Br. J. Surg. 1992, Vol. 79, October 1095-1 101
A. P. Savage and R. A. M a l t Surgical Services, Massachusetts General Hospital and Department of Surgery, Harvard Medical School, Boston, Massachusetts, USA Correspondence to: Mr A. P. Savage, Department of
Surgery, J o h n Radcliffe Hospital, Oxford OX3 9DU, UK
Survival after hepatic resection for malignant tumours A retrospective analysis of 194 patients who underwent hepatic resection .for primary or metastatic malignant disease f r o m January 1962 to December 1988 was undertaken to determine variables that might aid the selection of patients f o r hepatic resection. Hepatic metastases were the indication .for resection in 126 patients. The 5-year survival rate was I7 per cent. For patients with resected metastases f r o m colorectal cancer ( n = 1041, the survival rate at 5 years was 18 per cent. The 5-year survival rate was 27 per cent when the resection margin was > 5 mm compared with 9 per cent when the margin was 6 5 mni ( P < 0.01). No patient with extrahepatic invasion, lymphatic spread, involvement of the resection margin or gross residual disease survived to 5 years, compared with a 23 per cent 5-year survival rate f o r patients undergoing curative resection ( P < 0.02). The survival rate of patients with poorly differentiated primary tumours was nil at 3 years compared with a 20 per cent 5-year survival rate f o r patients with well or moderately diflerentiated fumours ( P not signijicant). The site and Dukes ' classification of the primary tumour, the ses and preoperative carcinoembryonic antigen level of the patient, and the number and size o f hepatic metastases did not afect the prognosis. The 5-year survival ratef o r patients with hepatocellular carcinoma ( n = 4 2 ) was 25per cent. An improved survival rate was .found .for patients whose wfetoprotein level was normal (37 per cent at 5 years) compared with those having a raised level (nil at 3 years) ( P < 0.01). Involvement of the resection margin, e.utrahepatic spread and spread to regional lymph nodes were associated with an 8 per cent 5-year survival rate versus 44 per cent .for curative resection ( P < 0.005). The presence of cirrhosis, the presence of symptoms, and the multiplicity and size of the tumour did not affect the prognosis. The 5-year survival rate of I I patients with hepatic sarcoma was 25 per cent. No patient with peripheral cholangiocarcinoma survived to I year in contrast to patients with hilar cholangiocarcinoma, all.four o f whom survived.for more than 14 months.
A previous study' has shown a reduction in overall operative mortality rates following hepatic resection from 19 per cent between 1962 and 1979 to 9.7 per cent after 1980. Further, the operative mortality rate after resection of benign hepatic neoplasms was nil compared with 4.8 per cent for resection of hepatic metastases and 20.4 per cent for resection of primary hepatic neoplasms performed after 1980'. This study has been extended to analyse the determinants of long-term survival in patients who have undergone hepatic resection for malignant disease.
Patients and methods Patients ( J I = 478) who underwent liver resection from January 1962 to December 1988 were identified by computer search of all records in the medical records and pathology departments. Of 194 patients included in the study. 68 had undergone major resection of primary malignant hepatic tumours and 126 resection of metastases. The results of segmental or wedge excision of tumours were included if surgery was curative. Couinaud's classification of hepatic resection was employed regardless of the nomenclature used in the operative report'-3. The reasons for exclusion of 284 charts were minor liver resection ('!I h/oc with other abdominal malignancy ( J I = 40), resection or biopsy of benign hepatic neoplasms ( n = 96). post-traumatic liver resection ( J I = 33). wedge excision of one of many hepatic metastases ( J I = 41 ), hepatic transplantation ( n = 49). miscellaneous resection or drainage
oOo7-1323/92/101G95-07
it''
1992 Butterworth-Heinemann Ltd
of an abscess, excision of cyst or other non-malignant condition ( 1 1 = 21) and donor hepatectomy ( J I = 4 ) . Surgery was classified as non-curative if the operation record described gross residual disease. direct spread to adjacent structures. or if there was involvement of the resection margin on histopathological examination or spread to the lymph nodes of the portal triad. Lymph node involvement regional to the primary tumour for those patients undergoing synchronous resection of a primary tumour and hepatic metastases did not preclude a curative hepatic resection. All other resections were classified as curative. Demographic. clinical and operative data were collected by review of the hospital records and surgeons' office records. Follow-up data were obtained by postal and telephone enquiry from the referring hospital or surgeon. from the tumour registry. or from the town clerk's office. Five of the 478 records (1.04 per cent) were unobtainable. Complete follow-up data were obtained for all but five patients who had undergone hepatic resection (97.4 per cent): three were lost to follow-up between 2 and 4 years after surgery. and no trace was found of t w o patients. Strrristicrrl rimilrsis
The endpoint for survival analysis was death of the patient. Analyses were performed by the life-table method': differences between survival curves were assessed using the log rank test'. The odds ratios and their 95 per cent confidence interval (c.i.) for pairs oisurvival analyses were also calculatedb. Patients who died after operation were included in the life-table analyses. Statistical significance was accepted for a two-tailed P value ~ 0 . 0 5 .
1095
Survival after hepatic resection: A. P. Savage and R. A. M a l t
Table 1 Ages qfputient.s who undmvent rcJ.sectionqf malignant ltr~patic~ neoplusms
Table 2 E.~tentof hepatic resection Primary
Metastases
No. of patients
Indication Primary hepatic malignancy Hepatocellular carcinoma Cholangiocarcinoma or gallbladder carcinoma Hepatoblastoma Sarcoma Metastases Colorectal cancer Other primary sites
Mean (range ) age (years)
68 42 12
52.4 (2-82) 54.2 (6-82) 59.0 (30-77)
3 I1 126 104 22
8.0 (2-13) 50.4 (12-71 ) 59.1 (6-80) 60.1 (28-79) 53.9 (6-80)
Colorectal Right hepatectomy Left hepatectomy Extended right hepatectomy Extended left hepatectomy Segmentectom y Wedge resection Total
Other
HCC
31 8 26
9 5 4
10
5
0
18 16
.. 7
104
Other
6
Total
15
9 3 6
55 26 51
3
0
8
2
5 3
8 0
33 21
22
42
26
194
HCC, hepatocellular carcinoma
Table 3 Outcome of.22 patients ivho undernetit hepatic resection Patient no.
of' metu.sta.se,s,from primarj. sites
Primary
Operation
Curative
Gall bladder Kidney Uterus Pancreas Lung Stomach Uterus Ileal carcinoid Ovary Kidney Kidney. Wilms' tumour Unknown (pancreas) Ileal carcinoid Unknown (pancreas) Gallbladder Kidney Breast Ocular melanoma Unknown (colon) Thyroid Ileal carcinoid Broad ligament embryonal cell tumour
Right hepatectomy Segmentectomy Right hepatectomy Right hepatectomy Left hepatectomy Segmentect om y Right hepatectomy Left hepatectomy Left hepatectomy Right hepatectomy Segmentectomy Right hepatectomy Right hepatectomy Segmentect omy Right hepdtectomy Right hepdtectomy Right hepatectomy Left hepatectomy Right hepatectomy Left hepatectomy Right hepatectomy Right hepatectomy
Yes No Yes No Yes No No No Yes No No No No No Yes Yes Yes Yes No Yes No No
otl~crthan the colorectiim
Survival (months )
Status
Operative death Lost to follow-up Operative death Died from disease Died from disease Died from disease Died from disease Died from disease Alive and well Died from disease Died from disease Died from disease Alive with recurrence Died from disease Died from disease Died from disease Alive and well Died from disease Died from disease Died from disease Alive with recurrence Alive and well
~-
I 7
3 4 5 6 7 8 9 10
11
12 13 14
15 16 17 18 19 20 21
77 --
0.0 0.4 1.1
2.6 2.7 5.8 6.2 7.7 8.8 9.3 9.5 14.9 15.9 18.1 24.4 28.0 33.5 35.7 38.7 58.1 87.8 130.0
100 80 W e m
60 40 20
t
I-_-___
I 0
12
24
36
40
60
0
Time a f t e r o p e r a t i o n [ m o n t h s )
I
I
36
48
I
1
60
No. a t r i s k 74 13
44 8
26 5
15 4
10 3
Figure 1 Survival curves f o r patients who underwent hepatic resection for metastases from colorectal cancer (-) and for metastases from primary carcinoma at other sites (-----). P not significant (log rank test)
1096
I
24
Time a f t e r o p e r a t i o n ( m o n t h s )
No. a t risk Colorectal 104 primary Other site 2 2
t
12
Curative Noncurative
76
58
37
24
14
28
18
9
3
2
9
Figure 2 Surriral curre.s,for patients iiho uncieriimt curatire (-) and rior~-curntire(-----) hepatic resection of mr1astarse.s f r o m colorecttrl cat7ccr. P < 0.02 (log r~t7litc.Yt)
Br. J. Surg., Vol. 79, No. 10, October 1992
S U,rvival after hepatic resection: A. P. Savage and R. A. Malt incidental laparotomy. Routine follow-up of patients with colorectal cancer disclosed hepatic metastases on investigation of a raised carcinoembryonic antigen (CEA ) level ( 37 patients ), abnormal liver function test results (five), routine radionuclide scintigraphy or ultrasonographic liver scan (ten), or hepatomegaly discovered on routine clinical examination (two ). Women had a survival advantage over men ( P = 0.06). The survival rate of patients with poorly differentiated primary tumours was nil at 3 years uersus a 5-year survival rate of 20 per cent for patients with well or moderately differentiated tumours. These differences failed to reach statistical significance because of the small numbers of patients with poorly differentiated primary tumours (Table 4 ).
Results Resection of metastases was the indication for hepatectomy in 126 patients; the site of the primary was the colon or rectum in 104. Of 68 patients who underwent resection of malignant tumours originating in the liver, histopathological examination showed hepatocellular carcinoma ( H C C ) in 42. Of these, three patients showed mixed HCC and cholangiocarcinoma, and cirrhosis was present in nine. Histopathological examination of the remaining 26 primary hepatic malignant tumours revealed cholangiocarcinoma ( 12 ) including carcinoma of the gallbladder (four ), hepatic sarcoma ( 1 1 ) and hepatoblastoma (three). Patients’ ages are shown in Tuble f and the extent of hepatic resection in Table 2. Metcisrases The overall 5-year survival rate of patients who underwent resection of the liver for metastatic disease was 17 per cent. The 5-year survival rate was 18 per cent for patients with metastases from colorectal cancer versus 14 per cent for those with metastases from other primary sites ( F i g u r e 1 ). The site of the primary tumour and the outcome of patients with metastases from primary sites other than the colorectum is shown in Tuhle3. For three patients, the site of the primary was unknown at the time of hepatic resection. Subsequent investigation showed that these were metastases from the pancreas (two patients) and colon (one ). M e rustuses ,from c.olor.ec.tci1cuncer The 5-year survival rate for 76 patients who underwent curative operation was 23 per cent. Resection of metastases was non-curative in 28 patients: eight cases showed involvement of the resection margin, eight had gross residual disease, five showed direct spread to adjacent structures, and seven spread to lymph nodes of the portal triad. The 5-year survival rate for patients undergoing non-curative resection was zero ( P < 0.02 ) (Figure 2). The resection margin was the most significant prognostic indicator. A margin > 5 mm was associated with a 5-year survival rate of 27 per cent compared with 9 per cent for lesser margins ( P < 0.01 ) ( F i y u r e 3 ). The proportional hazard of survival was 0.48 (95 per cent c.i. 0.29-0%1 ) if the resection margin was < 5 mm cornpared with a margin > 5 mm. Asymptomatic patients had a better survival rate than symptomatic patients ( P < 0.05) ( T u h l p 4 ) .Only 14 patients presented with symptoms from liver metastases (abdominal pain in 11, and fever, weight loss and malaise in three patients). Hepatic metastases were.discovered during investigation or at operation for the primary tumour in 33 patients. Unsuspected hepatic metastases were discovered in three patients undergoing
f f e p a t o c e ~ / u ~ acrirciiiomu r.
The overall median survival of patients who had undergone resection of HCC was 12 months, and the 5-year survival rate 25 per cent (Tuble5). There were no differences in survival rates between patients operated on before or after 1980. Hepatectomy for HCC was classified as non-curative in 27 patients: 15 showed involvement of the resection margin on histopathological examination, five showed the presence of gross residual disease, three exhibited direct invasion ofadjacent structures, and regional lymph node involvement was present in four. Curative operations were associated with a 44 per cent 5-year survival rate compared with 8 per cent for non-curative operations ( P < 0.001 ) (Figure 4). Only nine patients had a resection margin > 5 mm; their 5-year survival rate was 40 per cent compared with a 17 per cent survival rate for patients with a margin < 5 mm ( P not significant ). However. if patients who had undergone non-curative operations were excluded, the 5-year survival rate was 54 per cent if the margin of resection was > 5 mm versus 43 per cent for patients with a lesser margin ( P not significant ). Preoperative r-fetoprotein estimation was performed for 25 patients. The median serum level was 19 ng/ml (range 9-447984 ng/ml; normal < 12.5 ng,’ml). Patients whose preoperative estimation was within the normal range had a 37 per cent 5-year survival rate compared with none of the patients with high levels ( P < 0.01 ) ( F i g u r e 5 ). The odds ratio for survival was 0.26 (95 per cent c.i. 0.11-0.68) for patients with high r-fetoprotein levels. There was no significant difference in the survival rate between patients who were symptomatic and those who were not ( Table 5 ). HCC was an incidental finding in seven patients; the diagnosis was made at incidental laparotomy in three. Thirty-five patients were symptomatic: pain was the presenting complaint for 16, six discovered a mass, one had jaundice. six had malaise, fever and weight loss, and six underwent emergency resection for ruptured HCC.
80 I-__
I_ Q)
Y
m
m
20
I-
$-_---
I-
‘-----
> .t_
-I--
0
2
-
m
I
I
I
I
I
12
24
36
48
60
L
Time a f t e r o p e r a t i o n ( m o n t h s )
0
No. a t r i s k
Margin > 0.5 cm Margin \< 0.5 cm
I
I
I
I
I
12
24
36
48
60
Time a f t e r o p e r a t i o n ( m o n t h s ) No. a t risk
51
43
25
17
8
7
46
27
17
8
6
3
Figure 3 Survival curves f o r patients who underwent hepatic resection for metastases from colorectal cancer with a resection margin > 5 mm (-) and 2 years
59 37
26 24
75 78
44 20
23
0.157
0.68 (0.40- 1 . l 6 )
Synchronous disease Metachronous disease
33 71
43 23
75 77
58 25
25 16
0.1 I3
0.67 ( 0 4 - 1 . 1 0 )
Asymptomatic patients Symptomatic patients
90 14
26 I1
81 47
38 14
19 14
0.043
0.42 (0.18-0.97)
Normal CEA Raised CEA
19 62
25 24
61 87
33 31
22
0.479
0.78 (0.40- I '53 )
Female Male
51 53 57 47
31 23
84 71
42 30
24 13
0.058
0.64 (0.39- 1.02 )
31 17
78 76
41 31
22 13
0,168
0.7 1 (0.43- 1.1 5 )
Solitary metastasis Multiple metastases
10
18
-
No. of metastases 1-3 24 Size of metastases < 5 cm >5cm
87 17
26 13
79 62
36 28
19 14
0.343
0.70 (0.34-1.46)
52 43
26 24
84 66
41
31
22 9
0.058
0.60 (0.36- 1.01 )
Curative resection Non-curative resection
76 28
30 21
80 66
42 17
23 0
0.0 I0
0.43 (023-0.82)
Resection margin > 5 mm Resection margin < 5 mm
51 46
35 20
88 62
46 24
27 9
0.006
0.48 (0.29-0.8 1 )
~
CEA, carcinoembryonic antigen; c.i., confidence interval. *Log rank test
Table 5 Suriirul rute ,follrm~ingreseciion of hepuiocellulur curcinoniu No. of patients
Median survival (months)
1 year
All patients
42
12
Resection before 1980 Resection after 1980
14 28
Male Female
Survival rate ( Y O ) P*
Odds ratio (95% c.i.)
28 I5
0.168
0.60 (0.29- 1.23)
42 14
31 14
0.527
0.80 (040- 1.60)
50 57
24 28
19 28
0,479
1.37 (058-3.24)
50 5
67 33
56 0
37 0
0.005
0.26 ( 0.1 I -0.68 )
33 9
16 6
54 38
26
22 0
0. I63
0.49 (017-1.36)
Solitary tumour Multifocal tumour
22 20
15
50 53
31 18
23
14
0.764
0.90 (0.45- 1.8 1 )
Tumour size < 5 cm Tumour size > 5 cm
10
32
24 12
60 48
25 24
25 19
0.60 I
0.8 I (0.36- I '78 )
Curative resection Non-curative resection
15 27
79 37
56 8
44 8
0.00 I
0.30 (015-0.62)
Resection margin > 5 m m Resection margin < 5 m m
9 33
50 6 27 12
67 47
40 20
40 17
0,180
0.58 (0.26- I .28 )
3 years
5years
51
25
25
24 8
64 41
28 23
20 22
12
16
48 54
Symptomatic patients Asymptomatic patients
35 7
14 24
Normal z-fetoprotein Raised r-fetoprotein
10
15
Patients without cirrhosis Patients with cirrhosis
_
_
_
_
_
~
~
.
.
19
18
c.i., Confidence interval. *Log rank test
1098
Br. J. Surg., Vol. 79, No. 10, October 1992
Survival after hepatic resection: A. P. Savage and R. A. M a l t
1 '-
L
I
'-1
7
0
I
I
12
24
I
I
I
I
36
48
60
Time a f t e r o p e r a t i o n ( m o n t h s )
No. a t r i s k
a -fetoprotein (12nglml 10
7
6
6
4
5
4
2
a -fetoprotein > 1 2 ngfml
15
Figure 5
Surr.iro1 curves ,/i)r prrrierirs will1 resecleci hepu~ocrllulcrr ii.it/i preoperative r:feroprotein leuel < 12 t i g l t l l l (-) atid > 12 t i g ; t i i l (-----). P < 0 4 1 (log runk irst)
ccrrcitioriicr
Clioluny ioccrscinomu Hepatic resection was performed for five patients with hilar cholangiocarcinoma, three with peripheral cholangiocarcinoma and four with carcinoma of the gallbladder involving the fossa. The median survival of this group of patients was 10 months; one patient survived to 5 years (5-year survival rate 22 per cent). Although the small number of patients in this group precluded life-table analysis of subgroups, the three patients with peripheral cholangiocarcinoma all died within 12 months while all four with hilar cholangiocarcinoma who survived the operation were alive at 14 months. All patients with carcinoma of the gallbladder died within 2 years, three within the first year. Hepcitic sLircotmi Histopathological examination of the 11 patients with hepatic sarcoma showed leiomyosarcoma (five patients ), fibrosarcoma (two ), malignant histiocytoma (one ), malignant schwannoma (one), lymphoma ( o n e ) and undifferentiated sarcoma (one). The tumour originated in the liver in six patients and metastasized to the liver from the stomach (two), retroperitoneum ( t w o )and ileum (one ) in five. One of the 1 1 patients was alive at 5 years (5-year survival rate 25 per cent ). Three children underwent resection of a hepatoblastoma. Although two were alive and well at 55 and 150 months after operation, the third had recurrent disease at 39 months.
Discussion This study examines whether it is possible to improve the long-term survival rates of patients with hepatic malignancy by defining criteria for the selection of patients for hepatic resection. While the operative mortality rate for hepatic resection has fallen in recent years', the present data show no improvement in the overall 5-year survival rate. Resection of' I ~ ~ P ~ N S ~ fkorn C I S ~ colosect~~l S , cmcu The efficacy of resection of liver metastases from tumours of the colon and rectum has recently been questioned7. In previous studies, the 5-year survival rate after resection of hepatic metastasess-I6 has been between 10 and 34 per cent, which is consistent with the present data. The survival of patients with untreated metastases to the liver from primary tumours of the is between 5.6 and 13 months. However, colon and patients with multiple, but potentially resectable, metastases survive a mean of 10.6 months, and those with solitary hepatic m e t a ~ t a s e survive s ~ ~ 16.7 months. There are no controlled trials
Br. J. Surg., Vol. 79, No. 10. October 1992
comparing hepatic resection with non-surgical management for potentially resectable hepatic metastases. None the less, patients in this series who had undergone hepatic resection for metastases had a better survival rate than historical patients treated without o p e r a t i ~ n ' ~ ~ ~ ~ * ~ ~ . The histological grade of the primary tumour may be important in the selection of patients for resection of liver metastases. Few patients with poorly differentiated primary neoplasms underwent resection of liver metastases; therefore the worse survival of patients with poorly differentiated tumours compared with those with moderately or well differentiated tumours failed to reach statistical significance. The present data accord with a previous report showing a mean survival of 7.1 months for patients with resected metastases from poorly differentiated tumours compared with 17.9 months for those with resected metastases from moderately or well differentiated tumours", and suggest that patients with metastases from poorly differentiated tumours may not benefit from hepatic resection. Consistent with other reports, the present data show that the Dukes' classification of the primary cancer is not important in the selection of patients for resection of hepatic m e t a s t a ~ e s ' ~ . ~Other ~ - ~ ~groups, . however, have reported 5-year survival rates between 7 and 45 per cent higher after resection of hepatic metastases from Dukes' B compared with Dukes' C lesion^'^^'^^^^-^^. None the less, the difference in survival rate was insufficient to deny patients with Dukes' C primary tumours the potential for cure. The interval between resection of the primary tumour and resection of hepatic metastases is not important in the selection of patients for resection of colorectal liver metastases. The present data show a trend towards a better survival rate if this interval is g 2 years, and for patients who have undergone resection of metastases discovered during surgery for the primary tumour. However, the improved survival rate of patients with synchronous metastases, or of those whose close follow-up resulted in the early detection of metastases. may represent a lead-time bias. other^^^.^^ have reported a longer median survival for patients with synchronous metastases of between 4 and 9 months and a 66 per cent 5-year survival rate for patients with synchronous metastases compared with zero for patients with metachronous metastases' I . Conversely, others have reported a 5-year survival rate between 10 and 26 per cent better for patients with metachronous m e t a s t a ~ e s ~An ~ .improved ~~. 5-year survival rate of 40 per cent for patients with a disease-free interval > 1 year, compared with 28 per cent for a shorter disease-free interval, has also been reported' possibly reflecting a lower biological activity of tumours showing a long disease-free interval. However, most studies have shown no significant differences in the survival rate after resection of liver metastases when analysed according to the interval between resection of the primary and the discovery of hepatic metastasess.'0~'b.2s.29.40-43; further, this interval is of no account in the selection of patients for hepatic resection or in the determination of prognosis. The presence of symptoms from hepatic metastases may not be a contraindication to resection. The present data agree with those from a multicentre study that reported a 32 per cent 5-year survival rate for patients with symptoms from hepatic metastases compared with 45 per cent when metastases were ~. discovered on routine follow-up or i n v e ~ t i g a t i o n ~Because symptomatic metastases may be considered to be at a later stage in their growth. a lead-time bias may account for the better survival of patients with asymptomatic metastases. The findings at operation are of great importance in determining the prognosis after resection of metastases from colorectal cancer. The present data confirm a previous report' that non-curative resection of liver metastases is associated with a 5-year survival rate of less than 5 per cent. The presence of neoplasia at the resection margin'5.27.2yand of margins of at least 5 mm (Reference 44), 10 mm (References 9, 15. 34) and 20 mm (Reference 45) have been analysed with respect to the long-term survival rate. These reports support the present data,
',
1099
Survival after hepatic resection: A. P. Savage and R. A. Malt
which show a significantly improved 5-year survival rate for resection per se is not important in the prognosis of HCC. The patients with a margin of resection of > 5 mm. These data also 5-year survival rate was similar for patients undergoing curative show the futility of resection of metastases in the presence of resection whether or not the resection margin was > 5 mm. lymphatic involvement regional to the hepatic metastases, Although one study has reported a 533-day disease-free survival extrahepatic spread or in circumstances resulting in gross for patients with a negative margin of resection iwws 160 days residual disease after resection. when the margin was involved by tumour5’, another6’ has The number of metastases may not be relevant in the reported no significant difference in recurrence of tumour in selection of patients for hepatic resection, provided that there the liver after limited resection with a resection margin < 10 mm is no extrahepatic spread and an adequate margin of resection compared with patients whose tumours were resected with a can be achieved. These data support previous reports showing margin > 10 mm. Similarly, for patients with small HCC no worse survival in patients with multiple m e t a s t a s e ~ ’ ~ - ’ ~ . ~complicating ~-~~. cirrhosis, limited hepatic resection was associated Other reports have suggested that the survival rate is better with a trend towards better survival than for those undergoing following resection of solitary metastases9~’5~27~2’~30~34~37~41.62~46. major hepatic resection61. N o discrimination was made in the present study between The median survival of patients with HCC complicating multifocal and solitary metastases with satellite lesions. One cirrhosis is 1.2 uersus 13 months for those with HCC in study, however, reported that satellite lesions around solitary non-cirrhotic liver”. Increasing severity of cirrhosis, as defined metastases d o not affect prognosis4’, while another showed a by Pugh’s modification of Child’s classification, adversely worse prognosis with satellite lesions’. influences both the operative mortality rate (3.7 per cent for Child’s grade A, 16.7 per cent for grade B / C ) and the long-term survival rate (56.7 per cent 5-year survival for Child’s grade A, Resection of meinstuses .from other prinzury sites 12.2 per cent for grade B / C ) of patients undergoing resection Although no significant differences were found when patients of HCC complicating cirrhosis62. The safe limits of hepatic with metastases from sites other than the colorectum were resection are dependent not only on the volume uf liver resected compared with those undergoing resection of metastases from but also on preoperative liver f ~ n c t i o n ~ ’None . ~ ~ . the less, one the colorectum, the histology of the primary tumour was report showed no statistical difference in the survival rates of important in determining the prognosis. Of the five surviving patients undergoing resection of HCC from a cirrhotic liver patients, two with slowly growing metastases from ileal compared with those without cirrhosis”. However, 58 per cent carcinoid tumours who underwent palliative resection are alive of the patients with HCC in a cirrhotic liver underwent minor in addition to three with gynaecological or breast carcinomas. resection compared with 24 per cent of those without Patients with metastases from carcinoma of the stomach and cirrhosis. Only nine of the present patients with HCC had pancreas fared badly. Other reports with similar results’ 394b cirrhosis; the trend towards worse survival in these patients also d o not recommend resection of metastases from primary was not statistically significant. sites other than the colorectum with intent to cure, although These data indicate that the selection of patients for resection palliative resection may be indicated for metastases from some of primary and secondary liver tumours may improve survival. tumours, for example ileal carcinoid tumours and those from The major contraindications to resection are finding that a rare sites such as mesonephric duct carcinoma. tumour has extrahepatic spread or regional lymph node involvement, or both, or that it is not resectable without leaving Pritnury liver tuniours gross residual neoplasia. For metastases from colorectal cancer, Patients with untreated HCC survive between 1 and 4 the margin of resection and possibly the histological grade of months after d i a g n o s i ~ ~ ’ - Those ~ ~ . fit enough to undergo the tumour are important. For HCC an r-fetoprotein level palliative surgery survive between 7 (References 48, 53 ) and > 1 2 5 ng/ml may be a contraindication to resection. 10.8 (Reference 50) months while patients undergoing curative resections survive between 23 (Reference 53) and 42 (Reference 50) months. In selected patients, surgical resection is superior to systemic or regional ~ h e m o t h e r a p y ~ ’ . ~ ’ . ~ References radiotherapy’ or hepatic artery ligation or dearterializ1 . Savage AP, Malt RA. Elective and emergency hepatic resection: results of non-surgical therapy support a t i ~ n ~ ’ The . ~ ~ poor . determinants of operative mortality and morbidity. Ann Surg the case for surgery despite the higher operative risk of resection 1991; 214: 689-95. for HCC than for metastatic liver disease’. 2. Couinaud C. Le Foie: Etudes Anaiomiques el Chirurgicales. Paris: The r-fetoprotein level may be useful in the preoperative Masson, 1957: 283-9. selection of patients for hepatic resection for HCC. The present 3. Malt RA. Surgery for hepatic neoplasms. N Engl J Med 1985; data show a worse survival rate for patients with a raised level 313: 1591-6. of r-fetoprotein before operation. One report” has shown a 4. Kaplan EL, Meier P. Nonparametricestimation from incomplete median disease-free survival of 531 days for patients whose observations. J Am Statist Assoc 1958; 53: 457-81. 5. Mantel N . Evaluation of survival data and two new rank order r-fetoprotein level was 200ng/ml, and a second study has 50: 163-70. confirmed this better survival5h. 6. Machin D, Gardner MJ. Calculating confidence intervals for The feasibility of performing a curative resection is most survival time analyses. BMJ 1988; 296: 1369-71. relevant in the selection of patients for resection of HCC. The I . Silen W. Hepatic resection for metastases from colorectal present data agree with a previous study5’ which reported a carcinoma is of dubious value. Arch Surg 1989; 124: 1021-2. 70 per cent 5-year survival rate for those undergoing curative 8. Fegiz G, Bezzi M, de Angelis R et al. Surgical treatment of liver resection but no survivors amongst patients undergoing metastases from colorectal cancer. Zial J Surg Sci 1985; 15: palliative resection, defined as the absence of a capsule around 259-65. 9. Ekberg H, Tranberg K-G, Andersson R et al. Determinants of the tumour, the presence of satellite lesions, the presence of survival in liver resection for colorectal secondaries. Br J Surg vascular or bile duct invasion, or a margin of resection 1986; 13: 121-31. < 10 mm. Similarly, other studies have reported a 50 per cent 10. Nordlinger B, Parc R, Delva E, Quilichini M-A, Hannoun L, 5-year survival rate in the absence of extrahepatic spread Huguet C. Hepatic resection for colorectal liver metastases. Ann compared with no 5-year survivors among patients with a Surg 1987; 205: 256-63. locally invasive tumour5’, and a 75 per cent 5-year survival 11. Coppa GF, Eng K, Ranson JHC, Gouge TH, Localio A. Hepatic rate for patients with completely resected tumours compared resection for metastatic colon and rectal cancer. Ann Surg 1985; with 11 per cent for those whose tumours had spread locally 202: 203-8. or to distant sites59. 12. Wilson SM, Adson MA. Surgical treatment of hepatic metastases However, the present data also suggest that the margin of from colorectal cancers. Arch Surg 1976. 111: 330-4.
1100
Br. J. Surg., Vol. 79, No. 10. October 1992
Survival after hepatic resection: A. P. Savage and 13. 14.
15. 16. 17. 18.
19.
20. 21.
22. 23. 24. 25.
26. 27. 28. 29. 30. 31. 32. 33.
34. 35. 36. 37. 38.
Iwatsuki S, Starzl TE. Personal experience with 411 hepatic resections. Ann Surq 1988; 208: 421-34. Stephenson KR. Steinberg SM, Hughes KS, Vetto JT, Sugarbaker PH, Chang AE. Perioperative blood transfusions are associated with decreased time t o recurrence and decreased survival after resection of colorectal liver metastases. Ann Sury 1988; 208: 679-87. Hughes K, Scheele J, Sugarbaker PH. Surgery for colorectal cancer metastatic to the liver. Sury Clin North Ani 1989; 69: 339-59. Butler J, Attiyeh FF, Daly JM. Hepatic resection for metastases ofthecolonand rectum. Sury GynecolOhster 1986; 162: 109-13. Bengtsson G, Carlsson G, Hafstrom L, Jonsson P-E. Natural history of patients with untreated liver metastases from colorectal cancer. Am J Sur{/ 1981; 141: 586-9. Bengmark S. Hafstrom L. The natural history of primary and secondary malignant tumors of the liver. Cmcer 1969; 23: 198-202. Lahr CJ, Soong S-J, Cloud G, Smith JW, Urist M M , Balch C M . A multifactorial analysis of prognostic factors in patients with liver metastases from colorectal carcinoma. J Clin O i d 1983; 1: 720-6. Pestana C, Reitmeier RJ, Moertel CG. Judd ES, Dockerty MB. The natural history of carcinoma of the colon and rectum. Ani J Surg 1964; 108: 826-9. Finan PJ, Marshall RJ, Cooper EH, Giles GR. Factors affecting survival in patients presenting with synchronous hepatic metastases from colorectal cancer: a clinical and computer analysis. Br J Sury 1985; 72: 373-7. Palmer M, Petrelli NJ, Herrera L. No treatment option for liver metastases from colorectal adenocarcinoma. Dis Cokm Recmni 1989; 32: 698-701. Cady B, Monson DO, Swinton NW Sr. Survival of patients after colonic resection for carcinoma with simultaneous liver metastases. Sury Gjmecol Ohstet 1970; 131 : 697-700. Wood CB, Gillis CR, Blumgart LH. A retrospective study of the natural history of patients with liver metastases from colorectal cancer. Clin Oncol 1976; 2: 285-8. Wagner JS, Adson MA, van Heerden JA. Adson MH. Ilstrup DM. The natural history of hepatic metastases from colorectal cancer: a comparison with resective treatment. A m Sury 1984; 199: 502-8. Adson MA, van Heerden JA, Adson MH, Wagner JS. Ilstrup D M . Resection of hepatic metastases from colorectal cancer. Arch Sury 1984; 119: 647-51. Holm A, Bradley E. Aldrete JS. Hepatic resection of metastasis from colorectal carcinoma: morbidity, mortality and pattern of recurrence. Ann Sury 1989; 209: 428-34. Gennari L, Doci R, Bozzetti F, Bignami P. Surgical treatment of hepatic metastases from colorectal cancer. Ann Surg 1986; 203:49-54. August DA. Sugarbaker PH, Ottow RT, Gianola FJ, Schneider PD. Hepatic resection of colorectal metastases. Ann Sury 1985; 201: 210-18. Logan S, Meier SJ, Ramming KP, Morton DL, Longmire W P Jr. Hepatic resection of metastatic colorectal carcinoma. Arch Sury 1982; 117: 25-8. Nims TA. Resection of the liver for metastatic cancer. Sury G J I I C VOb.s/c’t ~ 1984; 158: 46-8. Thompson HH, Tompkins RK. Longmire W P Jr. Major hepatic resection: a 25-year experience. Ann Sury 1983; 197: 375-88. Fortner JG, Silva JS, Golbey RB, Cox EB. Maclean BJ. Multivariate analysis of a personal series of 247 consecutive patients with liver metastases from colorectal cancer: I. Treatment by hepatic resection. Ann Sury 1984; 199: 306-16. Registry of Hepatic Metastases. Resection of the liver for colorectal carcinoma metastases: a multi-institutional study of indications for resection. Surgery 1988; 103: 278-88. Morrow CE, Grage TB, Sutherland DER, Najarian JS. Hepatic resection for secondary neoplasms. Suryery 1982; 92: 610-14. Petrelli NJ, Nambisan RN, Herrera L, Mittleman A. Hepatic resection for isolated metastasis from colorectal carcinoma. Am J Sury 1985; 149: 205-9. Bengmark S, Hafstrom L, Jeppsson B, Jonsson P-E, Ryden S, Sundqvist K. Metastatic disease in the liver from colorectal cancer:anappraisalofliver surgery. WorldJSurg 1982;6:61-5. Sesto ME, Vogt DP. Hermann RE. Hepatic resection in 128
Br. J . Surg., Vol. 79, No. 10, October 1992
R. A. M a l t
patients: a 24-year experience. Surgery 1987; 102: 846-51. Tomas-de la Vega JE, Donahue EJ. Doolas A ct ( I / . A ten year experience with hepatic resection. Surg Gjwc,col Ohstcv 1984; 159: 223-8. 40. Kortz WJ, Meyers WC, Hanks JB. Schirmer BD. Jones RS. Hepatic resection for metastatic cancer. A m Sury 1984; 199: 182-6. 41. Taylor B, Langer B, Falk RE. Arnbus U. Role of resection in the management of metastases to the liver. Ccui J Simq 1983: 26: 215- 17. 42. Cobourn CS, Makowka L, Langer B, Taylor BR, Falk RE. Examination of patient selection and outcome for hepatic resection for metastatic disease. Sury G ! m d Ohstc,f 1987; 165: 239-46. 43. Vogt P. Raab R, Ring B, Pichlmayr R. Resection of synchronous liver metastases from colorectal cancer. World J S m / 1991 ; 15: 62-7. 44. Bradpiece HA, Benjamin IS, Halevj A, Blumgart LH. Major hepatic resection for colorectal liver metastases. Br J Sury 1987; 74: 324-6. 45. Rajpal S, Dasmahapatra KS, Lecesma EJ, Mittelman A. Extensive resections of isolated metastasis from carcinoma of the colon and rectum. Sury Gj.neco/ Ohsir/ 1982; 155: 813-16. 46. Foster JH. Survival after liver resection for secondary tumors. A m J Sury 1978; 135: 389-94. 47. Lee NW. Wong J, Ong GB. The surgical management of primary carcinoma of the liver. World J Sury 1982; 6: 66-75. 48. Bengmark S, Hafstrom L, Jeppsson B, Sundqvist K. Primary carcinoma of the liver: improvement in sight? WorltlJSu,y 1982; 6: 54-60. 49. Inouye AA, Whelan TJ Jr. Primary liver cancer: a review of 205 cases in Hawaii. Ani J Sury 1979; 138: 53-61. 50. Honjo I. Mizumoto R. Primary carcinoma of the liver. A m J SUU/ 1974; 128: 31-6. 51. Davidson AR, Tomlinson S, Calne R k , Williams R. The variable course of primary hepatocellular carcinoma. Br J Surg 1974; 61 : 349-52. 52. Nagasue N, Yukaya H, Hamada T, Hirose S, Kanashima R. Inokuchi K. The natural history of hepatocellular carcinoma: a study of 100 untreated cases. Cuncer 1984; 54: 1461-5. 53. Nagorney DM, van Heerden JA, Ilstrup DM. Adson MA. Primary hepatic malignancy: surgical management and determinants of survival. Suryerj, 1989; 106: 740-9. 54. Appelqvist P. Primary carcinoma of the liver: clinical course and therapeutic results. J Sury Oncol 1982; 21: 87-93. 55. Lai EC-S, Ng 10-L. Ng MM-T c t a/. Long-term results of resection for large hepatocellular carcinoma: a multivariate analysis of clinicopathological features. Heptfology 1990; 11: 815-18. 56. Nagao T. Goto S. Kawano N (’1 d.Hepatic resection for hepatocellular carcinoma. Ann Sury 1987; 205: 33-40. 57. Okamoto E, Tanaka N , Yamanaka K , Toyosaka A. Results of surgical treatments of primary hepatocellular carcinoma: some aspects to improve long-term survival. World J Surg 1984; 8: 360-6. 58. Smalley SR. Moertel CG, Hilton J F e f a/. Hepatoma in the noncirrhotic liver. Cuncer 1988; 62: 1414-24. 59. Fortner JG, Maclean BJ, Kim D K et ti/. The seventies evolution in liver surgery for cancer. Cuncer 1981; 47: 2162-6. 60. Yoshida Y, Kanematsu T, Matsumata T, Takenaka K, Sugimachi K. Surgical margin and recurrence after resection of hepatocellular carcinoma in patients with cirrhosis. Ann Sury 1989; 209: 297-301. 61. Kanematsu T, Takenaka K, Matsumata T , Furuta T , Sugimachi K. Inokuchi K. Limited hepatic resection effective for selected cirrhotic patients with primary liver cancer. Ann Sury 1984; 199: 51-6. 62. Franco D, Capussotti L, Smadja C er NI. Resection of hepatocellular carcinomas: results in 72 European patients with cirrhosis. Gusrroenteroloyy 1990; 98: 733-8. 63. Okamoto E, Kyo A, Yamanaka N, Tanaka N, Kuwata K. Prediction of the safe limits of hepatectomy by combined volumetric and functional measurements in patients with impaired hepatic function. Surgery 1984; 95: 586-92.
39.
Paper accepted 28 March 1992
1101
A. P. Savage and R. A. M a l t Surgical Services, Massachusetts General Hospital and Department of Surgery, Harvard Medical School, Boston, Massachusetts, USA Correspondence to: Mr A. P. Savage, Department of
Surgery, J o h n Radcliffe Hospital, Oxford OX3 9DU, UK
Survival after hepatic resection for malignant tumours A retrospective analysis of 194 patients who underwent hepatic resection .for primary or metastatic malignant disease f r o m January 1962 to December 1988 was undertaken to determine variables that might aid the selection of patients f o r hepatic resection. Hepatic metastases were the indication .for resection in 126 patients. The 5-year survival rate was I7 per cent. For patients with resected metastases f r o m colorectal cancer ( n = 1041, the survival rate at 5 years was 18 per cent. The 5-year survival rate was 27 per cent when the resection margin was > 5 mm compared with 9 per cent when the margin was 6 5 mni ( P < 0.01). No patient with extrahepatic invasion, lymphatic spread, involvement of the resection margin or gross residual disease survived to 5 years, compared with a 23 per cent 5-year survival rate f o r patients undergoing curative resection ( P < 0.02). The survival rate of patients with poorly differentiated primary tumours was nil at 3 years compared with a 20 per cent 5-year survival rate f o r patients with well or moderately diflerentiated fumours ( P not signijicant). The site and Dukes ' classification of the primary tumour, the ses and preoperative carcinoembryonic antigen level of the patient, and the number and size o f hepatic metastases did not afect the prognosis. The 5-year survival ratef o r patients with hepatocellular carcinoma ( n = 4 2 ) was 25per cent. An improved survival rate was .found .for patients whose wfetoprotein level was normal (37 per cent at 5 years) compared with those having a raised level (nil at 3 years) ( P < 0.01). Involvement of the resection margin, e.utrahepatic spread and spread to regional lymph nodes were associated with an 8 per cent 5-year survival rate versus 44 per cent .for curative resection ( P < 0.005). The presence of cirrhosis, the presence of symptoms, and the multiplicity and size of the tumour did not affect the prognosis. The 5-year survival rate of I I patients with hepatic sarcoma was 25 per cent. No patient with peripheral cholangiocarcinoma survived to I year in contrast to patients with hilar cholangiocarcinoma, all.four o f whom survived.for more than 14 months.
A previous study' has shown a reduction in overall operative mortality rates following hepatic resection from 19 per cent between 1962 and 1979 to 9.7 per cent after 1980. Further, the operative mortality rate after resection of benign hepatic neoplasms was nil compared with 4.8 per cent for resection of hepatic metastases and 20.4 per cent for resection of primary hepatic neoplasms performed after 1980'. This study has been extended to analyse the determinants of long-term survival in patients who have undergone hepatic resection for malignant disease.
Patients and methods Patients ( J I = 478) who underwent liver resection from January 1962 to December 1988 were identified by computer search of all records in the medical records and pathology departments. Of 194 patients included in the study. 68 had undergone major resection of primary malignant hepatic tumours and 126 resection of metastases. The results of segmental or wedge excision of tumours were included if surgery was curative. Couinaud's classification of hepatic resection was employed regardless of the nomenclature used in the operative report'-3. The reasons for exclusion of 284 charts were minor liver resection ('!I h/oc with other abdominal malignancy ( J I = 40), resection or biopsy of benign hepatic neoplasms ( n = 96). post-traumatic liver resection ( J I = 33). wedge excision of one of many hepatic metastases ( J I = 41 ), hepatic transplantation ( n = 49). miscellaneous resection or drainage
oOo7-1323/92/101G95-07
it''
1992 Butterworth-Heinemann Ltd
of an abscess, excision of cyst or other non-malignant condition ( 1 1 = 21) and donor hepatectomy ( J I = 4 ) . Surgery was classified as non-curative if the operation record described gross residual disease. direct spread to adjacent structures. or if there was involvement of the resection margin on histopathological examination or spread to the lymph nodes of the portal triad. Lymph node involvement regional to the primary tumour for those patients undergoing synchronous resection of a primary tumour and hepatic metastases did not preclude a curative hepatic resection. All other resections were classified as curative. Demographic. clinical and operative data were collected by review of the hospital records and surgeons' office records. Follow-up data were obtained by postal and telephone enquiry from the referring hospital or surgeon. from the tumour registry. or from the town clerk's office. Five of the 478 records (1.04 per cent) were unobtainable. Complete follow-up data were obtained for all but five patients who had undergone hepatic resection (97.4 per cent): three were lost to follow-up between 2 and 4 years after surgery. and no trace was found of t w o patients. Strrristicrrl rimilrsis
The endpoint for survival analysis was death of the patient. Analyses were performed by the life-table method': differences between survival curves were assessed using the log rank test'. The odds ratios and their 95 per cent confidence interval (c.i.) for pairs oisurvival analyses were also calculatedb. Patients who died after operation were included in the life-table analyses. Statistical significance was accepted for a two-tailed P value ~ 0 . 0 5 .
1095
Survival after hepatic resection: A. P. Savage and R. A. M a l t
Table 1 Ages qfputient.s who undmvent rcJ.sectionqf malignant ltr~patic~ neoplusms
Table 2 E.~tentof hepatic resection Primary
Metastases
No. of patients
Indication Primary hepatic malignancy Hepatocellular carcinoma Cholangiocarcinoma or gallbladder carcinoma Hepatoblastoma Sarcoma Metastases Colorectal cancer Other primary sites
Mean (range ) age (years)
68 42 12
52.4 (2-82) 54.2 (6-82) 59.0 (30-77)
3 I1 126 104 22
8.0 (2-13) 50.4 (12-71 ) 59.1 (6-80) 60.1 (28-79) 53.9 (6-80)
Colorectal Right hepatectomy Left hepatectomy Extended right hepatectomy Extended left hepatectomy Segmentectom y Wedge resection Total
Other
HCC
31 8 26
9 5 4
10
5
0
18 16
.. 7
104
Other
6
Total
15
9 3 6
55 26 51
3
0
8
2
5 3
8 0
33 21
22
42
26
194
HCC, hepatocellular carcinoma
Table 3 Outcome of.22 patients ivho undernetit hepatic resection Patient no.
of' metu.sta.se,s,from primarj. sites
Primary
Operation
Curative
Gall bladder Kidney Uterus Pancreas Lung Stomach Uterus Ileal carcinoid Ovary Kidney Kidney. Wilms' tumour Unknown (pancreas) Ileal carcinoid Unknown (pancreas) Gallbladder Kidney Breast Ocular melanoma Unknown (colon) Thyroid Ileal carcinoid Broad ligament embryonal cell tumour
Right hepatectomy Segmentectomy Right hepatectomy Right hepatectomy Left hepatectomy Segmentect om y Right hepatectomy Left hepatectomy Left hepatectomy Right hepatectomy Segmentectomy Right hepatectomy Right hepatectomy Segmentect omy Right hepdtectomy Right hepdtectomy Right hepatectomy Left hepatectomy Right hepatectomy Left hepatectomy Right hepatectomy Right hepatectomy
Yes No Yes No Yes No No No Yes No No No No No Yes Yes Yes Yes No Yes No No
otl~crthan the colorectiim
Survival (months )
Status
Operative death Lost to follow-up Operative death Died from disease Died from disease Died from disease Died from disease Died from disease Alive and well Died from disease Died from disease Died from disease Alive with recurrence Died from disease Died from disease Died from disease Alive and well Died from disease Died from disease Died from disease Alive with recurrence Alive and well
~-
I 7
3 4 5 6 7 8 9 10
11
12 13 14
15 16 17 18 19 20 21
77 --
0.0 0.4 1.1
2.6 2.7 5.8 6.2 7.7 8.8 9.3 9.5 14.9 15.9 18.1 24.4 28.0 33.5 35.7 38.7 58.1 87.8 130.0
100 80 W e m
60 40 20
t
I-_-___
I 0
12
24
36
40
60
0
Time a f t e r o p e r a t i o n [ m o n t h s )
I
I
36
48
I
1
60
No. a t r i s k 74 13
44 8
26 5
15 4
10 3
Figure 1 Survival curves f o r patients who underwent hepatic resection for metastases from colorectal cancer (-) and for metastases from primary carcinoma at other sites (-----). P not significant (log rank test)
1096
I
24
Time a f t e r o p e r a t i o n ( m o n t h s )
No. a t risk Colorectal 104 primary Other site 2 2
t
12
Curative Noncurative
76
58
37
24
14
28
18
9
3
2
9
Figure 2 Surriral curre.s,for patients iiho uncieriimt curatire (-) and rior~-curntire(-----) hepatic resection of mr1astarse.s f r o m colorecttrl cat7ccr. P < 0.02 (log r~t7litc.Yt)
Br. J. Surg., Vol. 79, No. 10, October 1992
S U,rvival after hepatic resection: A. P. Savage and R. A. Malt incidental laparotomy. Routine follow-up of patients with colorectal cancer disclosed hepatic metastases on investigation of a raised carcinoembryonic antigen (CEA ) level ( 37 patients ), abnormal liver function test results (five), routine radionuclide scintigraphy or ultrasonographic liver scan (ten), or hepatomegaly discovered on routine clinical examination (two ). Women had a survival advantage over men ( P = 0.06). The survival rate of patients with poorly differentiated primary tumours was nil at 3 years uersus a 5-year survival rate of 20 per cent for patients with well or moderately differentiated tumours. These differences failed to reach statistical significance because of the small numbers of patients with poorly differentiated primary tumours (Table 4 ).
Results Resection of metastases was the indication for hepatectomy in 126 patients; the site of the primary was the colon or rectum in 104. Of 68 patients who underwent resection of malignant tumours originating in the liver, histopathological examination showed hepatocellular carcinoma ( H C C ) in 42. Of these, three patients showed mixed HCC and cholangiocarcinoma, and cirrhosis was present in nine. Histopathological examination of the remaining 26 primary hepatic malignant tumours revealed cholangiocarcinoma ( 12 ) including carcinoma of the gallbladder (four ), hepatic sarcoma ( 1 1 ) and hepatoblastoma (three). Patients’ ages are shown in Tuble f and the extent of hepatic resection in Table 2. Metcisrases The overall 5-year survival rate of patients who underwent resection of the liver for metastatic disease was 17 per cent. The 5-year survival rate was 18 per cent for patients with metastases from colorectal cancer versus 14 per cent for those with metastases from other primary sites ( F i g u r e 1 ). The site of the primary tumour and the outcome of patients with metastases from primary sites other than the colorectum is shown in Tuhle3. For three patients, the site of the primary was unknown at the time of hepatic resection. Subsequent investigation showed that these were metastases from the pancreas (two patients) and colon (one ). M e rustuses ,from c.olor.ec.tci1cuncer The 5-year survival rate for 76 patients who underwent curative operation was 23 per cent. Resection of metastases was non-curative in 28 patients: eight cases showed involvement of the resection margin, eight had gross residual disease, five showed direct spread to adjacent structures, and seven spread to lymph nodes of the portal triad. The 5-year survival rate for patients undergoing non-curative resection was zero ( P < 0.02 ) (Figure 2). The resection margin was the most significant prognostic indicator. A margin > 5 mm was associated with a 5-year survival rate of 27 per cent compared with 9 per cent for lesser margins ( P < 0.01 ) ( F i y u r e 3 ). The proportional hazard of survival was 0.48 (95 per cent c.i. 0.29-0%1 ) if the resection margin was < 5 mm cornpared with a margin > 5 mm. Asymptomatic patients had a better survival rate than symptomatic patients ( P < 0.05) ( T u h l p 4 ) .Only 14 patients presented with symptoms from liver metastases (abdominal pain in 11, and fever, weight loss and malaise in three patients). Hepatic metastases were.discovered during investigation or at operation for the primary tumour in 33 patients. Unsuspected hepatic metastases were discovered in three patients undergoing
f f e p a t o c e ~ / u ~ acrirciiiomu r.
The overall median survival of patients who had undergone resection of HCC was 12 months, and the 5-year survival rate 25 per cent (Tuble5). There were no differences in survival rates between patients operated on before or after 1980. Hepatectomy for HCC was classified as non-curative in 27 patients: 15 showed involvement of the resection margin on histopathological examination, five showed the presence of gross residual disease, three exhibited direct invasion ofadjacent structures, and regional lymph node involvement was present in four. Curative operations were associated with a 44 per cent 5-year survival rate compared with 8 per cent for non-curative operations ( P < 0.001 ) (Figure 4). Only nine patients had a resection margin > 5 mm; their 5-year survival rate was 40 per cent compared with a 17 per cent survival rate for patients with a margin < 5 mm ( P not significant ). However. if patients who had undergone non-curative operations were excluded, the 5-year survival rate was 54 per cent if the margin of resection was > 5 mm versus 43 per cent for patients with a lesser margin ( P not significant ). Preoperative r-fetoprotein estimation was performed for 25 patients. The median serum level was 19 ng/ml (range 9-447984 ng/ml; normal < 12.5 ng,’ml). Patients whose preoperative estimation was within the normal range had a 37 per cent 5-year survival rate compared with none of the patients with high levels ( P < 0.01 ) ( F i g u r e 5 ). The odds ratio for survival was 0.26 (95 per cent c.i. 0.11-0.68) for patients with high r-fetoprotein levels. There was no significant difference in the survival rate between patients who were symptomatic and those who were not ( Table 5 ). HCC was an incidental finding in seven patients; the diagnosis was made at incidental laparotomy in three. Thirty-five patients were symptomatic: pain was the presenting complaint for 16, six discovered a mass, one had jaundice. six had malaise, fever and weight loss, and six underwent emergency resection for ruptured HCC.
80 I-__
I_ Q)
Y
m
m
20
I-
$-_---
I-
‘-----
> .t_
-I--
0
2
-
m
I
I
I
I
I
12
24
36
48
60
L
Time a f t e r o p e r a t i o n ( m o n t h s )
0
No. a t r i s k
Margin > 0.5 cm Margin \< 0.5 cm
I
I
I
I
I
12
24
36
48
60
Time a f t e r o p e r a t i o n ( m o n t h s ) No. a t risk
51
43
25
17
8
7
46
27
17
8
6
3
Figure 3 Survival curves f o r patients who underwent hepatic resection for metastases from colorectal cancer with a resection margin > 5 mm (-) and 2 years
59 37
26 24
75 78
44 20
23
0.157
0.68 (0.40- 1 . l 6 )
Synchronous disease Metachronous disease
33 71
43 23
75 77
58 25
25 16
0.1 I3
0.67 ( 0 4 - 1 . 1 0 )
Asymptomatic patients Symptomatic patients
90 14
26 I1
81 47
38 14
19 14
0.043
0.42 (0.18-0.97)
Normal CEA Raised CEA
19 62
25 24
61 87
33 31
22
0.479
0.78 (0.40- I '53 )
Female Male
51 53 57 47
31 23
84 71
42 30
24 13
0.058
0.64 (0.39- 1.02 )
31 17
78 76
41 31
22 13
0,168
0.7 1 (0.43- 1.1 5 )
Solitary metastasis Multiple metastases
10
18
-
No. of metastases 1-3 24 Size of metastases < 5 cm >5cm
87 17
26 13
79 62
36 28
19 14
0.343
0.70 (0.34-1.46)
52 43
26 24
84 66
41
31
22 9
0.058
0.60 (0.36- 1.01 )
Curative resection Non-curative resection
76 28
30 21
80 66
42 17
23 0
0.0 I0
0.43 (023-0.82)
Resection margin > 5 mm Resection margin < 5 mm
51 46
35 20
88 62
46 24
27 9
0.006
0.48 (0.29-0.8 1 )
~
CEA, carcinoembryonic antigen; c.i., confidence interval. *Log rank test
Table 5 Suriirul rute ,follrm~ingreseciion of hepuiocellulur curcinoniu No. of patients
Median survival (months)
1 year
All patients
42
12
Resection before 1980 Resection after 1980
14 28
Male Female
Survival rate ( Y O ) P*
Odds ratio (95% c.i.)
28 I5
0.168
0.60 (0.29- 1.23)
42 14
31 14
0.527
0.80 (040- 1.60)
50 57
24 28
19 28
0,479
1.37 (058-3.24)
50 5
67 33
56 0
37 0
0.005
0.26 ( 0.1 I -0.68 )
33 9
16 6
54 38
26
22 0
0. I63
0.49 (017-1.36)
Solitary tumour Multifocal tumour
22 20
15
50 53
31 18
23
14
0.764
0.90 (0.45- 1.8 1 )
Tumour size < 5 cm Tumour size > 5 cm
10
32
24 12
60 48
25 24
25 19
0.60 I
0.8 I (0.36- I '78 )
Curative resection Non-curative resection
15 27
79 37
56 8
44 8
0.00 I
0.30 (015-0.62)
Resection margin > 5 m m Resection margin < 5 m m
9 33
50 6 27 12
67 47
40 20
40 17
0,180
0.58 (0.26- I .28 )
3 years
5years
51
25
25
24 8
64 41
28 23
20 22
12
16
48 54
Symptomatic patients Asymptomatic patients
35 7
14 24
Normal z-fetoprotein Raised r-fetoprotein
10
15
Patients without cirrhosis Patients with cirrhosis
_
_
_
_
_
~
~
.
.
19
18
c.i., Confidence interval. *Log rank test
1098
Br. J. Surg., Vol. 79, No. 10, October 1992
Survival after hepatic resection: A. P. Savage and R. A. M a l t
1 '-
L
I
'-1
7
0
I
I
12
24
I
I
I
I
36
48
60
Time a f t e r o p e r a t i o n ( m o n t h s )
No. a t r i s k
a -fetoprotein (12nglml 10
7
6
6
4
5
4
2
a -fetoprotein > 1 2 ngfml
15
Figure 5
Surr.iro1 curves ,/i)r prrrierirs will1 resecleci hepu~ocrllulcrr ii.it/i preoperative r:feroprotein leuel < 12 t i g l t l l l (-) atid > 12 t i g ; t i i l (-----). P < 0 4 1 (log runk irst)
ccrrcitioriicr
Clioluny ioccrscinomu Hepatic resection was performed for five patients with hilar cholangiocarcinoma, three with peripheral cholangiocarcinoma and four with carcinoma of the gallbladder involving the fossa. The median survival of this group of patients was 10 months; one patient survived to 5 years (5-year survival rate 22 per cent). Although the small number of patients in this group precluded life-table analysis of subgroups, the three patients with peripheral cholangiocarcinoma all died within 12 months while all four with hilar cholangiocarcinoma who survived the operation were alive at 14 months. All patients with carcinoma of the gallbladder died within 2 years, three within the first year. Hepcitic sLircotmi Histopathological examination of the 11 patients with hepatic sarcoma showed leiomyosarcoma (five patients ), fibrosarcoma (two ), malignant histiocytoma (one ), malignant schwannoma (one), lymphoma ( o n e ) and undifferentiated sarcoma (one). The tumour originated in the liver in six patients and metastasized to the liver from the stomach (two), retroperitoneum ( t w o )and ileum (one ) in five. One of the 1 1 patients was alive at 5 years (5-year survival rate 25 per cent ). Three children underwent resection of a hepatoblastoma. Although two were alive and well at 55 and 150 months after operation, the third had recurrent disease at 39 months.
Discussion This study examines whether it is possible to improve the long-term survival rates of patients with hepatic malignancy by defining criteria for the selection of patients for hepatic resection. While the operative mortality rate for hepatic resection has fallen in recent years', the present data show no improvement in the overall 5-year survival rate. Resection of' I ~ ~ P ~ N S ~ fkorn C I S ~ colosect~~l S , cmcu The efficacy of resection of liver metastases from tumours of the colon and rectum has recently been questioned7. In previous studies, the 5-year survival rate after resection of hepatic metastasess-I6 has been between 10 and 34 per cent, which is consistent with the present data. The survival of patients with untreated metastases to the liver from primary tumours of the is between 5.6 and 13 months. However, colon and patients with multiple, but potentially resectable, metastases survive a mean of 10.6 months, and those with solitary hepatic m e t a ~ t a s e survive s ~ ~ 16.7 months. There are no controlled trials
Br. J. Surg., Vol. 79, No. 10. October 1992
comparing hepatic resection with non-surgical management for potentially resectable hepatic metastases. None the less, patients in this series who had undergone hepatic resection for metastases had a better survival rate than historical patients treated without o p e r a t i ~ n ' ~ ~ ~ ~ * ~ ~ . The histological grade of the primary tumour may be important in the selection of patients for resection of liver metastases. Few patients with poorly differentiated primary neoplasms underwent resection of liver metastases; therefore the worse survival of patients with poorly differentiated tumours compared with those with moderately or well differentiated tumours failed to reach statistical significance. The present data accord with a previous report showing a mean survival of 7.1 months for patients with resected metastases from poorly differentiated tumours compared with 17.9 months for those with resected metastases from moderately or well differentiated tumours", and suggest that patients with metastases from poorly differentiated tumours may not benefit from hepatic resection. Consistent with other reports, the present data show that the Dukes' classification of the primary cancer is not important in the selection of patients for resection of hepatic m e t a s t a ~ e s ' ~ . ~Other ~ - ~ ~groups, . however, have reported 5-year survival rates between 7 and 45 per cent higher after resection of hepatic metastases from Dukes' B compared with Dukes' C lesion^'^^'^^^^-^^. None the less, the difference in survival rate was insufficient to deny patients with Dukes' C primary tumours the potential for cure. The interval between resection of the primary tumour and resection of hepatic metastases is not important in the selection of patients for resection of colorectal liver metastases. The present data show a trend towards a better survival rate if this interval is g 2 years, and for patients who have undergone resection of metastases discovered during surgery for the primary tumour. However, the improved survival rate of patients with synchronous metastases, or of those whose close follow-up resulted in the early detection of metastases. may represent a lead-time bias. other^^^.^^ have reported a longer median survival for patients with synchronous metastases of between 4 and 9 months and a 66 per cent 5-year survival rate for patients with synchronous metastases compared with zero for patients with metachronous metastases' I . Conversely, others have reported a 5-year survival rate between 10 and 26 per cent better for patients with metachronous m e t a s t a ~ e s ~An ~ .improved ~~. 5-year survival rate of 40 per cent for patients with a disease-free interval > 1 year, compared with 28 per cent for a shorter disease-free interval, has also been reported' possibly reflecting a lower biological activity of tumours showing a long disease-free interval. However, most studies have shown no significant differences in the survival rate after resection of liver metastases when analysed according to the interval between resection of the primary and the discovery of hepatic metastasess.'0~'b.2s.29.40-43; further, this interval is of no account in the selection of patients for hepatic resection or in the determination of prognosis. The presence of symptoms from hepatic metastases may not be a contraindication to resection. The present data agree with those from a multicentre study that reported a 32 per cent 5-year survival rate for patients with symptoms from hepatic metastases compared with 45 per cent when metastases were ~. discovered on routine follow-up or i n v e ~ t i g a t i o n ~Because symptomatic metastases may be considered to be at a later stage in their growth. a lead-time bias may account for the better survival of patients with asymptomatic metastases. The findings at operation are of great importance in determining the prognosis after resection of metastases from colorectal cancer. The present data confirm a previous report' that non-curative resection of liver metastases is associated with a 5-year survival rate of less than 5 per cent. The presence of neoplasia at the resection margin'5.27.2yand of margins of at least 5 mm (Reference 44), 10 mm (References 9, 15. 34) and 20 mm (Reference 45) have been analysed with respect to the long-term survival rate. These reports support the present data,
',
1099
Survival after hepatic resection: A. P. Savage and R. A. Malt
which show a significantly improved 5-year survival rate for resection per se is not important in the prognosis of HCC. The patients with a margin of resection of > 5 mm. These data also 5-year survival rate was similar for patients undergoing curative show the futility of resection of metastases in the presence of resection whether or not the resection margin was > 5 mm. lymphatic involvement regional to the hepatic metastases, Although one study has reported a 533-day disease-free survival extrahepatic spread or in circumstances resulting in gross for patients with a negative margin of resection iwws 160 days residual disease after resection. when the margin was involved by tumour5’, another6’ has The number of metastases may not be relevant in the reported no significant difference in recurrence of tumour in selection of patients for hepatic resection, provided that there the liver after limited resection with a resection margin < 10 mm is no extrahepatic spread and an adequate margin of resection compared with patients whose tumours were resected with a can be achieved. These data support previous reports showing margin > 10 mm. Similarly, for patients with small HCC no worse survival in patients with multiple m e t a s t a s e ~ ’ ~ - ’ ~ . ~complicating ~-~~. cirrhosis, limited hepatic resection was associated Other reports have suggested that the survival rate is better with a trend towards better survival than for those undergoing following resection of solitary metastases9~’5~27~2’~30~34~37~41.62~46. major hepatic resection61. N o discrimination was made in the present study between The median survival of patients with HCC complicating multifocal and solitary metastases with satellite lesions. One cirrhosis is 1.2 uersus 13 months for those with HCC in study, however, reported that satellite lesions around solitary non-cirrhotic liver”. Increasing severity of cirrhosis, as defined metastases d o not affect prognosis4’, while another showed a by Pugh’s modification of Child’s classification, adversely worse prognosis with satellite lesions’. influences both the operative mortality rate (3.7 per cent for Child’s grade A, 16.7 per cent for grade B / C ) and the long-term survival rate (56.7 per cent 5-year survival for Child’s grade A, Resection of meinstuses .from other prinzury sites 12.2 per cent for grade B / C ) of patients undergoing resection Although no significant differences were found when patients of HCC complicating cirrhosis62. The safe limits of hepatic with metastases from sites other than the colorectum were resection are dependent not only on the volume uf liver resected compared with those undergoing resection of metastases from but also on preoperative liver f ~ n c t i o n ~ ’None . ~ ~ . the less, one the colorectum, the histology of the primary tumour was report showed no statistical difference in the survival rates of important in determining the prognosis. Of the five surviving patients undergoing resection of HCC from a cirrhotic liver patients, two with slowly growing metastases from ileal compared with those without cirrhosis”. However, 58 per cent carcinoid tumours who underwent palliative resection are alive of the patients with HCC in a cirrhotic liver underwent minor in addition to three with gynaecological or breast carcinomas. resection compared with 24 per cent of those without Patients with metastases from carcinoma of the stomach and cirrhosis. Only nine of the present patients with HCC had pancreas fared badly. Other reports with similar results’ 394b cirrhosis; the trend towards worse survival in these patients also d o not recommend resection of metastases from primary was not statistically significant. sites other than the colorectum with intent to cure, although These data indicate that the selection of patients for resection palliative resection may be indicated for metastases from some of primary and secondary liver tumours may improve survival. tumours, for example ileal carcinoid tumours and those from The major contraindications to resection are finding that a rare sites such as mesonephric duct carcinoma. tumour has extrahepatic spread or regional lymph node involvement, or both, or that it is not resectable without leaving Pritnury liver tuniours gross residual neoplasia. For metastases from colorectal cancer, Patients with untreated HCC survive between 1 and 4 the margin of resection and possibly the histological grade of months after d i a g n o s i ~ ~ ’ - Those ~ ~ . fit enough to undergo the tumour are important. For HCC an r-fetoprotein level palliative surgery survive between 7 (References 48, 53 ) and > 1 2 5 ng/ml may be a contraindication to resection. 10.8 (Reference 50) months while patients undergoing curative resections survive between 23 (Reference 53) and 42 (Reference 50) months. In selected patients, surgical resection is superior to systemic or regional ~ h e m o t h e r a p y ~ ’ . ~ ’ . ~ References radiotherapy’ or hepatic artery ligation or dearterializ1 . Savage AP, Malt RA. Elective and emergency hepatic resection: results of non-surgical therapy support a t i ~ n ~ ’ The . ~ ~ poor . determinants of operative mortality and morbidity. Ann Surg the case for surgery despite the higher operative risk of resection 1991; 214: 689-95. for HCC than for metastatic liver disease’. 2. Couinaud C. Le Foie: Etudes Anaiomiques el Chirurgicales. Paris: The r-fetoprotein level may be useful in the preoperative Masson, 1957: 283-9. selection of patients for hepatic resection for HCC. The present 3. Malt RA. Surgery for hepatic neoplasms. N Engl J Med 1985; data show a worse survival rate for patients with a raised level 313: 1591-6. of r-fetoprotein before operation. One report” has shown a 4. Kaplan EL, Meier P. Nonparametricestimation from incomplete median disease-free survival of 531 days for patients whose observations. J Am Statist Assoc 1958; 53: 457-81. 5. Mantel N . Evaluation of survival data and two new rank order r-fetoprotein level was 200ng/ml, and a second study has 50: 163-70. confirmed this better survival5h. 6. Machin D, Gardner MJ. Calculating confidence intervals for The feasibility of performing a curative resection is most survival time analyses. BMJ 1988; 296: 1369-71. relevant in the selection of patients for resection of HCC. The I . Silen W. Hepatic resection for metastases from colorectal present data agree with a previous study5’ which reported a carcinoma is of dubious value. Arch Surg 1989; 124: 1021-2. 70 per cent 5-year survival rate for those undergoing curative 8. Fegiz G, Bezzi M, de Angelis R et al. Surgical treatment of liver resection but no survivors amongst patients undergoing metastases from colorectal cancer. Zial J Surg Sci 1985; 15: palliative resection, defined as the absence of a capsule around 259-65. 9. Ekberg H, Tranberg K-G, Andersson R et al. Determinants of the tumour, the presence of satellite lesions, the presence of survival in liver resection for colorectal secondaries. Br J Surg vascular or bile duct invasion, or a margin of resection 1986; 13: 121-31. < 10 mm. Similarly, other studies have reported a 50 per cent 10. Nordlinger B, Parc R, Delva E, Quilichini M-A, Hannoun L, 5-year survival rate in the absence of extrahepatic spread Huguet C. Hepatic resection for colorectal liver metastases. Ann compared with no 5-year survivors among patients with a Surg 1987; 205: 256-63. locally invasive tumour5’, and a 75 per cent 5-year survival 11. Coppa GF, Eng K, Ranson JHC, Gouge TH, Localio A. Hepatic rate for patients with completely resected tumours compared resection for metastatic colon and rectal cancer. Ann Surg 1985; with 11 per cent for those whose tumours had spread locally 202: 203-8. or to distant sites59. 12. Wilson SM, Adson MA. Surgical treatment of hepatic metastases However, the present data also suggest that the margin of from colorectal cancers. Arch Surg 1976. 111: 330-4.
1100
Br. J. Surg., Vol. 79, No. 10. October 1992
Survival after hepatic resection: A. P. Savage and 13. 14.
15. 16. 17. 18.
19.
20. 21.
22. 23. 24. 25.
26. 27. 28. 29. 30. 31. 32. 33.
34. 35. 36. 37. 38.
Iwatsuki S, Starzl TE. Personal experience with 411 hepatic resections. Ann Surq 1988; 208: 421-34. Stephenson KR. Steinberg SM, Hughes KS, Vetto JT, Sugarbaker PH, Chang AE. Perioperative blood transfusions are associated with decreased time t o recurrence and decreased survival after resection of colorectal liver metastases. Ann Sury 1988; 208: 679-87. Hughes K, Scheele J, Sugarbaker PH. Surgery for colorectal cancer metastatic to the liver. Sury Clin North Ani 1989; 69: 339-59. Butler J, Attiyeh FF, Daly JM. Hepatic resection for metastases ofthecolonand rectum. Sury GynecolOhster 1986; 162: 109-13. Bengtsson G, Carlsson G, Hafstrom L, Jonsson P-E. Natural history of patients with untreated liver metastases from colorectal cancer. Am J Sur{/ 1981; 141: 586-9. Bengmark S. Hafstrom L. The natural history of primary and secondary malignant tumors of the liver. Cmcer 1969; 23: 198-202. Lahr CJ, Soong S-J, Cloud G, Smith JW, Urist M M , Balch C M . A multifactorial analysis of prognostic factors in patients with liver metastases from colorectal carcinoma. J Clin O i d 1983; 1: 720-6. Pestana C, Reitmeier RJ, Moertel CG. Judd ES, Dockerty MB. The natural history of carcinoma of the colon and rectum. Ani J Surg 1964; 108: 826-9. Finan PJ, Marshall RJ, Cooper EH, Giles GR. Factors affecting survival in patients presenting with synchronous hepatic metastases from colorectal cancer: a clinical and computer analysis. Br J Sury 1985; 72: 373-7. Palmer M, Petrelli NJ, Herrera L. No treatment option for liver metastases from colorectal adenocarcinoma. Dis Cokm Recmni 1989; 32: 698-701. Cady B, Monson DO, Swinton NW Sr. Survival of patients after colonic resection for carcinoma with simultaneous liver metastases. Sury Gjmecol Ohstet 1970; 131 : 697-700. Wood CB, Gillis CR, Blumgart LH. A retrospective study of the natural history of patients with liver metastases from colorectal cancer. Clin Oncol 1976; 2: 285-8. Wagner JS, Adson MA, van Heerden JA. Adson MH. Ilstrup DM. The natural history of hepatic metastases from colorectal cancer: a comparison with resective treatment. A m Sury 1984; 199: 502-8. Adson MA, van Heerden JA, Adson MH, Wagner JS. Ilstrup D M . Resection of hepatic metastases from colorectal cancer. Arch Sury 1984; 119: 647-51. Holm A, Bradley E. Aldrete JS. Hepatic resection of metastasis from colorectal carcinoma: morbidity, mortality and pattern of recurrence. Ann Sury 1989; 209: 428-34. Gennari L, Doci R, Bozzetti F, Bignami P. Surgical treatment of hepatic metastases from colorectal cancer. Ann Surg 1986; 203:49-54. August DA. Sugarbaker PH, Ottow RT, Gianola FJ, Schneider PD. Hepatic resection of colorectal metastases. Ann Sury 1985; 201: 210-18. Logan S, Meier SJ, Ramming KP, Morton DL, Longmire W P Jr. Hepatic resection of metastatic colorectal carcinoma. Arch Sury 1982; 117: 25-8. Nims TA. Resection of the liver for metastatic cancer. Sury G J I I C VOb.s/c’t ~ 1984; 158: 46-8. Thompson HH, Tompkins RK. Longmire W P Jr. Major hepatic resection: a 25-year experience. Ann Sury 1983; 197: 375-88. Fortner JG, Silva JS, Golbey RB, Cox EB. Maclean BJ. Multivariate analysis of a personal series of 247 consecutive patients with liver metastases from colorectal cancer: I. Treatment by hepatic resection. Ann Sury 1984; 199: 306-16. Registry of Hepatic Metastases. Resection of the liver for colorectal carcinoma metastases: a multi-institutional study of indications for resection. Surgery 1988; 103: 278-88. Morrow CE, Grage TB, Sutherland DER, Najarian JS. Hepatic resection for secondary neoplasms. Suryery 1982; 92: 610-14. Petrelli NJ, Nambisan RN, Herrera L, Mittleman A. Hepatic resection for isolated metastasis from colorectal carcinoma. Am J Sury 1985; 149: 205-9. Bengmark S, Hafstrom L, Jeppsson B, Jonsson P-E, Ryden S, Sundqvist K. Metastatic disease in the liver from colorectal cancer:anappraisalofliver surgery. WorldJSurg 1982;6:61-5. Sesto ME, Vogt DP. Hermann RE. Hepatic resection in 128
Br. J . Surg., Vol. 79, No. 10, October 1992
R. A. M a l t
patients: a 24-year experience. Surgery 1987; 102: 846-51. Tomas-de la Vega JE, Donahue EJ. Doolas A ct ( I / . A ten year experience with hepatic resection. Surg Gjwc,col Ohstcv 1984; 159: 223-8. 40. Kortz WJ, Meyers WC, Hanks JB. Schirmer BD. Jones RS. Hepatic resection for metastatic cancer. A m Sury 1984; 199: 182-6. 41. Taylor B, Langer B, Falk RE. Arnbus U. Role of resection in the management of metastases to the liver. Ccui J Simq 1983: 26: 215- 17. 42. Cobourn CS, Makowka L, Langer B, Taylor BR, Falk RE. Examination of patient selection and outcome for hepatic resection for metastatic disease. Sury G ! m d Ohstc,f 1987; 165: 239-46. 43. Vogt P. Raab R, Ring B, Pichlmayr R. Resection of synchronous liver metastases from colorectal cancer. World J S m / 1991 ; 15: 62-7. 44. Bradpiece HA, Benjamin IS, Halevj A, Blumgart LH. Major hepatic resection for colorectal liver metastases. Br J Sury 1987; 74: 324-6. 45. Rajpal S, Dasmahapatra KS, Lecesma EJ, Mittelman A. Extensive resections of isolated metastasis from carcinoma of the colon and rectum. Sury Gj.neco/ Ohsir/ 1982; 155: 813-16. 46. Foster JH. Survival after liver resection for secondary tumors. A m J Sury 1978; 135: 389-94. 47. Lee NW. Wong J, Ong GB. The surgical management of primary carcinoma of the liver. World J Sury 1982; 6: 66-75. 48. Bengmark S, Hafstrom L, Jeppsson B, Sundqvist K. Primary carcinoma of the liver: improvement in sight? WorltlJSu,y 1982; 6: 54-60. 49. Inouye AA, Whelan TJ Jr. Primary liver cancer: a review of 205 cases in Hawaii. Ani J Sury 1979; 138: 53-61. 50. Honjo I. Mizumoto R. Primary carcinoma of the liver. A m J SUU/ 1974; 128: 31-6. 51. Davidson AR, Tomlinson S, Calne R k , Williams R. The variable course of primary hepatocellular carcinoma. Br J Surg 1974; 61 : 349-52. 52. Nagasue N, Yukaya H, Hamada T, Hirose S, Kanashima R. Inokuchi K. The natural history of hepatocellular carcinoma: a study of 100 untreated cases. Cuncer 1984; 54: 1461-5. 53. Nagorney DM, van Heerden JA, Ilstrup DM. Adson MA. Primary hepatic malignancy: surgical management and determinants of survival. Suryerj, 1989; 106: 740-9. 54. Appelqvist P. Primary carcinoma of the liver: clinical course and therapeutic results. J Sury Oncol 1982; 21: 87-93. 55. Lai EC-S, Ng 10-L. Ng MM-T c t a/. Long-term results of resection for large hepatocellular carcinoma: a multivariate analysis of clinicopathological features. Heptfology 1990; 11: 815-18. 56. Nagao T. Goto S. Kawano N (’1 d.Hepatic resection for hepatocellular carcinoma. Ann Sury 1987; 205: 33-40. 57. Okamoto E, Tanaka N , Yamanaka K , Toyosaka A. Results of surgical treatments of primary hepatocellular carcinoma: some aspects to improve long-term survival. World J Surg 1984; 8: 360-6. 58. Smalley SR. Moertel CG, Hilton J F e f a/. Hepatoma in the noncirrhotic liver. Cuncer 1988; 62: 1414-24. 59. Fortner JG, Maclean BJ, Kim D K et ti/. The seventies evolution in liver surgery for cancer. Cuncer 1981; 47: 2162-6. 60. Yoshida Y, Kanematsu T, Matsumata T, Takenaka K, Sugimachi K. Surgical margin and recurrence after resection of hepatocellular carcinoma in patients with cirrhosis. Ann Sury 1989; 209: 297-301. 61. Kanematsu T, Takenaka K, Matsumata T , Furuta T , Sugimachi K. Inokuchi K. Limited hepatic resection effective for selected cirrhotic patients with primary liver cancer. Ann Sury 1984; 199: 51-6. 62. Franco D, Capussotti L, Smadja C er NI. Resection of hepatocellular carcinomas: results in 72 European patients with cirrhosis. Gusrroenteroloyy 1990; 98: 733-8. 63. Okamoto E, Kyo A, Yamanaka N, Tanaka N, Kuwata K. Prediction of the safe limits of hepatectomy by combined volumetric and functional measurements in patients with impaired hepatic function. Surgery 1984; 95: 586-92.
39.
Paper accepted 28 March 1992
1101
