Value of alpha-foetoprotein for screening of recurrence in hepatocellular carcinoma post resection.

INTRODUCTION: The aim of this study was to establish the value of alpha-foetoprotein (AFP) for the screening of recurrences in hepatocellular carcinoma (HCC) in patients who have undergone curative hepatic resection. METHODS: 72 HCC patients who had curative resection/liver transplant in 2000-2006 were monitored for recurrence by evaluating the three- or six-monthly AFP and computed tomography images. Patients without recurrence were followed up for a mean duration of 7.27 years. RESULTS: Out of the 72 patients, 34 (47.2%) suffered from HCC recurrence. 65.4% of recurrent cases had AFP values showing an upward trend. Patients with recurrence had higher AFP values than those without at last follow-up (119.45 µg/L vs. 3.1 µg/L, p < 0.001). AFP at recurrence was independent of gender, race, history of alcohol consumption and hepatitis C or cirrhosis status. Patient with hepatitis B or those with tumours larger than 5 cm had higher AFP values. The best cut-off AFP indicative of HCC recurrence was 5.45 µg/L (sensitivity 84.4%; specificity 77.1%). High preoperative AFP was associated with high AFP at recurrence (correlation coefficient 0.553, p = 0.01). CONCLUSION: AFP alone is an inadequate screening test for HCC recurrence since only about two-thirds of patients showed upward AFP trend on recurrence. Our study found a relatively low cut-off point for detection of recurrence (5.54 µg/L). Patients with high preoperative AFP tended to have high AFP on recurrence. Imaging is recommended for patients with AFP levels > 5.45 µg/L, especially when AFP shows a rising trend.

Successful living donor liver transplantation in portomesenteric thrombosis.

Chronic portal vein thrombosis (PVT) is often considered a relative contraindication for living donor liver transplantation due to the risks involved and higher morbidity. In this report, we describe a surgical strategy for living donor liver transplant in patient with complete PVT using venovenous bypass from the inferior mesenteric vein (IMV) and then using a jump graft from the IMV for portal inflow into the graft. IMV is a potential source for portal inflow in orthotopic liver transplant.

The selective use of magnetic resonance cholangiopancreatography in the imaging of the axial biliary tree in patients with acute gallstone pancreatitis.

BACKGROUND: Magnetic resonance cholangiopancreatography (MRCP) is an emerging modality in the management of acute gallstone pancreatitis (AGP). The aim of this study was to assess the impact following the introduction of MRCP in the management of AGP in a tertiary referral unit. METHODS: Patients presenting with AGP from January 2002 to December 2004 were reviewed to assess the impact of the introduction of MRCP in June 2003. The indication for MRCP was suspected common bile duct (CBD) stones in the absence of biliary sepsis. Definitive treatment for AGP was laparoscopic cholecystectomy, with endoscopic retrograde cholangiopancreatography (ERCP) and sphincterotomy reserved for patients unfit for cholecystectomy and those with biliary sepsis. RESULTS: 249 patients were identified of whom 36 (14.5%) underwent ERCP and sphincterotomy as definitive treatment. 96 patients with a non-dilated CBD and normal or resolving liver function tests proceeded to laparosocopic cholecystectomy and intraoperative cholangiogram (IOC), 8 (8.5%) of whom had CBD stones intraoperatively. Eleven patients underwent cholecystectomy during pancreatic necrosectomy. Of those undergoing preoperative diagnostic biliary tract imaging, ERCP was undertaken in 57 patients and MRCP in 49 patients. There was no significant difference in serum bilirubin levels [ERCP 43 mmol/l (18-204) vs. MRCP 39 mmol/l (24-180), p = NS] or the proportion of patients with CBD stones [ERCP 10 (17.5%) vs. MRCP 7 (14.2%), p = NS] between the two groups. Patients who underwent MRCP had a shorter median hospital stay [MRCP 5 days (range: 3-14) vs. ERCP 9 days (range: 4-20), p < 0.01] and higher rate of cholecystectomy during the index admission (MRCP 83.3% vs. ERCP 67.2%, p < 0.05). There was a high degree of correlation between preoperative MRCP results and findings of subsequent IOC or therapeutic ERCP (area under ROC curve: 0.94). CONCLUSIONS: MRCP is an accurate modality for imaging the axial biliary tree in patients with AGP. Selective use of MRCP reduces the need for ERCP and results in shorter hospital stay. and IAP.

Liver resection and transplantation in the management of iatrogenic biliary injury.

BACKGROUND: Biliary injury during cholecystectomy can be managed successfully by biliary reconstruction in the majority of patients; however, a proportion of patients may require hepatic resection or even liver transplantation. METHODS: Data on all patients referred with biliary injuries were recorded prospectively. The details of patients who required hepatic resection or transplantation were analyzed and compared to those patients managed with biliary reconstruction alone. RESULTS: From November 1984 until November 2003 there were 119 patients referred with Strasberg grade E injuries to the biliary tree, 14 of whom (9 women, 5 men) required hepatic resection or transplantation. The median age of these 14 patients was 48 (range: 30-81) years. Nine patients were considered for hepatic resection, and of these six underwent right hepatectomy, two had a left lateral sectionectomy, and one patient was deemed unfit for surgery and underwent metal stenting of the right hepatic duct. All patients are alive and remain well. Five patients developed hepatic failure and were considered for liver transplantation. Two patients who were unfit for transplantation died, and another died while on the waiting list for transplantation. The remaining two patients underwent liver transplantation, and one of them died from overwhelming sepsis. Concomitant vascular injury was demonstrated in 8 of the 14 patients (57%), and in 3 of the 4 (75%) patients that died. CONCLUSIONS: Hepatic atrophy or sepsis after biliary injury can be managed successfully with hepatic resection. Liver transplantation is required occasionally for patients with secondary biliary cirrhosis, but is rarely successful for early hepatic failure following iatrogenic biliary injury.

Prognostic factors in patients undergoing surgery for severe necrotizing pancreatitis.

Pancreatic necrosectomy remains an important treatment modality for the management of infected pancreatic necrosis but is associated with significant mortality. The aim of this study was to identify factors associated with mortality following pancreatic necrosectomy. Patients who underwent pancreatic necrosectomy from January 1995 to December 2004 were reviewed. The association between admission, preoperative and postoperative variables, and mortality was assessed using logistic regression analysis. A total of 1248 patients presented with acute pancreatitis, of whom 94 (7.5%) underwent pancreatic necrosectomy (51 men, 43 women). The preoperative median Acute Physiology, Age, and Chronic Health Evaluation (APACHE II) score was 9 (range 2-19). The median cumulative organ dysfunction score was 2 (0-9) preoperatively and 4 (1-11) postoperatively. In all, 23 patients (24.5%) died. Those who died were older than the survivors; the ages (median and range) were 69 years (40-80 years) versus 52 years (19-79 years) (p < 0.05). They also had higher admission APACHE II scores (median and range): 14 (12-19) versus 9 (2-22) (p < 0.001). There were significant associations between preoperative (p < 0.01) and postoperative (p < 0.01) Marshall scores and mortality following pancreatic necrosectomy. The presence of the systemic inflammatory response syndrome (SIRS) during the first 48 hours (p < 0.01) and the time between presentation and necrosectomy (p < 0.01) were independent predictors of survival. Pancreatic necrosectomy is associated with higher mortality in patients with increased APACHE II scores, early persistent SIRS, and unresolved multiorgan dysfunction. Necrosectomy is associated with poorer outcome when performed within 2 weeks of presentation.

An audit of the management of patients with acute pancreatitis against national standards of practice.

BACKGROUND: The aim of this study was to audit the management of patients with acute pancreatitis against the standards of practice in the British Society of Gastroenterology guidelines. METHODS: The study assessed consecutive patients with acute pancreatitis over 5 years. Audit targets were overall mortality below 10 per cent, mortality for severe acute pancreatitis below 30 per cent, correct diagnosis and severity stratification within 48 h, aetiology determined in more than 80 per cent, availability of computed tomography and high-dependency or intensive therapy units when indicated and definitive treatment of gallstone pancreatitis within 2 weeks. RESULTS: Of 759 patients with acute pancreatitis, 219 (28.9 per cent) had severe acute pancreatitis (SAP). Overall mortality was 5.9 per cent, and 19.6 per cent in those with SAP. Acute pancreatitis was diagnosed within 48 h of presentation in 96.3 per cent of patients. The definitive aetiology was classified in 87.5 per cent. Of patients with SAP, 95.9 per cent underwent computed tomography within 6-10 days of admission. Of 93 patients with severe gallstone pancreatitis, 48 per cent had urgent endoscopic retrograde cholangiopancreatography, and 89.6 per cent of 359 patients with acute gallstone pancreatitis underwent definitive management within 2 weeks of admission. CONCLUSION: Patients with acute pancreatitis can be managed according to revised guidelines with a low associated mortality.

Palliation for suspected unresectable hilar cholangiocarcinoma.

AIM: The aim of this study was to evaluate the outcome of different techniques of palliation for patients with hilar cholangiocarcinoma. METHOD: All patients treated with palliative intent between 1988 and 2004 at the Royal Infirmary of Edinburgh were reviewed. Patients were analysed on an intention to treat basis. Demographics, procedure and outcome (including re-admissions) were recorded. RESULTS: Two hundred and thirty-three patients underwent palliative treatment for suspected hilar cholangiocarcinoma. The diagnosis was confirmed histologically in 109 patients. The procedure related morbidity and mortality was 54/225 and 18/207 respectively. Seventy-one patients required re-admission. Twenty patients underwent surgical biliary bypass for jaundice. Those undergoing surgical palliation had a longer median (95% CI) time to re-admission (16 (0-36) vs.7 (2-12) weeks, p=0.001). Endoscopic retrograde cholangio-pancreatography (ERCP) and stenting was only successful in 28 patients and was associated with a significantly higher re-admission rate compared to patients in whom ERCP was not performed (60/179 vs. 4/27, p=0.050). The overall median (95% CI) survival was 145 (124-185) days. CONCLUSION: Current options for palliation of hilar cholangiocarcinoma provide good short term success but are all associated with significant early and late morbidity. Due to its low success and association with an increased re-admission rate, ERCP for definitive palliation should not be used in the first line staging and management of these patients.

Follow-up and outcomes for resection of colorectal liver metastases in Edinburgh.

AIM: The aim of this study was to assess the value of a defined follow-up protocol for patients undergoing potentially curative hepatic resection for colorectal hepatic metastases. METHODS: A standard protocol for the duration of the study consisted of clinical assessment, serum carcinoembryonic antigen (CEA) and computed tomography. Patterns of recurrence, method and timing of diagnosis and outcome were recorded. RESULTS: One hundred and ninety-one patients underwent potentially curative resection from 1989 to 2004 of whom 103 developed recurrence. The median (inter-quartile range) follow-up was 24.4 (6.5-42.3) months. The median (IQR) time to recurrence and overall survival was 25.0 (10 -not yet reached) and 45.2 (21-123) months, respectively. Seventeen patients (8.9%) underwent further surgery with curative intent. Fifty-five patients (57.9%) had recurrence diagnosed at routine follow-up with 71% (44/62) being diagnosed by CEA and CT. The CEA was elevated in 85.7% (72/84 patients) at the time of diagnosis of recurrence. CONCLUSION: Although the detection of recurrent disease is common during follow-up after hepatic resection for colorectal metastases, few patients will be suitable for further intervention with curative intent. The exact nature of the follow-up protocol remains to be determined but if it is going to be performed it should be most intensive within the first 3 years.

Predicting survival in patients with hepatocellular carcinoma: a UK perspective.

BACKGROUND AND AIMS: Hepatocellular carcinoma (HCC) is a cancer of rising incidence in the UK. The aim of this study was to compare the Okuda, Cancer of the Liver Italian Program (CLIP), and Barcelona Clinic Liver Cancer (BCLC) classifications as predictors of survival in UK patients with HCC. METHODS: Data were analysed from a prospective database maintained in a specialist hepatobiliary unit from 1998 to 2003. Each system was assessed for its discriminatory power, monotonicity of gradient, and independent contribution to prediction of mortality status based on a multivariate model. RESULTS: One hundred and two patients (77 males, 25 females) were identified with a median age of 65 (range, 14-87) years. The overall median survival time was 13 months and the one- and five-year survival rates were 52.9% (95% CI: 43.2%, 62.6%) and 35.3% (95% CI: 26.0%, 44.6%), respectively. All three classification systems had the capacity to differentiate between patient survival times across different stages. The Okuda system was superior in overall discriminatory power and in strength of monotonicity. The BCLC system, however, made the highest independent contribution of all three systems in predicting survival in the Cox regression model. CONCLUSIONS: All three classification systems were effective in predicting survival for patients with HCC in a UK population.

Preoperative lymphocyte count as a prognostic factor in resected pancreatic ductal adenocarcinoma.

BACKGROUND AND AIMS: Recognized prognostic factors for resected pancreatic ductal adenocarcinoma (PDAC) include tumour size, differentiation, resection margin involvement and lymph node metastases. A further prognostic factor of less certain significance is lymphocyte count. The aim of this study was to investigate whether preoperative lymphocyte count is a prognostic indicator in patients with PDAC. MATERIAL AND METHODS: Patients who had undergone a potentially curative pancreaticoduodenectomy (PD) for PDAC between 1998 and 2005 were analysed. Standard prognostic factors, preoperative lymphocyte count, preoperative neutrophil count and survival data were collected. RESULTS: Of the 44 patients studied, univariate analysis identified predictors of a poor survival as lymph node status (node positive (+ve) 10.3 [5.4-20.9] months versus node negative (-ve) 14.2 [10.9-31.4] months; p=0.038), posterior resection margin invasion (margin +ve 7.0 [5.1-15.0] months versus margin -ve 13.1 [10.0-28.3] months; p=0.025) and lymphocyte count below the reference range (<1.5 x 10(9)/litre 8.8 [7.0-13.1] months versus > or = 1.5 x 10(9)/litre 14.3 [7.0-28.3] months; p=0.029). Low preoperative lymphocyte count (p=0.027) and posterior margin invasion (p=0.023) retained significance on multivariate analysis. Preoperative neutrophil to lymphocyte ratio was not a significant prognostic factor. CONCLUSION: Preoperative lymphocyte count is a significant prognostic factor in patients with PDAC.

Perioperative transfusion for pancreaticoduodenectomy and its impact on prognosis in resected pancreatic ductal adenocarcinoma.

BACKGROUND AND AIMS: Pancreaticoduodenectomy (PD) is a major operative intervention performed most commonly for malignancy in the head of pancreas. The aim of this study was to evaluate the utilization of blood transfusion for PD and to determine whether this had prognostic significance in a subset of patients with pancreatic ductal adenocarcinoma (PDAC). MATERIAL AND METHODS: Data on blood transfusion requirement were retrospectively collected for patients undergoing PD from 1998 to 2005. Standard prognostic factors and survival data were also collected in patients with PDAC. RESULTS: One-hundred-and-seventy patients underwent PD. Seventy-six patients (45%) received transfusion. The median (interquartile range) number of units of red cell concentrate (RCC) transfused perioperatively (intraoperatively and within 24 h of surgery) was 1.5 (0.5-2.5). The median preoperative haemoglobin (Hb) was 126 g/dl. The median number of units of RCC transfused perioperatively in patients with Hb <126 g/dl was 2 (1-3); for those with Hb > or = 126 g/dl the median was 0 (0-1); p=0.003. Forty-nine patients who were resected for PDAC were subjected to survival analysis. Univariate and multivariate analyses showed that only posterior resection margin invasion was associated with an adverse outcome (margin positive 198 [143-470] days vs margin negative 398 [303-859] days; p=0.02). Perioperative RCC transfusion requirement was not a significant predictor of survival (transfusion 408 [214-769] days vs no transfusion 331 [217-391] days; p=0.18). Furthermore, RCC transfusion within 30 days of operation was not a significant predictor of poor survival (transfusion 331 [201-459] days vs no transfusion 317 [196-769] days; p=0.43). CONCLUSIONS: PD can be performed with a moderately low requirement for RCC transfusion; however, low preoperative haemoglobin is a predictor for the requirement of RCC transfusion. Administration of RCC transfusion does not appear to be a significant adverse prognostic factor in patients with resected PDAC.

Results of decompression surgery for pain in chronic pancreatitis.

INTRODUCTION: A vast majority of patients with chronic pancreatitis require regular opiate/opioid analgesia and recurrent hospital admission for pain. However, the role and timing of operative strategies for pain in chronic pancreatitis is controversial. This study hypothesized that pancreatic decompression surgery reduces analgesia requirement and hospital readmission for pain in selected patients. PATIENTS AND METHODS: This was a retrospective review of patients undergoing longitudinal pancreatico-jejunostomy (LPJ), with or without coring of the pancreatic head (Frey's procedure), between 1995 and 2007 in a single UK centre. Surgery was performed for chronic pain with clinical/radiological evidence of chronic pancreatitis amenable to decompression/head coring. RESULTS: Fifty patients were identified. Thirty-six were male with a median age of 46 years and median follow-up of 30 months. Twenty-eight underwent LPJ and 22 underwent Frey's procedure. No significant difference in reduction of analgesia requirement (71% vs 64%, p=0.761) or hospital readmission for pain (21% vs 23%, p=1.000) was observed when comparing LPJ and Frey's procedure. Patients were significantly more likely to be pain-free following surgery if they required non-opiate rather than opiate analgesia preoperatively (75% vs 19%, p=0.0002). Fewer patients required subsequent hospital readmission for pain if taking non-opiate rather than opiate analgesia preoperatively (12.5% vs 31%, p=0.175). CONCLUSIONS: In selected patients, LPJ and Frey's procedure have equivalent benefit in short-term pain reduction. Patients should be selected for surgery before the commencement of opiate analgesia.

Association between early systemic inflammatory response, severity of multiorgan dysfunction and death in acute pancreatitis.

BACKGROUND: Mortality in patients with acute pancreatitis is associated with the number of failing organs and the severity and reversibility of organ dysfunction. The aim of this study was to assess the significance of early systemic inflammatory response syndrome (SIRS) in the development of multiorgan dysfunction syndrome (MODS) and death from acute pancreatitis. METHODS: Data for all patients with a diagnosis of acute pancreatitis between January 2000 and December 2004 were reviewed. Serum C-reactive protein (CRP), Acute Physiology And Chronic Health Evaluation (APACHE) II scores and presence of SIRS were recorded on admission and at 48 h. Marshall organ dysfunction scores were calculated during the first week of presentation. Presence of SIRS and raised serum CRP levels on admission and at 48 h were correlated with the cumulative organ dysfunction scores in the first week. RESULTS: A total of 759 patients with acute pancreatitis were identified, of whom 45 (5.9 per cent) died during the index admission. SIRS was identified in 162 patients on admission and was persistent in 138 at 48 h. The median (range) cumulative Marshall score in patients with persistent SIRS was significantly higher than that in patients in whom SIRS resolved and in those with no SIRS (4 (0-12), 3 (0-7) and 0 (0-9) respectively; P < 0.001). Thirty-five patients (25.4 per cent) with persistent SIRS died from acute pancreatitis, compared with six patients (8 per cent) with transient SIRS and four (0.7 per cent) without SIRS (P < 0.001). No correlation was observed between CRP level on admission and Marshall score (P = 0.810); however, there was a close correlation between CRP level at 48 h and Marshall score (P < 0.001). CONCLUSION: Persistent SIRS is associated with MODS and death in patients with acute pancreatitis and is an early indicator of the likely severity of acute pancreatitis.

Refining the role of laparoscopy and laparoscopic ultrasound in the staging of presumed pancreatic head and ampullary tumours.

Laparoscopy and laparoscopic ultrasound have been validated previously as staging tools for pancreatic cancer. The aim of this study was to identify if assessment of vascular involvement with abdominal computed tomography (CT) would allow refinement of the selection criteria for laparoscopy and laparoscopic ultrasound (LUS). The details of patients staged with LUS and abdominal CT were obtained from the unit's pancreatic cancer database. A CT grade (O, A-F) of vascular involvement was recorded by a single radiologist. Of 152 patients, who underwent a LUS, 56 (37%) had unresectable disease. Three of 26 (12%) patients with CT grade O, 27 of 88 (31%) patients with CT grade A to D, 17 of 29 (59%) patients with CT grade E and all nine patients with CT grade F were found to have unresectable disease. In all, 24% of patients with tumours <3 cm were found to have unresectable disease. In those patients with tumours considered unresectable, local vascular involvement was found in 56% of patients and vascular involvement with metastatic disease in 17%, while 20% of patients had liver metastases alone and 5% had isolated peritoneal metastases. The remaining patient was deemed unfit for resection. Selective use of laparoscopic ultrasound is indicated in the staging of periampullary tumours with CT grades A to D.

Early specialist repair of biliary injury.

BACKGROUND: Considerable debate surrounds the timing of repair of injury to the common bile duct following cholecystectomy. In the absence of sepsis or significant peritoneal soiling, repair within the first week may be optimal. This study compared the outcome of early (within the first 2 weeks) and delayed (between 2 weeks and 6 months) repair. METHODS: Data on all patients referred with biliary injuries were recorded prospectively. In the absence of sepsis or significant peritoneal soiling, repair was considered within 2 weeks. RESULTS: Between November 1988 and November 2003, 123 patients were referred with injury to the biliary tree. Repair of the injury had been attempted in 55 patients (44.7 per cent) before referral. Of the 68 patients with no previous repair, nine were managed without surgery and 59 required subsequent surgical reconstruction of the ductal injury. Within the first 2 weeks after injury, 22 patients underwent primary biliary repair and three had revision of a failed biliary repair. Between 2 weeks and 6 months, a further 22 injuries were repaired. Successful repair was possible in 22 of 25 early repairs compared with 20 of 22 delayed repairs (P = 0.615). The overall operative mortality rate for patients undergoing repair was 4 per cent (two of 47 patients). CONCLUSION: A successful outcome was achieved in a high proportion of patients (42 of 47) when repair of the bile duct injury was undertaken in a unit experienced in the management of biliary injury. In selected patients, early repair within the first 2 weeks resulted in a similar outcome to that of delayed repair.

Surgical palliation for unresectable hilar cholangiocarcinoma.

The majority of patients who present with hilar cholangiocarcinoma will have incurable disease and require only palliation. Efficient relief of disabling symptoms is required with minimal morbidity and mortality and can be achieved by either surgical or non-operative options. A review of the indications, anatomical considerations and surgical techniques is presented. Segment III cholangio-jejunostomy is the most frequently used surgical bypass procedure and in those patients with an expected survival of more than 6 months, surgical palliation offers good quality and long-lasting palliation. There is a need for randomized controlled data to define the optimal role of surgical palliation in this difficult disease.

Presentation, treatment and outcome in patients with ampullary tumours.

BACKGROUND: Ampullary tumours are relatively rare, and few large single-centre reports provide information on their treatment and outcome. The aim of this study was to analyse outcome and determine predictors of survival for patients with ampullary tumours treated in a specialist centre. METHODS: Over an 11-year period, 561 patients were treated for periampullary tumours, 88 of whom had a histologically proven ampullary neoplasm. Prospectively gathered data were analysed to assess predictors of survival. RESULTS: The overall resection rate was 92 per cent; there were no postoperative deaths. Median survival was 45.8 months for patients with resectable tumours and 8.0 months for those with irresectable disease (P < 0.001). On univariate analysis, age less than 70 years (P = 0.015) and a bilirubin level of 75 micromol/l or less (P = 0.012) favoured long-term survival. Among 70 patients who underwent cancer resection, factors associated with significantly worse long-term survival on univariate analysis included poorly differentiated tumour (P < 0.001), positive nodes (P < 0.001), perineural invasion (P = 0.001) and invasion of the pancreas (P = 0.018). Multivariate analysis identified positive nodes and bilirubin concentration as independent predictors of survival. CONCLUSION: An aggressive surgical approach to ampullary tumours is justified by the low proportion of benign lesions, the absence of postoperative mortality and improved long-term survival.

The extent of resection influences outcome following hepatectomy for colorectal liver metastases.

BACKGROUND: The acceptable indications for liver resection in patients with colorectal metastases have increased significantly in the last decade. It is thus becoming more difficult to ascertain the limitations for selection as the boundaries have been greatly extended. This has resulted in not only more extensive resections, but more atypical and bilobar resections. The aim of this study was to compare the outcome of patients undergoing different extent of liver resection in a specialist unit. METHODS: All patients undergoing liver resection for colorectal metastases at the Royal Infirmary of Edinburgh between October 1988 and April 2001 were reviewed. Patients were allocated into one of three groups: standard group, extended group, and segmental group. Patient information was collected from a prospectively completed database. RESULTS: One hundred and thirty-seven patients had liver resections for colorectal metastases during the study period. There were 69 standard hepatectomies, 41 extended resections and 27 segmental resections. CEA level was significantly lower in the segmental group(p = 0.012). There was a significant difference between the groups in terms of median operating time (p < 0.0001, Kruskal-Wallis test), operative blood loss (p = 0.006, Kruskal-Wallis test) and post-operative stay ( p = 0.036, Kruskal-Wallis test). Major post-operative complications were similar between standard and extended resections but less following segmental resection (p = 0.050. Predicted median survival was 51 months following standard resection, 23 months following extended resection and 59 months after segmental resection ( p = 0.037, log rank test), however, there was no difference between the three groups for actual 5-year survival (p = 0.662, Pearson chi-square test). CONCLUSION: Morbidity and mortality rates were comparable with other previous studies as was overall survival, although survival in patients undergoing extended resections was reduced. There was an acceptable level of morbidity and mortality for all three groups. Patients undergoing segmental resection had fewer complications, shorter length of stay, and the longest median survival suggesting adequate oncological clearance. Segmental resection has a role for favourably placed tumour deposits if oncological clearance can be ensured. Extended liver resections have a role for selected patients with bilobar colorectal metastases or large solitary deposits close to the hepatic vein confluence.