Abdominal Masaaki

Ebara,

Nobuyuki

Sakae

MD

Sugiura,

#{149}

MD

Watanabe, Masao Ohto,

MD MD

#{149}

and

#{149} Kazuhiko

Gastrointestinal

Kita,

Fukuo

MD

Kondo,

#{149}

MD

#{149} Masaharu

Yoichiro

#{149}

Radiology

Yoshikawa,

Kondo,

MD

MD

MR Imaging of Small Hepatocellular Carcinoma: Effect of Intratumoral Copper Content on Signal Intensity’ Small ( < 5 cm) hepatocellular carcinomas (HCCs) in 45 of 112 patients (40.2%) had a high-signal-intensity pattern (relative to that of liver) on Ti-weighted magnetic resonance (MR) images. To identify the cause of this pattern, specimens of histologically defined HCC with high- (n = 21), iso- (n = 6), and low- (n = 22) intensity

patterns

were

histologically

evaluated

for paramagnetic ionic forms of metals. The incidence of steatosis, clear cell formation, and copper accumulation was statistically (P < .01) higher in tumors with the high-intensity pattern than in those with other paHerns. Of 17 HCCs that stained positive for copper, 16 (94%) had a high-intensity pattern; the pattern of one tumor (6%) was isointense. All 2i tumors with the highintensity pattern had at least one of the findings of steatosis, clear cell formation, or copper accumulation. These features were evident in only four of 28 tumors (14%) with an isoor low-intensity pattern. Thus, copper accumulation might be a cause of the high-intensity pattern on Tiweighted images of small HCCs. Index terms: Copper Liver, MR studies, 761.1214 #{149}Liver neoplasms, 761.321 #{149} Liver neoplasms, MR studies, 761.1214 #{149} Neoplasms, MR studies, 761.1214 Radiology

1991;

M

resonance (MR) imaging is now considered to be (and sometimes superior) to

equal computed diagnosis

tomography of liver tumors

signal-intensity

images

Tissue

levels

might

have

intensity patterns in MR imaging. In some cases of HCC, copper accumulates more in the tumor than in the surrounding liver parenchyma. This

compelled

us to consider

The

purpose

was to determine pattern of HCC ages is attributable

hancement contained liver

of this

March

1985

and

March

contrast-enhanced

and angiography, Imaging

CT

in

104

by Kitagawa

et al (pp

623-

The

Methods

early

diagnosis increase

of HCC and

rapid

was decline

coil with Sagittal,

26

coro-

(US)

was

to deterthe space-ocPatients were

performed

the area that included cupying lesions in the liver. supine and ing imaging.

were respiring Sections were

and the raw data 128

or 256

matrix

x 256

matrix

size was

elements.

The

durthick,

128 display

x

size was 512 x 512, and a radio freof 21 MHz was used. Spin-echo or T2-weighted (300 or 1,500/24,

time [TEl msec) Sections ous

normally 10 mm

were

sections

time

images 10 mm

had

[TR]

were thick,

a 2-mm

mseclecho

obtained. and

gap

contigu-

between

them. From the same section in which a focal lesion was visualized most clearly, additional SE images were obtained with inversion recovery (IR) (1,500/40/400 or 500) (TR/TE/inversion time [TIJ msec). The

signal

intensity

of HCCs

patterns:

low,

were

tients

with

hepatic

either

a CTIF9800

on

MR

im-

that of the surroundwas classified into iso,

high,

and

performed mass

mixed

in all pa-

lesions

by

using

(GE Medical

Systems, Version 7 scanner

or Quantex Medical Systems,

and scans

Tokyo).

Con-

non-conat i-cm

in-

tervals through the liver were usually obtained. Contrast medium (60% iothalamate sodium [Conray; Daiichi Pharmaceutical, Tokyoj) was administered as a rapid intravenous bolus at a rate of approximately 7.5 mlisec and a dose of 0.7 mLlkg to all patients. mass lesions lesion

was

In patients in the

liver,

with only

multiple the

main

evaluated.

(92.9%).

1991

See also the article 628) in this issue.

axis

had a (Oxford Carteret,

nal, and axial images could be acquired. Prior to MR imaging, real-time ultra-

trast-material--enhanced trast-material-enhanced

1990,

scanning

patients

a whole-body of 53 cm.

included diameter

CT studies

age ranged from 37 to 75 years. The diagnosis of HCC was established by means of biopsy with a 21-gauge needle guided with sonography in 108 patients (96.4%) and/or by use of diagnostic imaging, including

NJ) and an inside

Milwaukee) (Yokogawa

the longest (86 men, by use of MR imaging. Patient

dy-

re-

Ohio)

magnet Technology,

four (6).

METHODS

than 5 cm in in 112 patients

smaller detected

Heights,

ages, compared with ing liver parenchyma,

of paramagnetic metals in HCCs and surrounding

AND

Highland

0.5-T superconducting Superconducting

study

if the high-intensity on Ti-weighted imto contrast en-

parenchyma.

women)

RSNA,

intratumoral

copper as a cause of the high-intensity pattern of HCC on Ti-weighted

MR images.

tional,

40, or 80) (repetition

signal-

with

vealed arterial neovasculature (8,9). The MR imaging system used (Picker Interna-

quency (SE) Ti

ions

in creating

observed

angiography

mine

sign

of paramagnetic

a role

was

CT and hepatic

sonography

Ti-

ring

ble.

Between

©

on

(6). The

corresponds to a fibrous capsule (6,7); the high-intensity pattern on Tiweighted images is considered to be caused in part by steatosis (6), but other causes for this pattern are possi-

HCCs were

‘From the First Department of Medicine (ME., SW., K.K., MY., N.S., MO.) and Second Department of Pathology (F.K., Y.K.), School of Medicine, Chiba University Hospital, 1-8-1 Inohana Chiba-Shi, Japan (280). Received December 18, 1990; revision requested January21, 1991; revision received May 6; accepted May 14. Supported in part by a grant-in-aid for cancer research from the Ministry of Health and Welfare of Japan. Address reprint requests to M.E.

(CT) for the such as he-

pattern

weighted

enhancement

namic

mangioma (1), metastatic liver cancer (2-5), and hepatocellular carcinoma (HCC) (6). Characteristic findings of HCC demonstrated with MR imaging include a ring sign (6,7) and a high-

MATERIALS

180:617-621

trast

AGNETIC

made

when in con-

Abbreviations: CBP = copper-binding tein, HCC = hepatocellular carcinoma, version recovery, SE = spin echo, TE time, TR = repetition time.

=

proIR = inecho

617

Histologic

Evaluation

Table

Biopsy specimens of HCCs and noncancerous liver parenchyma were obtained with

a 21-gauge

modified

Menghini

dle (Sonopsy; fled Tru-Cut

Hakko, Tokyo) needle (Quick-Cut;

respectively,

with

tissue

was formaldehyde

and

then

fixed

US

specimens

were

evaluation

of staining

These

tumors

of quantity

Tumor

>ibut >2but3 >3but4

a wide

patterns

on

on TI-weighted

IR Images

in

Pattern

No. of Patients

Low

Iso-

High

35 35 20

14(40) 13(37) 8(40)

4(h) 7(20) 2(10)

0 1(3)

22

8(37)

3(14)

17(49) 14(40) 9(45) 5(23) 45 (40.2)

8(7.1)

2

>4but5

whose

various

Size

(cm)

Histologic

sufficient

for

of HCC

Mixed

Liver

in 49 patients

covered

nal-intensity

(10).

in paraffin.

was assessed

Patterns

Intensity

a modiHakko),

in a 10% neutral buffered solution, was dehydrated,

embedded

material

of Signal-Intensity to Tumor Size

Relation

nee-

and

guidance

1

Prevalence

Total

for

112

(38.4)

43

16 (14.3)

1(5) 6(27)

metals.

range

of sig-

TI-weighted

im-

Note.-Intensity sea are percentages,

is defined in comparison which may not equal

with surrounding liver parenchyma. 100% because of rounding

Numbers

in parenthe-

ages, but none were of mixed intensity. Sections of HCC and noncancerous liver

parenchyma

from these

49 patients

were

stained with hematoxylin-eosin and also with rubeanic acid (II), Prussian blue (12), and the Okamoto-Aoyama method (13) for the identification of copper, iron, and manganese, respectively. In rubeanic acid staining, greenish to black pigments mdicate a copper-rubeanic complex. Each slide was examined by two pathologists (F.K., Y.K.) without prior information about the MR signal intensity of the tumors. The degree of staining with the

rubeanic

acid method

to 2+,

was scored

from

positive.

ing was

The grade

correlated

with

of cancerous

Pathologic

of copper the signal

tissue

on

MR

Patterns

of

on 12-weighted

HCC

Patients

Low

Iso-

High

24 29 18 20

0 1 (3) 1(6) 0

2(8) 2 (7) 1(6) 0

22(92) 26 (90) 15(83) 14(70)

0 0 1(6) 6(30)

91

2 (2)

5 (5)

77 (85)

7(8)

>ibut2 > 2 but 3 >3but4 >4but5 Total Note.-Numbers

in parentheses

are percentages,

which

may

not equal

tern and pathologic ated

for 21 HCCs

pattern

between Findings

in

Signal-Intensity Small HCCs

Pattern

on Ti-weighted

with

was evalu-

Intensity Pattern

the isoin(n = 22)

Phantom

Study

To demonstrate per can increase

that signal

on MR images,

test tubes

ter)

with

various

low levels of copintensity of HCC

(13-mm

complexed to cerunder the same

conditions

same

with

the

MR

equip-

ment used in the studies of patients. TIand T2-weighted images were obtained (at 24#{176}C) of ii test tubes, one with distilled water and the the rest with dilutions of

CuSO4

(0.1-4

mol

of copper

per milliliter

of water), and of six test tubes, one with distilled water and the rest with dilutions of ceruloplasmin (0.125-2 imol of ceruloplasmin per miffiliter of water). TI and T2 times were calculated. Ceruloplasmin (Sigma Chemical, St Louis) is a lyophiized powder containing approximately 30% protein;

it consists

of a balance

nly sodium chloride and sodium and has 1.5-2.5 p.g of copper per

618

Radiology

#{149}