Resection of hepatocellular carcinomas

GASTROENTEROLOGY 1990;91:733-733

Resection of Hepatocellular Carcinomas Results in 72 European Patients With Cirrhosis DOMINIQUE FRANCO, LORENZO CAPUSSOTTI, CLAUDE SMADJA, HEDAYAT BOUZARI, JONATHAN MEAKINS, FRANCOIS KEMENY, DIDIER GRANGE, and MARIO

DELLEPIANE

Groupe de Recherche sur la Chirurgie du Foie et de 1’Hypertension Portale. Hapital Louise Michel. Evrv, Hbital Paul Brousse. Villeiuif France, et I” Divisione di Chirurgia Generale, ’ Ospedale ~aurtz~ano, Torino, Italia

This study was undertaken to determine the results of resection of hepatocellular carcinoma in cirrhotic patients in Europe, using the same criteria as in the Orient for detection, surgical management, and pathology of the tumors. Seventy-two patients had a liver resection. One- and 3-yr survival rates were 66 % and 61% , respectively. Survival rate was significantly higher in Child’s/Pugh’s class A than in class B-C patients. Patients with a thickly encapsulated tumor lived longer than those with an infiltrating tumor and had in addition a significantly lower incidence of cancer recurrence. Class A patients with a thickly encapsulated hepatocellular carcinoma had a 77% &year survival rate. There was no relation between the size of the tumor or the presence of symptoms and survival. These data suggest that good results can be achieved by resection of hepatocellular carcinomas in European cirrhotic patients. A thickly encapsulated tumor and an adequate liver function are the main determinants of low cancer recurrence and high survival. The clinical results in this series are similar to those reported from the Orient.

verthe

last 7 yr an increasing number of reports 0 from Japan and Taiwan (l-5) have shown that in patients having a surgical resection of hepatocellular carcinoma (HCC) in cirrhosis, operative risk was low with an operative mortality from 0% to 7.69’0, and that Z-yr survival was high, from 55.4 to 72%. Reports from the West have been fewer and less optimistic (6-8) because of a higher operative mortality (more than lo%], significant recurrence rates, and low survival rates. These discrepancies between Oriental and Western results were ascribed to differences in both the

type of cirrhosis and the nature of HCC. The rate of encapsulation was in particular estimated to be much lower in the West than in the Orient (6,7). In 1983, a cooperative study between two European centers experienced in liver resection and surgery in cirrhotic patients was launched to evaluate the surgical treatment of HCC in cirrhotic patients, using the same criteria as those in Eastern countries to detect and treat HCC, and to assess tumor pathology. Cooperation between 2 distant centers using the same diagnostic and operative techniques avoided bias resulting from patient recruitment, increased the number of patients studied, and would establish results applicable to other Western countries without the disadvantages of multicenter trials in complicated surgical problems (9). Diagnostic, surgical, and pathological techniques were determined at the beginning of the study. They were then reassessed twice yearly on the basis of the experience gained and of progress reports from other centers. The purpose of this article is to present clinical pathological correlates and results of resection of HCC in 72 patients operated during a 5-yr period. Patients and Methods Between June 1,1983,and May 31,1988,72patients with cirrhosis underwent liver resection for a hepatocellular carcinoma. Thirty-eight patients were operated at Ospedale Mauriziano [L.C. and M.D.), 30 at HBpital Louise Michel (D.F., C.S., and D.G.) and 4 at HBpital Paul Brousse (D.F.).

Abbreviation used in this paper: HCC, hepatocellular carcinoma. o 1990 by the American Gastroenterological Association 0016-5035/90R3.00

734

FRANC0

ET AL.

GASTROENTEROLOGY

Table 2. Circumstances of the Diagnosis of Hepatocellular Carcinoma in 72 Patients With Cirrhosis

Patients There were 68 white, 2 Oriental, and 2 black patients. Mean age of the patients was 61.8 yr (range, 39-78 yr). Cirrhosis was alcoholic in 38 patients (52.8%), postnecrotic in 31 patients (43.0%). and from another origin in 3 patients (4.2%). Previous complications of cirrhosis, results of clinical examination and biochemical liver tests, and endoscopic findings at time of surgery are indicated in Table 1. According to the classification of Child modified by Pugh et al. (lo), 54 patients (75%) were in class A, 14 (19.4%) were in class B, and 4 (5.6%) were in class C. During the study period, 6 patients with cirrhosis and a localized HCC did not have resection because of end-stage liver disease.

Vol. 98, No. 3

Circumstances

No. of patients

of diagnosis

Routine examination in an asymptomatic patient Ultrasonography Increase of serum concentration of (YFP Symptoms, signs, or complications Right-upper-quadrant pain Fever and fatigue Palpable abdominal mass Self-limited jaundice Hemoperitoneum

54 51 3 18 8 2 3 2 3

aFP, alphafetoprotein.

Hepatocellular

Carcinomas

The diagnosis of an HCC was most frequently made during surveillance of an asymptomatic patient with cirrhosis (Table 21. Preoperatively, the HCC was further defined with three liver imaging procedures: ultrasonography, dynamic computed tomography scan images before and immediately after bolus injection of contrast medium, and selective angiography. The serum concentration of alphafetoprotein was above 100 ng/ml in 34 patients (47.2%). A preoperative pathological diagnosis of HCC was confirmed by fine-needle ultrasonographically guided biopsy in 17 patients before referral to the surgical unit.

Operations The detailed surgical protocol has been previously published (11). Classification of the resections was made according to the surgical liver anatomy of Couinaud (12), which divides the liver into 8 independent segments. There were 14 major liver resections (19.4%) and 58 limited liver resections (80.6%) either segmental (48 patients) or nonanatomical(l0 patients).

Table 1. Previous Complications Patients With Cirrhosis

Pathology All resected specimens were analyzed for the gross appearance of the tumors according to Nakashima et al. (13). dividing HCC into expanding and infiltrating tumors. The free margin of noncancerous liver parenchyma was specifically measured in the 43 last patients. The presence of a capsule, its thickness 1 mm or more or less than 1 mm, the presence of satellite nodules, and tumor extension to distal portal branches were carefully noted on histological sections. Detailed results of the pathology of resected specimens of a smaller group of patients operated on have already been reported (141, showing clearly that the rate of encapsulation, differentiation, and histological grading were similar in European and Oriental patients.

All patients were assessed every 6 mo by clinical examination, liver biochemical tests, measurement of the serum concentration of alphafetoprotein, and ultrasono-

of Cirrhosis and Preoperative Clinical, and a Hepatocellular Carcinoma Number present, or increased above or decreased below normal value

History Ascites Variceal bleeding Preoperative clinical findings Ascites Encephalopatby Palpable tumor Preoperative endoscopic findings Esophageal varices Preoperative biological findings Serum bilirubin (~mof/L] Serum albumin [g/IJ Prothrombin time [% J ASAT (IV]

Follow-up

Endoscopic,

and Biochemical

Mean * 1 SD

Findings

in 72

Range

14(19.4%) 7(9.7%] 7 (9.7%) 5 (6.9%) 3 (4.2%) 31(43%) 32 (44.4%) 27(37.5%) 27(37.5%) 41 (56.9%]

21.6zt15.5 35.8f 6.0 79.0* 19.3 52.3 f 39.6

5-83 23-49 38-100 7-100

RESECTION OF HCC IN CIRRHOSIS

March 1990

735

graphic examination of the liver. No patient has been lost to follow-up. All patients were evaluated in July or August 1988.

Statistical Analysis Results are expressed as means kl standard deviation (m + 1 SD). Comparison between means was done by Student’s t-test. Comparison between groups was done by the x2 test. Survival and tumor recurrences were expressed according to Kaplan-Meier curves. Comparisons between survival curves and recurrence curves were done by the log-rank test. Results Pathological both centers.

Operative

2

3

4

5

6

7

6

9 1011

13

14 15

ems Figure 1. Mean diameter of the tumor in 72 patients with cirrhosis and an HCC. Open bars indicate patients without symptoms. Hatched bars indicate patients with symptoms or complications.

and clinical results were similar in

Morbidity

and Mortality

Thirty-seven patients (51.4%) had an uneventful postoperative recovery. Beside transient changes in biochemical liver tests, 43 complications occurred in 35 patients (48.6%) during the postoperative course (2 mo). The most frequent complication was an ascitic leak through abdominal drains (19.4%), causing infection and/or requiring intensive therapy. Five patients (6.9%) died during the postoperative period. Variceal bleeding was the most common cause of death (3 patients). Operative mortality was more frequent in class B-C patients (16.7%) than in class A patients (3.7%). The difference was, however, not significant.

Pathological

1

Results

Sixty-eight (94.4%) patients had a single tumor and 4 (5.6%) patients had 2 or 3 separate lesions. The largest diameter of the tumors is indicated in Figure 1. Mean tumor diameter was significantly larger in symptomatic (79.8 k 43.2 mm) than in asymptomatic patients (42.9 + 24.6 mm; p < 0.01). There were 19 infiltrating tumors (26.4%) and 53 expanding tumors (73.6%). All expanding tumors were surrounded by a microscopically defined capsule. The capsule was thin (less than 1 mm] in 22 patients (41.5%) and was thick (1 mm or more] in 31 patients (58.5%]. The mean tumor diameter, extension to distal portal branches, and presence of satellite nodules are indicated in Table 3. There were 43 tumors with a diameter less than 5 cm (59.7%) and 29 tumors with a diameter of 5 cm or more (40.3%). Extension to distal portal branches and satellite nodules was significantly less frequent in expanding than in infiltrating tumors,

and particularly low (35.5%) in expanding tumors with a thick capsule. The mean free margin of noncancerous liver parenchyma was accurately measured in 43 patients and was 10 mm or more in 25, and between 1 and 10 mm in 10 others. In 8 patients, there was no free margin. Cancer Recurrences During surveillance, the HCC recurred in 16 patients (22.2%). In 3, only the lungs or bones were involved. Figure 2 shows the cumulative onset of cancer recurrence according to Kaplan-Meier. Significantly more patients with expanding tumors and either a thick capsule (p < O.Ol] or a thin capsule (p < 0.05) were recurrence-free compared with those with infiltrating tumors (Figure 2A). Although the recurrence rate was higher in patients with a big tumor than in those with a small one, the difference was not significant [Figure 2B). Two patients with a liver recurrence were reresected 6 and 12 mo after the original operation and were alive at 26 and 30 mo. Another patient was reresected twice (24th and 58th mo] and eventually died from liver recurrence in the 62nd postoperative mo. Among the 43 patients in whom the free margin of noncancerous liver parenchyma was measured, 7 had cancer recurrence: 5 of the 8 patients with no free margin (62.5%) and 2 of the 36 patients with a free margin of liver parenchyma more than 1 mm (5.6%). The difference was significant (p < 0.011. Survival When considering all 72 patients, l-, 2-, and 3-yr survival rates were respectively 67.570, 55‘70, and 51%. In class A patients survival rates were 80.2%, 76.5%, and 56.7%, respectively. In class B-C patients they were 55.670, 12.2%, and 12.2%. Survival curves

736 FRANC0

ET AL.

GASTROENTEROLOGY

Vol. 98, No. 3

Table 3. Pathological Features of 72 Hepatocellular Carcinomas Resected in Patients With Cirrhosis Expanding tumors (53) Infiltrating tumors (19) Mean tumor diameter [mm) Distal portal venous extension Satellite modules

With a thin capsule (221

64 f 40 18 (94.7%) 18(94.7%)

With a thick capsule (31)

48 + 40 15 (68.2%) 10 (45.5%)b

46 i 20” 11(35.5%) 11(35.5%)

Significantly different from values of infiltrating tumors: “p < 0.05; bp < 0.01; ‘p < 0.001.

were significantly different (p < 0.01) between the two groups (Figure 3A-D). In patients with an infiltrating tumor, survival rates were 50.3‘70, 33.570, and 0%. In patients with an expanding tumor and a thin capsule, they were 73.870, 58.370, and 29.1%. In those with an expanding tumor and a thick capsule, they were 82.870, 6570, and 65%. Survival curves of patients with an infiltrating tumor and of those with an expanding tumor and a thick capsule are significantly different (p < 0.05). Survival rates of patients with a tumor less than 5 cm were 66.8%, 53.8%, and 41.9%. Those of patients with a tumor 5 cm or more were 69%, 52.770, and 53.2%. These curves are not significantly different. Survival rates of patients with an asymptomatic tumor were 69.3%, 51.4?70,and 45.770, not statistically different from patients with symptoms (55.4%, 48.570, and 48.5%). In 25 class A patients with an expanding tumor and a thick capsule, l-, 2-, and 3-yr survivals were 77%. Fifty-five percent (13 patients) of the 24 late deaths were related to cancer recurrence and 45% (11 patients] to a complication of cirrhosis, without recurrent HCC. Death resulting from HCC itself was significantly more frequent in patients with an infiltrating

tumor than (p < 0.05).

3 0

E

20

pi,j,

t

:I

I

I_________

***++++

??

i !,

3

.Z

? 40 2 20

:

: :Ii-L -!. I___ 7

20i_______________

I

I

12

loo

24

36

C

I____. :__

_~

_____ __

-Led12

24

36

Time 20

, mos

t

I’.& 12

12 24 38

Figure 3. Survival rate after liver resection for HCC in patients with cirrhosis.

24

Time , mos

Figure 2. Recurrence patients with cirrhosis.

+“I

a .z 40 ? z 20 t

I

40

After Liver Resection in

i

P)

‘-1

t

\

@60

3

E”

40

tumor

Liver failure (6 patients] and intractable ascites (6 patients) were undoubtedly triggered by hepatic resection. All of the patients with intractable ascites were successfully treated by a peritoneovenous shunt.

*

5

with an expanding

Late Complications Cirrhosis

ml

P =

in patients

ratea after liver resection

for HCC in

A. According to encapsulation of the tumor: [-----I infiltrating tumors: (+ + + + +) expansive tumors and a thin capsule: (-) expansive tumors and a thick capsule. B. According to the size of the tumor: (-) tumors less than 5 cm in diameter: (-----I tu mors 5 cm or greater in diameter.

A. According class B-C.

to Child’s/Pugh’s

class: (+ + + + +) class A; (-----I

B. According to encapsulation of the tumor: (-----I infiltrating tumors; (+ + + + +] expansive tumors and a thin capsule; (-1 expansive tumors and a thick capsule. C. According to the size of the tumor: (-----I tumors less than 5 cm in diameter: (--I tumors 5 cm or more in diameter. D. According to the presence (-----) or absence (-1 of symptoms.

RESECTION

March 1990

Postoperative cantly more than in class patients died

aggravation of liver failure was signififrequent in class B-C patients (27.8%) A patients (1.9%; p < 0.01); all of these within 11/2yr.

Discussion The most important finding of this study is that excellent survival is observed following resection of an HCC in a selected group of cirrhotic patients from two European centers. Good preoperative liver function and the presence of a thick capsule around the tumor are the 2 most important factors in identifying long-term survivors. The quality of liver function was predictive of both early and late mortality. Operative mortality was higher, and postoperative deterioration of liver function was significantly more frequent in class B-C than in A patients. In the former group, more than half the patients died of progression of their liver disease within 1*/z yr. Liver resection could have contributed to the aggravation of their liver failure. Other investigators have attempted to define more precisely which cirrhotic patients can best support liver resection (6,15,16). These data suggest that Pugh’s modification of Child’s classification provides a reproducible basis for preoperative selection of patients, as has already been reported for Child’s classification (17). In class B-C patients with an HCC, the choice of liver resection for primary treatment should be made very cautiously. Survival was better in patients with an expanding encapsulated tumor than in those with an infiltrating tumor. This was particularly significant in patients whose tumor was surrounded by a thick capsule. This confirms data from Nakashima et al. (13) and Kanai et al. (18). Using a slight modification of Nakashima’s classification, Kanai et al. have shown a 100% 3-yr survival in a small group of patients with completely encapsulated tumors. In the present study, cancer recurrence was significantly less in patients with encapsulated tumors than in those with infiltrating ones. The recurrence rate was particularly low in patients whose tumors were surrounded by thick capsules. This supports previous suggestions that with a thick capsule there is less invasion of adjacent parenchyma, and decreased extension to distal portal branches (14). It was previously thought that the clinical differences between the Orient and the West resulted from differences in the pathology of the tumors (6,7). We have already shown using Japanese pathological criteria that the rate of encapsulation, differentiation, and histological grading in our patients were similar to the same data from Japanese patients (14). The present data confirm both pathologically and

OF HCC IN CIRRHOSIS

737

clinically the similarities of HCC between the Orient and Western Europe. Patients with an infiltrating tumor had a high recurrence rate with no patient free of recurrence at 2 yr, suggesting that this lesion may not be suitable for resection. It would be of great importance to detect the presence or the absence of a capsule before HCC resection is undertaken in a cirrhotic patient. It has been shown that preoperative ultrasonography can make the capsule visible (19). In a retrospective study, a dynamic computed tomography scan with repeated bolus injection of 30 ml of contrast and thin repeated section planes predicted the presence of a thick capsule in 99% of cases and gave no false negative results (20). Preoperative morphological studies of the liver should be refined to detect those encapsulated tumors, which are the most suitable for surgical resection. It has been suggested that cancer recurrence and survival were related to the size of the tumor (7,17,21). Our data do not confirm this relationship. The cancer recurrence rate was slightly higher in patients with tumors 5 cm or more in diameter. However, the difference was not significant. It is noteworthy that early recurrences were observed in 4 patients with tumors of 1,3,3, and 4 cm. Therefore, encapsulation is much more important than the size of the tumor in predicting survival. Similarly, there was no difference in survival rates of patients with tumor-related symptoms and asymptomatic patients. The free margin of parenchyma is also important for postoperative recurrence and survival (4). Although it was clearly measured in only 43 resected specimens in this series, the results suggest that the risk of recurrence is greater in patients with a very thin margin. The resolution of the conflicting demands of conserving non-tumor-bearing liver parenchyma to avoid liver failure and obtaining an adequate cancer-free margin to reduce recurrence is essential to good results. Resections based on Couinaud’s classification of liver segmentation have permitted formal anatomic resection in 86% of patients with an HCC. Therefore, most operations can be anatomically based and should be designed to provide a cancer-free margin of parenchyma. In summary, the present data suggest that results of resection of HCC in cirrhosis are similar in Europe and in the Orient. In patients with good liver function, the indications for resection should be broadened, especially in those with a thickly encapsulated tumor. On the other hand, our results raise strong doubts about the usefulness of resection in patients with poor liver function or an infiltrating tumor, for whom other types of treatment such as alcohol injection (22) or, in highly selected cases, liver transplantation (23) might be more suitable.

738 FRANC0

GASTROENTEROLOGY

ET AL.

References 1. Kanematsu T, Takenaka K, Matsmnata T, Furuta T, Sugimachi K. Inokuchi K. Limited hepatic resection effective for selected cirrhotic patients with primary liver cancer. Ann Surg 1984;199: 51-56. 2. Okuda K, Ohtsuki T, Obata H, Tomimatsu M, Okazaki N, Hasegawa H, Nakajima Y, Ohnishi K. Natural history of hepatocellular carcinoma and prognosis in relation to treatment. Study of 850 patients. Cancer 1985;56:918-928. 3. Makuuchi M, Hasegawa H, Yamazaki S. Ultrasonically guided subsegmentectomy. Surg Gynecol Obstet 1985;161:346-350. 4. Lee CS, Sung JL, Hwang LY, Sheu JC, Chen DS, Lin TY, Beasley RP. Surgical treatment of 109 patients with symptomatic and asymptomatic hepatocellular carcinoma. Surgery 1986;99:481490. 5. Nagasue N. Yukaya H, Ogawa Y, Sasaki Y, Chang YC, Niimi K. Clinical experience with 118 hepatic resections for hepatocellular carcinoma. Surgery 1986;99:694-701. 6. Bismuth H, Houssin D, Ornowski J. Meriggi F. Liver resections in cirrhotic patients: a western experience. World J Surg 1986;10:311-317. 7. Belghiti J, Menu Y, Cherqui D, Nahum H, F&k&e F. Traitement chirurgical des carcinomes hepatocellulaires sur cirrhose. Inter& de l’echotomographie per-operatoire. Gastroenterol Clin Biol1986;10:224-227. 8. Gozzetti G, Mazziotti A, Cavallari A, Belhrsci R, Bolondi L, Grigioni N, Bragaglia R, Grazi GL, de Raffele E. Clinical experience with hepatic resections for hepatocellular carcinoma in patients with cirrhosis. Surg Gynecol Obstet 1988;166: 503-510. 9. Fielding LP, Stewart-Brown S, Dudley HAF. Surgeon related variables and the clinical trial. Lancet 1978:2:778-779. 10. Pugh RNH. Murray-Lyon IM, Dawson JL. Pietroni ME, Williams R. Transection of the esophagus for bleeding esophageal varices. Br J Surg 1973;60:646-649. 11. Franc0 D, Smadja C, Meakins JL, Wu A, Berthoux L, Grange D. The operative risk of elective hepatic resection for liver tumors: results of a series of one hundred consecutive hepatectomies in cirrhotic and non cirrhotic patients. Arch Surg 1989;124:10331037. 12. Couinaud C. Le foie. Etudes anatomiques et chirurgicales. Paris: Masson, 1957. 13. Nakashima T, Okuda K, Kojiro M, Jimi A, Yamaguchi R, Sakamoto K, Ikari T. Pathology of hepatocellular carcinoma in Japan. 232 consecutive cases autopsied in ten years. Cancer 1983:51:863-877. 14. Kemeny F, Vadrot J, Wu A. Smadja C, Meakins JL, Franc0 D.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Vol. 98, No. 3

Morphological and histological features of resected hepatocellular carcinoma in cirrhotic patients in the west. Hepatology 1989;9:253-257. Yamanaka N, Okamoto E, Kuwata K, Tanaka N. A multiple regression equation for prediction of posthepatectomy liver failure. Ann Surg 1984;200:658-663. Hasegawa H, Yamazaki S, Makuuchi M, Elias D. Hepatectomiespour hepatocarcinome sur foie cirrhotique: shemas decisionnels et principes de reanimation p&i-operatoire. Experience de 204 cas. J Chir 1987;124:425-431. Ohnishi K, Tanabe Y, Ryu M. Isono K, Yamamoto Y, Usui S, Hiyama Y, Goto N, Iwana S, Sugita S, Nomura F, Okuda K. Prognosis of hepatocellular carcinoma smaller than 5 cm in relation to treatment: study of 100 patients. Hepatology 1987;7: 1285-1290. Kanai T, Hirohashi S, Upton MP, Noguchi M, Kishi K, Makuuchi M, Yamasaki S, Hasegawa H, Takayasu K, Moriyama N, Shimosato Y. Pathology of small hepatocellular carcinoma. A proposal for a new gross classification. Cancer 1987;60:810-819. Sheu JC, Sung JL, Chen DS, Yu JY, Wang TH, Su CT, Tsang YM. Ultrasonography of small hepatic tumors using high resolution linear-array real-time instruments. Radiology 1984;150:797-802. Lejeune R, Meakins JL, Kemeny F. Smadja C, Franc0 D. Predictibilite pre-operatoire de l’encapsulation des carcinomes hepatocellulaires (Abstr). Gastroenterol Clin Biol 1989;13:A194. Hsu HC, Sheu JC, Chen DS, Lee CS, Hwang LY, Beasley RP. Prognostic histologic features of resected small hepatocellular carcinoma in Taiwan. A comparison with resected large HCC. Cancer 1985;56:672-680. Livraghi T, Salmi A, Bolondi L, Marin G, Arienti V. Monti F. Vettori C. Small hepatocellular carcinoma: percutaneous alcohol injection. Results in 23 patients. Radiology 1988;168:313-317. Iwatsuki S, Gordon RD. Shaw BW, Starzl TE. Role of liver transplantation in cancer therapy. Ann Surg 1985;202:401-407.

Received December 23.1988. Accepted August 18.1989. Address requests for reprints to: D. France, M.D., HBpital Paul Brousse. F-94804 Villejuif Cedex, France. This work was supported by grants from University Paris-Sud, 1’Association Claude Bernard and from Fondation pour la Recherthe Medicale. The authors thank Dr. D. Dhumeaux for carefully revising this manuscript. J. Meakins’ present address is McGill University, Department of Surgery, Royal Victoria Hospital, 687 Pine Avenue West, Montreal, Quebec H3A lA1, Canada.