Preoperative Assessment of Resectability of Hepatic ...

741

Preoperative Resectability Metastases Carcinoma: Sonography

Philippe Soyer1 Marc Levesque1 Dominique EIias2 Guy Zeitoun3 Alain Roche4

Assessment of Hepatic from Colonic CT Portography and Dynamic CT

OBJECTIVE. A retrospective portography

vs sonography

resectability

of hepatic

of

study was and dynamic

metastases

from

vs

performed to determine CT on the preoperative

colorectal

the influence assessment

of CT of the

cancer.

MATERIALS

AND METHODS. Results of sonography, bolus dynamic CT, and CT portography in 28 patients who underwent surgical exploration (resection or intraartrlal catheter placement) for hepatic metastases from colorectal cancer were retrospectively reviewed by two abdominal radiologists and one hepatic surgeon. For each patient, the resectability and surgical approach were decided sonography-bolus dynamic CT and compared

portographic

on the basis of the results of combined with the decision made from the CT

results alone. The final approach suggested

with the surgical procedure actually performed. RESULTS. Sixty-nine metastases were identified

Combined (93%)

sonography-bolus

metastases,

CT portography.

at surgery

patients,

Twelve

metastases

CT portography

proved.

showed 52 (75%) and 64

in five patients depicted

retrospectively

and pathologically

dynamic CT and CT portography

respectively. In four

was compared

were

additional

seen only with

metastases,

which

changed the surgical approach that had been chosen on the basis of results of sonography and bolus dynamic CT. In one patient, CT portography showed four additional metastases, CONCLUSION. sonography and

resectability AJR

precluding hepatic resection. Findings from CT portography provide vital data unattainable bolus dynamic CT that improve the preoperative assessment of liver metastases from colonic carcinoma.

159:741-744,

Partial

hepatic

October

resection

with of the

1992

is the

best

treatment

currently

available

for

hepatic

metastases from colorectal cancer. Without hepatic resection, only 5% of patients with limited hepatic metastases from colorectal cancer survive for more than 2 years after the diagnosis [1 ]. On the other hand, the 5-year survival rates after curative resection are up to 30% [2]. Unfortunately, estimates of the percentage of patients who may undergo hepatic resection range only from 5% [3] to 10% [4]. Careful preoperative selection of patients through the use of imaging studies is viobnApnil February 13, 1992; accepted after reI Department of Radiology, H#{244}pital Louis Mour-

crucial to avoid unnecessary surgical explorations. It is well established portography (CTP) is the most sensitive preoperative imaging technique

ier, 178, rue des Renouillers, 92701 Colombes Cedex, France. Address reprint requests to P.

tecting

hepatlc

hepatic

oncologic

Soyer.

in such patients Institut

ROUSSy,VdIeJUif, 3

Department

of Surgery,

-

.

Hopital

Gustave .

.

Louis

Mouner,

Radiology,

Institut

92701 Colombes Cedex, France. 4

Department

of Interventional

Gustave Aoussy, Villejuif, France. 0361-803X/92/1594-0741 C American Roentgen Ray Society

metastases

surgery more critical.

from

colorectal

cancers

[5-7].

have made the diagnostic Hepatic

resection

Recent

imaging

should

advances

methods

be attempted

that CT for den

[8, 9] used

only in patients

with four or fewer metastases [1 0] located in a distribution that permits resection. Thus, imaging studies are used to determine which patients have resectable metastases and to plan the operation (e.g., left or nght lobe resection, segmental .

or multisegmental

resection,

.

or subsegmental

the

impact

of

CTP on surgical

decision

resection).

making

However,

to our knowl-

has not been analyzed. The purpose of this study was to determine the influence of CTP compared with that of sonography combined with bolus dynamic CT on the preoperative assess-

edge,

742

SOYER

ment of the resectability cancer

of hepatic metastases

and on the choice

of surgical

from colorectal

ET AL.

AJR:159, October1992

(Medrad).

(Toshiba,

Materials

35 consecutive patients with suspected colorectal cancer were referred for CTP before possible surgical resection. Seven patients with metastases that were considered inoperable on the basis of the results of prooperative imaging techniques did not have surgery and were not included in this study. Twenty-eight patients with hepatic metastases from colorectal cancer, 16 men and 1 2 women 20-72 years old (mean age, 57 years), underwent surgical exploration and form the basis of this retrospective study. Twenty-five patients underwent partial hepatectomy, and three patients had a catheter surgically placed in the hepatic artery. The metastases originated from the colon in 23 patients and the rectum in five patients. Three patients referred for recurrent disease had previously had hepatic surgery (left lobectomy, n 1 ; right lobectomy, n 2). All patients underwent period, from

=

preoperative

=

real-time

was begun 40 sec after beginning

sonography

Tokyo,

Japan)

was

sonography,

dynamic

CT

after

bolus

injection

sterilized

and placed

was

in one

held

liver after extensive

The examination lobe. Thereafter, directions. the round

and

on the surface

was

applied

mobilization

parts

palpation

with the same technique.

obtained

by two

experienced

All intraoperative liver surgeons who bolus dynamic CT, and

were aware of the results of sonography, CTP. At the time of surgical exploration, the entire liver was inspected and bimanually examined for the presence of intrahepatic metastases. Intraoperative sonography was performed whenever a partial hepatic resection had already been planned in order to determine whether additional lesions were present. When resection was not performed,

ington,

experienced hepatic surgeon reviewed the sonographic,

mesentenc

artery,

the patency

of the portal vein was assessed by delayed filming after injection of 30 ml of a nonionic contrast medium (30 g I/i 00 ml; lopamiron 300, Schering, Berlin, Germany) at a rate of 6 mI/sec. For each patient, the interval between the beginning of the injection and the beginning of portal opacification was noted and used to plan the scan delay time for CTP. Thereafter, patients were transferred immediately to the CT scanner. CT was first performed without contrast material to identify hepatic cysts. Intraarterial injection of 40 mg of papaverine hydrochloride

(Papavenve,

Laboratoire

Aguettant,

Lyon,

France)

was

done to increase portal blood flow [1 1], after which nonionic contrast medium (lopamiron 300) was injected through the catheter with a power

injector

(Medrad,

Pittsburgh,

PA) at a rate of 2 mI/sec.

Contig-

uous 10-mm-thick slices were obtained. The scan delay was based on the interval (mean, 12 sec) until partial venous opacification was seen angiographically. CTP was performed with either an Exel 2400 scanner (Elscint, Haifa, Israel) with a 340 x 340 matrix, 2.1-sec scan time, and 3.2-sec

interscan

delay or a CT Pace Plus scanner

(General

Electric, Milwaukee, WI) with a 51 2 x 512 matrix, 2-sec scan time, and 3-sec interscan delay. Approximately 1 1 sections per minute were obtained. The total volume of contrast medium used varied with the number of sections needed to image the entire liver and was determined by multiplying the number of sections by the cycle time and the rate of injection. The mean number of CTP sections necessary to image the whole liver was 15.5 ± 2.0, but the mean volume of contrast material used was only 1 47 ± 6 ml because we never injected more than 150 ml. Sonography was performed by a medical sonographer with a realtime

sonographic

77 SAL-A, scan

planes

unit (Radius,

Toshiba, and

Tokyo, hard-copy

General

Electric,

Milwaukee,

Japan) and a 3.5-MHz images

were

obtained

probe. during

of 5.3 sec and 5.0 sec, respectively.

1 00

ml of contrast

radiologists

and one bolus

dy-

namic CT, and CTP examinations of the 28 patients. The original interpretations rendered at the time the studies were performed were unknown to the reviewers. The radiologists and the surgeon identified all visible

viewing

lesions

and

sonographic

selected

a surgical

and bolus dynamic

during a first interpretation. Two pretation, the CTP examinations

planning

approach

CT examinations

after

together

weeks later, during a second interwere reviewed alone and analyzed

in a random manner. Lesions were identified, and a surgical approach was determined. On all CTP images, a rounded, well-delineated, lowattenuation space-occupying lesion was diagnosed as a metastasis. Triangular segments of low attenuation were considered to be artifacts of portal blood flow and were not counted as lesions. The intraoperative sonographic findings of these areas were evaluated to confirm that they reflected perfusion abnormalities rather than metastases or perfusion abnormalities associated with a more proximally located tumor. The operative approaches chosen retrospectively on the basis of the two sets of imaging findings were compared with the surgical procedures actually performed and correlated with the intraoperative findings and pathologic data obtained after examination of

the resected specimens. The subsegmental location of the metastases was categorized according to the Couinaud numbering system [12]. The sizes of the metastases to the nearest millimeter were determined from measurements made by the pathologist after surgory or from intraoperative sonography when metastases were not removed.

The sensitivities

ofthe

imaging

techniques

were statistically

compared by using the McNemar test for paired data [13]. A value less than .05 was considered statistically significant.

WI; or these

Both

CT were performed after a bolus injection medium containing 38 g of iodine per 100 ml (Telebnx 38, Guerbet, Aulnay-sous-Bois, France) into an antecubital vein at a rate of 2 mI/sec. Injections were administered via an 18- or 20-gauge plastic cannula either manually or through a power injector of

biopsy specimens from each lesion detected with intraoperative sonography were obtained with intraoperative sonographic guidance

Routine

examinations. All bolus dynamic CT scans were obtained with an Exel 2400 scanner or a CT Pace Plus scanner, with 10-mm-thick contiguous sections and cycle times unenhanced and enhanced

of the

of the liver.

laterally and moved to the left to segment and the undersurface of

to confirm their malignant natures. For this retrospective review, two abdominal

in the superior

38

of the liver. The probe

of contrast medium, and CTP within 1 0 days of surgery (mean, 4 days). All CTP examinations were performed by using the same procedure. After a 5-French end-hole angiographic catheter (Cook, BloomIN) was placed

an SAL

to all accessible

and bimanual

The probe was placed ligament. The left lateral were

either

was started on the posterior surface of the right the probe was passed in both longitudinal and axial

the liver were imaged sonograms

with

a Scannel

directly

hand

performed

contrast

300 (CGR-General Electric, unit and a high-resolution 5.0- or 7.5-MHz probe. The T-shaped transducer was

or

Pans, France) sonographic intraoperative sonographic

and Methods

During a 2-year hepatic metastases

CT scanning

injection. Intraoperative

approach.

Results Fifty-six

metastases

to the liver were

resected,

and a total

of 69 metastases were identified by means of palpation and intraoperative sonography with intraoperative biopsy or pathologic

examination

of the

resected

specimens.

The

mean

number of metastases per patient was 2.46 (range, one to seven metastases), and the mean tumor diameter was 4.2 ± 2.9 cm (range, 0.4-1 1 .2 cm). One biliary cyst was correctly

AJR:159,

October

LIVER

1992

METASTASES

FROM

identified with all preoperative imaging techniques and confirmed at surgery. Sonography showed 46 of 69 metastases and bolus dynamic CT showed 49 of 69. By combining the results of sonography and bolus dynamic CT, a sensitivity of 75% in the detection of metastases (52 of 69 metastases) was achieved. Three metastases located in the left lobe of the liver were

depicted

with

transabdominal

sonography

but not

with bolus dynamic CT. No lesions were false-positive on the basis of either sonography or bolus dynamic CT (specificity 100%). The overall sensitivity for detection of metastases was 93% for CTP (64 of 69 metastases). Five liver metastases were not seen with CTP. Two of them, located deep in the parenchyma, were smaller than 5 mm in diameter, were not detected with intraoperative sonography, and were found incidentally in the resected specimens. The three other metastases measured 8, 1 0, and 1 2 mm in diameter and were =

located

on the surface

of the liver under

the hemidiaphragm.

These metastases, not visible with sonography or bolus dynamic CT, were detected only by palpation and intraoperative sonography. Twelve metastases (1 7%) in five patients were detected preoperatively with CTP alone. The mean size of these metastases was 1 .4 ± 0.3 cm (range, 0.8-2.0 cm). No lesions ificity

were false-positive

on the basis of CTP findings

(spec-

100%). The difference in sensitivity between combined sonography-bolus dynamic CT and CTP was statistically significant (p < .01). The confidence interval of the differences in the =

sensitivities

between

CTP

and combined

sonography-bolus

dynamic CT was 0.08-0.28, indicating that CTP can be used to detect between 8% and 28% more metastases from colorectal cancer than the combination of sonography and bolus dynamic CT can. The operative approaches suggested on the basis of combined sonography-bolus dynamic CT and CTP are outlined in Table 1 CTP findings altered the choice of operative approach in five of 28 patients. One case was considered .

Unresectable

CT Portography

sided

Left sided

were

seen

in

decision

was changed

from

right

lobectomy

to trisegmentec-

tomy in one patient because two additional metastases were detected in subsegment IV on CTP. The initial decision was changed from left lateral segmentectomy to left lateral segmentectomy with additional right inferior subsegmentectomy in one patient when two additional metastases were detected in subsegments II and VI. The initial decision was changed from left lateral segmentectomy to left lateral segmentectomy

with additional additional

wedge

metastasis

resection was found

in one patient

when

in subsegment

one

VIII.

In all patients, the operative approaches chosen in retrospect on the basis of CTP findings correlated with the surgical procedures actually performed. At the time surgery was performed, findings on intraoperative sonography did not change the operative procedure selected after viewing sonographic, bolus dynamic CT, and CTP examinations simultaneously.

Discussion

Recent

advances

an increase

resection

in hepatic

in the number

for metastatic

surgery

have resulted

of candidates

disease

for partial

and a decrease

in both hepatic

in operative

morbidity and mortality. Nevertheless, since the number of patients who are candidates for surgery is low, a major role of imaging techniques in patients with liver metastases from colorectal cancer is to assist the surgeon in the preoperative decision regarding the feasibility of and approach to hepatic resection. It is critical that the surgeon have a clear and detailed map of the hepatic involvement with metastases. The ability to perform multisegmentectomies or multisubsegmen-

tectomies

underscores

the need to accurately

determine

the

number and segmental locations of hepatic metastases. Currently, patients with four or fewer hepatic metastases are considered candidates for resection [1 0]. Since resection usually is not performed if five or more metastases are

it is vital to detect

even small metastases in noninvasive bolus dynamic

(less than

imaging techCT, delayed CT,

and MRI imaging, have improved the detection of hepatic metastases, the detection rate still remains less than about 85% [6, 14, 1 5]. The sensitivities namic CT, and MR imaging make

of sonography, them unsuitable

bolus dyfor preop-

3

erative imaging when used alone, especially when metastases are small [5, 6, 15]. Previous studies have shown that CTP

5

6

is the most

9

9

Segmentectomy

6 4

7 2

Subsegmentectomy Wedge resection

7 4

8 5

26

25b

resections

metastases

4

Subtotal

Note-Combined

additional

3

Trisegmentectomy

Total

four

2

Resectable Lobectomy Right

because

Couinaud subsegments II, IV, V, and VIII on CTP. The initial decision was changed from left lateral segmentectomy to left lobectomy in one patient because three additional metastases were found in subsegments Ill and IV on CTP. The initial

Although recent advances niques, including sonography,

No. of Patients

Combined SonographyBolus Dynamic CT

743

CARCINOMA

inoperable

present, 1 cm).

TABLE 1: Proposed Surgical Approaches Based on Retrospective Evaluation of Imaging Studies

Type of Resection

COLONIC

were performed

in four

and sixb patients.

sensitive

technique

for detecting

focal

hepatic

lesions [5-7, 14-17]. In our study, sonography did not show 23 (33%) of 69 metastases and bolus dynamic CT did not show 20 (29%) of 69 metastases; the majority of the missed lesions were around 1 cm or less in diameter. CTP did not show five metastases (7%), three of which were located at the hepatic

metastases and probably

dome.

located

This

lack of sensitivity

high in the hepatic

is due to artifacts

caused

for the detection

of

dome is well known by respiratory

motion.

744

SOYER

CTP showed 1 2 metastases in five patients that were not seen with sonography or bolus dynamic CT, changing the surgical

planning

approach

in each

case.

In the one

case

considered inoperable owing to CTP findings, the choice was changed from a complex surgical procedure with associated morbidity to a more appropriate and simple surgical procedure (intraarterial catheter placement for chemotherapy). Our study confirms that sonography and bolus dynamic CT are significantly less valuable than CTP in the preoperative management of such cases. Furthermore, in our study, the final decisions based on CTP alone were the same as the surgical decisions actually made for each patient. Since the surgical decisions actually made were based on a combination of the findings of sonography, bolus dynamic CT, CTP, and intra-

operative only

sonography,

imaging

this result

examination

tients with hepatic

suggests

necessary

before

that CTP is the surgery

in pa-

from colorectal cancer. Recently, intraoperative sonography has been used to dotect small metastases not identified by direct palpation or by preoperative imaging techniques [1 8, 1 9]. In a recent article, Charnley et al. [1 8] reported that intraoperative sonography was used to detect additional metastases in patients in whom findings on liver palpation were normal and changed the operative approach based on conventional imaging studies. However, none of these patients were examined preoperatively with CTP. In our study, intraoperative sonography did not show any lesions not seen on CTP that altered the surgical approach. On the other hand, in order to detect additional metastases, intraoperative sonography necessitates wide surgical exposure of the liver. This is of importance when a patient is subjected to the wide surgical exposure necessary for intraoperative discovered that

metastases

sonography only to have it subsequently the metastases are unresectable and that

unnecessary surgery has been performed. In such cases, detection of multiple small metastatic nodules before rather than after opening the abdomen with wide exposure of the liver is definitely to the patient’s benefit. Proper preoperative assessment is important because placement of a catheter in the hepatic artery requires a smaller surgical incision and is associated with lower morbidity and mortality. A high false-positive rate has sometimes been considered a disadvantage we performed

of CTP [6]. In accord with another CTP with a higher injection rate

series [17], of contrast

material and with prior intraarterial injection of papaverine to increase portal blood flow [1 1]. Because no findings in this study were falsely positive for metastasis, the use of papaverine to improve the homogeneity of hepatic enhancement may have been beneficial in avoiding such errors in diagnosis. In conclusion,

this

study

shows

data unattainable

with combined

CT that improve ability of hepatic

the preoperative metastases from

that

CTP

provides

sonography-bolus

vital

dynamic

assessment of the resectcolorectal cancer. Further-

ET AL.

more,

AJA:159,

CTP findings

gical approach spective dynamic

more

accurately

in such patients.

predict

October

the optimal

1992

sur-

On the basis of our retro-

study, we recommend that CT no longer be performed

sonography because

and bolus CTP alone

provides information sufficient for deciding to what extent the liver should be resected in patients with hepatic metastases from colorectal

cancer.

REFERENCES 1 . Adson MA, Van Heerden of hepatic metastases 647-651

JA, Adson MH, WagnerJS, from colorectal cancer.

llstrup DM. Resection Arch Surg 1984;1 19:

2. Morrow CE, Grage TB, Suthaland DER, Najanan JS. Hepatic resection of secondary neoplasms. Surgery 1982;92:610-614 3. Fortner JG, Silva JS, Golbey RB, et al. Multivanate 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 4. August DA, Ottow AT, Sugarbaker PH. Clinical perspective on human colorectal cancer metastases. Cancer Metastasis Rev 1984;3:303-324 5. Matsui 0, Takashima T, Kadoya M, et al. Liver metastases from colorectal cancers: detection with CT during arterial portography. Radiology

1987;165:65-69 6. Heiken JP, Weyman PJ, Lee JKT, et al. Detection offocal hepatic masses: prospective evaluation with CT, delayed CT, CT during arterial portography, and MA imaging. Radiology 1989;171 :47-51 7. Nelson AC, Chezmar JL, Sugarbaker PH, Bernardino ME. Hepatic tumors: comparison of CT during arterial portography, delayed CT, and MR imaging for preoperative evaluation. Radiology 1989;172:27-34 8. Chamley AM, Binch C, Morris DL, Hardcastle JD. visualization of detailed hepatic anatomy by intraoperative ultrasound. Br J Radio! 1988;61: 769-770

9. Ravikumar TS, Kane R, Cady B, et al. Hepatic cryosurgery with intraoperative ultrasound monitoring for metastatic colon carcinoma. Arch Surg 1987;122:403-409 10. Sugarbaker PH. Surgical decision making for large bowel cancer metastatic to the liver. Radiology 1990;174:621 -626 11 . Widnch WC, Nordahi DL, Aobbins AH. Contrast enhancement of the mesenteric and portal veins 1974;121 :374-379 12. Couinaud C. Le foie. In: Couinaud cales. Pans: Masson, 1957

using

C, ed. Etudes

13. Dwyer AJ. Matchmaking and McNemar modalities. Radiology 1991;178:328-330 14.

Gunven

P, Makuuchi

M, Takayasu

intra-arterial

papaverine.

anatomiques

in the comparison

K, Moriyama

N, Yamasaki

AJR

et chirurgi-

of diagnostic 5, Hasegawa

H. Preoperative

imaging of liver metastases comparison of angiography, CT scan and ultrasonography. Ann Surg 1985;202:573-579 15. Yamaguchi A, Ishida T, Nishimura G, et al. Detection by CT during arterial portography of colorectal cancer metastases to liver. Dis Colon Rectum 1991;34:37-41

16. Miller DL, Simmons JT, Chang A, et al. Hepatic metastases comparison of three CT contrast enhancement methods.

detection: Radiology

1987;1 65:785-790

17. Nelson AC, Chezmar JL, Sugarbaker PH, Murray DA, Bemardino ME. Preoperative localization of focal liver lesions to specific liver segments: utility of CT during arterial portography. Radiology 1990;176:89-94 18. Chamley AM, Moms DL, Dennison AR, Amar 55, Hardcastle JD. Detection of colorectal liver metastases using intraoperative ultrasonography. Br J Surg 1991;78:45-48

19. Igawa 5, Sakai K, Kinoshita H, Hirohashi K. Intraoperative sonography: clinical usefulness in liver surgery. Radiology 1985;156:473-478