Sandro

Sironi,

#{149} Tito

MD

Livraghi,

MD

#{149} Alessandro

DelMaschio,

MD

Small Hepatocellular Carcinoma Treated with Percutaneous Ethanol Injection: MR Imaging Findings’ Fifty-seven magnetic resonance (MR) imaging examinations were obtained at 0.5 T in 19 patients before and after percutaneous ethanol injection (PEI) for 23 hepatocellular carcinoma (HCC) lesions less than 3.5 cm in diameter. Seventeen patients also underwent MR imaging 6 months after completion of therapy. In 11 patients, computed tomography was performed before and after treatment. After PEI, fine-needle biopsy spedmens were obtained in all cases. Before treatment, HCC lesions had low signal intensity on Ti-weighted finages in 13 cases, had the same signal intensity as normal liver parenchyma in six, and had high signal intensity in four; all 23 tumors had high signal intensity on T2-weighted images. After treatment and at 6-month followup, all 21 lesions that contained no malignant cells at fine-needle biopsy had high signal intensity on Tiweighted images and had low signal intensity on T2-weighted images. The remaining two HCC lesions in which tumor necrosis was not achieved with PEI displayed a different MR pattern, since the residual neoplastic tissue showed no change in signal intensity on either Ti- or T2-weighted

images.

The

authors

conclude that MR imaging may be useful for evaluating the effectiveness of PEI in achieving tumor regression. Index

terms:

Alcohol,

761.1299

ventional procedure, 761.1299 plasms, MR studies, 761.1214 plasms, therapy Radiology

1991;

the

#{149}Liver, interLiver neo#{149} Liver neo-

with US-guided

ultrasound fine-needle

CREENING

and opsy

of patients

chronic,

with

active

detection small,

(US) bi-

cirrhosis

hepatitis

of increasing

to

numbers

asymptomatic

of

hepatocellular

carcinomas (HCCs) (1). Patients who undergo surgical removal of tumors at this stage live significantly longer than do untreated patients (2). However, many patients with small HCC are not considered for surgery because of multifocality of the lesions or coexistence

of severe

liver

dysfunc-

tion (3). For this group of patients, percutaneous ethanol injection (PEI) therapy with US guidance has been adopted, and its clinical efficacy seems to be excellent (4-9). The effectiveness

of this

evaluated phy

treatment

with

(CT)

with

has

computed contrast

tomogra-

hancement, angiography, needle biopsy (4-9). Serum markers

such

been

material

en-

and finetumor

as a-fetoprotein

(AFP)

have been measured to evaluate the efficacy of the treatment. Measurements of AFP levels, however, have limited value because only a minority of patients with small HCC have an elevated AFP level (10,11). Recently, magnetic resonance (MR) imaging has become an important diagnostic technique in the evaluation of focal liver lesions (12-15). The purpose of this prospective study was to assess the potential role of MR imaging in the evaluation of HCC lesions

managed

with

PEI

MATERIALS

therapy.

Scien-

tific Institute San Raffaele, University Hospital, Olgettina 60,20132 Milan, Italy (S.S., A.D.), and the Department of Radiology, Ospedale Civile, Vimercate, Milan (T.L.). Received January 10, 1991; revision requested February 21; revision received March 18; accepted March25. Address reprint requests to S.S. ORSNA, 1991

The study

METHODS

patients with 23 HCC lesions with PEI with US guidance.

Nineteen were treated of Radiology,

AND

group induded 51-72 years.

women, aged for treatment

one

patient

subsequently

gery

(liver

transplantation).

serum

or

is leading

#{149}

180:333-336

Department

S

14 men and five Requirements

were

that the lesions have a of less than 3.5 cm and be focal

diameter and detectable with US. Four patients had two lesions and 15 had a single lesion. Maximum lesion diameter was 2.1-3.4 cm. In all cases the diagnosis of HCC was established with US-guided biopsy. Only

was

AFP

underwent

sur-

In all patients,

measured

before

and

after

treatment. The diameters of all tumors were measured after therapy. US procedures were performed with use of a convex, 3.5-MHz probe and a biopsy attachment (650; Aloka, Tokyo). Sterile 95% ethyl alcohol was injected with a 17.7-cm,

22-gauge

Dickinson,

spinal

needle

Rutherford,

anesthesia. The injected at each

(Becton-

NJ), without

total amount session was

local

of alcohol determined

according to lesion diameter: 2-5 mL of alcohol for lesions less than 3 cm and 6-8 mL for lesions 3.0-3.4 cm in diameter. needle was placed two or three times each session to inject alcohol at more

The at tu-

mor sites. As suggested in previous reports (7,9), the needle was left in place at the end of the injection for 30-60 seconds to reduce reflux of alcohol into the pentoneal

cavity, which sometimes causes pain. Treatment was performed two or three times per week. The total number of PEI sessions

for each

sion diameter: sions

for

patient

depended

In general,

lesions

less

on le-

six to eight ses3 cm and 10 ses-

than

sions for lesions 3.0-3.4 cm in diameter. In all cases, to determine the efficacy of ther-

apy, fine-needle was performed

biopsy with US guidance 1 month after the final in-

jection.

cases,

In nine

fine-needle

biopsy

specimens were also obtained 6 months after the completion of treatment. A 22gauge ratory,

cutting Tokyo)

needle (Surecut; TSK Labowas used for biopsy. Two

or three histologic specimens from different areas, especially at the periphery of the tumor, were obtained at each biopsy. Plain and contrast material-enhanced CT was performed in 11 of the 19 cases before treatment and within I month after the final injection. CT was performed with a CTIT 8800 (GE Medical Systems, Milwaukee) or a T 60 A scanner (Toshiba Medical Systems, Tokyo). Contrast-enhanced studies were performed with use of intravenous rapid-drip infusion of 300-350 mL

of 30% meglumine iothalamate Bracco, Milan, Italy). In seven

Abbreviations:

AFP carcinoma,

hepatocellular

ethanol time,

injection, TR

=

repetition

SE

=

a-fetoprotein, PEI

spin time.

=

(Conray; of the

=

echo,

11

HCC

=

percutaneous

TE

=

echo

333

.4

a.

b.

c.

d.

e.

f.

g.

h.

Figure 1. Radiologic images of HCC in the right lobe that became necrotic after PEt. (a) Pretreatment sonogram shows slightly hypoechoic lesion. (b) Pretreatment contrast-enhanced CT scan demonstrates the tumor as a low-attenuation lesion (arrow). (c) Pretreatment SE MR image (500/20) shows mass with same signal intensity as normal liver parenchyma. Pseudocapsule is seen as a low-signal-intensity band (arrows). The high-signal-intensity band (arrowhead) was interpreted as artifact. (d) On pretreatment SE MR image (2,000/120) the tumor (arrows) is of high signal intensity. (e) On posttreatment sonogram, the mass appears isoechoic. (0 On posttreatment contrast-enhanced CT scan, the lesion is identified as a cornpletely nonenhanced area (arrow). (g) Posttreatment SE MR image (500/20) shows lesion as area with high signal intensity (arrows). (h) Posttreatment SE MR image (2,000/120) clearly reveals the tumor as an area of low signal intensity (arrows). (i) SE MR image (2,000/120) obtamed at 6-month follow-up examination shows no changes in signal intensity of the tumor (arrows).

patients, tration

CT scanning after of contrast material

after

treatment.

tamed

before

contrast

and

after

19 patients

before

treatment

MR

and

I month

within

months

after

completion

imaging

was

performed

perconducting iba Medical use

displayed thickness

section ages

gap were

was

7.5

of 2.5

mm.

and

matrix.

mm,

with

In all

an

cases,

in only

the

inter-

(SE)

technique (TR)

echo time weighted weighted were used

(TE) of 20 msec (500/20) for Tiimages, and 2,000/40, 120 for T2images. Four signal acquisitions for Ti-weighted images and

preted (S.S.,

334

in conference T.L.,

A.D.).

Radiology

#{149}

images. techniques MR images by

used,

im-

plane.

time

msec

three

with

and

an

No motionwere used. were interradiologists

the

(over 320 pg/L) before two cases, AFP levels

After

repetition

two, for T2-weighted artifact reduction The US, CT, and

was

these values retreatment. In

therapy

level

was

cases

to under

became

(18 of 23), the smaller

after

a

PEI,

all lesions

70

HCC

treat-

showed

changes in echo pattern on US scans. Fifteen of 16 hypoechoic tumors became hyperechoic, and all seven hyperechoic lesions became inhomogeneous. One lesion that was slightly hypoechoic before ethanol injection became isoechoic after treatment (Fig la,le) In 11 patients

who

underwent

1.

mark-

ment.

Spin-echo

of 500

nor-

increased In these

lesions

secMR

were

patients,

after

AFP

levels

edly PEI.

In many

were

axial

4

in 17 of 19 paAFP

.Lg/L.

su-

The

serum

two

decreased

MR

a 0.5-T,

matrix

x 512

acquired

pa6

(MRT 50-A; ToshData were collected

x 256

in a 512

after

of therapy. with

magnet Systems). of a 256

imaging

Seventeen examination

treatment the

mal (under 20 pg/L); mained normal after

ob-

of

underwent

V.

-‘--

tients,

of

administration

the final ethanol injection. tients also underwent MR

tion

Before

of

material.

All

with

CT scans

with a section thickness intersection gap, were

4

RESULTS

bolus adminiswas performed

In all cases,

the entire liver, 10 mm and no

#{149}1

con-

trast-enhanced CT examinations before treatment, six HCCs were low attenuating and five were of the same attenuation as normal tissue (Fig ib).

After mors

treatment, none of the 11 tuwas enhanced after administra-

tion of contrast material evidence of tumor was of these

Before weighted

lesions

(Fig found

at fine-needle

biopsy.

treatment, on SE TiMR images (500/20),

had

low

signal

had mal

the same signal liver parenchyma

slightly high All 23 lesions

intensity

if). No in any

tumors

in 13 cases,

intensity as norin six, and had

signal intensity had high signal

in four. inten-

sity

on SE T2-weighted images (2,000/ (Fig ic, id). After treatment, 21 of 23 lesions had high signal intensity on Ti-weighted images and had low signal intensity on T2-weighted images (Fig ig, ih). All 21 lesions were found to be necrotic at fine-needle biopsy. At 120)

6-month follow-up MR imaging features

had

not

changed

examination, of these

noticeably

the tumors

cornAugust

1991

a.

b.

c.

d.

Figure 2. Radiologic images of HCC in which tumor tissue was found after PEI. (a) Pretreatment area of slightly high signal intensity (arrows). (b) Pretreatment SE MR image (2,000/120) shows (arrows) and right renal cyst (arrowhead). (c) After treatment, on SE MR image (2,000/120), the strates high signal intensity (arrows). (d) Surgical specimen shows the area of residual neoplastic Pathologically evident tumor corresponds to area of high signal intensity seen on posttreatment,

pared

with

MR imaging

results

after

treatment (Fig 1). Results of another fine-needle opsy performed in nine of these

confirmed

the

of tumor

evidence

of recur-

rence. After treatment, was found at fine-needle two

of the

plastic peared

23 HCC

could

ne-

mor apy

tumor tissue biopsy in

lesions.

The

sity on Ti-weighted area of high signal weighted images,

images intensity showing

and as an on T2no

changes

intensity

pattern

compared treatment

two

cases,

ment

were

performed after (Fig 2d).

number

such

liver

of HCC

to result The

however, mainly

has

be good

have

survival

rate

of HCC,

rather

of underlying

performed

has been who are

with

for surgery,

survival

(4-9,16). assessed the

low,

US

rates

seem

to

and to be

Previous studies toxic effects of etha-

nol injected into neoplastic lesions (17,18). Alcohol enters the cells by diffusion and produces coagulation necrosis, which is followed by thrombo-

Volume

180

elevated

#{149} Number

2

AFP

study,

the

levels

sonographic

tumor changed In general,

be-

became previous

fea-

and

hy-

in homogestudies

sion after PEI, showing a nonenhanced, low-attenuation area, which suggests complete necrosis of the lesion (5,7-9). In our study, CT was helpful for evaluation of tumor regression after treatment. In patients who underwent CT examinations af-

that

were not adminis-

tration of contrast material were found to be necrotic; moreover, the decrease in attenuation of the tumor at CT corresponded to the changes in signal intensity within these lesions at both Ti- and T2-weighted MR imaging. In our series, on Ti-weighted images before treatment, many HCCs appeared to have low signal intensity compared with normal liver parenchyma; in the other cases, HCCs were

(2,000/40)

displays

same

lesion

as

high signal intensity

neoplastic tissue and intratumoral image.

or slightly probably

degeneration

still

demonnecrosis

higher because

(i2-i4).

(N).

signal of fatty

Before

ment, all tumors showed intensity on T2-weighted

treat-

high signal images corn-

pared with the surrounding normal liver. These MR imaging features confirm the findings reported by Ebara et al (13), Itoh et al (12), and Rummeny etal(i4). After

after ethahypoechoic

hyperechoic,

ter treatment, all tumors enhanced after intravenous

liver

adopted recently not considered

candidates

the 3-year excellent

B

liver tumor has technique most

remained

cirrhosis PEI therapy, guidance, in patients

of hepatitis

in extended

because (3).

le-

patients with (i). Surgical re-

resectability

small

is found

(5,8) suggested that such changes in US pattern do not reliably indicate necrosis of the tumor, and therefore sonography cannot be used to confirm tumor regression. Contrast-enhanced CT proved to be useful for evaluation of tumor regres-

at an early of high-risk

as carriers

section of the primary been the therapeutic (2).

had

with

level

treatment.

tumors However,

antigen and liver disease

likely

patients

for

percentage of cases series, a minority of

became

are being detected with US screening

surface chronic

of AFP

perechoic neous.

DISCUSSION

patients,

elevation

lesions

increasing

sions stage

HCC,

a small In our

useful

of treat-

in patients

In our

treat-

intensity,

of the efficacy

MR findings

after

of the

however,

in only (iO,ii).

area of residual tissue (arrows) T2-weighted

denaturwhich decrease in hiafter ther-

ment;

fore

at pathologic

the

image

lesion as area of markedly

phenomena

size that is observed (4,7,17,i8). AFP measurements are

tures of the nol injection.

confirmed

These of protein dehydration,

explain

with MR results before (Fig 2a-2c). In one of these

examination transplantation

An

also

evaluation

neo-

tissue of these lesions still apas an area of low signal inten-

in signal

vessels.

are the outcome ation and cellular

bicases

presence

without

crosis,

sis of small

SE MR

low-up became

treatment

and

at 6-month

fol-

examinations, all lesions that necrotic showed an increase

in signal intensity on Ti-weighted images and showed markedly low signal intensity on T2-weighted images compared with normal liver parenchyma. These MR findings in the successfully treated lesions confirm the results recently reported by

Kubota et al (i9). However, when comparing our data with those of Kubota et al (19), it should be noted that in our study, Ti- and T2weighted

cases, quired patients

images

whereas

were

Kubota

T2-weighted who each

obtained

in all

et al only

ac-

images. In two had a single HCC

lesion, the results of fine-needle biopsy indicated that complete necrotic degeneration of the tumor was not achieved with PEI. Both of these lesions images

were readily identified because the neoplastic

tions tensity before

maintained pattern treatment.

cases,

pathologic

on

MR por-

the same signal inshown on MR images In one of these two

examination

per-

formed after liver transplantation revealed that the tumor comprised portions that were isolated by thin septas; this pathologic finding may

explain injection

the

partial

efficacy

of ethanol

(5).

We conclude that MR imaging may be useful in enabling differentiation of necrotic HCCs from lesions in which residual tumor is still present Radiology

#{149} 335

References

after ethanol injection. Furthermore, the low signal intensity on T2weighted images of necrotic HCCs after

PEI

seems

to be

a finding

1.

of low

signal

intensity

is 2.

on

T2-

weighted sequences. Tumor necrosis in HCCs not treated with PEI is characterized by high signal intensity on T2-weighted images (12-14,20). The mechanisms responsible for the low signal

intensity

ages

after

fully

understood.

on

T2-weighted

PEI in this

study

Previous

im-

make

it possible

to evaluate

advantage

over

CT because of CT on

inherent dependence iodinated contrast material for evaluation of tumor necrosis after treatment. If our results are confirmed by further investigations of large series of patients, we believe that use of MR imaging could limit the role of fine-needle biopsy in confirming tumor regression

336

after

#{149} Radiology

PEI.

4.

#{149}

5.

of early

K, et al.

detection

Prospecof hepatocel-

lular carcinoma in patients with cirrhosis. Hepatology 1990; 12:680-687. Okuda K, Ohtsuki T, Obata M, et aL Natural history of hepatocellular carcinoma

and prognosis in relation to treatment. Cancer 1985; 56:918-928. Ebara M, Ohto M, Shinagawa T, et al. Natural history of minute hepatocellular carcinoma smaller than three centimeters

Livraghi tori C.

13.

Gastroenterology

1986; 161:309-312. H, Muto H, et aL

6.

Sheu

JC, SungJL,

Huang

CT, et al.

injection

of absolute

ethanol

tumor ultrasound

guidance

Percu-

15.

16.

small hepatocellular 7.

8.

9.

10.

11.

Intra-

under

for the treatment

carcinoma.

sound and alpha-fetoprotein: a prospective study. Hepatogastroenterology 1988; 35: 101-103. Itoh K, Nishimura K, Togashi K, et al. Hepatocellular carcinoma: MR imaging. Radiology 1987; 164:21-25. Ebara M, Ohto M, Watanabe Y, et al. Diagnosis of small hepatocellular carcinoma: correlation of MR imaging and tumor histologic studies. Radiology 1986; 159:371-

377. 14.

T, Festi D, Monti F, Salmi A, VetUS-guided percutaneous ethanol of small hepatic and abdominal

injection tumors. Radiology Shiina 5, Yasuda

12.

taneous ethanol injection in the treatment of liver neoplasms. AJR 1987; 149:949-952.

the

of the use of

N, Kim

complicating cirrhosis. 1986; 90:289-298.

studies

effectiveness of treatment. In assessing PEI efficacy, MR imaging has a real

3.

are not

(19), however, suggested that this phenomenon could be related to cellular dehydration induced by ethanol. Specffic evaluation procedures are essential to determine whether a tumor has become necrotic after PEI. Our results suggest that MR imaging may

H, Kurioka

tive study

that

specific to this study. To our knowledge, HCCs have not been reported as areas

Oka

of

17.

Hepato-

gastroenterology 1987; 34:255-261. Livraghi T, Salmi A, Bolondi L, et al. Small hepatocellular carcinoma: percutaneous alcohol injection-results in 23 patients. Radiology 1988; 168:313-317. Shiina 5, Tagawa K, Unuma T, Terano A. Percutaneous ethanol injection therapy for

the treatment of hepatocellular carcinoma. AJR 1990; 154:947-951. Livraghi T, Vettori C. Percutaneous ethanol injection therapy of hepatoma. Cardiovasc Intervent Radiol 1990; 13:146-152. Livraghi T, Sangaffi C, Giordano F, et aL 240 hepatocellular carcinomas: ultrasound features, tumor size, cytologic and histologic patterns, serum alpha-fetoprotein and HBs Ag. Tumori 1987; 73:507-512. Cottone M, Turn M, Caltagirone M, et al. Early detection of hepatocellular carcinoma associated with cirrhosis by ultra-

Rummeny E, Weissleder R, Stark DD, et aL Primary liver tumors: diagnosis by MR imaging. AJR 1989; 152:63-72. WittenbergJ, Stark DD, Forman B, et al. Differentiation of hepatic metastases from hemangioma and cysts by MR. AJR 1988; 151:79-84.

Ebara M, Ohto M, Sugiura N, et al. Percutaneous ethanol injection for the treatment of small hepatocellular carcinoma: study of 95 patients. J Gastroenterol Hepatol 1990; 5:616-626. Burgener experimental

neous

FA, Steinmetz SD. adenocarcinomas

intratumor

injection

cohol. Invest Radiol 1986; 18.

19.

20.

Fujimoto studies therapy

Treatment of by percuta-

of absolute

al-

22:472-478.

T. The experimental and clinical of percutaneous ethanol injection (PElT) under ultrasonography for

small hepatocellular carcinoma. Ada HepatolJpn 1988; 29:52-59. Kubota Y, Nakano T, Seki T, et aL Validity of MR imaging for monitoring effects of percutaneous ethanol injection for HCC. Hepatogastroenterology 1989; 36:262-265. Ohtomo K, Itai Y, Furui 5, Yashiro N, Yoshikawa K, lio M. Hepatic tumors: differentiation by transverse relaxation time (T2) of magnetic resonance imaging. Radiology 1985; 155:421-423.

August

1991