Hugh K. Brown, PhD a Barbara S. Stoll, BS a Santo Patricia S. Hambley, MD #{149} Laurence P. Clarke, PhD

Uterine Histologic

Junctional Findings

Index

terms:

Uterus,

MR

Uterus, studies,

endometrium, 854.1214

a

854.92 Uterus,

Radiology

1991;

From

the

Medicine and the

pita!, Tampa. 10; accepted dent’s

Departments

Received December

Council

and

H.K.B. C

RSNA,

MDC

1991

of Anatomy

June 25, 1990; 18. Supported the

March

of Dimes

Birth

most

stnik-

pearance of the myometnium in hematoxylin-eosin-stained histologic sections is relatively homogeneous. Abso, the wide variation in thickness and visibility of the junctional zone caused by reproductive status, phase of menstrual cycle, use of oral contraceptives, and individual variation makes it difficult to describe the “normal”

uterine

zones

in MR

im-

tion

appears

to be the

inner

third

of sagittally

sectioned

human

and has been referred to as the junctionab zone (5). This strategically located feature has become a significant landmark for diagnosis of uterine pathologic conditions (2,6-8). However, histologic and ubtrastnuctural correlates of

uteri employed various stains and morphometnic studies. Also, uterine surgical specimens were prepared for

this feature fined. An

fens from other zones tnium. Also, monphometric

have remained early suggestion

junctional the basal

zone stratum

has

dismissed

been

(5).

undethat the

may correspond to of the endometnium because

locations

Another

tionab uterine sumed

the

do not

theory

histo-

match

is that

the

junc-

zone is a manifestation of the vascular stratum due to an ashigher rate of blood perfusion

However,

the

zone

is well

visual-

ized in images of excised uteri (9) and is actually bess vascubar than the outer myometnium (10). The hypothesis that the junctional zone consists of dense subendometrial bundles of smooth muscle (11) has not been thoroughly investigated, partly because the superficial ap-

and

requested by grants Defects

between

tnium

Radiology

of South Florida, and Gynecologic

revision in part

The

a

ages. However, some morphometnic studies have shown that the ultrasound appearance of the junctional zone is significantly thinner than that depicted in MR images (8). In this study, histologic examina-

(5,9).

a

(H.K.B.)

images.

MD

ing of these features is the inner bamma of myometnium, which is observed as a bow-intensity band on T2weighted spin-echo (Fig 1) and lowflip-angle gradient-refocused MR images. This portion of the myome-

pathologic

myome-

Box 6, University of Pathology (S.V.N.)

Correlation MR Imaging’

HE

or ultrasound

179:409-413

(B.S.S.), Departments

MD a James V. Fiorica, L. Silbiger, MD

diagnosis and staging of cervical and endometnial carcinoma, adenomyosis, and !eiomyomas have improved significantly with the use of magnetic resonance (MR) imaging to visualize the uterus and adnexa (14). Certain features of the soft tissues are neveabed in ways that are less evident in computed tomognaphic scans

854.92

trium,

Zone: and

T

Uterine zonal anatomy as visualized on T2-weighted (repetition time, 2,500 msec; echo time, 80 msec) magnetic resonance (MR) images consists of a high-intensity central (endometrial) zone, a subjacent lowintensity junctional zone of rnyometrium, a moderately intense zone of myometrium, and a thin, low-intensity subserosal zone of myometrium. To better define the histologic correlates of these diagnostically significant zones, T2-weighted MR images of 17 in vivo and 13 extirpated human uteri were compared with histologic sections of 17 uteri stained with hematoxylin-eosin, Mallory trichrome, and immunofluorescence staining for actin. Morphometric and electron microscopic observations of uterine surgical specimens were also made. The observations indicate that both the junctional zone and the subserosal zone consist of compact smooth muscle fibers with little extracellular matrix compared with the myometrium proper. Also, the junctional zone is divided into a compact region and a transitional region. The compact region correlates well with the hypointense MR appearance of the junctional zone.

V. Nicosia, a Martin

(P.S.H.,

L.P.C.,

M.L.S.),

12901 N 30th St. Tampa, Oncology (J.V.F.), H. August from the

Foundation.

15; revision University Address

received of South reprint

College

of

FL 33612-4799; Lee Moffitt HosDecember Florida Presirequests

to

transmission

and

scanning

microscopy. Results scribe the microscopic the junctional zone

electron

of this study deanatomy of and how it difof the

myomedata from

histologic slides were compared with measurements obtained from MR images in an effort to define the structunal correlates of this diagnostically significant landmark. MATERIALS

AND

MR images

and histopathologic

mens

from

a total

METHODS speci-

of 47 women

were

or-

ganized

into three groups: Group 1 consisted of sagittal MR images of the fresh uterus in vitro acquired within 1 hour after surgery with a 1.5-T imager (Signa; GE Medical Systems, Milwaukee) set for Ti- (300/20 [repetition time

msec/echo

time

msec])

and

T2-

weighted (2,500/80) spin-echo pulse sequences. These images were matched to corresponding histopathologic sections in the sagittal orientation of the anterior wall of the same uterus. A plastic marker

was used

to ensure

that

corresponding

MR sections and histopathologic slide sections matched. None of the 13 women represented by the images in group 1 were using birth control pills, four were 409

in

late

luteal

menses,

phase,

and

four

five

had

were

scope

irregular

postmenopausal.

Group 2a consisted of in vivo MR images obtained in 17 healthy women of meproductive age. These were subdivided into a subgroup of images women aged 25-39 years using

birth

control

pills

obtained who were

pills.

Images

imager

were

acquired

(Magnetom;

tems,

Iselin,

(Signa;

NJ)

and

pulse

sequences

gap,

with

gittal

plane

Group slides

were

and

specimens

tron microscopy 17 women. These

from

(300/20)

sections,

50%

in the midsa-

(Use ords

of patient

for into

of reproducand seven (aged

staining

was

pared from the fundic portion that was most representative uterus; (c) sections including from

endometrium

tact layers tears, logic

sity

of

(d)

in-

folds,

tron microscopy, stored in phosphatebuffered saline solution to be used for hematoxylin-eosin or Mallory tnichmome staining, or frozen in sections for actin staining, transmission electron microscoRoutine histologic hematoxylin-eosin staining

was

(12).

Specific

performed

on frozen

of para-

with

rho-

(i3).

for transmission

microscopy were hyde in 0.i-mol/L and

sections

phalloidmn

electron

fixed in 2.5% glutaraldephosphate buffer (pH

postfixed

in

1% osmium

tetrox-

ide, dehydrated through graded alcohol, infiltrated with epon 812 (Polysciences, Wamnington, Pa), sectioned at 90-nm thickness,

stained

uranyl

acetate,

electron

with

and

examined

microscope

Electronic

lead

(model

Instruments,

For scanning specimens

gen-cooled

electron

were

frozen

2-methylbutane,

citrate

with

410

a

with

Radiology

a scanning

an Ga).

microscopy, liquid

nitro-

fractured

and postfixed in 2.5% glutamaldehyde, hydrated through graded alcohols hexamethyldisilazine (Polysciences),

examined

and

301; Philips

Nomcross, with

of South

MR images scale. All in milli-

made and

specimens, with

Florida

College

Review

Board.)

the

by perage. and

nec-

tient tion

1. facing

Sagittal viewer’s

(arrow)

MR image (2,500/80) (paleft) shows the transi-

between

compact

junctional

zone and outer myometmium of this midcycle uterus. Chemical shift artifact obscures the subsemosal zone of myometmium of the anterior (A) and posterior (P) walls of the uterus.

Univer-

of Medicine

Examination

Macroscopic examination of the slices of fundic portions of the uterus, which included endometnium, myometrium, and serosa, showed a longitudinal directionality of the dense, smooth muscle bundles making up the area corresponding to the junctional zone. The area of transition from junctional zone to the myometrium proper exhibited the classically described interwoven arrangement of smooth muscle bundles, collagen bundles, and larger vessels. Smooth muscle bundles in the subserosal region were arranged parallel to the surface of the uterus. Observations obtained by means of bow-magnification scanning electron microscopy confirmed these macroscopic observations (Fig 2).

of actin

specimens

damine-linked

Specimens

included tnichrome

staining

formaldehyde-fixed

7.4)

microscopy.

methods and Mallory

were history

in compliance

Macroscopic of Specimens

of uterus of corpus all layers

rips,

electron

from

calibrated expressed

Figure

RESULTS

Wedges of anterior wall of the fundic uterus were appropriately fixed for elec-

scanning

images,

slides

of reproducwomen.

measurements of anterior uterine

directly

to clinical

and

hema-

60-74

and major artifacts); foci not interfering

and

was

with

performed

to semosa;

(ie, free of major

proper,

made

Measurements

Institutional

(e) any pathowith nor distorting normal tissue lamination; and (f) area of tissue sample being measured free of cysts, adenomyosis, endometniosis, carcinoma, or other pathologic conditions.

py,

made

blind

with hematoxylmn-eosin. Mallory tnchrome staining was also available for seven women of reproductive age and three postmenopausal women. Selection criteria included (a) presence of definite endometnium; (b) slides pre-

tissue

were

of

transitional

Mallory-stained

with a superimposed, measurements were

meters.

women

Histologic

wall

sons

that were available were subdivided

were

zonal sectioned

of histopathologic for elec-

zone, myometnium

and

Corresponding the sagittally

widths

endometrium,

of specimens from women tive age and postmenopausal

processed

postmenopausal

years).

Sys-

slides

a micrommagnifica-

mean

including

toxylin-eosin-

a l.0-T

Ti-

10 specimens from women tive age (aged 29-35 years)

of the

zones

zones

used.

2b consisted

Comparisons

subserosal

spin-echo

5-mm

tion. uterine

performed

uterine wall with at 40 times original

zone,

imager

and a 256 X 256 matrix

of anterior eter stage

were

of histopathologic

junctional

Medical

(500/80)

measurement

compact

Systems).

and T2-weighted

in 10 not

by

Mass).

studies

junctional

with

a i.5-T

Peabody,

a subgroup

Siemens

GE Medical

USA,

women aged birth control

and

of images obtained in seven 26-34 years who were using

(Jeol

Momphometnic

electron

deand and

micro-

Morphometric Correlations Images Matched to Histologic Slices of the Same Uteri

of MR

Direct morphometnic correlations of in vitro MR images and histopathologic slices of the anterior wall of the same uteri are reported in Table 1. The thickness of the endometrium was consistently less in the histobogic slices than in the MR images due to the absence of luminal mucus. However, the position and thickness of the junctional zone and outer myometrium correlated well, with the exception of the junctional zone of women with irregular menses. The compact portion of the junctional zone in this group tended to be

Figure

2.

Scanning

electron

micrograph

of

a uterine specimen provides a macroscopic view of the uterine zones: endometrium (EN), junctional zone (JZ), and vascular straturn (SV) (original magnification, X31.5).

thicker (mean, 5.3 mm) than those in women with regular cycles in the bate buteab phase (mean, 3.4 mm) and postmenopausab women (mean, 2.9 mm). Morphometric Images

Studies

of MR

The junctional zone in women reproductive age was observed most

prominently

visible

of to be

in the

T2-

weighted MR images. Within the laminar appearance of the uterus at MR imaging, the endometnium was observed to have a mean thickness of 3.2 mm, the subjacent junctional zone thickness maining,

mm

was outer

(Table

significantly

5.4 mm, and myometnium

2). The thinner

the

rewas 10.5

endometnium

was

in women

who

were using birth control pills. The subserosab bamina was too thin to be observed in the MR images due to chemical shift artifact. While the transition

between

and endometnium transition between and myometnium tinct.

This

transition

junctional

zone

was distinct, junctional proper was was

the zone indis-

graded

May

and

1991

Table 1 Mean Thickness Histologic Slices

of Uterine Zones of in Vitro Uteri

Measured

from

MR Images

Thickness Source

of

and

Corresponding

(mm)

Junctional

Uteri

Endometnium

Women in late luteal phase (n = 4) (aged 39-48 y) MR images H-E-stained slices Women with irregular menses (n = 4) (aged 27-46 y) MR images H-E-stained slices

Outer

Zone

Myometrium

2.7 ± 0.9 1.4 ± 0.5

8.0 8.7 (3.4 5.3

0.6 0.3 0.7, 0.8)

6.6 ± 2.5 7.0 ± 3.1

2.1 ± 0.6

8.0 ± 1.18 10.9 ± 2.1

9.4 ± 2.9 8.1 ± 2.9

1.1 ± 0.7

bundles. More collagen was observed between the individual muscle bundles in this region of junctional zone with interwoven bundles running along diagonal, horizontal, and ventical pbanes in the cephabocaudab axis of the uterus (Fig 3b). Just below the less organized or transitional region of the junctional zone was a thin band where the muscle bundles again ran circumferential to the endometnium before they became more random in orientation in the more external region below. There, a much higher proportion of cobbagenous fibers and intercellular space separated the bundles of smooth muscle. In this transitional region of the junctional zone, arteries were dispersed within fields of cobbagen fibers and loosely defined muscle bundles and individual muscle fibers (Fig 3c). The uterine arcuate yes-

± ± ± ±

(5.3 ± 0.9, 5.6 ± 2.4) Postmenopausal

women (n = 5) (aged 50-60 y) MR images H-E-stained slices

2.4 ± 1.0 0.8 ± 0.4

7.9 ± 4.8

7.4 ± 2.24 6.3 ± 2.2

8.1 ± 2.4 (2.9 ± 0.9,

5.2

1.8)

Note-Measurements, shown as the mean ± standard deviation, are one-half the width of the hyperintense luminal portion of the sagittal uterus on MR images. This includes some luminal mucus, making the MR measurements of the endometmium thicker than histologically observed. Measurements in parentheses are those of the compact and transitional portions, respectively, of the junctional zone. H-E = hematoxylin-eosin.

Table 2 Thicknesses Histopathologic

of Uterine Slides

Zones

Measured

from

MR Images

Not Matched

Thickness Source

Nonusers

(a

of birth

Endometnium

control

(n

Note.-Measurements

are mean

5.44 ± 2.46

10.51 ± 2.02

1.36 ± 0.41

4.95 ± 1.77

11.83

* standard

of the

smooth

toxybin-eosin and tnichnome, the combined junctional zone observed in tnichnome-stained slices generally included more of the myometnium proper. This was probably due to the improved contrast and visual discnimination between smooth muscle and connective tissue in tnichromestained slices. As a result, a more ac-

made with made direction-

muscle

bundles

of

the junctional zone was primarily parable! to the endometnium or stratum basale (Fig 3a). The homogeneous appearance displayed by the myometnium stained with hematoxy!in-eosin made the laminar zonation difficult to distinguish. staining

laminar While

However, revealed

composition there

phometnic Volume

Mallory a more

was

of the no

difference 179

a

tnichrome distinct

Number

significant

between 2

uterus. mor-

hema-

± 4.07

deviation.

Examination

Histologic observations hematoxylin-eosin staining clean only that the general ality

Outer

Myometrium

3.22 ± 0.99

intermediate in signal intensity between the low-intensity junctional zone and the higher-intensity myometnium proper (Fig 1). Histologic

Zone

pills

10) (aged 25-39 y) of birth control pills 7) (aged 25-39 y)

Users

(mm)

Junctional

of

MR Images

to

curate possible.

measurement

of zonation

was

With trichnome stain, the bamina immediately subjacent to the endometnium appeared as a compact negion

of longitudinally

oriented

smooth muscle corresponding to the junctional zone. This compact, subendometnial region of the junctional zone runs almost exclusively parallel to the surface of the endometnium. Beneath the compact region, the junctional zone loses its tightly compact organization of smooth muscle

sebs usuabby served as the boundary of the transitional

outer portion

of the junctional zone and the myometnium proper. Fluorescence staining for actin fibaments revealed densely packed smooth muscle fibers in the junctional zone and little interceblubar space. Progressively more intercellular space, between both individual musc!e fibers and fiber bundles, was noted as observations were made in the direction of the subsenosa. The thin (0.5-mm) subsenosab bamina of smooth muscle bundles were tightly packed, with little intercellular space between fibers or bundles that were predominantly oriented panabbel to the serosal surface. The histologic appeanance of the subserosa! “capsule” was often similar to that of the junctionab zone (Fig 3d). Morphometnic observations of histologic specimens are reported in Table 3. Ultrastructural

Examination

Transmission electron microscopy confirmed the observations of minima! extnacellulan space between the individual fibers and bundles in the junctional zone and subserosal lamina. The area of transition between the junctional zone and myometnium proper showed increased space between individual muscle cell fibers. In the outer myometnium, both the extracelbular space and interbundbe space were dramatically increased. DISCUSSION The uterus consists

zonal anatomy of the as classically described of (a) the endometnium Radiology

human (14,15) com-

a

411

posed of the mucosal stratum functionalis and the region of glandular interdigitation into the myometrium or stratum basale; (b) the subendome-

Figure 3. (a) Light micrograph shows the basal stratum of endometrium (E) and the subjacent compact myometrium of the junctional zone (JZ) (original magnification,

trial

tional muscle

region

of densely

packed

smooth muscle bundles with an onentation primarily parallel to the endometnial stratum basale (14); (c) the stratum vascubane, where the arcuate arteries are branching through the randomly oriented and loosely organized smooth muscle bundles of the myometnium proper; and (d) the thin subserosal zone of dense circumferentialby oriented smooth muscle fibers forming an irregular capsule or external uterine boundary that becomes continuous with the myosabpinx. Our observations confirm this description at the microscopic and histologic levels and correlate them with observations and measurements of MR images in this and previous studies (5,8,16-20). A high degree of variance was to be expected in normal human uterus due to the cyclic nature of the tissue and various ages of the women. Even so,

our

observations

generally

con-

curred with other previously reported observations (5,8,16,19). The divergent observations indicate that the normal thickness of these uterine zones is as yet undetermined because of the wide variation of imaging conditions,

menstrual

phases,

and

meth-

ods of measurement. It is interesting that the junctional zone appears to maintain the same thickness after menopause, while the other three zones are reduced. The increased thickness of the junctional zone in women with irregular cycles is also interesting and deserves further study. The correlation of the anatomic region referred to as subendometnial myometnium and the junctional zone visualized with MR imaging is evidenced by corresponding histologic location and thicknesses. Also, the contrast behavior of the junctional zone seen at MR imaging is consistent with that which would be predicted for an area of compact smooth muscle bundles, which tend to have a short T2 (21). The decreased extracebbubar space of the junctional zone presumably corresponds to a bower water content, which has been reported by McCarthy

et al (9) and

which

could

dramat-

icalby affect the Ti value of this negion, since small changes in water content cause barge variations in Ti values of tissue (22). Orientation of fibers may also contribute to the appearance of the junctional zone on 412

a

Radiology

X470).

(b) Transitional

fication, dles

portion

zone shows more bundle orientation

X470).

are

randomly

extensive

areas

of the junc-

random smooth (original magni-

(c) Smooth

muscle

oriented

and

of connective

cell bunseparated

tissue

by

space

in

this light micrograph of the outer myometrium (original magnification, X470). (d) Subserosal lamina (SL) forms a thin capsule of smooth muscle (original magnification, X470).

a.

T2-weighted images (23,24). Understanding the histologic conrelation of features such as the junctional zone is important for know!edgeable interpretations in diagnostic imaging. It may be useful to realize that, were it not for chemical shift

artifact,

the

thin

subserosab

uter-

me zone might be visualized on T2weighted images, since its histologic composition is similar to that of the junctional zone. Mitchell et al (8) have described the apparent discrepancy between the thickness of the junctional zone observed at MR imaging and that observed at ultrasound. Our morphometric correlations lead us to speculate that the inner compact portion of the junctional portion could comespond to the subendotheliab hypoechoic area seen on ultrasound images, while a broader zone, including a transitional zone that blends into the outer myometnium, is visualized on MR images. In summary, the myometnium exhibits an inner one-third representing the junctional zone, which consists of two subdivisions: an inner compact portion and an outer transitiona! portion that blends into the myometnium proper. The outer onethird

of the

uterus

consists

Acknowledgments: BS,

Margaret

Todd Bryant,

Hazelton, MS.

C.

of smooth

muscle and collagen bundles that are more widely separated. This region also has the greatest vasculanity (iO). The outermost, thin subserosab megion of compact myometrium forms a capsular boundary to the myometnium. Further microscopic, physiologic, biophysicab, and morphometric correlations of the uterine zones will be required for a fuller understanding of these significant uterine featunes. U

Slater,

b.

BS, Alicia

d.

References 1.

ing

S.

in

Reson 2.

Hricak LG,

3.

and

Imaging

1988;

H,

Stern

Winkler

4.

ML,

at

Radiology Togashi

1.5 1,

imagMagn

4:59-66. Fisher

Lacey by

JL,

with

1987; K, Ozasa

aging. Radiology Lee JKT, Gersell ton

resonance gynecology. MR. CC. MR

Shapeero Endometrial

imaging.

Radi-

1987; 162:319-324. DC, Mintz MC, Spritzer CE, masses: MR imaging observa-

larged uterus: adenomyosis 5.

JL,

staging

ology Mitchell Adnexa! tions

Magnetic

obstetrics

carcinoma

Amanda

Garces, Karl Muffly, PhD, Mitchell Hoffman, MD, Teresa Dula, RT, Cindy Philips, RT, Anna Parsons, MD, Donald Diebel, BS, James Lefler, BS, and Angelia Hayes, BS.

McCarthy

Picus

US

and

CT

162:319-324. H, Konishi

et a!.

correlation. I, et a!.

differentiation and leiomyoma

En-

between with MR

im-

1989; 171:531-534. DJ, Balfe DM, WorthingD, Gapp

C.

The

uterus:

May

in

1991

15.

Finn

CA,

Porter

DG.

gy handbooks. Mass: Publishing

16.

134. McCarthy pelvic

the

and

abnormal

specimens.

Radiology

10.

1985;

157:175-179. 6.

Worthington Uterine gy

JL,

Balfe

neoplasms:

1986;

MR

JKT,

imaging.

et a!.

11.

Radiolo-

8.

Mitchell

ing. Radiology

1988;

DC,

Pennell

RG,

Blum

12.

L, Hilpert

L, Rifkin

discrepancy

images.

K, et a!. MR imag-

166:111-114.

Schonholz

uterus:

MR

K, Itoh with

PL,

MD. between

Radiology

US

1990;

174:827-

13.

831. 9.

McCarthy Uterine

S. Scott junctional

C,

Majumdar

zone:

S. et a!.

MR

ter content and relaxation diology 1989; 171:241-243.

study

of wa-

properties.

Ra-

14.

a

Number

ture

2

JOW,

Tarbit

spin-echo

images

of the

of the

Armed

BE,

Hricak

imaging: . Radiology

Haynor

Institute

of the

musculature

muscles

Gynecol

and

of the connective

1966;

H, Jaffe

Mack

LA,

Montana appearance the

Soules

MA,

1986;

stimulaMR.

Shu-

Moss AA. of the normal

menstrual

Radiology

1986;

Uterine

effects of hormonal 1 986; 159:123-126.

DR.

with

cycle:

uter-

MR

stud-

161:459-462.

20.

Janus CL, ic resonance

Wiczyk HP, imaging

cle.

Reson

21.

Hricak

Magn H,

Dooms

Laufer of the

Imaging GC.

N. Magnetmenstrual cy-

1988;

6:669-674.

McNea!

JE.

et a!.

MR imaging of the prostate gland: anatomy. AJR 1987; 148:51-58. 22.

Fleischer

AC,

Entman

Kalemeris

55, James

piction

of normal

trium

with

GC,

AE Jr. and

Med

Machin

abnormal

1986;

23.

Mansfield P. Morris PG. in biomedicine. London:

24.

Fullerton entation

normal JE,

Sonographic

histopathologic

Ultrasound

de-

endomecorrelation.

5:445-452. NMR imaging Academic Press,

1982; 28-30.

stain-

Forces

and

Wiczyk HP, Janus CL, Richards CJ, et a!. Comparison of MRI and US in evaluating fo!!icular and endometrial development throughout the normal cycle. Ferti! Steril 1988; 49:969-972.

uterus:

study

cycle

19.

MH.

of Pathology. 3rd ed. New York: McGrawHill, 1968; 75-76. Borland K, Ehrlich HP, Muffly K, et a!. Interaction of rat sertoli cells with a collagen lattice in vitro. In Vitro Cell Dev Biol 1986; 22:661-669. Schwalm H, Dubrauszky V. The struc-

J Obstet

179

on

methods

uterus:

Volume

Beilby

an anatomic and clinicopathologic (abstr). Radiology 1985; 157(P):310. Luna LG, ed. Manual of histologic ing

Zones

C,

The blood supply of the uterus. J Obstet Gynaec Br Comm 1970; 77:673-689. Worthington BS, Powell M, Womack C, Buckley J, Symonds EM. The low intensity band

Togashi K, Nishimura Adenomyosis: diagnosis

and

Lee

159:725-730.

7.

of the

DM,

Farrer-Brown

of varia-

menstrual

RB.

ies.

of normal

Female

Demas

us during

correlation

J.

assessment

Radiology

man WP, Changing

MR-anatomic

Acton, 1975;

use of oral contraceptives. 160:119-123.

tion

vitro

biolo-

uterus. Group,

C, Gore

MR

during

MR 18.

Reproductive

1, The Sciences

5, Tauber anatomy:

tions

17.

Vol

GD, Cameron of tendons in

and

its effect

ology

1985;

IL, Ord VA. the magnetic

on T2 relaxation

Onfield

times.

Radi-

155:433-435.

human tissue.

Am

94:391-404.

Radiology

a

413