Human Reproduction Update, Vol.9, No.6 pp. 505±514, 2003 DOI: 10.

1093/humupd/dmg044

Ultrasound assessment of the polycystic ovary:

international consensus de®nitions

Adam H.Balen1,5, Joop S.E.Laven2, Seang-Lin Tan3 and Didier Dewailly4

1
Department of Reproductive Medicine, Leeds General In®rmary, Leeds, UK, 2Department of Reproductive Medicine, Erasmus

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

Medical Centre, Rotterdam, The Netherlands, 3Department of Obstetrics and Gynecology, McGill University Hospital, Montreal,
Canada and 4Department of Endocrine Gynecology and Reproductive Medicine, Lille, France
5
To whom correspondence should be addressed at: Department of Reproductive Medicine, Leeds General In®rmary, Leeds, LS2 9NS,
UK. E-mail: [email protected]

The polycystic ovary syndrome (PCOS) is a heterogeneous condition, the pathophysiology of which appears to be
both multifactorial and polygenic. The de®nition of the syndrome has been much debated. Key features include
menstrual cycle disturbance, hyperandrogenism and obesity. There are many extra-ovarian aspects to the
pathophysiology of PCOS, yet ovarian dysfunction is central. At a recent joint ASRM/ESHRE consensus meeting, a
re®ned de®nition of the PCOS was agreed, encompassing a description of the morphology of the polycystic ovary
(PCO). According to the available literature, the criteria ful®lling suf®cient speci®city and sensitivity to de®ne the
PCO should have at least one of the following: either 12 or more follicles measuring 2±9 mm in diameter, or
increased ovarian volume (>10 cm3). If there is a follicle >10 mm in diameter, the scan should be repeated at a time
of ovarian quiescence in order to calculate volume and area. The presence of a single PCO is suf®cient to provide
the diagnosis. The distribution of follicles and a description of the stroma are not required in the diagnosis.
Increased stromal echogenicity and/or stromal volume are speci®c to PCO, but it has been shown that the
measurement of ovarian volume (or area) is a good surrogate for quanti®cation of the stroma in clinical practice. A
woman having PCO in the absence of an ovulation disorder or hyperandrogenism (`asymptomatic PCO') should not
be considered as having PCOS, until more is known about this situation. Three-dimensional and Doppler ultrasound
studies may be useful research tools but are not required in the de®nition of PCO. This review outlines evidence for
the current ultrasound de®nition of the polycystic ovary and technical speci®cations.

Key words: consensus de®nition/PCOS/polycystic ovary/ultrasound

Introduction et al., 1988; Michelmore et al., 1999). Likewise, the biochemical

features associated with PCOS were not consistent in all women
Historically, detection of the polycystic ovary required visualiza- (Pache et al., 1993; Balen et al., 1995). There is considerable
tion of the ovaries at laparotomy and histological con®rmation heterogeneity of symptoms and signs amongst women with PCOS,
following biopsy (Stein and Leventhal, 1935). As further studies and for an individual these may change over time (Balen et al.,
identi®ed the association of certain endocrine abnormalities in 1995; Elting et al., 2000). Thus, consensus on a single biochemical
women with histological evidence of polycystic ovaries, biochem- or clinical de®nition for PCOS was thwarted by the heterogeneity
ical criteria became the mainstay for diagnosis. Raised serum of presentation of the disorder.
levels of LH, testosterone and androstenedione, in association with Presentation of the syndrome is so varied that one, all or any
low or normal levels of FSH, described an endocrine pro®le which combination of the above features may be present in association
many believed to be diagnostic of polycystic ovary syndrome with an ultrasound picture of polycystic ovariesÐthe de®ning
(PCOS) (Franks, 1995). Well-recognized clinical presentations features of PCOS in the UK and much of Europe (Balen, 1999).
included menstrual cycle disturbances (oligo/amenorrhoea), The 1990 National Institute of Health Conference on PCOS,
obesity and hyperandrogenism manifesting as hirsutism, acne or however, recommended that diagnostic criteria should include
androgen-dependent alopecia. These de®nitions proved inconsist- evidence of hyperandrogenism and ovulatory dysfunction, in the
ent, however, as clinical features were noted to vary considerably absence of non-classic adrenal hyperplasia, and that evidence of
between women, and indeed some women with polycystic ovaries polycystic ovarian morphology was not essential (Zawadzki and
do not appear to display any of the common symptoms (Polson Dunaif, 1992; this reference is in book form only and no longer

Human Reproduction Update Vol. 9 No. 6 ã European Society of Human Reproduction and Embryology 2003; all rights reserved 505
A.H.Balen et al.

For many years, wedge resection was the only treatment for
PCOS, and histological assessment of the ovaries was therefore
routine practice. Wedge resection is, however, an outdated
operation, and so histological specimens of polycystic ovaries
are no longer readily available.
The histopathological criteria have been de®ned as the obser-
vation of: atretic follicles and/or degenerating granulosa cells;
hypertrophy and luteinization of the inner theca cell layer; and
thickened ovarian tunica. A good correlation has been shown
between ultrasound diagnoses of polycystic morphology and the
histopathological criteria for polycystic ovaries by studies exam-
ining ovarian tissue obtained at hysterectomy or after wedge

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

Figure 1. Polcystic ovaries (B mode, transvaginal route). In both left and
resection (Saxton et al., 1990; Takahashi et al., 1994). The
right ovaries, the ovarian length and width are increased as well as the literature on correlations between ultrasound and histology is
ovarian area. The follicle number, with a diameter mainly between 2 and sparse, as histological assessment of the ovary became obsolete
5 mm, is more than 12. The distribution within the ovaries is mainly before ultrasound became common practice.
peripheral. The increased and hyperechoic stroma occupies the centre of the The histological data of Hughesdon (Hughesdon, 1982) indi-
ovaries.
cated a 2- to 3-fold increase of follicle number in the polycystic
ovary, from the stage of primary follicles up to tertiary follicles,
and identi®ed the cystic structures as follicles as opposed to
readily available). It has been considered necessary to rede®ne
PCOS and to include within it an appropriate de®nition of the pathological cysts.
polycystic ovary (Balen and Michelmore, 2002; Homburg, 2002).
Transabdominal ultrasound
Herein, the literature on ultrasound characteristics of the
polycystic ovary has been reviewed. A Medline search was In recent years, transabdominal and/or transvaginal ultrasound
performed of all reports of polycystic ovaries and PCOS published have become the most commonly used diagnostic methods for the
since 1970. Papers were included which attempted to correlate identi®cation of polycystic ovaries. Although the ultrasound
features of the PCOS with speci®c quanti®able measurements of criteria for the diagnosis of polycystic ovaries have never been
the ovary in order to best de®ne the polycystic ovary. The present universally agreed, the characteristic features are accepted as
report was ®rst presented at the joint ASRM/ESHRE consensus being an increase in the size (volume) of the ovary due to a greater
meeting on PCOS held in Rotterdam, May 1±3, 2003. At this number of follicles and volume of stroma as compared with
meeting, a re®ned de®nition of the PCOS was agreed (Fauser et al., normal ovaries.
2004) which, for the ®rst time, included a description of the One group (Swanson et al., 1981) were among the ®rst to use
morphology of the polycystic ovary (see Table III and Figure 1). high-resolution real-time ultrasound (static B-scanner, 3.5 MHz,
The new de®nition required the presence of two from the transabdominal) to describe polycystic ovaries. Prior to this, it was
following three criteria: (i) oligo- and/or anovulation; (ii) thought that the tiny cysts/follicles of the polycystic ovary could
hyperandrogenism (clinical and/or biochemical); and (iii) poly- not be detected by ultrasound. The follicles were noted to be
cystic ovaries, with the exclusion of other aetiologies (Fauser et al., 2±6 mm in diameter, but their number was neither recorded nor
2003). de®ned, nor were stromal characteristics described.
These early studies were hampered by the limitations of static
B-scanners, but these were superseded in the early 1980s by high-
Historical overview of polycystic ovary imaging: resolution, real-time sector scanners (Campbell et al., 1982; Orsini
technical breakthroughs et al., 1983). Ultrasound was used to describe the ovarian
appearance in women classi®ed as having PCOS (by symptoms
Appearance and histology of the polycystic ovary and serum endocrinology) rather than to make the diagnosis.
Numerous descriptions have been made of the morphology of the The transabdominal ultrasound criteria of another group
polycystic ovary, and these have been re®ned over time, alongside (Adams et al., 1985) attempted to de®ne a polycystic ovary as
advances in imaging technology. It has been suggested that the one which contains, in one plane, at least 10 follicles (usually
®rst description of the anatomy and pathology of the polycystic between 2 and 8 mm in diameter) arranged peripherally around a
ovary and features of the condition was made during the 18th dense core of ovarian stroma, or scattered throughout an increased
century (Vallisneri, 1721). In more recent times, Stein and amount of stroma. This was a seminal paper which has been most
Leventhal described the features of seven hirsute, amenorrhoeic often quoted in the literature on PCOS.
women based on the characteristic ovarian morphology from The Adams' criteria have been adopted by many subsequent
histological specimens taken at wedge resection of the ovaries. studies which have used ultrasound scanning to detect polycystic
The histology of the polycystic ovary was of an ovary with ovaries (Polson et al., 1988; Conway et al., 1989; Kiddy et al.,
prominent theca, ®brotic thickening of the tunica albuginea and 1990; Fox et al., 1991; Abdel Gadir et al., 1992; Clayton et al.,
multiple cystic follicles (Stein and Leventhal, 1935). The number 1992; Farquhar et al., 1994; Balen et al., 1995). In common with
of antral follicles (2±6 mm in diameter) was described as many authors, one of these groups (Abdel Gadir et al., 1992) found
`excessive' (Goldzieher and Green, 1962), but not quanti®ed. that the visualization of polycystic ovaries supported the diagnosis

506
 Consensus de®nition of the polycystic ovary

of the syndrome in women with signs and symptoms, rather than Table I. Correlation of ultrasound formulaic methods with 3-D ultrasound
being key in making the diagnosis. volume measurements (Nardo et al., 2003)

Transvaginal ultrasound Method Correlation coef®cient

Transabdominal ultrasound has been largely superseded by p/6 tv 3 ap 3 long 0.70

transvaginal scanning because of greater resolution and in many p/6 (tv)3 0.55
p/6 (ap)3 0.61
cases patient preference, as the need for a full bladder is avoidedÐ
p/6 (long)3 0.10
which saves time and may be more comfortable (Goldstein, 1990). p/6 [(tv + ap)/2]3 0.72
Whilst this may be the case in the context of infertility clinics, p/6 [(long + ap)/2]3 0.49
where women are used to having repeated scans, it was found that p/6 [(tv + long)/2]3 0.61
20% of those undergoing routine screening declined a transvaginal p/6 [(tv + ap + long)/3])3 0.73
scan after ®rst having had a transabdominal scan (Farquhar et al., p/6 (tv)2(ap) 0.67

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

1994). p/6 (ap)2(tv) 0.73
p/6 (tv)2(long) 0.61
The transvaginal approach provides a more accurate view of the
p/6 (ap)2(long) 0.51
internal structure of the ovaries, avoiding apparently homogeneous p/6 (long)2(tv) 0.49
ovaries as described with transabdominal scans, particularly in p/6 (long)2(ap) 0.30
obese patients. With the transvaginal route, high-frequency probes
(>6 MHz), which have a better spatial resolution but less
ap = anteroposterior diameter; long = longitudinal diameter; tv = transverse
examination depth, can be used because the ovaries are close to diameter.
the vagina and/or the uterus and because the presence of fatty
tissue is usually less disruptive (except when very abundant).
an ellipse is drawn around the outline of the ovary. The ultrasound
Three-dimensional ultrasound and magnetic resonance imaging software for this calculation appears to be accurate.
The recent innovation of three-dimensional (3-D) ultrasound, as Traditionally, the calculation of ovarian volume has been
well as colour- and pulsed-Doppler ultrasound, may further performed using the formula for a prolate ellipsoid (p/6 3
enhance the detection of polycystic ovaries, and may be more maximal longitudinal, anteroposterior and transverse diameters)
commonly employed in time (Zaidi et al., 1995a; Kyei-Mensah (Sample et al., 1977; Adams et al., 1985; Orsini et al., 1985). As p/
et al., 1996a). Although 3-D ultrasound requires a longer time for 6 = 0.5233, a simpli®ed formula for a prolate ellipse is (0.5 3
storage and data analysis, increased training and more expensive length 3 width 3 thickness) (Swanson et al., 1981; Hann et al.,
equipment, good correlations were found between 2-D and 3-D 1984; Saxton et al., 1990; Pache et al., 1991; Fulghesu et al.,
ultrasound measurements of ovarian volume and polycystic ovary 2001). In practice, this formula is both easy to use and of practical
morphology (Nardo et al., 2003) (Table I). value.
The use of magnetic resonance imaging (MRI) to visualize the A large number of different ultrasound formulae with different
structure of pelvic organs has been claimed to have even greater weightings for the different diameters were used to calculate
sensitivity than ultrasound for the detection of polycystic ovaries ovarian volume, and the prolate spheroid formula (p/6 3
(Mitchell et al., 1986; Faure et al., 1989). However, the substantial anteroposterior diameter2 3 transverse diameter) was found to
cost and practical problems involved with this imaging technique correlate well with ovarian volume as assessed by 3-D ultrasound
may limit its use as an easily accessible diagnostic tool in general (Nardo et al., 2003). A similar correlation was found with the
clinical practice. The early reports of MRI were also made at a spherical volume method {[p/6 3 (transverse diameter +
time when high-resolution transvaginal ultrasound was emerging anteroposterior diameter + longitudinal diameter)/3]}3. However,
as a valuable tool, and time has con®rmed the place of the latter as polycystic ovaries appear to be more spherical than ovoid, it was
and limited further interest in MRI. suggested that the formula should be modi®ed (Nardo et al., 2003).
Three means have been proposed for calculating ovarian area:
1. Using the formula for an ellipse (length 3 width 3 p/4). As p/4
Examination of the polycystic ovary: technical aspects = 0.78, a simpli®ed formula for an ellipse is (0.8 3 length 3
and normative data width).
2. Fitting an ellipse to the ovary, the area of which is calculated by
External features of the polycystic ovary
the ultrasound machine.
Surface area and volume 3. Outlining by hand the ovary with automatic calculation of the
Technical aspects: It is necessary to identify each ovary and outlined area.
measure the maximum diameter in each of three planes (longitu- This last technique is preferred in cases of non-ellipsoid ovaries,
dinal, anteroposterior and transverse). It is recognized that, as may sometimes be observed.
because of the irregular shape of the ovary, any calculation of Normative data: In the ®rst study to assess ovarian volume, the
the volume of a sphere or prolate ellipse is, at best, an estimate. simpli®ed formula for a prolate ellipse was used for the calculation
The left ovary may be more dif®cult to measure because of the and found on average to be 12.5 cm3 (range 6±30 cm3) (Swanson
overlying sigmoid colon, particularly if there is distension with et al., 1981). This formula was also used by others (Hann et al.,
¯atus in the bowel. Modern ultrasound machines can calculate 1984), who reported considerable variety in ultrasound character-
volume once the callipers have been used to measure the ovary and istics in women with PCOS. These authors took the upper limit of

507
A.H.Balen et al.

ovarian volume to be 5.7 cm3 based on data from another group of cysts/follicles. Follicle number should be estimated in two
(Sample et al., 1977). In the latter study, ovarian volume was planes of the ovary in order to estimate their size and their position.
calculated using the more accurate formula for a prolate ellipsoid The diameter of follicles is measured as the mean of three
(0.5233 3 maximal longitudinal, anteroposterior and transverse diameters (longitudinal, transverse and antero-posterior).
diameters). Women with PCOS were compared with normal Normative data: One group (Sample et al., 1977) described the
controls and found to have signi®cantly greater ovarian volume follicles as being <8 mm in size, whilst others (Swanson et al.,
(14.04 6 7.36 versus 7.94 6 2.34 cm3) and smaller uterine 1981) noted the follicles to be 2±6 mm in diameter, though a
volumes. However, no record was made of the timing of the scan prerequisite number was neither recorded nor de®ned. Ovaries
in relation to the menstrual cycle in either the PCOS or control were also described as either being predominantly solid if fewer
subjects. than four small (<9 mm) cystic structures were detected in the
In another report (Adams et al., 1985), polycystic ovaries were ovary, or predominantly cystic if multiple small (neither quanti-
found to have a higher volume (14.6 6 1.1 cm3) than both ®ed) cystic structures or at least one large (>10 mm) cyst were

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

multicystic (8.0 6 0.8 cm3) and normal ovaries (6.4 6 0.4 cm3). present (Orsini et al., 1985). Patients with PCOS usually had
Uterine cross-sectional area was also greater in women with PCOS follicles of between 4 and 10 mm, but occasionally follicles of
than in those with multicystic or normal ovaries (26.0 6 1.4 versus 15 mm were identi®ed, presumably indicative of follicular
13.1 6 0.9 versus 22.4 6 1.0 cm2), which is a re¯ection of the recruitment. A seminal paper (Adams et al., 1985) described the
degree of estrogenization. polycystic ovary as having, in one plane, at least 10 follicles
A large study of 80 oligo-/amenorrhoeic women with PCOS was (usually between 2 and 8 mm in diameter), usually arranged
compared with a control group of 30 using a 6.5 MHz transvaginal peripherallyÐalthough when scattered through the stroma it was
probe (Fulghesu et al., 2001). Based on mean 6 2 SD data from suggested that the follicles were usually 2±4 mm in diameter
the control group, the cut-off values were calculated for ovarian (Adams et al., 1985). Others claimed that the transvaginal
volume (13.21 cm3), ovarian total area (7.00 cm2), ovarian stromal de®nition of a polycystic ovary should require the presence of at
area (1.95 cm2) and stromal/area ratio (0.34). The sensitivity of least 15 follicles (2±10 mm in diameter) in a single plane (Fox
these parameters for the diagnosis of PCOS was 21, 4, 62 and et al., 1991).
100% respectively, suggesting that a stromal/area ratio >0.34 is In a study of 214 women with PCOS (oligo-/amenorrhoea,
diagnostic of PCOS (Fulghesu et al., 2001). Whilst these data may elevated serum LH and/or testosterone, and/or ovarian area
be useful in a research setting, the measurement of ovarian stromal >5.5 cm2) and 112 with normal ovaries, the aim was to determine
area is not easily achieved in routine daily practice. the importance of follicle number per ovary (FNPO) (Jonard et al.,
Thus, the consensus de®nition for a polycystic ovary includes an 2003). These authors performed a 7 MHz transvaginal ultrasound
ovarian volume >10 cm3. It is recognized that not all polycystic scan, and three different categories of follicle size were analysed
ovaries will be enlarged to this size or greater, and that the separately (2±5, 6±9 and 2±9 mm). The size range of follicles has
consensus is based on the synthesis of evidence from many studies been considered important by some authors, with polycystic
which have reported a greater mean ovarian volume for polycystic
ovaries tending to have smaller follicles than normal or multicystic
ovaries combined with a consistent ®nding of a smaller mean
ovaries (Hughesdon, 1982; Pache et al., 1993). The mean FNPO
volume than 10 cm3 for normal ovaries.
was similar between normal and polycystic ovaries in the 6±9 mm
The consensus view is that, until more data are collected and
range, but signi®cantly higher in the polycystic ovaries in both the
validated, the volume of the polycystic ovary should be calculated
2±5 and 2±9 mm ranges. A FNPO of >12 follicles of 2±9 mm gave
using the more widely accepted criterion of a prolate ellipsoid.
the best threshold for the diagnosis of PCOS (sensitivity 75%,
Uterine size and relationship to ovarian size speci®city 99%) (Jonard et al., 2003) (Table II). These authors
suggested that intra-ovarian hyperandrogenism promotes exces-
The size of the uterus is often enlarged in women with PCOS
sive early follicular growth up to 2±5 mm, with more follicles able
because of the increased degree of estrogenization (Adams et al.,
to enter the growing cohort, which then become arrested at the
1985; Balen et al., 1995). The ratio of ovarian:uterine volume had
6±9 mm size.
been suggested as never being higher than 1.0 in women with
Thus, the consensus de®nition for a polycystic ovary is one that
polycystic ovaries (Parisi et al., 1982). However, others (Orsini
et al., 1985) reported a wide range of ovarian:uterine volumes, and contains 12 or more follicles of 2±9 mm diameter. This should help
this diagnostic criterion was subsequently abandoned. to discriminate PCO from the other causes of multifollicular
ovaries.
Internal features of the polycystic ovary Multicystic and polycystic ovaries: The multicystic ovary is one
in which there are multiple (>6) follicles, usually 4±10 mm in
Follicles: size and number diameter, with normal stromal echogenicity (Adams et al., 1985).
Technical aspects: It is now known that it is oocyte-containing Almost no histological data about multicystic ovaries are avail-
follicles that were observed when describing the polycystic ovary, able. Again, the terminology might be better annotated as multi-
rather than pathological or atretic cystic structures. The early follicular rather than multi-cystic. The multi-follicular appearance
literature often refers to `cysts' rather than follicles, and as the is characteristically seen during puberty and in women recovering
latter are indeed small cystsÐthat is, a `sac containing ¯uid'Ðthe from hypothalamic amenorrhoeaÐboth situations being associ-
terminology polycystic ovary syndrome has remained. ated with follicular growth without consistent recruitment of a
Each ovary should be scanned in longitudinal cross-section dominant follicle (Venturoli et al., 1983; Stanhope et al., 1985).
from the inner to outer margins in order to count the total number There may be confusion among inexperienced ultrasonographers,

508
 Consensus de®nition of the polycystic ovary

Table II. Receiver operating characteristic (ROC) curve data for the assessment of polycystic ovaries
(Jonard et al., 2003)

FNPO (mm) Area under ROC curve Threshold Sensitivity (%) Speci®city (%)

2±5 0.924 10 65 97
12 57 99
15 42 100
6±9 0.502 3 42 69
4 32.5 80
5 24 89
2±9 0.937 10 86 90
12 75 99
15 58 100

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

FNPO = follicle number per ovary.

Table III. Ultrasound assessment of the polycystic ovary (PCO): international consensus de®nitions

De®nition

1. The PCO should have at least one of the following: either 12 or more follicles measuring 2±9 mm in diameter or increased ovarian volume (>10 cm3). If there is
evidence of a dominant follicle (>10 mm) or a corpus luteum, the scan should be repeated during the next cycle.
2. The subjective appearance of PCOs should not be substituted for this de®nition. The follicle distribution should be omitted as well as the increase in stromal
echogenicity and/or volume. Although the latter is speci®c to polycystic ovary, it has been shown that measurement of the ovarian volume is a good
surrogate for the quanti®cation of the stroma in clinical practice.
3. Only one ovary ®tting this de®nition or a single occurrence of one of the above criteria is suf®cient to de®ne the PCO. If there is evidence of a dominant
follicle (>10 mm) or corpus luteum, the scan should be repeated next cycle. The presence of an abnormal cyst or ovarian asymmetry, which may suggest a
homogeneous cyst, necessitates further investigation.
4. This de®nition does not apply to women taking the oral contraceptive pill, as ovarian size is reduced, even though the `polycystic' appearance may persist.
5. A woman having PCO in the absence of an ovulation disorder or hyperandrogenism (`asymptomatic PCO') should not be considered as having PCOS,
until more is known about this situation.
6. In addition to its role in the de®nition of PCO, ultrasound is helpful to predict fertility outcome in patients with PCOS (response to clomiphene citrate, risk
for ovarian hyperstimulation syndrome (OHSS), decision for in-vitro maturation of oocytes). It is recognized that the appearance of PCOs may be seen in
women undergoing ovarian stimulation for IVF in the absence of overt signs of PCOS. Ultrasound also provides the opportunity to screen for endometrial
hyperplasia.
7. The following technical recommendations should be respected:
d State-of-the-art equipment is required and should be operated by appropriately trained personnel.
d Whenever possible, the transvaginal approach should be preferred, particularly in obese patients.
d Regularly menstruating women should be scanned in the early follicular phase (days 3±5). Oligo-/amenorrhoeic women should be scanned either at random
or between days 3±5 after a progestogen-induced bleed.
d If there is evidence of a dominant follicle (>10mm) or a corpus luteum, the scan should be repeated the next cycle.
d Calculation of ovarian volume is performed using the simpli®ed formula for a prolate ellipsoid (0.5 3 length 3 width 3 thickness).
d Follicle number should be estimated both in longitudinal, transverse and antero-posterior cross-sections of the ovaries. Follicle size should be expressed as
the mean of the diameters measured in the three sections.

The usefulness of 3-D ultrasound, Doppler or MRI for the de®nition of PCO has not been suf®ciently ascertained to date, and should be con®ned to research
studies.

radiologists and gynaecologists; hence the need for careful the setting of the ultrasound machine. Stromal echogenicity has
consideration of the clinical picture and endocrinology. been described in a semi-quantitative manner with a score for
normal (= 1), moderately increased (= 2) or frankly increased (= 3)
Stroma: volume and echogenicity
(Pache et al., 1991). In the latter study, the total follicle number of
Stromal echogenicity: The increased echodensity of the polycystic both ovaries combined correlated signi®cantly with stromal
ovary is a key histological feature (Hughesdon, 1982), but is a echogenicity, and follicle number also correlated signi®cantly
subjective assessment that may vary depending upon the setting of with free androgen index. A further study comparing women with
the ultrasound machine and the patient's body habitus. In one PCOS with controls found that the sensitivity and speci®city of
study (Ardaens et al., 1991), subjectively increased stromal ovarian stromal echogenicity in the diagnosis of polycystic ovaries
hyperechogenicity, when assessed transvaginally, appeared exclu- were 94 and 90% respectively (Pache et al., 1992).
sively to be associated with PCOS. Echogenicity has been quanti®ed by one group (Al-Took et al.,
Normal stromal echogenicity is said to be less than that of the 1999) as the sum of the product of each intensity level (ranging
myometrium, which is a simple guide that will take into account from 0 to 63 on the scanner) and the number of pixels for that

509
A.H.Balen et al.

intensity level divided by the total number of pixels in the similar in all groups, indicating that increased stromal volume is
measured area: Mean = (S xi.fi)/n, where n = total number of pixels the main cause of ovarian enlargement in polycystic ovaries.
in the measured area, x = intensity level (0±63), and f = number of In summary, ovarian volume correlates well with ovarian
pixels corresponding with the level. The stromal index was function and is both more easily and reliably measured in routine
calculated by dividing the mean stromal echogenicity by the mean practice than ovarian stroma. Thus, in order to de®ne the
echogenicity of the entire ovary in order to correct for cases in polycystic ovary, neither qualitative or quantitative assessment
which the gain was adjusted to optimize image de®nition (Al-Took of the ovarian stroma is required.
et al., 1999). When using these measurements, the stromal index
Blood ¯ow
did not predict responsiveness to clomiphene citrate, and neither
did the stromal index differ after ovarian drilling (Al-Took et al., The combination of transvaginal ultrasound with colour Doppler
1999). measurements is beginning to provide a detailed picture of
Another approach used a 7.5 MHz transvaginal probe with follicular events around the time of ovulation, and also allows

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

histogram measurement of echogenicity (Buckett et al., 1999). assessment of the uterine blood ¯ow to predict endometrial
The mean echogenicity was de®ned as the sum of the product of receptivity (Zaidi et al., 1995b, 1996a). Blood ¯ow through the
each intensity level (0±63) using the same formula as others (Al- uterine and ovarian arteries has been extensively investigated in
Took et al., 1999; see above). Women with PCOS had greater total spontaneous and stimulated cycles (Tan et al., 1996). Colour (or
ovarian volume, stromal volume and peak stromal blood ¯ow `power') Doppler also allows assessment of the vascular network
compared with normal ovaries, yet mean stromal echogenicity was within the ovarian stroma. Intra-ovarian stromal blood ¯ow is
similar. The stromal index (mean stromal echogenicity:mean signi®cantly higher in polycystic ovaries than normal ovaries, and
echogenicity of entire ovary) was higher in PCOS, due to the its measurementÐeither in the early follicular phase or following
®nding of a reduced mean echogenicity of the entire ovary pituitary suppressionÐhas been found to be predictive of
(Buckett et al., 1999). The conclusion was that the subjective follicular response to ovarian stimulation for IVF (Zaidi et al.,
impression of increased stromal echogenicity was due both to 1996b; Engmann et al., 1999).
increased stromal volume alongside reduced echogenicity of the A number of studies of colour Doppler measurement of uterine
multiple follicles. and ovarian vessel blood ¯ow have demonstrated a low resistance
Stromal area or volume: One group (Dewailly et al., 1994) index in the stroma of polycystic ovaries (i.e. increased ¯ow) and
designed a computer-assisted method for standardizing the correlations with endocrine changes (Battaglia et al., 1995;
assessment of stromal hypertrophy. Patients with hyperandrogen- Loverro et al., 2001; Pan et al., 2002). One of these groups
ism (of whom 68% had menstrual cycle disturbances) were (Battaglia et al., 1999) reported a good correlation between serum
compared with a control group and a group with hypothalamic androstenedione concentrations and LH:FSH ratio with the
amenorrhoea. Transvaginal ultrasound (5 MHz) was used and number of small follicles; the LH:FSH ratio also correlated well
polycystic ovaries were de®ned as the presence of `abnormal with the stromal artery pulsatility index (PI). In another study, the
ovarian stroma and/or the presence of at least 10 round areas of blood ¯ow was more frequently visualized in PCOS (88%) than in
reduced echogenicity <8 mm in size on a single ovarian section normal patients (50%) in the early follicular phase and also
and/or an increased cross-sectional ovarian area (>10 cm2)' appeared to be increased (Battaglia et al., 1996).
(Ardaens et al., 1991; Dewailly et al., 1994). The computerized Both the resistance index (RI) and PI have been found to be
technique for reading the scans involved a longitudinal section in signi®cantly lower in PCOS than in normal patients, and the peak
the middle part of the ovary and a calculation of stromal area and systolic velocity (PSV) greater (Aleem and Predanic, 1996). No
the area of the follicles. Of 57 women with hyperandrogenism, correlation was found with the number of follicles and the ovarian
65% had an polycystic ovaries visualized on ultrasound, and volume, but there was a positive correlation between LH levels
elevated serum testosterone and LH concentrations were found in and increased PSV. One group (Zaidi et al., 1995a) found no
50 and 45% respectively. There was no correlation between LH signi®cant difference in PI values between the normal and PCOS
and androstenedione concentrations. Stromal area, however, groups, while the ovarian ¯owÐas re¯ected by the PSVÐwas
correlated signi®cantly with levels of androstenedione and 17- increased in the former.
hydroxy-progesterone, but not of testosterone, LH or insulin; Recent data have indicated that assessment of Doppler blood
follicle area did not correlate with endocrine parameters (Dewailly ¯ow may have some value in predicting the risk for ovarian
et al., 1994). Thus, it was suggested that in women with polycystic hyperstimulation during gonadotrophin therapy (Agrawal et al.,
ovaries an analysis of ovarian stromal area is better than 1998). Increased stromal blood ¯ow has also been suggested as a
quanti®cation of the follicles. more relevant predictor of ovarian response to hormonal stimu-
Three-dimensional ultrasound has been shown to be a good tool lation (Buckett et al., 1999; Engmann et al., 1999) than parameters
for the accurate measurement of ovarian volume, and to be more such as ovarian or stromal volume. However, the measurement of
precise than 2-D ultrasound (Kyei-Mensah et al., 1996b). Three Doppler blood ¯ow requires speci®c expertise and machinery, and
groups of patients were de®ned: (1) those with normal ovaries; (2) at the present time is not necessary as part of the diagnostic criteria
those with asymptomatic polycystic ovaries; and (3) those with for polycystic ovary.
PCOS (Kyei-Mensah et al., 1998). The ovarian and stromal
volumes were similar in groups 2 and 3, and both were greater than
De®nition of PCO
the volumes in group 1. Stromal volume was positively correlated
with serum androstenedione concentrations in group 3 only (Kyei- With all imaging systems, the ovarian size (i.e. volume) and
Mensah et al., 1998). The mean total volume of the follicles was number of pre-antral follicles are, when in combination, the key

510
 Consensus de®nition of the polycystic ovary

and consistent features of polycystic ovaries. In routine clinical De®nition of PCO in particular circumstances
practice it is transabdominal or transvaginal ultrasound alone that
There are circumstances where the above de®nition does not ®t,
suf®ces in assessment of the ovary.
until more data are collected:
Previously proposed de®nitions 1. In women taking the combined oral contraceptive pill, the
ovarian volume is suppressed but the appearance may still be
During the early 1990s, a series of studies was performed to
polycystic (Franks et al., 1985).
distinguish between normal and polycystic ovaries and to deter-
2. Polycystic ovaries can also be detected in post-menopausal
mine the key features of the polycystic ovary (Pache et al., 1991;
women and whilst, not surprisingly, smaller than in pre-meno-
1992, 1993). First, PCOS was de®ned (on the basis of elevated
pausal women with polycystic ovaries, they are still larger (6.4
serum testosterone or LH) and transvaginal ultrasound (5 MHz)
versus 3.7 cm3) with more follicles (9.0 versus 1.7) than normal
then used to compare those women with the syndrome to a control
post-menopausal ovaries (Birdsall and Farquhar, 1996). However,
group (Pache et al., 1992). Women with amenorrhoea had similar
no threshold is available.

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

ultrasound features to those with oligomenorrhoea. Control
3. Criteria to discriminate polycystic ovaries from multi-cystic
ovaries never had a volume of more than 8.0 cm3, or contained
ovaries in adolescent girls have not been established (Herter et al.,
more than 11 follicles. The mean number of follicles was 10 in
1996). Indeed, it appears that PCOS manifests for the ®rst time
polycystic ovaries and ®ve in normal ovaries. Median values for
during the adolescent years, which are critical for future ovarian
mean ovarian volume were 5.9 cm3 in controls and 9.8 cm3 in
and metabolic function (Gulekli et al., 1993; Balen and Dunger,
PCOS (P < 0.001), while mean follicular size and number were 5.1
1995).
versus 3.8 cm3 and 5.0 versus 9.8 for control and PCOS women
respectively. Stromal echogenicity was also signi®cantly increased Use of ultrasonography in the diagnostic strategy for PCOS
in the PCOS patients, based on a semi-quantitative assessment (see
Making the diagnosis of PCOS has been a matter of great debate,
below) (Pache et al., 1992). The greatest power of discrimination
between normal and polycystic ovaries was provided by a and in particular the use of ultrasound as a universal standard has
combined measurement of follicular size and ovarian volume been disputed. The PCOS phenotype can be structured into three
(sensitivity 92%, speci®city 97%). components: anovulation, hyperandrogenism, and obesity (with
A later study from the same group de®ned normal ovarian associated hyperinsulinaemia) (Dewailly, 1997). However, these
morphology in a control group of 48 normally cycling women, and components are neither constantly, necessarily nor equally asso-
compared both ultrasound and endocrine parameters with those in ciated, thus explaining the great variability in the clinical
patients with normogonadotrophic oligomenorrhoea or amenor- presentation of PCOS (Balen et al., 1995; Dewailly, 1997). In
rhoea (Van Santbrink et al., 1997). In the normal ovaries the mean some cases, only one or two components are present (e.g.
number of follicles per ovary was 7.0 6 1.7, and none of the `ovulatory PCOS' or `non-hirsute anovulatory PCOS')Ðhence
women had more than nine follicles or an ovarian volume the new consensus de®nition of the syndrome which requires the
>10.7 ml. Polycystic ovaries were therefore considered to have presence of two out of the following three criteria: (i) oligo- and/or
>10 follicles and a volume >10.8 ml. anovulation; (ii) hyperandrogenism (clinical and/or biochemical);
Based on their ®ndings, one group (Jonard et al., 2003) and (iii) polycystic ovaries, with the exclusion of other aetiologies
proposed a new de®nition of the polycystic ovary as being an (Fauser et al., 2004).
increased ovarian area (>5.5 cm2) or volume (>11 cm3) and/or the It is important to assess the ultrasound features of the ovary in
presence of >12 follicles of 2±9 mm diameter (as a mean of both all clinical presentations, alongside appropriate endocrine, bio-
ovaries). chemical and metabolic tests as indicated by the presentation. For
example, in patients who might be considered to have `obvious
Proposition for a consensus de®nition PCOS' when being referred for treatment of anovulatory infertil-
Based on the literature review and on discussions held at the joint ity, it is also important to remember that abnormalities of basal
ASRM/ESHRE consensus meeting on PCOS held in Rotterdam, serum prolactin or FSH levels may indicate a coexistent
May 1±3, 2003, a consensus de®nition may be provided (Table III). hypothalamic±pituitary disorder or incipient ovarian failure.
Ultrasound assessment of the ovaries will help in the prediction
Sensitivity and speci®city of response to stimulation.
A recent study set out to assess variability in the detection of In cases of an isolated menstrual disorder or so-called
polycystic and normal ovaries with four experienced practitioners, `idiopathic hirsutism' (i.e. with ovulatory menstrual cycles),
who independently reviewed recordings of 27 pairs of ovaries, PCOS is the most likely diagnosis. The clinical picture may not
and demonstrated intra-observer agreement of 69.4% and inter- clearly provide the diagnosis without appropriate hormonal assays
observer agreement of 51% (Amer et al., 2002). Polycystic ovaries together with ultrasound. However, the ®nding of polycystic
were de®ned as the presence of >10 follicles (2±8 mm diameter), ovaries at ultrasound does not exclude other diagnoses as
ovarian volume >12 cm3 and bright echogenic stroma. Thus, there polycystic ovaries may be coincidentally associated with other
was signi®cant intra-observer and inter-observer variability using conditions.
these criteria. This suggests either that these criteria are too The incidental discovery of polycystic ovaries at ultrasound is
subjective, or that their measurement is too insensitive. It was common in women undergoing investigation for any gynaecolo-
concluded (Amer et al., 2002) that the use of 3-D ultrasound might gical complaint, such as pelvic pain, unexplained bleeding or
provide a more reliable and reproducible diagnostic tool, although infertility. If polycystic ovaries are observed in ovulatory infertile
a similar evaluation of observer variability was not carried out. women (in whom PCOS is not the cause of infertility), the

511
A.H.Balen et al.

Table IV. The ultrasound scan report luteum will result in an increased ovarian volume and area, and
therefore the scan should be repeated during days 1±3 of the next
Name and age of patient cycle.
Identifying unique hospital record number The time of day needs to be recorded only if Doppler studies are
Date of scan
being performed, in which diurnal variation has been demonstrated
Relation to menstrual cycle
Relevant treatment (COCP, GnRHa, etc.) in uterine (Zaidi et al., 1995c) and ovarian (Zaidi et al., 1996a)
Type of scan (transabdominal/transvaginal, etc.) blood ¯ow. The Vmax in the ovary with a dominant follicle rises
Ovarian morphology: each ovary recorded separately during the day, while the PI is highest in morning and lowest in
Volume (and area ± optional) afternoon and evening (Zaidi et al., 1996a).
No. and size/range of cysts Unless there is evidence of a dominant follicle (>10 mm) or a
Stromal echogenicity (if local grading system exists) corpus luteum, it is probably not necessary to repeat the scan for
Doppler studies (if performed)
Uterine morphology, cross-sectional area and endometrial thickness
validation of the ®ndings as there is evidence of little change if the

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

Other features scan is repeated one or two times over a 9- to 12-day period (Pache
Hard copy and/or electronic copy et al., 1991), although few data are available from other studies on
Grade and signature of person performing scan successive ultrasound scans over time.
Grade and signature of person verifying scan (if relevant)
Transabdominal, transvaginal, or both routes?
The transabdominal route is of course required in girls and women
who are virgo intacta or for patients who decline a transvaginal
information is very important when designing a `superovulation' scan. A transabdominal scan offers a panoramic view of the pelvic
protocol because there is an increased risk of OHSS. Also, it may cavity, and so may be useful if there are associated uterine or
be useful to look for a family history of PCOS, as some siblings ovarian developmental abnormalities, or if the transvaginal scan
may have symptomatic, yet undiagnosed, PCOS. In addition, fails to visualize caudally displaced ovaries. Although a full
metabolic features of hyperinsulinism may be present and deserve bladder is required for visualization of the ovaries, one should be
careful evaluation as they could indicate risks for long-term health. cautious that an over®lled bladder can compress the ovaries,
yielding a falsely increased length. This emphasizes the need for
assessing the ovarian size by measuring the area or the volume (see
Conclusion below) and by repeating the measurement after partial micturition.
Although by its sensitivity (providing that suf®cient speci®city is If not found between the uterus and the iliac vessels, the ovaries
guaranteed), ultrasonography has widened the clinical spectrum of must be searched for upward, in the iliac fossa close to
PCOS, this has led to a reduction in the numbers of cases the abdominal wall, or downward and backward in the Douglas
diagnosed with `idiopathic hirsutism' and `idiopathic anovula- cul-de-sac.
tion'. Establishing the diagnosis of PCOS is, and will always be, a In one study (Farquhar et al., 1994), there was no signi®cant
matter of good clinical sense. difference in the detection of polycystic ovaries, and the same
The international consensus de®nitions for the ultrasound criteria for the number of follicles was felt to be appropriate for
assessment of the polycystic ovary are proposed in Table III. both types of scan. There was only a 78% agreement between
transabdomianl and transvaginal ultrasound for polycystic ovaries,
though this was 92% for normal ovaries (Farquhar et al., 1994).
Appendix 1: Practical considerations for ultrasound Another group (Ardaens et al., 1991) also compared transabdom-
assessment of ovaries in PCOS inal (3.5 MHz) with transvaginal (6.5 MHz) ultrasound and
reported that the latter was more consistent in achieving the
Timing of the ultrasound scan diagnosis of polycystic ovaries in women with PCOS. Increased
Few studies that describe the morphology and endocrinology of stromal echogenicity assessed transvaginally appeared exclusively
PCOS make reference to timing of the menstrual cycle. The to be associated with PCOS, but this was a subjective appearance
baseline ultrasound scan of the pelvis is best performed in the early rather than quanti®able measurement (Ardaens et al., 1991).
follicular phase (days 1±3), when the ovaries are relatively It has been argued that transvaginal ultrasound is a more
quiescent. This is the optimal approach in order to obtain sensitive method for the detection of polycystic ovaries. For
consistency in the measurement of ovarian volume and area. It example, polycystic ovaries were not detected in 30% of women
is recognized that in routine clinical practice it is not always with PCOS when a 3.5 MHz transabdominal transducer was used;
possible to schedule the scan to coincide with the ®rst 3 days of the however, a 7.5 MHz transvaginal probe was found to be more
menstrual cycle. It is important to be as precise as possible when reliable (Fox et al., 1991).
performing scans for research studies, in which consistency is
more relevant than the pragmatic approach that is often taken in
Appendix 2: Guidelines for ultrasound assessment of
day-to-day practice.
ovaries in PCOS
If the patient is oligo-/amenorrhoeic, the scan should be
performed at random. It is recognized that women with PCOS A pelvic ultrasound scan should be performed by appropriately
are usually oligo-ovulatory rather than totally anovulatory, and so trained personnel, who have obtained the relevant quali®cations
it is not uncommon to see a recruited follicle when assessing the and continue to participate in continuing professional development
ovaries. The presence of a follicle >10 mm diameter or a corpus and appraisal programmes. Only trained personnel should report

512
 Consensus de®nition of the polycystic ovary

on ultrasound scans. Assessment of inter-observer variation should Birdsall, M.A. and Farquhar, C.M. (1996) Polycystic ovaries in pre and post-
menopausal women. Clin. Endocrinol., 44, 269±276.
be performed on a regular (e.g. annual) basis and at the start of Buckett, W.M., Bouzayen, R., Watkin, K.L., Tulandi, T. and Tan, S.L. (1999)
scienti®c studies of ovarian function. Ovarian stromal echogenicity in women with normal and polycystic
State-of-the-art equipment is required, which should be up to ovaries. Hum Reprod., 14, 618±621.
date and serviced regularly. An appropriate selection of transab- Campbell, S., Goessens L., Goswamy R. and Whitehead, M. (1982) Real-time
ultrasonography for determination of ovarian morphology and volume.
dominal and transvaginal probes should be available for all body Lancet, 1, 425±428.
shapes/sizes. In addition to a real-time assessment of ovarian and Clayton, R.N., Ogden, V., Hodgkinson, J. et al. (1992) How common are
uterine morphology, images should be recorded as either hard polycystic ovaries in normal women and what is their signi®cance for the
copy or electronically. fertility of the population? Clin. Endocrinol., (Oxf) 37, 127±134.
Conway, G.S., Honour, J.W. and Jacobs, H.S. (1989) Heterogeneity of the
The scan should be performed with the patient's consent. She polycystic ovary syndrome: clinical, endocrine and ultrasound features in
should be accompanied by a relative, friend or her partner if she 556 patients. Clin. Endocrinol., (Oxf) 30, 459±470.
wishes. Due consideration should be taken for her need for privacy Dewailly, D. (1997) De®nition and signi®cance of polycystic ovaries. In

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

when changing. A chaperone should be present and should sign Rosen®eld, R.L. (ed.), Hyperandrogenic States and Hirsutism. Balliere's
Clin. Obstet. Gynaecol., 11, 349±368.
that the procedure has been witnessed. Dewailly, D., Robert, Y., Helin, I., Ardaens, Y., Thomas-Desrousseaux, P.,
The scan should be performed in a systematic fashion. Each Lemaitre L and Fossati P (1994) Ovarian stromal hypertrophy in
ovary should be scanned from the inner to outer margins in order to hyperandrogenic women. Clin. Endocrinol., 41, 557±562.
count the total number of cysts/follicles. Appropriate measure- Elting, M.W., Korsen, T.J.M., Rekers-Mombarg, L.T.M. and Schoemaker, J.
(2000) Women with polycystic ovary syndrome gain regular menstrual
ments should then be performed of the ovarian and uterine cycles when ageing. Hum. Reprod., 15, 24±28.
dimensions (see above). A suggested scheme for the ultrasound Engmann, L., Sladkevicius, P., Agrawal, L.R., Bekir, J.S., Campbell, S. and
report is presented in Table IV. Tan, S.L. (1999) Value of ovarian stromal blood ¯ow velocity
measurement after pituitary suppression in the prediction of ovarian
responsiveness and outcome of in vitro fertilization treatment. Fertil.
References Steril., 71, 22±29.
Farquhar, C.M., Birdsall, M., Manning, P., Mitchell, J.M. and France, J.T.
Abdel Gadir, A, Khatim, M.S., Mowa®, R.S., Alnaser, H.M., Muharib, N.S. (1994) The prevalence of polycystic ovaries on ultrasound scanning in a
and Shaw, R.W. (1992) Implications of ultrasonically diagnosed population of randomly selected women. Aust. N Z J. Obstet. Gynaecol.,
polycystic ovaries. I. Correlations with basal hormonal pro®les. Hum. 34, 67±72.
Reprod., 7, 453±457. Faure, N., Prat, X., Bastide, A. and Lemay, A. (1989) Assessment of ovaries
Adams, J, Polson, D.W., Abdulwahid, N., Morris, D.V., Franks, S., Mason, by magnetic resonance imaging in patients presenting with polycystic
H.D., Tucker, M., Price, J. and Jacobs, H.S. (1985) Multifollicular ovarian syndrome. Hum. Reprod., 4, 468±472.
ovaries: clinical and endocrine features and response to pulsatile Fauser, B., Tarlatzis, B., Chang, J. et al. (2004) The Rotterdam ESHRE/
gonadotropin releasing hormone. Lancet, ii, 1375±1379. ASRM sponsored PCOS consensus workshop. 2003 consensus on
Agrawal, R., Conway, G., Sladkevicius, P., Tan, S.L., Engmann, L., Payne, N., diagnostic criteria and long-term health risks related to Polycystic
Bekir, J., Campbell, S. and Jacobs, H. (1998) Serum vascular endothelial Ovary Syndrome (PCOS). Hum. Reprod. and Fertil. Steril. in press.
growth factor and Doppler blood ¯ow velocities in in vitro fertilization: Fox, R., Corrigan, E., Thomas, P.A. and Hull, M.G. (1991) The diagnosis of
relevance to ovarian hyperstimulation syndrome and polycystic ovaries. polycystic ovaries in women with oligo-amenorrhoea: predictive power of
Fertil. Steril., 70, 651±658. endocrine tests. Clin. Endocrinol., (Oxf) 34, 127±131.
Aleem, F.A., and Predanic, M.P., (1996) Transvaginal color Doppler
Franks, S. (1995) Polycystic ovary syndrome. N. Engl. J. Med. 333, 853±861.
determination of the ovarian and uterine blood ¯ow characteristics in
Franks, S, Adams J, Mason HD and Polson DW (1985) Ovulatory disorders in
polycystic ovary disease. Fertil. Steril., 65, 510±516.
women with polycystic ovary syndrome. Clin. Obstet. Gynecol., 12, 605±
Al-Took, S., Watkin, K., Tulandi, T. and Tan, S.L., (1999) Ovarian stromal
632.
echogenicity in women with clomiphene citrate-sensitive and clomiphene
Fulghesu, A.M., Ciampelli, M., Belosi, C., Apa, R., Pavone, V. and Lanzone,
citrate-resistant polycystic ovary syndrome. Fertil. Steril., 71, 952±954.
A. (2001) A new ultrasound criterion for the diagnosis of polycystic ovary
Amer, S.A.K.S., Li, T.C., Bygrave, C., Sprigg, A., Saravelos, H. and Cooke,
syndrome: the ovarian stroma:total area ratio. Fertil. Steril., 76, 326±331.
I.D. (2002) An evaluation of the inter-observer and intra-observer
variability of the ultrasound diagnosis of polycystic ovaries. Hum. Goldstein, G. (1990) Incorporating endovaginal ultrasonography into the
Reprod., 17, 1616±1622. overall gynaecologic examination. Am. J. Obstet. Gynecol., 160, 625±
Ardaens, Y., Robert, Y., Lemaitre, L, Fossati, P. and Dewailly, D. (1991) 632.
Polycystic ovarian disease: contribution of vaginal endosonography and Goldzieher, M.W. and Green, J.A. (1962) The polycystic ovary. I. Clinical and
reassessment of ultrasonic diagnosis. Fertil. Steril., 55, 1062±1068. histologic features. J. Clin. Endocrinol. Metab., 22, 325±338.
Balen, A.H. (1999) The pathogenesis of polycystic ovary syndrome: the Gulekli, B., Turhan, N.O., Senoz, S., Kukner, S, Oral, H. and Gokmen, O.
enigma unravels. Lancet, 354, 966±967. (1993) Endocrinological, ultrasonographic and clinical ®ndings in
Balen, A.H. and Dunger, D. (1995) Pubertal maturation of the internal adolescent and adult polycystic ovary patients: a comparative study.
genitalia. Ultrasound Obstet. Gynecol., 6, 164±165. Gynecol. Endocrinol., 7, 273±277.
Balen, A.H. and Michelmore, K. (2002) What is polycystic ovary syndrome? Hann, L.E., Hall, D.A., McArdle, C.R. and Seibel, M. (1984) Polycystic
Are national views important? (Debate). Hum. Reprod., 17, 2219±2227. ovarian disease: sonographic spectrum. Radiology, 150, 531±534.
Balen, A.H., Conway, G.S., Kaltsas, G. et al. (1995) Polycystic ovary Herter, L.D., Magalhaes, J.A. and Spritzer, P.M. (1996) Relevance of the
syndrome: the spectrum of the disorder in 1741 patients. Hum. Reprod., determination of ovarian volume in adolescent girls with menstrual
10, 2107±2111. disorders. J. Clin. Ultrasound., 24, 243±248.
Battaglia, C., Artini, P.G., D'Ambrogio, G., Genazzani, A.D. and Genazzani, Homburg, R. (2002) What is polycystic ovary syndrome? A proposal for a
A.R. (1995) The role of colour Doppler imaging in the diagnosis of consensus on the de®nition and diagnosis of polycystic ovary syndrome.
polycystic ovary syndrome. Am. J. Obstet. Gynecol., 172, 108±113. Hum. Reprod., 17, 2495±2499.
Battaglia, C., Artini, P.G., Genazzani, A.D., Sgherzi, M.R., Salvatori, M., Hughesdon, P.E. (1982) Morphology and morphogenesis of the Stein-
Giulini, S. and Volpe, A. (1996) Color Doppler analysis in lean and obese Leventhal ovary and of so-called `hyperthecosis'. Obstet. Gynecol.
women with polycystic ovaries. Ultrasound Obstet. Gynecol., 7, 342±346. Surv., 37, 59±77.
Battaglia, C., Genazzani, A.D., Salvatori, M., Giulini, S., Artini, P.G., Jonard, S., Robert, Y., Cortet-Rudelli, C., Decanter, C. and Dewailly, D.
Genazzani, A.R. and Volpe, A. (1999) Doppler, ultrasonographic and (2003) Ultrasound examination of polycystic ovaries: is it worth counting
endocrinological environment with regard to the number of small the follicles? Hum. Reprod., 18, 598±603.
subcapsular follicles in polycystic ovary syndrome. Gynecol. Kiddy, D.S,. Sharp, P.S., White, D.M. et al. (1990) Differences in clinical and
Endocrinol., 13, 123±129. endocrine features between obese and non-obese subjects with polycystic

513
A.H.Balen et al.

ovary syndrome: an analysis of 263 consecutive cases. Clin. Endocrinol., Sample, W.F., Lippe, B.M. and Gyepes, M.T. (1977) Grey-scale
(Oxf) 32, 213±220. ultrasonography of the normal female pelvis. Radiology. 125, 477±483.
Kyei-Mensah, A., Maconochie, N., Zaidi, J., Pittrof., R., Campbell, S. and Saxton, D.W., Farquhar, C.M., Rae, T., Beard, R.W., Anderson, M.C. and
Tan, S.L. (1996a) Transvaginal three-dimensional ultrasound: Wadsworth, J. (1990) Accuracy of ultrasound measurements of female
reproducibility of ovarian and endometrial volume measurements. pelvic organs. Br. J. Obstet. Gynaecol., 97, 695±699.
Fertil. Steril., 66, 718±722. Stanhope, R., Adams, J., Jacobs, H.S. and Brook, C.G. (1985) Ovarian
Kyei-Mensah, A., Maconochie, N., Zaidi, J., Pittrof, R., Campbell, S. and Tan, ultrasound assessment in normal children, idiopathic precocious puberty,
S.L. (1996b) Transvaginal three-dimensional ultrasound: accuracy of and during low dose pulsatile gonadotrophin releasing hormone treatment
ovarian follicular volume measurements. Fertil. Steril., 65, 371±376. of hypogonadotrophic hypogonadism. Arch. Dis. Child., 60, 116±119.
Kyei-Mensah, A., Tan, S.L., Zaidi, J. and Jacobs, H.S. (1998) Relationship of Stein, I.F. and Leventhal, M.L. (1935) Amenorrhea associated with bilateral
ovarian stromal volume to serum androgen concentrations in patients with polycystic ovaries. Am. J. Obstet. Gynecol., 29, 181±191.
polycystic ovary syndrome. Hum. Reprod., 13, 1437±1441. Swanson, M., Sauerbrei E.E. and Cooperberg, P.L. (1981) Medical
Loverro, G., Vicino, M., Lorusso, F., Vimercati, A., Greco, P. and Selvaggi, L. implications of ultrasonically detected polycystic ovaries. J. Clin.
(2001) Polycystic ovary syndrome: relationship between insulin Ultrasound, 9, 219±222.
sensitivity, sex hormone levels and ovarian stromal blood ¯ow. Takahashi, K., Eda, Y., Abu Musa, A., Okada, S., Yoshino, K. and Kitao, M.

Downloaded from https://academic.oup.com/humupd/article-abstract/9/6/505/638314 by guest on 15 October 2019

Gynecol. Endocrinol., 15, 142±149. (1994) Transvaginal ultrasound imaging, histopathology and
Michelmore, K.F., Balen, A.H., Dunger, D.B. and Vessey, M.P. (1999) endocrinopathy in patients with polycystic ovarian syndrome. Hum.
Polycystic ovaries and associated clinical and biochemical features in Reprod., 9, 1231±1236.
young women. Clin. Endocrinol., (Oxf) 51, 779±786. Tan, S.L., Zaidi, J., Campbell, S., Doyle, P. and Collins, W. (1996) Blood ¯ow
Mitchell, D.G., Gefter, W.B., Spritzer, C.E., Blasco, L., Nulson, J., Livolsi, V., changes in the ovarian and uterine arteries during the normal menstrual
Axel, L., Arger, P.H. and Kressel, H.Y. (1986) Polycystic ovaries: MR cycle. Am. J. Obstet. Gynecol., 175, 625±631.
imaging. Radiology, 160, 425±429. Vallisneri, A. (1721) Storzia della generazione dell'uomo e dell'animale.
Nardo, L.G., Buckett, W.M. and Khullar, V. (2003) Determination of the best- Cited in Cooke, I.D. and Lunenfeld, B. (1989) Res Clin Forums 11, 109±
®tting ultrasound formulaic method for ovarian volume measurement in 113.
women with polycystic ovary syndrome. Fertil. Steril., 79, 632±633.
Van Santbrink, E.J.P, Hop, W.C. and Fauser, B.C.J.M. (1997) Classi®cation of
Orsini, L.F., Rizzo, N., Calderoni, P., Pilu, G. and Bovicelli, L. (1983)
normogonadotropic infertility: polycystic ovaries diagnosed by ultrasound
Ultrasound monitoring of ovarian follicular development: a comparison
versus endocrine characteristics of polycystic ovary syndrome. Fertil.
of real-time and static scanning techniques. J. Clin. Ultrasound, 11, 207±
Steril., 67, 452±458.
211.
Venturoli, S., Porcu, E., Fabbri, R., Paradisi, R., Orsini, L.F. and Flamigni, C.
Orsini, L.F., Venturoli, S., Lorusso, R., Pluchinotta, V., Paradisi, R. and
Bovicelli, L. (1985) Ultrasonic ®ndings in polycystic ovarian disease. (1983) Ovaries and menstrual cycles in adolescence. Gynecol. Obstet.
Fertil. Steril., 43, 709±714. Invest., 17, 219±223.
Pache, T.D., Hop, W.C., Wladimiroff, J.W., Schipper, J. and Fauser, B.C.J.M. Zaidi, J., Campbell, S., Pittrof, R., Kyei-Mensah, A., Jacobs, H.S. and Tan,
(1991) Transvaginal sonography and abnormal ovarian appearance in S.L. (1995a) Ovarian stromal blood ¯ow in women with polycystic
menstrual cycle disturbances. Ultrasound Med. Biol., 17, 589±593. ovaries ± a possible new marker for diagnosis? Hum. Reprod., 10, 1992±
Pache, T.D., Wladimiroff, J.W., Hop, W.C. and Fauser, B.C.J.M. (1992) How 1996.
to discriminate between normal and polycystic ovaries: transvaginal Zaidi, J., Jurkovic, D., Campbell, S., Okokon, E. and Tan, S.L. (1995b)
ultrasound study. Radiology, 183, 421±423. Circadian variation in uterine artery blood ¯ow indices during the
Pache, T.D., de Jong, F.H., Hop, W.C. and Fauser, B.C.J.M. (1993) follicular phase of the menstrual cycle. Ultrasound Obstet. Gynecol., 5,
Association between ovarian changes assessed by transvaginal 406±410.
sonography and clinical and endocrine signs of the polycystic ovary Zaidi, J., Jurkovic, D., Campbell, S., Pitroff, R., McGregor, A. and Tan, S.L.
syndrome. Fertil. Steril., 59, 544±549. (1995c) Description of circadian rhythm in uterine artery blood ¯ow
Pan, H.-A., Wu, M.-H., Cheng, Y.-C., Li, C..-H. and Chang, F.-M. (2002) indices during the peri-ovulatory period. Human Reprod., 10, 1642±1646.
Quanti®cation of Doppler signal in polycystic ovary syndrome using 3-D Zaidi, J., Tan, S.L., Pitroff, R., Campbell, S. and Collins, W. (1996a) Blood
power Doppler ultrasonography: a possible new marker for diagnosis. ¯ow changes in the intra-ovarian arteries during the peri-ovulatory period
Hum. Reprod., 17, 201±206. ± relationship to the time of day. Ultrasound Obstet. Gynecol., 7, 135±
Parisi, L., Tramonti, M., Casciano, S., Zurli, A. and Gazzarini, O. (1982) The 140.
role of ultrasound in the study of polycystic ovarian disease. J. Clin. Zaidi, J., Barber, J., Kyei-Mensah, A., Bekir, J., Campbell, S. and Tan, S.L.
Ultrasound, 10, 167±172. (1996b) Relationship of ovarian stromal blood ¯ow at the baseline
Polson, D.W., Adams, J., Wadsworth, J. and Franks, S. (1988) Polycystic ultrasound scan to subsequent follicular response in an in vitro
ovaries ± a common ®nding in normal women. Lancet, 1, 870±872. fertilization program. Obstet. Gynecol., 88, 779±784.
The Rotterdam ESHRE/ASRM sponsored PCOS consensus workshop group Zawadzki, J.A. and Dunaif, A. (1992) Diagnostic criteria for polycystic ovary
(2004) Revised 2003 consensus on diagnostic criteria and long-term syndrome: towards a rational approach. In: Dunaif, A., Givens, J.R.,
health risks related to Polycystic Ovary Syndrome (PCOS). Hum. Haseltine, F.P., Merriam, G.R. (eds), Polycystic Ovary Syndrome.
Reprod., in press. Blackwell Scienti®c, Boston, pp. 377±384.

514