Krukenberg tumours of colorectal origin: Experience of a tertiary referral centre and review of the literature

Aim: Krukenberg tumours are tumours of the gastrointestinal tract that metastasize to the ovary. The condition is uncommon and accounts for 5 per cent of ovarian tumours. It is our objective to describe the outcome of patients after resection of Krukenberg tumours of colorectal origin. Patients and Methods: The present study is a retrospective review of 20 patients with resection performed for Krukenberg tumours of colorectal origin from November 1996 to April 2010 at Queen Elizabeth Hospital, Hong Kong, China. Results: The most common colonic primary site was sigmoid colon (40 per cent). Thirteen patients (65 per cent) had T4 tumours. Ten patients (50 per cent) had synchronous tumours. Seven patients (35 per cent) had bilateral ovarian involvement. Nine patients (45 per cent) had elevated serum carcinoembryonic antigen levels. The median carbohydrate antigen 125 (was 57 U/mL (range: 12–850 U/mL). Seven patients (35 per cent) developed metastases after ovarian resection. The most common sites were intra‐abdominal lymph nodes (15 per cent), bone (15 per cent) and liver (10 per cent). The overall 3‐ and 5‐year survival rates after ovarian resection were 40 per cent and 25 per cent, respectively. Right colon cancer (30 per cent right colon vs 60 per cent left colon vs 10 per cent rectum, P = 0.021) and T4 staging of the colonic primary (30 per cent T3 vs 65 per cent T4, P = 0.033) were found to be poor prognostic factors for survival. Conclusion: Although recurrence after resection of Krukenberg tumours is common, bilateral salpingo‐oophorectomy should still be considered. A more aggressive approach, such as debulking surgery or metastasectomy, is also recommended to improve the outcome of these patients.     

Outcome after emergency surgery for cancer of the large intestine

The data for 77 patients with colorectal cancer who underwent emergency surgery for acute intestinal obstruction (57 patients) or perforation (20 patients) within 24 h of admission were evaluated. The patients were older and had more advanced disease than patients undergoing elective surgery for colorectal cancer. Emergency surgery for carcinoma of the right colon consisted of primary resection in 95 per cent of cases and was followed by a 28 per cent mortality rate. Perforated tumours of the left colon and rectum were managed by primary resection in 82 per cent of cases with a 22 per cent mortality rate. In contrast, obstructing tumours of the left colon and rectum were treated by primary resection in 38 per cent of cases with a 6 per cent mortality rate, and by primary decompression in 62 per cent of cases with a 25 per cent mortality rate. The overall postoperative mortality rate was 23 per cent and increased with advanced tumour disease, perforation and peritonitis. Cardiac decompensation and intraabdominal sepsis were the major causes of death. Although the long-term survival rate following emergency surgery was worse than after elective surgery, improvements in outcome should be achieved by better management of the initial emergency situation.     

Survival benefit of high ligation of the inferior mesenteric artery in sigmoid colon or rectal cancer surgery

BACKGROUND: The aim of this study was to assess the impact of inferior mesenteric artery (IMA) root nodal dissection before high ligation of the artery on survival in patients with sigmoid colon or rectal cancer. METHODS: Data on 1188 consecutive patients who underwent resection for sigmoid colon or rectal cancer, with high ligation of the IMA, were identified from a prospective database (April 1965 to December 1999). Survival of patients with involvement of nodes along the IMA proximal to the origin of the left colic artery (root nodes, station 253) through the bifurcation of the superior rectal artery (trunk nodes, station 252) was determined. RESULTS: Twenty patients (1.7 per cent) had metastatic involvement of station 253 lymph nodes and 99 (8.3 per cent) had metastases to station 252. The 5- and 10-year survival rates of patients with metastases to station 253 were 40 and 21 per cent, and those for patients with metastases to station 252 were 50 and 35 per cent, respectively. CONCLUSION: High ligation of the IMA allows curative resection and long-term survival in patients with cancer of the sigmoid colon or rectum and nodal metastases at the origin of the IMA.     

A population-based study of the incidence, management and prognosis of hepatic metastases from colorectal cancer

BACKGROUND: The aim of this population-based study was to evaluate the incidence, management and prognosis of patients with hepatic metastases related to colorectal cancer using data from the Digestive Cancer Registry of Calvados, France. METHODS: Of 1325 patients with colorectal cancer registered between January 1994 and December 1999, 358 developed hepatic metastases. Logistic regression was used to analyse prognostic factors. Survival analysis was carried out with Cox's proportional hazards model. RESULTS: Some 18.8 per cent of patients had synchronous metastases, while 29.3 per cent developed metastases at 3 years. Of patients with hepatic metastases, 17.3 per cent had a surgical resection, 40.2 per cent were treated with palliative chemotherapy and 42.5 per cent had symptomatic treatment. Factors associated with receiving symptomatic treatment only were age over 75 years and more than one metastasis, but not place of treatment. Median survival after a diagnosis of hepatic metastases was 10.7 (range 4.6-23.1) months. Significant adverse prognostic factors were: age over 75 years (P = 0.001), lymph node invasion of primary tumour (P = 0.024), bilateral distribution of metastases (P = 0.001), other metastases (P = 0.004) and symptomatic treatment only (P = 0.041). CONCLUSION: Despite improvement in treatment for hepatic metastases, age and extent of disease remain limiting factors for surgical resection and palliative chemotherapy.     

Neoadjuvant chemotherapy and resection of advanced synchronous liver metastases before treatment of the colorectal primary

BACKGROUND: In many patients with advanced synchronous liver metastases from colorectal tumours, the metastases progress during treatment of the primary, precluding curative treatment. The authors have investigated a management strategy that involves high-impact chemotherapy first, resection of liver metastases second and finally removal of the primary tumour in patients with adverse prognostic factors. METHODS: Twenty consecutive patients with non-obstructive colonic (nine patients) or rectal (11 patients) cancer and advanced synchronous liver metastases were treated according to this strategy. Median age was 56 years. Patients received between two and six cycles of 5-fluorouracil, oxaliplatin and irinotecan-based chemotherapy. Data were collected prospectively. RESULTS: Overall survival rates at 1, 2, 3 and 4 years after the start of treatment were 85, 79, 71 and 56 per cent respectively, with a median survival of 46 months. Sixteen of the 20 patients had complete removal of liver metastases and colorectal tumours (resectability rate 80 per cent). CONCLUSION: This new strategy produced resectability and survival rates better than those expected from the published data on patients with disease of similar severity. It allows initial control and downstaging of liver metastases, and delivery of preoperative radiotherapy for rectal cancer without the fear that liver metastases will meanwhile progress beyond the possibility of cure.     

Prognostic factors in resection of pulmonary metastasis from colorectal cancer

BACKGROUND: The aim of this retrospective study was to evaluate characteristics of primary colorectal cancer and pulmonary metastases in order to identify prognostic factors for overall survival and risk factors for further intrapulmonary recurrence after resection of pulmonary metastases from colorectal cancer. METHODS: Forty-nine patients who underwent resection of pulmonary metastases from colorectal cancer were reviewed. The factors assessed were age, sex, pathological findings of the original colorectal cancer (depth, lymphatic invasion, venous invasion, lymph node metastasis, differentiation, Dukes' stage) and pulmonary metastasis (maximum tumour size, number of tumours, completeness of resection), serum carcinoembryonic antigen level, previous hepatectomy for liver metastases, and surgical procedure for resection of pulmonary metastasis. Overall survival and intrapulmonary recurrence were also reviewed. RESULTS: Survival rates after resection of pulmonary metastases were 78 per cent at 3 years and 56 per cent at 5 years. Solitary pulmonary metastases were significantly correlated with survival (P = 0.049). The pathological features of the primary colorectal cancer had no impact on survival. Histologically incomplete resection of pulmonary metastasis significantly correlated with pulmonary re-recurrence (P = 0.034). CONCLUSION: Long-term survival can be expected after complete resection of pulmonary metastases arising from colorectal cancer, especially in patients with a solitary pulmonary metastasis.     

Relationship between DNA ploidy and survival in patients with primary breast cancer

The DNA ploidy of breast cancer tissue from paraffin blocks was measured by flow cytometry in 117 patients whose disease had been detected and treated with surgery between 1974 and 1976. Patients with aneuploid tumours had positive axillary nodes and distant metastases more often than those with diploid tumours. Aneuploid tumours were more common in postmenopausal than premenopausal women. The S-phase fraction (SPF) was significantly higher in aneuploid than in diploid tumours and positive axillary lymph nodes were found in 26 per cent of the patients who had a tumour with a SPF below the median (4.8 per cent) and in 48 per cent of those with tumours with SPF values above the median. At the primary clinical investigation 2 per cent of the patients with diploid tumours and 6 per cent of those with aneuploid tumours had distant metastases. During the follow-up, the proportion of patients with distant metastases increased to 42 and 72 per cent, respectively. With a follow-up of 11.5 years, the DNA aneuploidy of the tumour showed a significant association with decreased survival. Thirty-three per cent of patients with diploid and 65 per cent of patients with aneuploid tumours had died from breast cancer during the follow-up (P less than 0.001). All patients with hypertetraploid or multiploid tumours died from breast cancer. High SPF values were associated more closely with distant metastases or death during the follow-up than low SPF values. Our results suggest that DNA ploidy measured by flow cytometry from paraffin embedded tissue blocks of human breast cancer can be used to predict the aggressiveness of the tumour and the survival of the patients.     

DNA ploidy pattern and tumour spread in gastric cancer

The DNA ploidy pattern of gastric cancer was studied in 58 patients to investigate the heterogeneity between primary tumour and metastases. In both primary tumours and lymph node metastases, diploid patterns accounted for 33 per cent, whereas all liver metastases were aneuploid. The percentage of polyploid cells was higher in the liver metastases than in primary tumours and lymph node metastases. When the heterogeneity of DNA ploidy pattern between primary tumour and metastasis was evaluated, diploid tumours had a significantly lower rate of lymph node metastasis heterogeneity than aneuploid tumours. When the DNA ploidy pattern and survival were evaluated, the patients who had a diploid pattern in both primary tumour and metastasis had a significantly higher survival rate than the patients who had an aneuploid pattern in the primary tumour and metastasis (57 per cent versus 26 per cent at 5 years). These data suggest that cell heterogeneity is a common phenomenon in gastric cancer, and this may be important in the evolution of the disease. Furthermore, the role of the DNA ploidy pattern as a prognostic factor is emphasized.     

Population-based study of short- and long-term survival from colorectal cancer in Denmark, 1977-1999

BACKGROUND: Survival rates for patients with colorectal cancer have been lower in Denmark than in other European countries. The aim of this study was to examine temporal trends in relative survival from colorectal cancer between 1977 and 1999. METHODS: All patients diagnosed with colorectal cancer between 1977 and 1999 were identified using the nationwide population-based Danish Cancer Registry. Patients were linked with the Danish Central Population Registry to obtain data on survival to December 2002, and to select ten population controls per patient. RESULTS: A total of 69 562 patients with colorectal cancer were identified, of whom 49.2 per cent were men. Six-month relative survival after diagnosis increased from 69.7 per cent in 1977-1982 to 77.7 per cent in 1995-1999. Five-year relative survival rates increased from 37.8 to 46.8 per cent respectively. Women had slightly higher 5-year relative survival than men throughout the study period. Rectal tumours were associated with better survival than colonic tumours until 2 years after diagnosis, after which tumour location had no impact on survival. CONCLUSION: Relative survival of patients with colorectal cancer has improved in Denmark between 1977 and 1999, most probably reflecting better management of the disease.     

Extrahepatic disease does not contraindicate hepatectomy for colorectal liver metastases

BACKGROUND: Extrahepatic disease has always been considered an absolute contraindication to hepatectomy for liver metastases. The present study reports the long-term outcome and prognostic factors of patients undergoing resection of extrahepatic disease simultaneously with hepatectomy for liver metastases. METHODS: From January 1987 to January 2001, 111 (30 per cent) of 376 patients who had hepatectomy for colorectal liver metastases underwent simultaneous resection of extrahepatic disease with curative intent. RESULTS: Surgery was considered R0 in 77 patients (69 per cent) and palliative (R1 or R2) in 34 patients (31 per cent). The mortality rate was 4 per cent and the morbidity rate 28 per cent. After a median follow-up of 4.9 years, the overall 3- and 5-year survival rates were 38 and 20 per cent respectively. The 5-year overall survival rate of patients with R0 resection only (n = 75) was 29 per cent. The difference in survival between patients with and without extrahepatic disease discovered incidentally at operation was significant, as was the number of liver metastases. CONCLUSION: Extrahepatic disease in patients with colorectal cancer who also have liver metastases should no longer be considered an absolute contraindication to hepatectomy. However, the presence of more than five liver metastases and the incidental intraoperative discovery of extrahepatic disease remain contraindications to hepatic resection.     

Resection rate and effect of postoperative chemotherapy on survival after surgery for colorectal liver metastases.

BACKGROUND: The purpose of this study was to investigate whether adjuvant therapy can improve survival after curative resection of colorectal liver metastases. METHODS: Some 235 patients had 256 liver resections for metastatic colorectal cancer. There were no predefined criteria for resectability with regard either to the number or size of the tumours or to locoregional invasion, except that resection had potentially to be complete and macroscopically curative. All patients who had curative hepatic resection were advised to start postoperative adjuvant chemotherapy. RESULTS: The resectability rate in screened patients was 91 per cent (235 of 259 patients); the postoperative mortality rate was 4 per cent. In 35 patients resection of the primary tumour was performed simultaneously with partial liver resection. Forty-four patients (19 per cent) developed intra-abdominal recurrence; 14 (6 per cent) underwent reoperation and the recurrent tumour was resected. Adjuvant chemotherapy was given to 99 patients (55 per cent), most treatments being based on 5-fluorouracil with folinic acid. The overall actuarial survival rates at 1, 3 and 5 years were 87, 60 and 36 per cent respectively. In a multivariate analysis, four or more metastases, preoperative carcinoembryonic antigen level higher than 5 ng/ml and a positive resection margin were independent predictors of poor outcome. Adjuvant chemotherapy improved the 5-year survival rate to 53 per cent. CONCLUSION: This study provides some evidence that postoperative chemotherapy is beneficial; however, prospective randomized studies are necessary to define its exact role.     

Peritoneal carcinomatosis from colorectal cancer

BACKGROUND: Aggressive therapeutic regimens have been advocated for the treatment of peritoneal carcinomatosis from colorectal cancer. It is essential to understand the clinical and histological features that govern the natural history of this condition if the efficacies of novel therapeutic approaches are to be assessed adequately. METHODS: A database of 3019 colorectal cancers was used to identify patients with synchronous peritoneal carcinomatosis, patients who developed metachronous peritoneal carcinomatosis, and those without carcinomatosis. Clinical, histological and survival data for the groups were collated and subjected to statistical analysis. RESULTS: Some 349 patients (13 per cent) with peritoneal carcinomatosis were identified; 214 had synchronous disease and 135 had metachronous carcinomatosis. Some 125 patients (58 per cent) in the synchronous group were free of systemic metastases; 80 of these patients had localized disease. Liver metastases, tumour (T) stage, nodal stage, and venous and perineural invasion were independent predictors of metachronous carcinomatosis. The median survival of patients with synchronous disease was 7 months; survival was adversely affected by the extent of peritoneal carcinomatosis and the T stage of the primary cancer. CONCLUSION: Peritoneal carcinomatosis is a common mode of disease progression in patients with colorectal cancer. For the majority of patients the prognosis is poor, but a small number with localized disease may be suitable for further aggressive therapy.     

Outcome of colorectal cancer

The outcome of 454 patients who presented with colorectal carcinoma during a 16 year period is reviewed: 54 per cent were males, 58 per cent were aged more than 60 and 10 per cent had an emergency admission, 42 per cent of tumours occurred in the rectum. A curative resection was possible in 68 per cent. Postoperative mortality was 7 per cent. The overall crude 5-year survival was 41 per cent. The mortality from local recurrence was significantly higher in rectal (11.7 per cent) than in colonic cancer (8.8 per cent; P less than 0.01). The rate of recurrence and metastases was higher in patients with low rectal cancer than in patients with cancer of the middle and the upper rectum (P less than 0.01). Distant metastases were the cause of death in 94 per cent of the patients who had a Miles' operation for cancer of the middle rectum, whereas local recurrence was responsible for late mortality in 80 per cent of patients who underwent an anterior resection. No difference in 5-year survival was found in the restorative and in the excisional group.     

Biopsy of resectable colorectal liver metastases causes tumour dissemination and adversely affects survival after liver resection

BACKGROUND: Liver resection is increasingly being performed for metastatic colorectal cancer. This study assessed the need for preoperative biopsy of suspected metastases and whether biopsy has any effect on long-term survival. METHODS: Prospectively collected data on patients who underwent liver resection for colorectal metastases between 1986 and 2003 were reviewed retrospectively. The endpoints of morbidity, operative mortality and long-term survival were compared between patients who had biopsy before referral (group 1) and those who did not (group 2). RESULTS: Patient demographics and disease distribution were similar for 90 patients in group 1 and 508 in group 2. Seventeen patients (19 per cent) who had undergone biopsy either at the time of colorectal resection or radiologically had evidence of needle-track deposits. Operative mortality and morbidity rates in the two groups were similar. The 4-year survival rate after liver resection was 32.5 (s.e. 5.5) per cent in group 1, compared with 46.7 (2.8) per cent in group 2 (P = 0.008). CONCLUSION: Needle-track deposits are common after biopsy of suspected colorectal liver metastases. Biopsy of metastases confers poorer long-term survival on patients after liver resection and cannot be justified in patients with potentially resectable disease.     

Multiple primary carcinomas of the colon and associated extracolonic primary malignant tumors.

A review of 308 cases of colorectal carcinoma showed 12 cases (3.9%) of colorectal multiple primary malignant tumors and 14 cases (4.5%) of colorectal primary malignant tumor associated with extracolonic primary malignant tumor. A total of 60 tumors was involved: 14 each in the sigmoid and transverse colon; eight in the cecum; six in the rectum; one each in the ascending and descending colons; three each in the bladder and prostate; two each in the breast, cervix, and lung; and one each in the skin, nasopharynx, kidney, and endometrium. Three of the four patients with multifocal cancer were found to have in situ cancer, and 12.7% of concomitant adenomas had malignant changes. Seven patients had regional lymph node metastases. Of the six patients with rectocolonic multiple tumors who had adequate follow-up data, survival in three ranged from 19.5 to 60 months; three were alive and well 15 to 51 months after resection. Survival of patients with primary rectocolonic and extracolonic tumors was similar. From these data and the pathologic study of the resected specimens, we recommend for any patient with cancer of the rectocolon that the entire colon be searched for multiple primary malignant tumors, that extracolonic tumors be considered second primary tumors unless proved to be metastases, and that follow-up be long and include frequent rectocolonic examinations for a second primary tumor.     

Survival after hepatic resection for malignant tumours.

A retrospective analysis of 194 patients who underwent hepatic resection for primary or metastatic malignant disease from January 1962 to December 1988 was undertaken to determine variables that might aid the selection of patients for hepatic resection. Hepatic metastases were the indication for resection in 126 patients. The 5-year survival rate was 17 per cent. For patients with resected metastases from colorectal cancer (n = 104), 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 < or = 5 mm (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 for 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 for patients with well or moderately differentiated tumours (P not significant). The site and Dukes' classification of the primary tumour, the sex and preoperative carcinoembryonic antigen level of the patient, and the number and size of hepatic metastases did not affect the prognosis. The 5-year survival rate for patients with hepatocellular carcinoma (n = 42) was 25 per cent. An improved survival rate was found for patients whose alpha-fetoprotein 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, extrahepatic 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 11 patients with hepatic sarcoma was 25 per cent. No patient with peripheral cholangiocarcinoma survived to 1 year in contrast to patients with hilar cholangiocarcinoma, all four of whom survived for more than 14 months.     

Clinicians' attitudes towards management of metastatic colorectal adenocarcinoma

Background: Despite being a common problem, the optimal management of in situ primary colorectal carcinomas in patients presenting with unresectable synchronous metastases remains unknown. To date, no prospective randomized studies have been conducted to evaluate the outcomes of different approaches to management. We studied the attitudes of clinicians involved in the management of this group of patients with a view to determine their treatment preferences and assess the feasibility of conducting a randomized study addressing the role of resection of primary colorectal carcinomas in patients presenting with unresectable synchronous metastases. Methods: A survey of Australian colorectal surgeons, Australian medical oncologists and Victorian rural general surgeons was conducted. Results: The results indicated that with regard to preferred treatment (i) for patients with asymptomatic sigmoid or caecal primary tumours, surgeons preferred surgery followed by chemotherapy, whereas oncologists preferred chemotherapy alone; (ii) for patients with symptomatic sigmoid or caecal primary tumours, both surgeons and oncologists preferred surgery followed by chemotherapy; and (iii) for patients with metastatic rectal carcinoma, whether asymptomatic or symptomatic, both surgeons and oncologists preferred a multimodality approach to treatment. Clinicians were accepting a broad range of treatment options for patients with both asymptomatic and symptomatic primary colorectal tumours. Conclusion: There is a high level of acceptability among Australian clinicians for both surgical and non‐surgical approaches to management of the in situ primary colorectal tumour in patients with unresectable synchronous metastases. Further research is warranted to determine the management strategy that will yield the best outcome for these patients.     

Percutaneous fine-needle aspiration cytology in the diagnosis and management of liver tumours

BACKGROUND: The aim of the study was to investigate the value of percutaneous fine-needle aspiration cytology (FNAC) in the diagnosis and management of liver tumours. METHODS: FNAC followed by histopathological examination was carried out in 216 patients with suspected liver tumours. The final diagnosis was primary liver cancer in 106, colorectal metastases in 51, non-colorectal metastases in 46, benign tumour in nine and no tumour in four patients. RESULTS: Cytology resulted in correct classification of the lesion as benign or malignant in 87 per cent of patients, correct discrimination between primary and secondary malignancy in half of the patients, and a correct diagnosis of tumour type in one-third of patients. The tumour was erroneously classified as benign or malignant in 22 patients (11 per cent) and four patients (2 per cent) respectively. When FNAC showed malignancy, the predictive value was 98 per cent, whereas the predictive value was 27 per cent when it did not. FNAC guided investigations and treatment in one-quarter of patients. Implantation metastases were recorded in seven patients (3 per cent), including five (10 per cent) of 51 patients with colorectal liver metastases, and caused major local problems and death in four patients. CONCLUSION: FNAC was valuable in about a quarter of patients with liver tumour. The risks of implantation metastases and a false-negative finding do not justify its use in candidates for curative therapy of liver tumours.     

Cancer of the large bowel in a defined population: Canterbury, New Zealand, 1970--4.

A total of 1024 new cases of cancer of the large bowel occurred in Canterbury, New Zealand (population 400,796), in the 5 years 1970--4. Of these, 992 were diagnosed before death and are reviewed in this paper. The high incidence of this disease in New Zealanders of European origin is illustrated. A significant difference in site distribution of primary tumours between the sexes was found, with a female preponderance of cancer of the proximal colon gradually changing to a male preponderance of cancer of the rectum. In all, 61.5 per cent of the patients had lymph node metastases or advanced disease at the time of diagnosis or treatment. Largely as a consequence of this, only 65 per cent were able to have potentially curative treatment. The estimated crude 5-year survival rate of the whole group was 32.7 per cent (relative rate 42.8 per cent) and the crude 5-year survival rate after potentially curative surgery was 48.4 per cent (relative rate 62.4 per cent). The results are compared with those of other authors. They emphasize the generally unsatisfactory outcome of treatment.     

Determinants of survival in liver resection for colorectal secondaries

All 72 resections for colorectal liver secondaries during the period 1971-1984 were analysed retrospectively. Liver tumours were single in 35 (49 per cent), unilateral in 55 (76 per cent) and associated with extrahepatic disease in 12 (18 per cent) patients. Operative mortality was 5.6 per cent. With respect to the disease in the liver, the presence or absence of four or more metastases was the predominant prognostic determinant with a 5 year survival rate of 20 per cent in patients with less than four liver tumours, and no 3 year survivor among patients with four or more tumours. When the number of liver tumours was less than four, the prognosis in patients with unilateral disease was not significantly better than in patients with bilateral disease (P = 0.19). No other liver disease variable seemed to play any role in the prognosis. Extrahepatic disease was associated with a poor prognosis and no 5 year survivor. The length of the tumour-free resection margin was the only treatment variable that varied with the outcome: a resection margin of less than 10 mm was followed by a poor survival. Variables that did not influence survival included uni- or bilateral disease, liver tumour volume, tumour size, type of liver resection, Dukes' classification, differentiation of the primary tumour and synchronous or metachronous disease. It is concluded that resection for liver colorectal secondaries is indicated when there are less than four liver tumours, even if bilateral, no extrahepatic disease is present, and a resection margin of at least 10 mm can be obtained. It should not be performed unless all of these requirements are met.