
nopause by 12–14 years (3). Reproductive aging in
men is due to ovarian oocyte depletion with approxi-
tely 1,000 oocytes remaining at the menopause (4). As-
sessment of ovarian reserve and reproductive age in healthy
women remains a challenge, but the recent application of
our solution to the Faddy-Gosden model of ovarian primor-
dial oocyte decline may allow an accurate assessment of
ovarian reserve by measurement of ovarian volume (5).
Radiotherapy may be used either alone or in combination
with surgery and chemotherapy to provide local disease
control for solid tumors. Because of its established late
sequelae on immature and developing tissues, irradiation is
used cautiously, especially in children and adolescents. To-
tal body, craniospinal axis, whole abdominal, or pelvic
irradiation potentially expose the ovaries to irradiation, and
may cause premature ovarian failure. The degree of im-
pairment is related to the volume treated, total radiation
dose, fractionation schedule, and age at time of treatment
(2, 6 – 8). The number of primordial oocytes present at the
time of treatment, together with the biologic dose of
radiotherapy received by the ovaries, will determine the
fertile “window” and influence the age at premature
ovarian failure.
eprint requests to: W. Hamish B. Wallace, M.D., 1LW, Scotland, UK; Tel: (
44) 131-536-0420; Fax: (
44) 131-doi:10.1016/j.ijr
INICAL INVESTIGATION
PREDICTING AGE OF OVARIAN F
FIELD THAT INCLU
W. HAMISH B. WALLACE, M.D., F.R.C.P.C.H.,* A
FRANK SARAN, M.D., F.R.C.R.,‡ AND
*Section of Child Life and Health, Department of Reproductive a
Scotland, United Kingdom; †Royal Hospital for Sick Children,
NHS Trust Foundation, Sutton, United Kingdom; §Sc
St. Andrews, Scotla
Purpose: To predict the age at which ovarian failure is
the ovary in women treated for cancer.
Methods and Materials: Modern computed tomograp
effective dose of radiation received by the ovaries. Toge
the human oocyte, the effective surviving fraction of prim
failure, with 95% confidence limits, predicted for any g
Results: The effective sterilizing dose (ESD: dose of frac.C.P.C.H., Department of Haematology/Oncology, Royal Hos-
al for Sick Children, 17 Millerfield Place, Edinburgh, EH9
536
R
73804.11.038
Ovary
URE AFTER RADIATION TO A
THE OVARIES
A B. THOMSON, B.SC., M.R.C.P.C.H., M.D.,†
W. KELSEY, B.SC., M.SC., PH.D.§
velopmental Sciences, University of Edinburgh, Edinburgh,
ill, Glasgow, Scotland, United Kingdom; ‡Royal Marsden
f Computer Science, University of St. Andrews,
ited Kingdom
to develop after radiation to a field that includes
diotherapy planning allows determination of the
ith our recent assessment of the radiosensitivity of
l oocytes can be determined and the age of ovarian
ose of radiotherapy.
ed radiotherapy [Gy] at which premature ovarian
Printed in the USA. All rights reserved
0360-3016/05/$–see front matter-0430; E-mail: Hamish.Wallace@ed.ac.uk
eceived Oct 20, 2004. Accepted for publication Nov 29, 2004.

pri
po
fol
sit
thi
fra
rad
int
em
Ra
F
hea
me
stu
res
wh
val
bir
W
mo
bie
fro
tha
the
mo
ooc
ova
sur
wh
So
Gy
bou
Pr
G
viv
fai
chr
is 5
cas
sur
pop
Th
and
giv
ing
age
exa
in
W
ave
the
ear
tha
the
(1.
me
(1.
exa
Ca
U
den
yea
est
the
at
pop
me
dos
fal
fai
rep
im
Ca
the
U
Fig
aft
dos
312
3.0
Th
the
yea
age
exp
sam
tre
Av
pau
739Predicting age of ovarian failure after radiation ● W. H. B. WALLACE et al.Assessing the extent of radiation-induced damage of the
mordial oocytes and predicting the impact on fertile
tential has been challenging. An understanding of ovarian
licle dynamics has allowed us to determine the radiosen-
ivity of the human oocyte to be 2 Gy (2). Application of
s estimate has made it possible to determine the surviving
ction of the primordial oocyte pool for a given dose of
iotherapy and therefore predict the age (with confidence
ervals) of premature ovarian failure by applying a math-
atical model of decay.
METHODS AND MATERIALS
diosensitivity of the human oocyte
addy and Gosden provide a model for natural follicle decline in
lthy women (1). The model is obtained by incorporating age at
nopause data into a least squares analysis of four histologic
dies, (3, 9–11) and a large study of ages at menopause (12). The
ult is the differential equation:
dy ⁄ day x
y


0.0595 3, 716 ⁄ (11, 780 y)

ere x denotes age, y(x) is population at age x, and with initial
ue y (0)  701,200. The initial value denotes population at
th.
e consider this to be the best model currently available. Early
dels were simple exponential (3, 6), and were superseded by a
xponential model (4), which included accelerated depletion
m age 37 years. The Faddy-Gosden model is the only model
t takes into account all available histologic data and combines
se with known ranges for the age of menopause. We solved this
del to revise our estimate of the radiosensitivity of the human
yte, based on data obtained from young women who developed
rian failure after total body irradiation (2). We obtained the
viving percentage function for a given dose:
log10 g(z)
2 0.15z
ich decreases exponentially from 100% at zero dosage (z  0).
lving g(z)  50 gives an LD50 for the human oocyte of about 2
. It should be noted that this estimate of the LD50 is an upper
nd and all modeling is based on the average case.
edicting ovarian failure
iven the solution to the Faddy-Gosden equation and the sur-
ing percentage function, we can predict the age of ovarian
lure for a patient given a known dose (z Gy) at a known
onologic age (xchron years). The average age of ovarian failure
0.4 years (12). Assuming that the patient is close to the average
e, her oocyte population before treatment will be y(xchron). The
viving fraction after treatment is g(z) / 100. Thus, the oocyte
ulation after treatment is the product of y(xchron) and g(z) / 100.
is new population corresponds to an age, xreprod, between xchron
50.4 years; xreprod is the age at which the Faddy-Gosden model
es the posttreatment population. The patient’s predicted remain-
reproductive life span is simply 50.4  xreprod. Her predicted
of ovarian failure is xchron
(50.4  xreprod) years. An
mple calculation with xchron  12 years and z 10 Gy is given
Fig. 1a. moe next consider a patient having a larger (or smaller) than
rage follicle population at age of treatment. This corresponds to
case that the patient is likely to have had a later (respectively
lier) than average menopause. In quantitative terms, this means
t the patient was toward the upper (respectively lower) limit of
95% confidence interval for age of menopause: 50.4 years 
96  3.9) years (12). The confidence interval used for age of
nopause after treatment is the predicted age of ovarian failure
96  3.9) years. This is illustrated in Fig. 1b, again using the
mple where xchron  12 years and z  10 Gy.
lculating the effective sterilizing dose
sing this model the age of ovarian failure, with 95% confi-
ce intervals, can be predicted given any age from birth to 50
rs and any practical dose. The predicted age is based on an
imate of the surviving follicle pool after treatment. We calculate
effective sterilizing dose (ESD) by considering those patients
the upper limit of the 95% confidence interval for follicle
ulation: equivalently, those patients expected to experience
nopause at 58 years (50.4 plus 1.96 SD years). The ESD is the
e after which the patients’ primordial oocyte population will
l below 1,000. Such a dose will induce immediate ovarian
lure in 97.5% of the female population, effectively taking their
roductive age to 58 years. The mean sterilizing dose will induce
mediate premature ovarian failure in 50% of women.
lculating the dose of radiotherapy received by
ovary
sing modern radiotherapy planning methods (computed to-
. 1. (a) An example of the method for predicting ovarian failure
er treatment. The chronologic age, xchron, is 12 years, and the
e, z, is 10 Gy. The average oocyte population at xchron is
,000. The surviving percentage after 10 Gy is log10 g (10) 
1%, corresponding to a population after treatment of 9,600.
is new population is the average population at age 42.9 years,
reproductive age, xreprod, after treatment. The average 42.9-
r-old patient is expected to have ovarian failure in 7.5 years, at
50.4 years. Hence, the 12-year-old patient is expected to
erience ovarian failure in 7.5 years, at age 19.5 years. (b) The
e example, with confidence intervals included. Average un-
ated menopause occurs at 50.4 years  (1.96  3.9) years.
erage menopause after treatment is the predicted age of meno-
se (19.5 years)  (1.96  3.9) years.graphy [CT] planning with or without image fusion) the