- Email: [email protected]
The result of liver resection for colorectal metastases in a small volume hospital in western Norway
Available online at www.sciencedirect.com
EJSO 33 (2007) S105eS110
www.ejso.com
Review
The result of liver resection for colorectal metastases in a small volume hospital in western Norway K. Søndenaa a,*, I. Nesvik b, K. Eiriksson b, M. Vetrhus b b
a Institute of Surgical Sciences, University of Bergen, Norway Department of Surgery, Stavanger University Hospital, Stavanger, Norway
Accepted 26 September 2007 Available online 5 November 2007
Abstract Aim: Our aim was to compare liver resection for colorectal metastases in a non-referral, small volume unit with a dedicated staff, with results from larger units. Methods: Thirty patients (15 men and 15 women) with a median age of 64 years (range 29e78) underwent hepatic resection during a 5-year period from 1997 to 2003 in a teaching hospital in western Norway. Results: Sixty-three percent (19/30) of the colorectal tumours were Dukes stage C (n ¼ 19) and CEA was increased in seven patients (23%), of which four (13%) had values above 50 mg/l. The metastases were synchronous with the colorectal tumours in 11 patients (37%). Nonanatomical (wedge) resections were the dominant type of surgeries and the resection margins were clear in all patients. A 77-year-old man (3%) died of MOF after right hemihepatectomy. Morbidity was encountered in eight other patients (28%). In 22 patients (76%) with recurrent disease, metastases first appeared in the liver in 18 (82%) of these patients. Seven patients (23%) have had resections for recurrences. Mean time to recurrence was 20 months (range 3e87). The actuarial 5-year survival rate was 42%. Six patients (20%) are currently disease free. Conclusion: Although our unit has treated a small number of patients compared with specialized units elsewhere, the survival rate, as well as morbidity and mortality, were comparable. However, 62% have had recurrent liver disease and this may suggest a role for neoadjuvant or adjuvant chemotherapy in selected cases. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Colorectal cancer; Liver metastases; Hepatic resection; Recurrence; Survival
Introduction Hepatic resection for colorectal liver metastases has become established as a reliable primary treatment modality.1e4 The common indication in a physically fit patient who has liver metastases and is able to tolerate general anaesthesia, is that the metastases are amenable to surgical resection provided a minimum of about 30% of functioning liver tissue is left behind without signs of extra-hepatic disease.5 Considering these provisions, complete surgical resection may result in a cure for some patients.3,5 The 5-year survival * Corresponding author: Department of Surgery, Haraldsplass Deaconal Hospital, PO Box, 6165 Postterminalen, N-5892 Bergen, Norway. E-mail address: [email protected] (K. Søndenaa). 0748-7983/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejso.2007.09.021
rate is generally considered to be above 30% or even 40% in favourable cases.1,5 These excellent figures from expert centres have set a standard that others should aim for. Colorectal liver metastasis is the commonest indication for liver resection in our country of 4.5 million people and is concentrated in a few regional hospitals. However, the number of hepatic resections is small compared with larger units abroad. It has been indicated that large volume units may achieve superior results compared with smaller units for several types of surgery.6,7 Low volume has been arbitrarily defined from 5 to 25 cases per year but may depend on the complexity of the operation.6,7 Still, surgery has been offered to all the patients in our catchment area whenever the possibility of radical resection was present and no one has been weeded out through a referral
S106
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
system. The question that remains is whether surgery should be concentrated in a few specialist hospitals in order to obtain optimal patient care and treatment. The geography of our country does not allow easy communication between hospitals by ground transport and we believe that treatment should be offered locally, given a minimum number of patients and a qualified and dedicated staff. Our aim was to present the results from a non-referral, small volume but dedicated unit, to examine if our patients have received standard treatment and been offered qualified care in an acceptable setting compared with that reported from larger units.
anaesthetic was administered regularly. It has been customary for our anaesthesiologists to use fluid administration freely during anaesthesia, but a CVP of less than 5 mmHg was intentionally used as a guideline in the latter part of the study. Every specimen was inspected and measured by the surgeon in the operating room. The surgically defined margin was used as a measure of the resection margin. Drains were used regularly but withdrawn as soon as fluid evacuation was below 100 ml/day. Results Surgery
Patients and methods Thirty patients (15 women and men) with a median age of 64 years (range 29e78), were operated by the first author during a 5.5-year period from 1997 through February 2003. Follow-up was concluded in April 2006. Another 6 patients (17%) were only explored because of undetected, peritoneal metastases (n ¼ 5) or nodes in the hepato-duodenal ligament (n ¼ 1) but were excluded from analysis. All patients belonged to the hospital’s catchment area of 265 000 inhabitants. Pathology The colorectal tumours were Dukes stage A (n ¼ 2), B (n ¼ 9) or C (n ¼ 19). Twenty-two tumours (75%) were located in the rectum (n ¼ 13) or the sigmoid colon (n ¼ 9), the rest were in the left (n ¼ 3) or right colon (n ¼ 5). The liver disease was bilateral in 14 patients (47%), located on the right side in another 14, and on the left side only in 2 (6%). It was not a policy of ours to biopsy the liver lesion but six patients had this done before surgical evaluation. Venous invasion by the primary tumour was observed at histological examination in 2 patients. Before the liver operation, CEA was increased (normal upper limit 9 mg/l before June 2002; 3 mg/l later) in 11 patients with a mean value of 57 mg/l, range 10e174. Four patients had values above 50 mg/l. Surgical technique All operations, except the first, one late second look and one late re-resection, were done by the first author. A J-shaped incision and intra-operative ultrasonography were used routinely. Liver tissue was divided using the forceps crushing technique but in the latter part of the study we used the Cavitron Ultrasonic Surgical Aspirator (CUSA, Tyco Health Care, Oslo, Norway). The Pringle manoeuvre was used whenever extensive resections were done. The portal triad was then clamped for 15 min interrupted by reperfusion periods of 5 min. Blood loss was substituted by packed RBC when Hb was anticipated to fall below 8 g/dl or to 10 g/dl in patients with heart disease. An epidural
Two patients had synchronous colorectal and hepatic surgery and hepatic surgery followed the primary, colorectal operation within 3 months in 9 other patients and the operation was therefore deemed synchronous in 37% (11/30). Resections were right hemi-hepatectomy or more than two segments in 9 patients (30%), segmentectomies or two or more non-anatomical parts (wedges) in 13, and a single part in 8. Extra-hepatic intra-abdominal disease was removed in three patients. Sequential Pringle’s manoeuvre was used in 26 patients (87%). The mean number of occlusion episodes was 3.5 (range 1e8). Additional chemotherapy, radiofrequency ablation (n ¼ 2) and lung surgery (n ¼ 2) were used in some patients. Morbidity and mortality A 77-year-old man (3%) died after a right hemi-hepatectomy. He had a cerebral haemorrhage during the first postoperative day and S. aureus in fluid drained from the abdomen. He subsequently succumbed to liver and multiorgan failure. Morbidity was encountered in 8 other patients (28%). Percutaneous drainage for an intra-abdominal fluid collection was carried out in three patients of whom one had a suspected bile leak that dried up without further intervention. One patient needed reoperation for a wound dehiscence. Further details can be seen in Table 1. The median duration of surgery was 335 min (range 210e495) in those with a complication, compared with 240 min (100e660) in those without complications. Blood transfusions were given to 4 patients with complications (50%), median 1375 ml (range 500e2100), compared with 250 ml (250e1500) in 5 patients (23%) without complications. Pathology The resection margin was clear in all instances, although it was 1 mm or less in one patient (but found to be clear surgically). Nine patients had a margin of 1 mm, 11 had between 1 and 9 mm, and 9 had a resection margin equal to or exceeding 10 mm. Micro-satellite deposits were not
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
Follow-up course
Table 1 Details of the 28% morbidity in 29 surviving patients Morbidity
Patients (n ¼ 8)
Splenic laceration (intraoperative splenectomy) Wound dehiscence, cerebral infarction Infected fluid collection/subphrenic abscess Lung infiltration, cardiac insufficiency, bile leak, abscess, septicaemia Ascites Cardiac arrhythmia, urinary tract infection Nerve injury right shoulder blade
1
a
S107
1a 2 1 1 1 1
Patient reoperated.
observed in routine pathological sections nor was evidence of such manifestations found intra-operatively. One deceased patient had a lymph node metastasis at the superior mesenteric vein removed before liver surgery and two had simultaneous removal of intra-abdominal metastases. Of these, one deceased patient had lymph nodes in the hepato-duodenal ligament that contained metastases and one had localized peritoneal seeding. In fact, this patient had a small peritoneal seeding just opposite a single liver metastasis after a preoperative liver biopsy. Full results are shown in Table 2.
First recurrent metastases in 22 patients were detected in the liver (n ¼ 12), liver and lung (n ¼ 3), lung (n ¼ 1), or multiple or other sites (n ¼ 6) of which three included the liver. In the 18 patients with liver recurrences, bilateral disease was eventually observed in 14 patients. It was confined to the right lobe in 4 patients, two with original disease in the right hemiliver only. Only one patient with recurrence had this at the sites of three previously removed liver metastases. Mean time to recurrence was 20 months (range 3e87). Seven patients (23%) were re-resected. Another patient with two recurrent liver metastases could have been a candidate for reoperation but for locally advanced prostate cancer recently diagnosed that required specific treatment felt to prevent further liver surgery. One renewed liver resection took place in an outside hospital after the patient moved and was followed up close to his new home. CEA became normal in 24 patients (two deceased were not analyzed) after the liver resections. In 4 deceased patients, the CEA value varied between 22 and 2468 mg/l, consistent with disease expression within a few months after hepatic surgery. Five patients were also treated later for a different malignancy in these organs: non-Hodgkin’s lymphoma (n ¼ 1, deceased), prostate (n ¼ 1, deceased), urinary bladder (n ¼ 2), and skin (n ¼ 1). Survival
Variable
Demographics Men/women (n ¼ 15/15) Ageb Tumour parameters Rectal tumours (n ¼ 13) Diff. CR tumour (n ¼ 29) Dukes stage Liver op. <3/<12 months Bilateral disease (n ¼ 14) Number of metastasesb Size (mm) of metastasesb Elevated CEA preop. (n ¼ 11/29) CEA >50 mg/l preop. a b
Deceased, n ¼ 19 (63%)a
AWD, n¼5 (17%)
NED, n¼6 (20%)
9/10
2/3
4/2
63 (29e78)
59 (53e68)
60 (45e71)
3
2
1H/15M/2L
5 moderate
6 moderate
1A/5B/13C 8/5 ¼ 13
1A/1B/3C 1/3 ¼ 4
3B/3C 2/1 ¼ 3
10
4
0
3 (1e5)
4.4 (1e7)
1.8 (1e3)
35 (20e100)
22 (8e60)
35 (15e55)
7
3
1
4
0
0
8
The 5-year survival rate was calculated as 42% (Fig. 1). Five patients are alive with disease (AWD) although recurrent disease is only suspected based on a CEA value of 28. Six patients are alive without signs of disease (NED) (see Table 3).
Kaplan-Meier Cum. Survival Plot for Survival Censor Variable: Censor 1 ,8
Cum. Survival
Table 2 Demographics and tumour specific data in 30 patients
,6 ,4
,2
Including one postoperative mortality. Values expressed as mean (range).
0 0
20
40
60
80
100
Time Figure 1. Survival following liver resection for 30 patients.
120
S108
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
Table 3 Surgical and follow-up data in 30 patients Variable Resection margin 0e1 mm (n ¼ 10) 2e9 mm (n ¼ 11) 10 mm (n ¼ 9) Pringle procedure (n ¼ 26) Pringle timea Duration of liver surgerya Blood transfusions (n ¼ 9; 30%) Transfused amounta Re-operations (n ¼ 10) Re-resections liver (n ¼ 7) Time to disease recurrence (n ¼ 23) Follow-up from liver operation a b c d e
Deceased, n ¼ 19 (63%)b
AWD, n ¼ 5 (17%)c
NED, n ¼ 6 (20%)
8 5 6 18 58 (10e157) 304 (120e660) 7 861 (250e1500) 11 7d 17 (3e53) 34 (1e102)
1 4 0 4 54 (18e85) 273 (195e390) 1 250 5 3e 36 (6e87)c 70 (43e97)
1 2 3 4 85 (15e137) 262 (150e495) 1 2100 0 0 71 (58e99)
Values are expressed as mean (range). Including one postoperative mortality. Including one patient with elevated CEA but no proven metastasis on imaging. Three patients had a third liver operation. One patient had an additional lung resection.
In the NED group, one patient had a preoperative liver biopsy, and one patient had venous invasion by his colorectal tumour. Discussion The most important intention of surgery for liver metastases is to prolong life and in some cases obtain long-term survival that should now exceed 30% at 5 years.2,4,5 In contrast to a couple of decades ago, less than 4 metastases and confinement to one hemi-liver are no longer contraindications to hepatic resection.5 The study population was recruited from our colorectal section. Thus, 70% of our patients were diagnosed within 1 year of the colorectal operation. In fact, two of our patients probably would not have been operated on elsewhere because rapid recurrent disease would have prevented their candidacy for hepatectomy. Others have found a longer time interval between the primary colorectal operation and the hepatic resection to be an important prognostic factor although no common agreement has been reached as to what constitutes a short interval. It has been given as 3 months5 or 12 months.2 In our opinion, appearance within 3 months obviously is equivalent to synchronous disease. But in the end, this may be a matter of semantics as metastases generally are regarded as originating from the primary tumour.5 Whether these are operated synchronously or by a deferred operation does not seem to influence prognosis.8 In large series, according to multivariate analysis, survival has been influenced by Dukes stage, the time period between the primary and the liver operation, number of lesions, a resection margin of more than 1 mm, and extrahepatic disease2,5 as well as tumour size and CEA.2 However, these parameters should not be considered as absolute contraindications if the hepatic lesions can be totally and
securely removed because there are long-term survivors after surgery who would otherwise succumb to the disease within a short time.3,5,9 Nevertheless, a score based on time-span, number of metastases and CEA level, proved to be quite reliable in predicting outcome.2 Dukes stage C was more prominent in our patients who died or were classified as AWD. The lymph node status of the colorectal tumour may perhaps indicate that the disease process is continuing and may minimize the effect of such variables as the time interval.10,11 It is remarkable that only two patients had definite signs of venous invasion in the colorectal tumour specimen. In a study from New York,1 the single most important surgical feature was the extent of the free resection margin. The authors speculated whether it is easier to achieve sufficient resection margins greater than 1 mm in anatomic resections1 but others have not reported similar importance of the resection margin.5,12 Although it is generally agreed that margins of 10 mm are advisable,12 it is debatable whether this is absolutely necessary as an oncological principle. We did not observe satellite lesions outside the boundary of any metastasis. Others have also reported this to be rare,13 whereas one author12 hypothesized that there is an increased possibility of microscopic deposits within 10 mm of the main lesion. However, it is mandatory to avoid exposure of the lesion on the cut surface of the liver.5 Unfortunately, exposure has usually not been addressed as a separate issue and it is unclear what the magnitude of the problem is. Nevertheless, it seems reasonable to conclude that exposure must be avoided but that the resection margin per se does not matter significantly.5 It is also a matter of principle how the margin should be measured. In a report by DeMatteo et al.,1 no mention was made of how it was done or by whom, although assumedly
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
it was done by the pathologist. Liver tissue shrinks by about 10% during the fixation process and this may be too little to influence the measurement.14 However, a narrow tissue band of 1 mm may easily become distorted or rupture during the specimen handling process and a false impression of exposure may be created. On the other hand, modern instruments used for dissection destroy tissue along the resection line actually making this wider. Pringle’s manoeuvre was used routinely by us with the aim of reducing blood loss. After discussions with the anaesthesiologists it was found acceptable to have a restrictive transfusion policy and only 30% of our patients had a transfusion. However, it has been their long-time policy to use fluid liberally to counteract the effect of anaesthesia and the surgical trauma and this, by itself, may lower the haemoglobin value. Such a policy has been shown to be unnecessary in most operations and even harmful.15e17 Although controversial, it has long been contended that blood transfusions affect morbidity and prognosis adversely.18,19 It has been our policy to remove the metastases completely but with the intention of preserving as much tissue as possible in accordance with the policy of others.5 Generally, this meant doing non-anatomical (wedge) resections, even if multiple. Whereas the number and sizes of the metastases did not offer any obvious clues to prognosis, bilateral distribution was apparently overrepresented in the two groups with the worst outcome. Perhaps it should be considered if anatomic resections would serve the patient better in some cases.1 Undiscovered metastases have the potential to grow after liver resection.20 Nevertheless, when the disease recurred in the liver, eventually it was bilateral in all instances (88%) except for two cases. Others have reported that recurrent disease was found in the opposite lobe in 50%.21 The presence of intra-abdominal extra-hepatic disease usually precludes liver resection but an acceptable number of patients still achieve long-term survival.5,9 We operated on three patients in whom such lesions were discovered. Gender has not been found to influence prognosis after treatment of colorectal cancer with liver metastases.5 Although our figures are small, there was a tendency for women to do worse. Four of our patients had a preoperative CEA value above 50 mg/l, which indicated a poor prognosis.5 An increased postoperative CEA was an ominous sign in our experience. In 7 re-resected patients, the recurrent tumour was found in a different location than in the primary operation. An interesting observation was made by Fong et al.20 who found distributions of recurrent disease survival patterns similar to ours after the primary operation. This may be in agreement with others5 who have suggested that recurrent disease may have a similar prognosis to the primary metastases and that some patients do better than others without obvious clinical signs to indicate this biological difference. Modern equipment like the PET scanner may nevertheless have an impact in the future.22
S109
Morbidity in a recent large series has been reported to be around 10e28%1,8,23 and mortality has been reduced to zero in the hands of experts.23 In comparison, in our series we had 3% mortality and 28% morbidity. It has been shown that the two variables that the surgeon is able to influence, namely the duration of the operation and the use of blood transfusions, may have a bearing on postoperative morbidity.18,24 It may well be argued that our operations might have been done within a shorter operating time. It is of practical interest to decide if trying to avoid large resections will instead lead to longer operating time and shorter resection margins thereby jeopardizing optimal surgical therapy.1 Blood transfusions may also reduce survival3 and should be kept at a minimum through meticulous surgery and avoidance of a too liberal transfusion practice.15e17 Our transfusion rate was similar to that from a large centre.15 Surgical resection for colorectal liver metastasis is an accepted treatment form and the only one able to offer longterm survival to a substantial number of patients. Although this is complicated surgery, our results show that a dedicated staff in a small volume hospital may provide good care similar to that of large volume institutions even though the opposite view has been advocated for surgical reasons.12 A recent study from high volume hospitals in Sweden25 found that a dedicated approach from specially trained surgeons might have increased the resection rate for colorectal hepatic metastases from 4% to 10% or even higher, underlining that the most important thing apart from maintaining low mortality and morbidity figures, is to assure that patients get the treatment they deserve from the medical community. Acknowledgements We are grateful to the late Arild K. Olsen, MD, who allowed us to include one of his patients in the study population and for his valuable contribution to liver surgery at our hospital. Conflicts of interest None of the authors has any ‘‘Conflict of interest’’. References 1. DeMatteo RP, Palese C, Jarnagin WR, et al. Anatomic segmental hepatic resection is superior to wedge resection as an oncological operation for colorectal liver metastases. J Gastrointest Surg 2000;4:178–84. 2. Fong Y, Fortner J, Sun RL, et al. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer. Analysis of 1001 consecutive cases. Ann Surg 1999;230:309–21. 3. Jamison RL, Donohue JH, Nagorney DM, et al. Hepatic resection for metastatic colorectal cancer results in cure for some patients. Arch Surg 1997;132:505–10. 4. Scheele J, Stang R, Altendorf-Hofmann A, Paul M. Resection of colorectal liver metastases. World J Surg 1995;19:59–71.
S110
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
5. Minagawa M, Makuuchi M, Torzilli G, et al. Extension of the frontiers of surgical indications in the treatment of liver metastases from colorectal cancer: long-term results. Ann Surg 2000;231:487–99. 6. Begg CB, Cramer LD, Hoskins WJ, et al. Impact of hospital volume on operative mortality for major cancer surgery. JAMA 1998;280: 1747–51. 7. Rabeneck L, Davila JA, Thompson M, et al. Surgical volume and long-term survival following surgery for colorectal cancer in the Veterans Affairs health-care system. Am J Gastroenterol 2004;99: 668–75. 8. Chua HK, Søndenaa K, Tsiotos GG, et al. Concurrent vs. staged colectomy and hepatectomy for primary colorectal cancer with synchronous hepatic metastases. Dis Colon Rectum 2004;47:1310–6. 9. Elias D, Ouellet JF, Bellon N, et al. Extrahepatic disease does not contraindicate hepatectomy for colorectal liver metastases. Br J Surg 2003;90:567–74. 10. Patel H, Le Marer N, Wharton RQ, et al. Clearance of circulating tumor cells after excision of primary colorectal cancer. Ann Surg 2002; 235:226–31. 11. Sugawara Y, Yamamoto J, Yamasaki S. et al Estimating the prognosis of hepatic resection in patients with metastatic liver tumors from colorectal cancer with special concern for the timing of hepatectomy. Surgery 2001;129:408–13. 12. Elias D, Cavalcanti A, Sabourin JC, et al. Resection of liver metastases from colorectal cancer: the real impact of the surgical margin. Eur J Surg Oncol 1998;24:174–9. 13. Yamamoto J, Sugihara K, Kosuge T, et al. Pathological support for limited hepatectomy in the treatment of liver metastases from colorectal cancer. Ann Surg 1995;221:74–8. 14. Rutherford EE, Karanjia ND. The measurement of liver resection margins. HPB 2004;6:18–20. 15. Melendez JA, Arslan V, Fischer ME, et al. Perioperative outcomes of major hepatic resections under low central venous pressure
16.
17.
18.
19. 20.
21. 22.
23.
24. 25.
anaesthesia: blood loss, blood transfusion, and the risk of postoperative renal dysfunction. J Am Coll Surg 1998;187:620–5. Brandstrup B, Tonnesen H, Beier-Holgersen R, et al. Effects of intravenous fluid restriction on postoperative complications: comparison of two perioperative fluid regimens. Ann Surg 2003;238:641–8. Johnson M, Mannar R, Wu AV. Correlation between blood loss and inferior vena caval pressure during liver resection. Br J Surg 1998;85: 188–90. Kooby DA, Stockman J, Ben-Porat L, et al. Influence of transfusions on perioperative and long-term outcome in patients following hepatic resection for colorectal metastases. Ann Surg 2003;237: 860–70. Houbiers JGA, Busch ORC, van de Watering LMG, et al. Blood transfusion in cancer surgery: a consensus statement. Eur J Surg 1995;161:307–14. Elias D, De Baere T, Roche A, et al. During liver regeneration following right portal embolization the growth rate of liver metastases is more rapid than that of the liver parenchyma. Br J Surg 1999;86: 784–8. Fong Y, Blumgart LH, Cohen A, et al. Repeat hepatic resections for metastatic colorectal cancer. Ann Surg 1994;220:657–62. Fernandez FG, Drebin JA, Linehan DC, Dehdashti F, Siegel BA, Strasberg SM. Five-year survival after resection of hepatic metastases from colorectal cancer patients screened by positron emission tomography with F-18 fluorodeoxyglucose (FDG-PET). Ann Surg 2004;240: 438–47. Imamura H, Seyama Y, Kokudo N, et al. One thousand fifty-six hepatectomies without mortality in 8 years. Arch Surg 2003;138: 1198–206. Fan ST, Lo CM, Liu CL, et al. Hepatectomy for hepatocellular carcinoma: toward zero hospital deaths. Ann Surg 1999;229:322–30. Sjo¨vall A, Ja¨rv V, Blomqvist L, et al. The potential for improved outcome in patients with hepatic metastases from colon cancer: a population-based study. Eur J Surg Ooncol 2004;30:834–41.
EJSO 33 (2007) S105eS110
www.ejso.com
Review
The result of liver resection for colorectal metastases in a small volume hospital in western Norway K. Søndenaa a,*, I. Nesvik b, K. Eiriksson b, M. Vetrhus b b
a Institute of Surgical Sciences, University of Bergen, Norway Department of Surgery, Stavanger University Hospital, Stavanger, Norway
Accepted 26 September 2007 Available online 5 November 2007
Abstract Aim: Our aim was to compare liver resection for colorectal metastases in a non-referral, small volume unit with a dedicated staff, with results from larger units. Methods: Thirty patients (15 men and 15 women) with a median age of 64 years (range 29e78) underwent hepatic resection during a 5-year period from 1997 to 2003 in a teaching hospital in western Norway. Results: Sixty-three percent (19/30) of the colorectal tumours were Dukes stage C (n ¼ 19) and CEA was increased in seven patients (23%), of which four (13%) had values above 50 mg/l. The metastases were synchronous with the colorectal tumours in 11 patients (37%). Nonanatomical (wedge) resections were the dominant type of surgeries and the resection margins were clear in all patients. A 77-year-old man (3%) died of MOF after right hemihepatectomy. Morbidity was encountered in eight other patients (28%). In 22 patients (76%) with recurrent disease, metastases first appeared in the liver in 18 (82%) of these patients. Seven patients (23%) have had resections for recurrences. Mean time to recurrence was 20 months (range 3e87). The actuarial 5-year survival rate was 42%. Six patients (20%) are currently disease free. Conclusion: Although our unit has treated a small number of patients compared with specialized units elsewhere, the survival rate, as well as morbidity and mortality, were comparable. However, 62% have had recurrent liver disease and this may suggest a role for neoadjuvant or adjuvant chemotherapy in selected cases. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Colorectal cancer; Liver metastases; Hepatic resection; Recurrence; Survival
Introduction Hepatic resection for colorectal liver metastases has become established as a reliable primary treatment modality.1e4 The common indication in a physically fit patient who has liver metastases and is able to tolerate general anaesthesia, is that the metastases are amenable to surgical resection provided a minimum of about 30% of functioning liver tissue is left behind without signs of extra-hepatic disease.5 Considering these provisions, complete surgical resection may result in a cure for some patients.3,5 The 5-year survival * Corresponding author: Department of Surgery, Haraldsplass Deaconal Hospital, PO Box, 6165 Postterminalen, N-5892 Bergen, Norway. E-mail address: [email protected] (K. Søndenaa). 0748-7983/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejso.2007.09.021
rate is generally considered to be above 30% or even 40% in favourable cases.1,5 These excellent figures from expert centres have set a standard that others should aim for. Colorectal liver metastasis is the commonest indication for liver resection in our country of 4.5 million people and is concentrated in a few regional hospitals. However, the number of hepatic resections is small compared with larger units abroad. It has been indicated that large volume units may achieve superior results compared with smaller units for several types of surgery.6,7 Low volume has been arbitrarily defined from 5 to 25 cases per year but may depend on the complexity of the operation.6,7 Still, surgery has been offered to all the patients in our catchment area whenever the possibility of radical resection was present and no one has been weeded out through a referral
S106
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
system. The question that remains is whether surgery should be concentrated in a few specialist hospitals in order to obtain optimal patient care and treatment. The geography of our country does not allow easy communication between hospitals by ground transport and we believe that treatment should be offered locally, given a minimum number of patients and a qualified and dedicated staff. Our aim was to present the results from a non-referral, small volume but dedicated unit, to examine if our patients have received standard treatment and been offered qualified care in an acceptable setting compared with that reported from larger units.
anaesthetic was administered regularly. It has been customary for our anaesthesiologists to use fluid administration freely during anaesthesia, but a CVP of less than 5 mmHg was intentionally used as a guideline in the latter part of the study. Every specimen was inspected and measured by the surgeon in the operating room. The surgically defined margin was used as a measure of the resection margin. Drains were used regularly but withdrawn as soon as fluid evacuation was below 100 ml/day. Results Surgery
Patients and methods Thirty patients (15 women and men) with a median age of 64 years (range 29e78), were operated by the first author during a 5.5-year period from 1997 through February 2003. Follow-up was concluded in April 2006. Another 6 patients (17%) were only explored because of undetected, peritoneal metastases (n ¼ 5) or nodes in the hepato-duodenal ligament (n ¼ 1) but were excluded from analysis. All patients belonged to the hospital’s catchment area of 265 000 inhabitants. Pathology The colorectal tumours were Dukes stage A (n ¼ 2), B (n ¼ 9) or C (n ¼ 19). Twenty-two tumours (75%) were located in the rectum (n ¼ 13) or the sigmoid colon (n ¼ 9), the rest were in the left (n ¼ 3) or right colon (n ¼ 5). The liver disease was bilateral in 14 patients (47%), located on the right side in another 14, and on the left side only in 2 (6%). It was not a policy of ours to biopsy the liver lesion but six patients had this done before surgical evaluation. Venous invasion by the primary tumour was observed at histological examination in 2 patients. Before the liver operation, CEA was increased (normal upper limit 9 mg/l before June 2002; 3 mg/l later) in 11 patients with a mean value of 57 mg/l, range 10e174. Four patients had values above 50 mg/l. Surgical technique All operations, except the first, one late second look and one late re-resection, were done by the first author. A J-shaped incision and intra-operative ultrasonography were used routinely. Liver tissue was divided using the forceps crushing technique but in the latter part of the study we used the Cavitron Ultrasonic Surgical Aspirator (CUSA, Tyco Health Care, Oslo, Norway). The Pringle manoeuvre was used whenever extensive resections were done. The portal triad was then clamped for 15 min interrupted by reperfusion periods of 5 min. Blood loss was substituted by packed RBC when Hb was anticipated to fall below 8 g/dl or to 10 g/dl in patients with heart disease. An epidural
Two patients had synchronous colorectal and hepatic surgery and hepatic surgery followed the primary, colorectal operation within 3 months in 9 other patients and the operation was therefore deemed synchronous in 37% (11/30). Resections were right hemi-hepatectomy or more than two segments in 9 patients (30%), segmentectomies or two or more non-anatomical parts (wedges) in 13, and a single part in 8. Extra-hepatic intra-abdominal disease was removed in three patients. Sequential Pringle’s manoeuvre was used in 26 patients (87%). The mean number of occlusion episodes was 3.5 (range 1e8). Additional chemotherapy, radiofrequency ablation (n ¼ 2) and lung surgery (n ¼ 2) were used in some patients. Morbidity and mortality A 77-year-old man (3%) died after a right hemi-hepatectomy. He had a cerebral haemorrhage during the first postoperative day and S. aureus in fluid drained from the abdomen. He subsequently succumbed to liver and multiorgan failure. Morbidity was encountered in 8 other patients (28%). Percutaneous drainage for an intra-abdominal fluid collection was carried out in three patients of whom one had a suspected bile leak that dried up without further intervention. One patient needed reoperation for a wound dehiscence. Further details can be seen in Table 1. The median duration of surgery was 335 min (range 210e495) in those with a complication, compared with 240 min (100e660) in those without complications. Blood transfusions were given to 4 patients with complications (50%), median 1375 ml (range 500e2100), compared with 250 ml (250e1500) in 5 patients (23%) without complications. Pathology The resection margin was clear in all instances, although it was 1 mm or less in one patient (but found to be clear surgically). Nine patients had a margin of 1 mm, 11 had between 1 and 9 mm, and 9 had a resection margin equal to or exceeding 10 mm. Micro-satellite deposits were not
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
Follow-up course
Table 1 Details of the 28% morbidity in 29 surviving patients Morbidity
Patients (n ¼ 8)
Splenic laceration (intraoperative splenectomy) Wound dehiscence, cerebral infarction Infected fluid collection/subphrenic abscess Lung infiltration, cardiac insufficiency, bile leak, abscess, septicaemia Ascites Cardiac arrhythmia, urinary tract infection Nerve injury right shoulder blade
1
a
S107
1a 2 1 1 1 1
Patient reoperated.
observed in routine pathological sections nor was evidence of such manifestations found intra-operatively. One deceased patient had a lymph node metastasis at the superior mesenteric vein removed before liver surgery and two had simultaneous removal of intra-abdominal metastases. Of these, one deceased patient had lymph nodes in the hepato-duodenal ligament that contained metastases and one had localized peritoneal seeding. In fact, this patient had a small peritoneal seeding just opposite a single liver metastasis after a preoperative liver biopsy. Full results are shown in Table 2.
First recurrent metastases in 22 patients were detected in the liver (n ¼ 12), liver and lung (n ¼ 3), lung (n ¼ 1), or multiple or other sites (n ¼ 6) of which three included the liver. In the 18 patients with liver recurrences, bilateral disease was eventually observed in 14 patients. It was confined to the right lobe in 4 patients, two with original disease in the right hemiliver only. Only one patient with recurrence had this at the sites of three previously removed liver metastases. Mean time to recurrence was 20 months (range 3e87). Seven patients (23%) were re-resected. Another patient with two recurrent liver metastases could have been a candidate for reoperation but for locally advanced prostate cancer recently diagnosed that required specific treatment felt to prevent further liver surgery. One renewed liver resection took place in an outside hospital after the patient moved and was followed up close to his new home. CEA became normal in 24 patients (two deceased were not analyzed) after the liver resections. In 4 deceased patients, the CEA value varied between 22 and 2468 mg/l, consistent with disease expression within a few months after hepatic surgery. Five patients were also treated later for a different malignancy in these organs: non-Hodgkin’s lymphoma (n ¼ 1, deceased), prostate (n ¼ 1, deceased), urinary bladder (n ¼ 2), and skin (n ¼ 1). Survival
Variable
Demographics Men/women (n ¼ 15/15) Ageb Tumour parameters Rectal tumours (n ¼ 13) Diff. CR tumour (n ¼ 29) Dukes stage Liver op. <3/<12 months Bilateral disease (n ¼ 14) Number of metastasesb Size (mm) of metastasesb Elevated CEA preop. (n ¼ 11/29) CEA >50 mg/l preop. a b
Deceased, n ¼ 19 (63%)a
AWD, n¼5 (17%)
NED, n¼6 (20%)
9/10
2/3
4/2
63 (29e78)
59 (53e68)
60 (45e71)
3
2
1H/15M/2L
5 moderate
6 moderate
1A/5B/13C 8/5 ¼ 13
1A/1B/3C 1/3 ¼ 4
3B/3C 2/1 ¼ 3
10
4
0
3 (1e5)
4.4 (1e7)
1.8 (1e3)
35 (20e100)
22 (8e60)
35 (15e55)
7
3
1
4
0
0
8
The 5-year survival rate was calculated as 42% (Fig. 1). Five patients are alive with disease (AWD) although recurrent disease is only suspected based on a CEA value of 28. Six patients are alive without signs of disease (NED) (see Table 3).
Kaplan-Meier Cum. Survival Plot for Survival Censor Variable: Censor 1 ,8
Cum. Survival
Table 2 Demographics and tumour specific data in 30 patients
,6 ,4
,2
Including one postoperative mortality. Values expressed as mean (range).
0 0
20
40
60
80
100
Time Figure 1. Survival following liver resection for 30 patients.
120
S108
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
Table 3 Surgical and follow-up data in 30 patients Variable Resection margin 0e1 mm (n ¼ 10) 2e9 mm (n ¼ 11) 10 mm (n ¼ 9) Pringle procedure (n ¼ 26) Pringle timea Duration of liver surgerya Blood transfusions (n ¼ 9; 30%) Transfused amounta Re-operations (n ¼ 10) Re-resections liver (n ¼ 7) Time to disease recurrence (n ¼ 23) Follow-up from liver operation a b c d e
Deceased, n ¼ 19 (63%)b
AWD, n ¼ 5 (17%)c
NED, n ¼ 6 (20%)
8 5 6 18 58 (10e157) 304 (120e660) 7 861 (250e1500) 11 7d 17 (3e53) 34 (1e102)
1 4 0 4 54 (18e85) 273 (195e390) 1 250 5 3e 36 (6e87)c 70 (43e97)
1 2 3 4 85 (15e137) 262 (150e495) 1 2100 0 0 71 (58e99)
Values are expressed as mean (range). Including one postoperative mortality. Including one patient with elevated CEA but no proven metastasis on imaging. Three patients had a third liver operation. One patient had an additional lung resection.
In the NED group, one patient had a preoperative liver biopsy, and one patient had venous invasion by his colorectal tumour. Discussion The most important intention of surgery for liver metastases is to prolong life and in some cases obtain long-term survival that should now exceed 30% at 5 years.2,4,5 In contrast to a couple of decades ago, less than 4 metastases and confinement to one hemi-liver are no longer contraindications to hepatic resection.5 The study population was recruited from our colorectal section. Thus, 70% of our patients were diagnosed within 1 year of the colorectal operation. In fact, two of our patients probably would not have been operated on elsewhere because rapid recurrent disease would have prevented their candidacy for hepatectomy. Others have found a longer time interval between the primary colorectal operation and the hepatic resection to be an important prognostic factor although no common agreement has been reached as to what constitutes a short interval. It has been given as 3 months5 or 12 months.2 In our opinion, appearance within 3 months obviously is equivalent to synchronous disease. But in the end, this may be a matter of semantics as metastases generally are regarded as originating from the primary tumour.5 Whether these are operated synchronously or by a deferred operation does not seem to influence prognosis.8 In large series, according to multivariate analysis, survival has been influenced by Dukes stage, the time period between the primary and the liver operation, number of lesions, a resection margin of more than 1 mm, and extrahepatic disease2,5 as well as tumour size and CEA.2 However, these parameters should not be considered as absolute contraindications if the hepatic lesions can be totally and
securely removed because there are long-term survivors after surgery who would otherwise succumb to the disease within a short time.3,5,9 Nevertheless, a score based on time-span, number of metastases and CEA level, proved to be quite reliable in predicting outcome.2 Dukes stage C was more prominent in our patients who died or were classified as AWD. The lymph node status of the colorectal tumour may perhaps indicate that the disease process is continuing and may minimize the effect of such variables as the time interval.10,11 It is remarkable that only two patients had definite signs of venous invasion in the colorectal tumour specimen. In a study from New York,1 the single most important surgical feature was the extent of the free resection margin. The authors speculated whether it is easier to achieve sufficient resection margins greater than 1 mm in anatomic resections1 but others have not reported similar importance of the resection margin.5,12 Although it is generally agreed that margins of 10 mm are advisable,12 it is debatable whether this is absolutely necessary as an oncological principle. We did not observe satellite lesions outside the boundary of any metastasis. Others have also reported this to be rare,13 whereas one author12 hypothesized that there is an increased possibility of microscopic deposits within 10 mm of the main lesion. However, it is mandatory to avoid exposure of the lesion on the cut surface of the liver.5 Unfortunately, exposure has usually not been addressed as a separate issue and it is unclear what the magnitude of the problem is. Nevertheless, it seems reasonable to conclude that exposure must be avoided but that the resection margin per se does not matter significantly.5 It is also a matter of principle how the margin should be measured. In a report by DeMatteo et al.,1 no mention was made of how it was done or by whom, although assumedly
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
it was done by the pathologist. Liver tissue shrinks by about 10% during the fixation process and this may be too little to influence the measurement.14 However, a narrow tissue band of 1 mm may easily become distorted or rupture during the specimen handling process and a false impression of exposure may be created. On the other hand, modern instruments used for dissection destroy tissue along the resection line actually making this wider. Pringle’s manoeuvre was used routinely by us with the aim of reducing blood loss. After discussions with the anaesthesiologists it was found acceptable to have a restrictive transfusion policy and only 30% of our patients had a transfusion. However, it has been their long-time policy to use fluid liberally to counteract the effect of anaesthesia and the surgical trauma and this, by itself, may lower the haemoglobin value. Such a policy has been shown to be unnecessary in most operations and even harmful.15e17 Although controversial, it has long been contended that blood transfusions affect morbidity and prognosis adversely.18,19 It has been our policy to remove the metastases completely but with the intention of preserving as much tissue as possible in accordance with the policy of others.5 Generally, this meant doing non-anatomical (wedge) resections, even if multiple. Whereas the number and sizes of the metastases did not offer any obvious clues to prognosis, bilateral distribution was apparently overrepresented in the two groups with the worst outcome. Perhaps it should be considered if anatomic resections would serve the patient better in some cases.1 Undiscovered metastases have the potential to grow after liver resection.20 Nevertheless, when the disease recurred in the liver, eventually it was bilateral in all instances (88%) except for two cases. Others have reported that recurrent disease was found in the opposite lobe in 50%.21 The presence of intra-abdominal extra-hepatic disease usually precludes liver resection but an acceptable number of patients still achieve long-term survival.5,9 We operated on three patients in whom such lesions were discovered. Gender has not been found to influence prognosis after treatment of colorectal cancer with liver metastases.5 Although our figures are small, there was a tendency for women to do worse. Four of our patients had a preoperative CEA value above 50 mg/l, which indicated a poor prognosis.5 An increased postoperative CEA was an ominous sign in our experience. In 7 re-resected patients, the recurrent tumour was found in a different location than in the primary operation. An interesting observation was made by Fong et al.20 who found distributions of recurrent disease survival patterns similar to ours after the primary operation. This may be in agreement with others5 who have suggested that recurrent disease may have a similar prognosis to the primary metastases and that some patients do better than others without obvious clinical signs to indicate this biological difference. Modern equipment like the PET scanner may nevertheless have an impact in the future.22
S109
Morbidity in a recent large series has been reported to be around 10e28%1,8,23 and mortality has been reduced to zero in the hands of experts.23 In comparison, in our series we had 3% mortality and 28% morbidity. It has been shown that the two variables that the surgeon is able to influence, namely the duration of the operation and the use of blood transfusions, may have a bearing on postoperative morbidity.18,24 It may well be argued that our operations might have been done within a shorter operating time. It is of practical interest to decide if trying to avoid large resections will instead lead to longer operating time and shorter resection margins thereby jeopardizing optimal surgical therapy.1 Blood transfusions may also reduce survival3 and should be kept at a minimum through meticulous surgery and avoidance of a too liberal transfusion practice.15e17 Our transfusion rate was similar to that from a large centre.15 Surgical resection for colorectal liver metastasis is an accepted treatment form and the only one able to offer longterm survival to a substantial number of patients. Although this is complicated surgery, our results show that a dedicated staff in a small volume hospital may provide good care similar to that of large volume institutions even though the opposite view has been advocated for surgical reasons.12 A recent study from high volume hospitals in Sweden25 found that a dedicated approach from specially trained surgeons might have increased the resection rate for colorectal hepatic metastases from 4% to 10% or even higher, underlining that the most important thing apart from maintaining low mortality and morbidity figures, is to assure that patients get the treatment they deserve from the medical community. Acknowledgements We are grateful to the late Arild K. Olsen, MD, who allowed us to include one of his patients in the study population and for his valuable contribution to liver surgery at our hospital. Conflicts of interest None of the authors has any ‘‘Conflict of interest’’. References 1. DeMatteo RP, Palese C, Jarnagin WR, et al. Anatomic segmental hepatic resection is superior to wedge resection as an oncological operation for colorectal liver metastases. J Gastrointest Surg 2000;4:178–84. 2. Fong Y, Fortner J, Sun RL, et al. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer. Analysis of 1001 consecutive cases. Ann Surg 1999;230:309–21. 3. Jamison RL, Donohue JH, Nagorney DM, et al. Hepatic resection for metastatic colorectal cancer results in cure for some patients. Arch Surg 1997;132:505–10. 4. Scheele J, Stang R, Altendorf-Hofmann A, Paul M. Resection of colorectal liver metastases. World J Surg 1995;19:59–71.
S110
K. Søndenaa et al. / EJSO 33 (2007) S105eS110
5. Minagawa M, Makuuchi M, Torzilli G, et al. Extension of the frontiers of surgical indications in the treatment of liver metastases from colorectal cancer: long-term results. Ann Surg 2000;231:487–99. 6. Begg CB, Cramer LD, Hoskins WJ, et al. Impact of hospital volume on operative mortality for major cancer surgery. JAMA 1998;280: 1747–51. 7. Rabeneck L, Davila JA, Thompson M, et al. Surgical volume and long-term survival following surgery for colorectal cancer in the Veterans Affairs health-care system. Am J Gastroenterol 2004;99: 668–75. 8. Chua HK, Søndenaa K, Tsiotos GG, et al. Concurrent vs. staged colectomy and hepatectomy for primary colorectal cancer with synchronous hepatic metastases. Dis Colon Rectum 2004;47:1310–6. 9. Elias D, Ouellet JF, Bellon N, et al. Extrahepatic disease does not contraindicate hepatectomy for colorectal liver metastases. Br J Surg 2003;90:567–74. 10. Patel H, Le Marer N, Wharton RQ, et al. Clearance of circulating tumor cells after excision of primary colorectal cancer. Ann Surg 2002; 235:226–31. 11. Sugawara Y, Yamamoto J, Yamasaki S. et al Estimating the prognosis of hepatic resection in patients with metastatic liver tumors from colorectal cancer with special concern for the timing of hepatectomy. Surgery 2001;129:408–13. 12. Elias D, Cavalcanti A, Sabourin JC, et al. Resection of liver metastases from colorectal cancer: the real impact of the surgical margin. Eur J Surg Oncol 1998;24:174–9. 13. Yamamoto J, Sugihara K, Kosuge T, et al. Pathological support for limited hepatectomy in the treatment of liver metastases from colorectal cancer. Ann Surg 1995;221:74–8. 14. Rutherford EE, Karanjia ND. The measurement of liver resection margins. HPB 2004;6:18–20. 15. Melendez JA, Arslan V, Fischer ME, et al. Perioperative outcomes of major hepatic resections under low central venous pressure
16.
17.
18.
19. 20.
21. 22.
23.
24. 25.
anaesthesia: blood loss, blood transfusion, and the risk of postoperative renal dysfunction. J Am Coll Surg 1998;187:620–5. Brandstrup B, Tonnesen H, Beier-Holgersen R, et al. Effects of intravenous fluid restriction on postoperative complications: comparison of two perioperative fluid regimens. Ann Surg 2003;238:641–8. Johnson M, Mannar R, Wu AV. Correlation between blood loss and inferior vena caval pressure during liver resection. Br J Surg 1998;85: 188–90. Kooby DA, Stockman J, Ben-Porat L, et al. Influence of transfusions on perioperative and long-term outcome in patients following hepatic resection for colorectal metastases. Ann Surg 2003;237: 860–70. Houbiers JGA, Busch ORC, van de Watering LMG, et al. Blood transfusion in cancer surgery: a consensus statement. Eur J Surg 1995;161:307–14. Elias D, De Baere T, Roche A, et al. During liver regeneration following right portal embolization the growth rate of liver metastases is more rapid than that of the liver parenchyma. Br J Surg 1999;86: 784–8. Fong Y, Blumgart LH, Cohen A, et al. Repeat hepatic resections for metastatic colorectal cancer. Ann Surg 1994;220:657–62. Fernandez FG, Drebin JA, Linehan DC, Dehdashti F, Siegel BA, Strasberg SM. Five-year survival after resection of hepatic metastases from colorectal cancer patients screened by positron emission tomography with F-18 fluorodeoxyglucose (FDG-PET). Ann Surg 2004;240: 438–47. Imamura H, Seyama Y, Kokudo N, et al. One thousand fifty-six hepatectomies without mortality in 8 years. Arch Surg 2003;138: 1198–206. Fan ST, Lo CM, Liu CL, et al. Hepatectomy for hepatocellular carcinoma: toward zero hospital deaths. Ann Surg 1999;229:322–30. Sjo¨vall A, Ja¨rv V, Blomqvist L, et al. The potential for improved outcome in patients with hepatic metastases from colon cancer: a population-based study. Eur J Surg Ooncol 2004;30:834–41.