Abdominal Philippe Soyer, MD Alain Roche, MD

Marc

#{149}

Levesque,

MD

and Dominique

Gastrointestinal

Elias,

#{149}

MD

Guy

#{149}

Detection ofLiver Metastases from Cancer: Comparison of Intraoperative and CT during Arterial Portography’ A prospective study was performed to compare the sensitivities of intraoperative ultrasound (US) and computed tomography during arterial portography (CTAP) in the depiction of hepatic metastases from colorectal cancer. Twenty-five patients with hepatic metastases from colorectal cancer were evaluated. All patients underwent partial hepatectomy, and 56 metastases were pathologically proved. Preoperatively, CTAP depicted 51 of the 56 metastases (91%). Intraoperative US depicted 54 of the 56 metastases (96%). Intraoperative US depicted three metastases (5%) that were not depicted with CTAP and two that were missed with palpation (3%). Furthermore, intraoperative US did not demonstrate any false-positive lesions. There was no statistically significant difference in sensitivity between the two techniques. The authors concluded that intraoperative US does not enable detection of more liver metastases from colorectal cancer when CTAP is considered as the preoperative standard of reference. Nevertheless, the results of the study suggest that intraoperative US and CTAP are complementary techniques, and the preoperative use of CTAP for determining the feasibility of hepatic resection cannot prevent the use of intraoperative US.

H

terms: Liver neoplasms, CT, 761.1211 Liver neoplasms, secondary, 761.3327 #{149} Liver neoplasms, US, 761.12982 #{149} Liver, surgery, 761.1299 #{149} Portography, 761.1242, 952.1242 Ultrasound (US), intraoperative, 761.1298

Radiology

1992;

183:541-544

is an

accepted

procedure in the management of secondary hepatic neoplasms from colonectal cancers. Patients with limited liver metastases from colorectal cancer who do not undergo hepatic resection rarely survive more than 2 years

after

diagnosis

(1).

On

the

other

hand, the 5-year survival rate after curative resection is 30% (2). Unfortunately, the fraction of patients that may

has

benefit

been

from

hepatic

estimated

resection

to range

from

5%

(3) to 10% (4). Careful preoperative selection of patients is crucial to avoid unnecessary explorations that considcrably reduce the quality of life that remains for patients with unresect-

able tumors. For those reasons, preopenative imaging techniques for the evaluation of hepatic metastases must be as accurate as possible. Currently, it is well established that computed tomography (CT) during arterial portography (CTAP) is the most

sensitive

preoperative

technique

for detecting

diagnostic

imaging

imaging

metastases from colonectal cancers (5-7). Recent advances in hepatic surgical oncology have changed the role of

hepatic

Zeitoun,

MD

Colorectal US

tant aid in the decision-making cess because it can depict static nodules that cannot

with conventional (11,14). To our racy

prosmall metabe detected

imaging knowledge,

of intraoperative

modalities the accu-

US in the

detec-

tion of hepatic metastases from cobrectal cancer has never been compared with that of CTAP. The purpose of this prospective study

was

to compare,

by

means

of a

lesion-to-lesion analysis, the accuracy of intraoperative US and CTAP in the detection of hepatic metastases from colorectal cancer. PATIENTS

AND

Twenty-five tastases

from

METHODS

patients

with

colorectal

hepatic

cancer

(14

memen,

11

women aged 20-70 years [mean, 56 years]) were prospectively evaluated during a 21-month period (December 1989 to August 1991). The patients were considered to be suitable candidates for hepatic resection on the basis of results of preoperative imaging.

All

patients

underwent

partial

fective, hepatic resection must be performed in patients with fewer than four metastases (13) located in an area

hepatectomy. During the same period, five patients with liver metastases that were determined to be inoperable at preoperative imaging were not included in this study. Three of those patients had undergone placement of a catheter into the intrahepatic artery. Fifty-six metastases to the liver were resected after intraoperative US was performed. The metastases originated from colon (20 patients), cecum (one patient),

that

and rectum

need

to know

hcpatic

of the Index

resection

EPATIC

Radiology

tases

the

metastases

permits

(8-12).

exact because,

resection.

number is crucial

Surgeons

number

and

of

to be ef-

Determination

location

to determine

of metaswhich

patients have nesectable metastases and to plan the type of surgical resection necessary (left on right lobe, segmental, on rnultisegmental). Recently, intraoperativc ultrasound (US)

has

been

advocated

as an

impor-

(four

patients).

ses

patients

was

determined

with

real-time

US,

dy-

namic CT after injection of a bolus of contrast material, and CTAP. All imaging examinations were performed within 10 days

of surgery

(mean,

4 days).

Intraoperative US was either an SAL 38 (Toshiba, From the Departments of Radiology (P.S., ML.) and Surgery (G.Z.), Hopital Louis Mourier, 92701 Colombes Cedex, France, and the Departments of Surgery (D.E.) and Radiology (AR.), Institut Gustave Roussy, Villejuif Cedex, France. Received September 13, 1991; revision requested November 11; revision received December 18; accepted December 30. Address reprint requests to P.S. ‘ RSNA, 1992

Three

referred for recurrence had previously undergone hepatic surgery (one with left hepatectomy, two with right hepatectomy). The detectability of these metasta-

Abbreviation:

ClAP

=

performed Tokyo)

CT during

with or a

arterial

por-

tography.

541

Scannel 300 (CGR-General Electric, Paris) unit and a high-resolution 5- or 7.5-MHz intraoperative US probe. The T-shaped transducer was sterilized and placed directly on the surface of the liver. The probe was held in one hand and was applied to all palpable parts of the liver after extensive mobilization and bimanual palpation of the liver. In some cases, a waterfilled sterile surgical glove was attached to the transducer to improve visualization of the superficial part of the liver. The posterior surface of the right lobe was examined first, and the probe was passed in both longitudinal and axial directions. The probe was placed laterally and moved toward the left to the round ligament. The left lateral segment and the undersurface of the liver were imaged with the same technique. Natural contact between the probe and the liver was usually obtained with peritoneal fluid. In a few cases, use of sterile saline was mandatory. When unsuspected nonpalpablc deep lesions were detected with mtraoperative US, a biopsy was performed with US guidance.

lntraoperative

formed

by

patic

surgeons

results

the

US was experienced

same

two

who

were

of US, CT, and

informed

CTAP.

of the

Complete

intraoperative US exploration added average of 12 minutes to the surgical cedure and was performed with the upper abdominal bilateral subcostal sion, with an extension to the xiphoid cess.

The

increase

imaging

time

number

of biopsies

surgery. All CTAP

was

in intraoperative

examinations

were

by multiplying

the

number

of sections

by

the cycle time and the rate of injection. The mean number and standard deviation of CTAP sections necessary to demonstrate

the

whole

liver

was

15.5

2, but

±

formed

and

prospectively

interpreted

No cavernous

hemangioma

was

be-

encoun-

perfusion

and

the study. The detectability intraoperative

were

excluded

from

of the metastasis

US

was

compared

with with

that with US, CT, and CTAP. The segmentectomy specimens were sliced 5 mm thick.

Small

metastatic

foci

were

carefully

time

electronic

Medical 3.5-MHz

performed

After the tip of a catheter (Cook,

tamed

equipment

(Radius,

GE

Systems or SAL, Toshiba) probe. All CT scans were with

scanner,

an

with

Exel

2400

10-mm

or a CT

contiguous

Pace

Plus

sections.

were

mesenteric artery to test the patency of the portal vein, the patients were transferred immediately to the CT scanncr. A set of 10-mm contiguous sections were obtained after intraarterial injection of 40 mg of papavenine hydrochloride (Pa-

injection

of a bolus

paverive;

mar test for paired data (15). A P value less than .05 was considered statistically

Ind)

was

placed

into

the

su-

perior

Laboratoire

France).

Papaverine

injected

to increase

Agucttant,

Lyon,

hydrochloride the

portal

was blood

pamiron

300;

2 mL/sec

Schening,

through

Berlin)

a power

at a rate

injector

of

(Med

Rad, Pittsburgh). The first section was obtamed a mean of 12 seconds after injecting the contrast medium to obtain images during the beginning of the portal phase of the arterial infusion. The scan delay was determined for each patient when testing the patency of the portal vein. CTAP was performed with either an Exel 2400 scanner (Elscint, Haifa, Israel) (300 mAs, 340 x 340 matrix, 2.1-second scan time, 3.2-second intenscan delay) on a CT Pace Plus scanner (GE Medical Systems,

Milwaukee) (260 mAs, 512 x 512 ma2-second scan time, 3-second interscan delay). The scanning parameters trix,

were as follows: 10-mm tion, 0#{176} gantry, variable 420 mm),

542

and

Radiology

#{149}

standard

table incrementafield of view (350algorithm.

Approxi-

before

of 100

mL

and after

containing 38 mg of iodine (Te38; Andre Guerbet, Aulnay-sous Bois, France) into the antecubital vein. Statistical review of the lesion-to-lesion lebnix

analysis

was

performed

and

was

with

an indication

the

McNe-

that

of the

difference between the imaging techniques was greater than that due to chance alone. Segmental location of the metastases was determined according to the Couinaud numbering system (16). Ninety-five percent confidence intervals were calculated for the difference between detection

CTAP

rates

of intraoperative

US

2.6 (range,

tastases

and

(15).

RESULTS The overall detection rate (sensitivity) was 68% (38 of 56 rnetastases) for US, 71 % (40 of 56 metastases) for CT, 91% (51 of 56 metastases) for CTAP, and 96% (54 of 56 metastases) for intraoperative US. The results of the lesion-to-lesion analysis are shown in the Figure, which demonstrates detection of metastases with intraopera-

0.4-11.2

missed

CTAP

ranged

were

less

cm).

with than

The

two

me-

intraoperative 5 mm

in diameter.

Those metastases, located deep in the parenchyma, were missed with other imaging modalities-even CTAPand were found incidentally in the nesected specimens. Of the five metastases not visible at CTAP, three lesions measuring 8, 10, and 12 mm in diameten, respectively, were located at the surface of the liver under the diaphragm

(two

in segment

VIII,

and

one in segment II). These metastases, not visible at US on CT, were detected intraopcratively with palpation and US.

Seven metastases were demonstrated with CTAP alone. ses were US; this

of contrast

medium

significant

flow

during tnanscathcter superior mesenteric infusion of a nonionic iodinated (30 mg of iodine per 100 mL) contrast medium (lo-

obtained

and a ob-

US and

from 0.0 to 0.10. Fifty-three of the 56 metastases (95%) were detected with palpation. The mean number of metastases per patient was 2.24 (range, 1-4), and the mean tumor diameter was 4.2 cm ±

US

tened. One cyst diagnosed with US and CT was excluded from the study and finally confirmed with surgery. Triangular, segmental areas of low attenuation were considered to be abnormalities or defects in portal

tive US and CTAP as a function of lesion size. The difference between the sensitivity of intraoperative US and CTAP in the detection of hepatic rnetastases was not statistically significant. The magnitude of the differences in sensitivities was .05 for the matched intraoperative US-CTAP data array. With a confidence of 95%, the confidence interval (15) for the difference in sensitivities between intraoperative

fore surgery by two abdominal radiologists who were blinded to the results of US and CT; they agreed on all CTAP findings. For all CTAP imaging procedures, a rounded, well-delineated space-occupying lesion with low attenuation (indicating a solid lesion) was diagnosed as metastasis.

CT scans

Bloomington,

the

mean volume of contrast medium used was only 147 mL ± 6 because we never injected more than 150 mL. The CTAP examinations were per-

searched for and compared with the imaging findings. US was performed with real-

US

with

performed

the same procedure. 5-F end-hole angiographic

an prousual mcipro-

the during

correlated

with

perhe-

mately 11 sections per minute were obtamed with these parameters. The total volume of contrast medium used varied with the number of sections needed to depict the entire liver and was obtained

detected did not

in three patients preopenatively All seven metastawith modify

intraoperative the surgical

decision making. The mean size of these metastases was 1.4 cm ± 0.3 (range, 0.8-2.0 cm). No false-positive findings occurred with intraoperative US;

however,

two

metastases

were

falsely diagnosed with CTAP in the same patient. Both intraopenativc US and CTAP depicted two metastases (in the same patient) that were missed with manual palpation. DISCUSSION Hepatic metastasis is one of the major problems in the treatment of cobrectal cancers. The major role of imaging techniques for evaluating hepatic metastases from coborectal cancer is to assist

the

surgeon

in selecting

candi-

dates for hepatic resection. Careful preoperative selection is needed to avoid unnecessary and discouraging surgical explorations. Presently, US and CT are routinely used for preoperative evaluation. The recent use of intraoperative US shows promise for hepatic surgery (10,17,18), and this technique is becoming important in May

1992

8

al (14) demonstrate that intraopenative US is a vital technique for imaging of the liver; the major drawback of their study is the fact that no patient was investigated preoperatively with CTAP. Although recent advances in noninvasive imaging techniques, including US, dynamic CT, delayed CT, and magnetic resonance (MR) imaging have improved the detection rate of hepatic metastases, the rate barely reaches 85% (6). The sensitivities of US, CT, and MR imaging make them unsuitable for preoperative imaging when used alone. This lack of sensitivity is obvi-

7 6 0

5

U m

b e

r

2

0

f

U

m

8

e t a

ous

6

S

t

when

small

metastases

are

present (5-7,28). Previous studies have proved that CTAP is the most sensitive preoperative technique in detecting focal hepatic lesions

S

a

(5-7,29,30).

S

tage of intraoperative US over preopcrative imaging techniques is not so evident when CTAP is included. The sensitivities of intraoperative US and CTAP in detecting hepatic metastases are 96% and 91%, respectively, and the difference between the two techniques is not statistically significant. The three small additional metastases detected at intnaoperative US were

e

3

S 2

0 co

-

eJ

-

I

I

cD

U)

Li

Q

U)

C%J

F4

t4

$

cI

I

U)

co

co

U)

U,

LI) t

I

LD

I

I

from

C)

U,

U,

of CTAP

colorectal

the

(top) 0

cancer.

and

intraoperative

undetected

=

surgeon

can

delineated,

cleanly

which

extensive

multiple can take anatomy scans.

US

define

the

resection

the

basis

tionships among veins and between tases,

the

a metastasis

point

re-

tunes

Volume

of large

or

of the

should

can

be located

the

venous

be nesccted

can

Intraopenative 183

spatial

Number

#{149}

rela-

hepatic and portal vessels and metas-

at which

determined.

c

U)

U)

iD

LI)

I

I

I

Q

LI)

c

U)

C)

U,

I

U,

C)

crci

I

and

strucbe easily

US is an 2

C)

U,

dco I

U)

I

#{149} I

I

D

located metastases

(cm)

(bottom)

U

in the

=

detection

detected

of liver

metastases

metastases.

exceptional tool for depicting the vascular structures that are the landmarks of the complex, surgically relevant, segmental anatomy of the liver (23-27). Therefore, at our institution, US-guided hepatectomy has largely replaced conventional hepatectomy for the treatment of a tumor located deep in the parenchyma. In our series, intnaopenative US depicted only three more metastases (5%)

permits

metastases because surgeons advantage of the hepatic as seen on intraopenative US

On

c,

metastases,

moval of the metastases with the most adequate tumor-free margin of resection. Intraopcrative US is also used to monitor hepatic cryotherapy (12,22). Furthermore, intnaoperative US permits

Q

diameter

hepatic venous structures, avoiding injury to the vessels and minimizing blood loss; the extent of the disease is well

I

co

the peroperative management of patients with hepatic metastases from coborectal cancer (8,14,19). This technique is considered an important peroperative aid because the anatomic venous structures (hepatic veins and portal branches) are very well outlined (9,20,21). With intraoperative US,

U)

I

r.:N I

e%JnJNpo greatest

Comparison

ID

U) I

than

were

depicted

with

CTAP

and two more than were found with palpation (3%). In the study by Charnlcy et al (14), 50% of patients with palpable metastases had additional liver metastases detected with intraopcrative US. Furthermore, intraoperative US depicted metastases in six patients in whom liver palpation was normal. In the same study, two of three patients would have been amenable to hepatic resection if, preoperatively, no further mctastases had been found. Furthermore, because additional metastases were detected at intraoperativc

that patient able. Results

US in one

was of the

deemed study

by Charnley

et

study,

in segments detected

the

advan-

involved by other with CTAP and

did not modify the decision to perform surgery. Furthermore, histologic examination of resccted specimens revealed two metastases undetected with intraopenative US. Of the five metastases missed at CTAP, three were located at the hepatic dome. This lack of sensitivity for the detection of metastases located high in the hepatic dome is well known and is probably due to rcspiratory movements. In our study, intraoperative US was of little help, since these metastases were located at the surface of the liver and were detected with palpation. One advantage of intnaoperativc US in our false-positive

study

was findings.

the

absence Conversely,

even though only two metastases our study were falsely diagnosed with CTAP, this technique had times

been

considered

of

in some-

as unaccept-

able because of a high false-positive rate (29). However, in a recent article (31), as in ours, CTAP was performed with a higher injection rate of contrast

medium;

this

correlated

with

an

acceptable false-positive rate. Furthermore, in our study, CTAP was penformed

patient,

to be moper-

In our

after

transcatheter

injection

of

papaverine hydrochloride, which increased the portal blood flow (32). Because false-positive findings are Radiology

543

#{149}

due to uneven hepatic perfusion, which results in uneven distribution of contrast medium within the hepatic parenchyma, the use of papavenine hydrochloride (which provides homogeneous enhancement) is recommended.

Several criticisms may be raised with respect to this study. The first is that the surgeons who performed intraoperative US were informed of the results of CTAP. It was impossible to perform a blinded comparative study because of ethical considerations. The preoperative knowledge of the number and location of hepatic metastases might have guided the surgeon in performing intraoperative US and helped improve the true sensitivity of this technique. The second criticism is due to the fact that pathologic examination of the healthy liver that was left in place was not possible. The more accurate standard of reference would be the results of pathologic examination of all of the liver, but a series of patients with such resections is only

feasible

with

patients

before

liver transplantation. In our study, we were unable to demonstrate a difference between intraoperative US and CTAP. The 95%

confidence

interval

(which

pro-

vides a range of plausible differences in the sensitivities) demonstrates that it can be inferred that the sensitivity of intraoperative US in the population is between

0%

and

10%

greater

can

prevent

an

study

demonstrates

that

CTAP

for

#{149} Radiology

2.

3.

4.

5.

Adson MA, Van Heerden

JA, Adson MH, WagnerJS, Ilstrup DM. Resection of hepatic metastases from colorectal cancer. Arch Surg 1984; 119:647-651. Morrow CE, Grage TB, Suthaland DER, Najarian JS. Hepatic resection of secondary neoplasms. Surgery 1982; 92:610-614. FortnerJG, SilvaJS, Golbey RB, et al. Multivariate analysis of a personal series of 247 consecutive patients with liver metastases from colorectal cancer. I. Treatment by hepatic resection. Ann Surg 1984; 199:306-316. August DA, Ottow RT, Sugarbaker PH. Clinical perspective on human colorectal cancer metastases. Cancer Metastasis Rev 1984; 3:303324. Matsui 0, Takashima T, Kadoya M, et al. Liver metastases from colorectal cancers: de-

tection with CT during

per-

fectly fills this role. In the patients studied, the use of intraopenative US did not modify the preoperatively determined surgical procedure because the additional metastases detected were located in segments involved by metastases detected at CTAP. Nevertheless, we believe that it is to be expected that

544

I.

open-and-close

surgical event that is discouraging the surgeon and the patient. Our

9.

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22. 23.

24.

References

than

that of CTAP. According to our clinical experience, we believe that intraoperative US is a sensitive technique for depicting hepatic metastases from colorectal cancer. However, intraoperative US necessitates a wide surgical exposure of the liver. Thus, a highly sensitive preoperative technique that provides reliable detection of metastases

the detection of additional metastases that preoperatively modify the surgical decision making will occur in the future. For these reasons, the routine use of intraoperativc US is recommended because it helps perform safe hepatic resection. We believe that intraoperative US is a simple technique that shows promise in hepatic surgery because it can enable a safer hepatic resection. Nevertheless, because we were unable to demonstrate a significant difference in sensitivities between intraopcrativc US and CTAP in our group of patients, we conclude that intraoperative US does not increase the number of detected liver metastases from coborectal cancer when CTAP is considered as the preoperative standard of reference. Nevertheless, the results of our study suggest that intraoperative US and CTAP are complementary techniques and that the preoperative use of CTAP for determining the feasibility of hepatic resection cannot prevent the use of intraoperative US. Conversely, since preoperative imaging is performed to avoid unnecessary surgical explorations, the use of intraoperative US cannot prevent the use of CTAP. The influence of intraoperative US on surgical decision making must be evaluated by means of additional clinical investigations. U

6.

7.

8.

arterial

25. 26. 27.

28.

29.

30.

portography.

Radiology 1987; 165:65-69. Heiken JP, Weyman PJ, Lee JKT, et al. Detection of focal hepatic masses: prospective evaluation with CT, delayed CT, O during arterial portography, and MR imaging. Radiology 1989; 171:47-51. Nelson RC, ChezmarJL, Sugarbaker PH, Bernardino ME. Hepatic tumors: comparison of CT during arterialportography, delayed CT, and MR imaging for preoperative evaluation. Radiology 1989; 172:27-34. Thomas WM, Morris DL, Hardcastle JD. Contact ultrasonography in the detection of liver metastases from colorectal cancer: an in vitro study. Br J Surg 1987; 74:955-956.

31.

32.

Charnley RM, Binch C, Morris DL, Hardcastle JD. Visualization of detailed hepatic anatomy by intraoperative ultrasound. Br J Radiol 1988; 61:769-770. Castaing D, Edmond J, Kunstlinger F, Bismuth H. Utility of operative ultrasound in the surgical management of liver tumors. Ann Surg 1986; 204:600-605. Makuuchi M, Hasegawa H, Yamazaki 5, Takayasu K, Moriyama N. The use of operative ultrasound as an aid to liver resection in patients with hepatocellular carcinoma. World Sur6 1987; 11:615-621. Ravikumar TS, Kane R, Cady B, et al. Hepatic cryosurery with intraoperative ultrasound monitonng for metastatic colon carcinoma. Arch Surg 1987; 122:403-409. Sugarbaker P1-I. Surgical decision making for large bowel cancer metastatic to the liver. Radiology 1990; 174:621-626. Charnley RM, Morris DL, Dennison AR, Amar SS, Hardcastle JD. Detection of colorectal liver metastases using intraoperative ultrasonography. Br J Surg 1991; 78:45-48. Dwyer AJ. Matchmaking and McNemar in the comparison of diagnostic modalities. Radiology 1991; 178:328-330. Couinaud C. Le foie. In: Couinaud C, ed. Etudes anatomiques et chirurgicales. Paris: Masson, 1957. Plainfoss#{233}MC, Merran S. Work in progress: intraoperative abdominal ultrasound. Radiology 1q83; 147:829-832. Gozzetti G, Mazziotti A, Bolondi L, et al. Intraoperative ultrasonography in surgery for liver tumors. Surgery 1986; 99:523-530. Castaing D, Kunstlinger F, Habib N, Bismuth H. Intraoperative ultrasonogra h study of the liver. Am J Surg 1985; 149:67 82. Igawa 5, Sakai K, Kinoshita H, Flirohashi K. Intraoperative sonography: clinical usefulness in liver surgery. Radiology 1985; 156:473-478. Rifkin MD, Rosato FE, Branch HM, et al. Intraoperative ultrasound of the liver: an important adjunctive tool for decision making in the qperating room. Ann Surg 1987; 205:466-472. Charnley RM, Doran J, Morris DL. Cryotherapy for liver metastases: a new approach. Br Surg 1989; 76:1040-1041. Bismuth H, Houssin D, Castain D. Major and minor segmentectomies “regl#{233}es”in liver surgery. World J Surg 1982; 6:10-24. Makuuchi M, Hasegawa H, Yamazaki S. Four hepatectomy procedures for resection of the right hepatic vein and preservation of the inferior right hepatic vein. Surg Gynecol Obstet 1987; 164:68-72. Couinaud C. Principles directeurs des h#{233}patectomies r#{233}glees.Chirurgie 1980; 106:8-10. Bismuth H. Surgical anatomy and anatomical surgery of the liver. World J Surg 1982; 6:3-9. Sugarbaker PH. En bloc resection for hepatic segments 4, 5 and 6 by transverse hepatectomy. Surg Gynecol Obstet 1990; 170:250-252. Gunven P. Makuuchi M, Takayasu K, Moriyama N, Yamasaki S, Hasegawa H. Preoperative imaging of liver metastases: comparison of angiography, CT scan and ultrasonography. Ann Surg 1985; 202:573-579. Miller DL, Simmons JT, Chang R, et al. Hepatic metastases detection: comparison of three CT contrast enhancement methods. Radiology 1987; 165:785-790. Matsui 0, Takashima T, Kadoya M, et al. D namic computed tomography during arteria portography: the most sensitive examination for smallhepatocellular carcinomas. J Comput Assist Tomogr 1985; 9:19-24. Nelson RC, ChezmarJL, Sugarbaker PH, Murray DR, Bernardino ME. Preoperative localization of focal liver lesions to specific liver segments: utility of CT during arterial portography. Radiology 1990; 176:89-94. Widrich WC, t”fordahl DL, Robbins AH. Contrast enhancement of the mesenteric and portal veins using intra-arterial papaverine. AJR

new

1974;

121:374-379.

May 1992