The effect of a primary tumour resection on the progression of synchronous colorectal liver metastases: an exploratory study.

AIM: The objective of this study was to determine the effect of an upfront primary tumour resection on the progression of synchronous colorectal liver metastases. MATERIALS AND METHODS: Patients with synchronous colorectal liver metastases referred between 2005 and 2010 were identified. Patients were analysed according to the following two groups: 1) an upfront primary tumour resection and 2) neo-adjuvant chemotherapy. A univariate and multivariate analysis was performed to identify factors significantly contributing to progressive disease. Cox regression analysis was undertaken to determine the effect of management on overall survival (OS) and time to tumour progression (TTP). RESULTS: A total of 116 patients with synchronous colorectal liver metastases were identified of which 49 patients received an upfront primary tumour resection and 67 received neo-adjuvant chemotherapy. Liver resections were performed in 18 (36.7%) and 14 (20.9%) of the patients in the upfront and neo-adjuvant groups respectively (P 0.06). On multivariate analysis, an upfront primary tumour resection significantly affected progressive disease (p < 0.001, OR 5.67; 95% CI 2.71-11.79). An upfront tumour resection was not a significant predictor of overall survival (P = 0.83; HR 1.10; 95% CI 0.48-2.52). CONCLUSION: Our findings suggest that an upfront primary tumour resection in patients with synchronous colorectal liver metastases results in progressive disease. These preliminary findings need to be validated in a future multi-centre independent study.

Bevacizumab plus mFOLFOX-6 or FOLFOXIRI in patients with initially unresectable liver metastases from colorectal cancer: the OLIVIA multinational randomised phase II trial.

BACKGROUND: For patients with initially unresectable liver metastases from colorectal cancer, chemotherapy can downsize metastases and facilitate secondary resection. We assessed the efficacy of bevacizumab plus modified FOLFOX-6 (5-fluorouracil/folinic acid, oxaliplatin) or FOLFOXIRI (5-fluorouracil/folinic acid, oxaliplatin, irinotecan) in this setting. PATIENTS AND METHODS: OLIVIA was a multinational open-label phase II study conducted at 16 centres in Austria, France, Spain, and the UK. Patients with unresectable liver metastases were randomised to bevacizumab (5 mg/kg) plus mFOLFOX-6 [oxaliplatin 85 mg/m(2), folinic acid 400 mg/m(2), 5-fluorouracil 400 mg/m(2) (bolus) then 2400 mg/m(2) (46-h infusion)] or FOLFOXIRI [oxaliplatin 85 mg/m(2), irinotecan 165 mg/m(2), folinic acid 200 mg/m(2), 5-fluorouracil 3200 mg/m(2) (46-h infusion)] every 2 weeks. Unresectability was defined as ≥1 of the following criteria: no possibility of upfront R0/R1 resection of all lesions; <30% residual liver volume after resection; metastases in contact with major vessels of the remnant liver. Resectability was evaluated by multidisciplinary review. The primary end point was overall resection rate (R0/R1/R2). Efficacy end points were analysed by intention-to-treat analysis. RESULTS: In patients assigned to bevacizumab-FOLFOXIRI (n = 41) or bevacizumab-mFOLFOX-6 (n = 39), the overall resection rate was 61% [95% confidence interval (CI) 45% to 76%] and 49% (95% CI 32% to 65%), respectively (difference 12%; 95% CI -11% to 36%). R0 resection rates were 49% and 23%, respectively. Overall tumour response rates were 81% (95% CI 65% to 91%) with bevacizumab-FOLFOXIRI and 62% (95% CI 45% to 77%) with bevacizumab-mFOLFOX-6. Median progression-free survival (PFS) was 18·6 (95% CI 12.9-22.3) months and 11·5 (95% CI 9.6-13.6) months, respectively. The most common grade 3-5 adverse events were neutropenia (bevacizumab-FOLFOXIRI, 50%; bevacizumab-mFOLFOX-6, 35%) and diarrhoea (30% and 14%, respectively). CONCLUSIONS: Bevacizumab-FOLFOXIRI was associated with higher response and resection rates and prolonged PFS versus bevacizumab-mFOLFOX-6 in patients with initially unresectable liver metastases from colorectal cancer. Toxicity was increased but manageable with bevacizumab-FOLFOXIRI. CLINICALTRIALSGOV: NCT00778102.

1000 Port-A-Cath ® placements by subclavian vein approach: single surgeon experience.

INTRODUCTION: Totally implantable venous access ports are widely used for the administration of chemotherapy in patients with cancer. Although there are several approaches to implantation, here we describe Port-A-Cath(®) (PAC) placement by percutaneous puncture of the subclavian vein with ultrasonographic guidance. PATIENTS AND METHODS: Data on our vascular access service were collected prospectively from June 2004. This service included port-a-caths and Hickman lines. Once 1000 consecutive port-a-caths(®) had been reached the study was closed and data analysed for the port-a-caths(®) alone. The left subclavian vein was the preferred site for venous access, with the right subclavian and jugular veins being the alternative choices if the initial approach failed. Patients were followed up in the short-term, and all the procedures were carried out by a single surgeon at each one of two institutions. RESULTS: Venous access by PAC was established in 100 % of the 1,000 cases. Of the 952 patients where the left subclavian vein was chosen for the first attempt of puncture, the success rate of PAC placement was 95 % (n = 904). Pneumothorax occurred in 12 patients (1.2 %), and a wound haematoma occurred in 4 (0.4 %) out of the total 1,000 patients. No infections were recorded during the immediate post-operative period but only in the long-term post-operative use with 8 patients requiring removal of the PAC due to infection following administration of chemotherapy. CONCLUSION: This is a very large series of PAC placement with an ultrasound-guided approach for left subclavian vein and X-ray confirmation, performed by a single surgeon, demonstrating both the safety and effectiveness of the procedure.

Outcomes of simultaneous resections for patients with synchronous colorectal liver metastases.

INTRODUCTION: The aim of this study was to determine the outcomes associated with simultaneous resections compared to patients undergoing sequential resections for synchronous colorectal liver metastases. METHOD: Consecutive patients undergoing hepatic resections between 2000 and 2012 for synchronous colorectal liver metastases were identified from a prospectively maintained database. RESULTS: Of the 112 hepatic resections that were performed, 36 were simultaneous resections and 76 were sequential resections. There was no difference in disease severity: number of metastases (P 0.228), metastatic size (P 0.58), the primary tumour nodal status (P 0.283), CEA (P 0.387) or the presence of extra-hepatic metastases (P 1.0). Major hepatic resections were performed in 23 (64%) and 60 (79%) of patients in the simultaneous and sequential groups respectively (P 0.089). Intra-operatively no differences were found in blood loss (P 1.0), duration of surgery (P 0.284) or number of adverse events (P 1.0). There were no differences in post-operative complications (P 0.161) or post-operative mortality (P 0.241). The length of hospital stay was 14 (95% CI 12.0-18.0) and 18.5 (95% CI 16.0-23.0) days in the simultaneous and sequential groups respectively (P 0.03). The 3-year overall survival was 75% and 64% in the simultaneous and sequential groups respectively (P 0.379). The 3-year hepatic recurrence free survival was 61% and 46% in the simultaneous and sequential groups respectively (P 0.254). CONCLUSION: Simultaneous resections result in similar short-term and long-term outcomes as patients receiving sequential resections with comparable metastatic disease and are associated with a significant reduction in the length of stay.

Multidisciplinary care of gastrointestinal stromal tumour: a review and a proposal for a pre-treatment classification.

The introduction of receptor tyrosine kinase inhibitors (TKIs) has revolutionized the management of gastrointestinal stromal tumour (GIST). Strong evidence supports the use of imatinib as first-line treatment in metastatic or unresectable tumours and its efficacy in the post-operative adjuvant setting has been confirmed by phase III trials. There are a number of reports concerning the administration of imatinib in the pre-operative setting, however, the heterogeneity of the terminology used and the indications for pre-operative treatment make it difficult to determine the true value of pre-operative imatinib. Larger studies, or a phase III trial could be helpful but patient accrual and standardization of care could be difficult. We propose a pre-treatment classification of GIST in order to facilitate the comparison and collection of data from different institutions, and overcome the difficulties related to accrual. Moreover, in the current era of multidisciplinary treatment of GIST, an appropriate classification is mandatory to properly design clinical trials and plan stage-adapted treatment.

The management of rectal cancer with synchronous liver metastases: a modern surgical dilemma.

INTRODUCTION: One in ten patients with rectal cancer presents with synchronous colorectal liver metastases. We present an up-to-date review of the different surgical strategies available for rectal cancer patients with synchronous colorectal liver metastases. METHOD: A literature review of MEDLINE, Cochrane and Google scholar was performed. RESULTS: Twenty retrospective studies comparing staged versus simultaneous resections were found. Overall survival was similar for both approaches whilst the length of stay was decreased in simultaneous resections. Only two studies comparing the 'reverse' versus staged or simultaneous resections were found. The studies investigating resection versus non-resection for rectal primaries with unresectable liver metastases were limited. CONCLUSION: Simultaneous resections are a reasonable alternative to staged resections for either advanced rectal cancers with limited liver disease or early rectal cancers with extensive liver disease. Currently, staged resections are favoured over simultaneous resections in patients with locally advanced rectal cancers with extensive liver disease. There are too few studies to determine the safety of reverse resections in the context of locally advanced rectal cancers. A resection of the primary tumour or a non-surgical intervention can be justified in the management of the rectal cancer primary in the presence of unresectable liver metastases.

The tumour biology of synchronous and metachronous colorectal liver metastases: a systematic review.

Forty to fifty percent of colorectal cancer (CRC) patients develop colorectal liver metastases (CLM) that are either synchronous or metachronous in presentation. Clarifying whether there is a biological difference between the two groups of liver metastases or their primaries could have important clinical implications. A systematic review was performed using the following resources: MEDLINE from PubMed (1950 to present), Embase, Cochrane and the Web of Knowledge. Thirty-one articles met the inclusion criteria. The review demonstrated that the majority of studies found differences in molecular marker expression between colorectal liver metastases and their respective primaries in both the synchronous and metachronous groups. Studies investigating genetic aberrations demonstrated that the majority of changes in the primary tumour were 'maintained' in the colorectal liver metastases. A limited number of studies compared the primary tumours of the synchronous and metachronous groups and generally demonstrated no differences in marker expression. Although there were conflicting results, the colorectal liver metastases in the synchronous and metachronous groups demonstrated some differences in keeping with a more aggressive tumour subtype in the synchronous group. This review suggests that biological differences may exist between the liver metastases of the synchronous and metachronous groups. Whether there are biological differences between the primaries of the synchronous and metachronous groups remains undetermined due to the limited number of studies available. Future research is required to determine whether differences exist between the two groups and should include comparisons of the primary tumours.

A meta-analysis comparing simultaneous versus delayed resections in patients with synchronous colorectal liver metastases.

INTRODUCTION: The traditional surgical management for patients presenting with synchronous colorectal liver metastases (SCLM) has been a delayed resection. However, in some centres, there has been a shift in favour of 'simultaneous' resections. The aim of this study was to use a meta-analytical model to compare the short-term and long-term outcomes in patients with synchronous colorectal liver metastases (SCLM) undergoing simultaneous resections versus delayed resections. METHOD: Comparative studies published between 1991 and 2010 were included. Evaluated endpoints were intra-operative parameters, post-operative parameters, post-operative adverse events and survival. A random-effects meta-analytical model was used and sensitivity analysis performed to account for bias in patient selection. RESULTS: Twenty-four non-randomized studies were included, reporting on 3159 patients of which 1381 (43.7%) had simultaneous resections and 1778 (56.3%) had delayed resections. The bilobar distribution (P = 0.01), size of liver metastases (P < 0.001) and the proportion of major liver resections (P < 0.001) was found to be higher in the delayed resection group compared to the simultaneous resection group. There was no significant difference in operative blood loss (95% CI, -279.28, 22.53; P = 0.1) or duration of surgery (WMD -23.83, 95% CI, -85.04, 37.38; P = 0.45). Duration of hospital stay was significantly reduced in simultaneous resections by 5.6 days (95% CI: 2.4-8.9 days, P = 0.007) No significant differences in post-operative complications (36% vs 37%, P = 0.27), overall survival (HR 1.00, 95% CI 0.86-1.15, P = 0.96) or disease free survival (HR 0.85, 95% CI 0.71-1.02, P = 0.08) were found. Sensitivity analysis revealed that these findings were consistent for the duration of hospital stay, post-operative complications, overall survival and disease free survival. CONCLUSION: This study demonstrates that the selection criteria for patients undergoing simultaneous or delayed resections differs resulting in a discrepancy in the metastatic disease severity being compared between the two groups. The comparable intra-operative parameters, post-operative complications and survival found between the two groups suggest that delayed resections may result in better outcomes. Similarly, the reduced length of hospital stay in simultaneous resections may only be as a result of the reduced disease severity in this group. Simultaneous resections can only be recommended in patients with limited hepatic disease until prospective studies comparing similar disease burdens between the two resection groups are available.

Laparoscopic adrenalectomy for co-secreting aldosterone and cortisol adenomas.

There are few published data on aldosterone and cortisol co-secreting adrenal tumours. Failure to perform comprehensive preoperative endocrine investigations in patients with adrenal "incidentalomas" or in those thought to be secreting only one hormone may account for this. Clinically patients with such lesions may have evidence of hypertension and hypokalaemia with no features of cortisol excess. Preoperative diagnosis of such lesions with accurate endocrinological work up is essential to prevent adrenal insufficiency and haemodynamic crises following removal of such glands. We present a series of 4 patients with co-secreting tumours treated by laparoscopic adrenalectomy between September 2010 and March 2011. Our experience suggests that dual secretors are more common than originally thought. A high index of suspicion and adequate endocrine work up is paramount in diagnosing such tumours and in experienced hands, laparoscopic adrenalectomy with appropriate substitutive steroid cover is safe, feasible and curative for these functioning adrenal tumours.

An intra-patient placebo-controlled phase I trial to evaluate the safety and tolerability of intradermal IMM-101 in melanoma.

BACKGROUND: IMM-101 is a heat-killed innate and adaptive immune-activating mycobacterial product; a phase I study aimed to determine its safety and tolerability in individuals with melanoma. PATIENTS AND METHODS: An intra-patient placebo-controlled study evaluated the safety and tolerability of three doses, namely, 0.1 (1 mg/ml), 0.5 (5 mg/ml) and 1.0 mg (10 mg/ml) of IMM-101 in stage III or IV melanoma. Each dose was administered in ascending order to one of the three cohorts. RESULTS: Based on observations from patients administered the 0.1-mg dose, it was considered appropriate to proceed with dosing the patients in the 0.5-mg dose cohort and then the 1.0-mg cohort (n = 6 per cohort). Treatment-emergent adverse events that would be considered typical of a post-vaccination state (including joint pains/aches, headaches and influenza-like symptoms) occurred at all dose levels, along with injection site reactions. These were mainly mild in intensity, resolved in a matter of days and responded well to supportive care. During post-study follow-up, two clinical responses (15%) were observed in patients with stage IV disease. CONCLUSION: IMM-101 is safe and well tolerated and there is a rationale for studying IMM-101 at a nominal 1.0-mg dose to complement conventional cytotoxic therapy for patients with advanced cancer.

A multicentre study of capecitabine, oxaliplatin plus bevacizumab as perioperative treatment of patients with poor-risk colorectal liver-only metastases not selected for upfront resection.

BACKGROUND: Perioperative chemotherapy improves outcome in resectable colorectal liver-only metastasis (CLM). This study aimed to evaluate perioperative CAPOX (capecitabine-oxaliplatin) plus bevacizumab in patients with poor-risk CLM not selected for upfront resection. PATIENTS AND METHODS: Poor-risk CLM was defined as follows: more than four metastases, diameter >5 cm, R0 resection unlikely, inadequate viable liver function if undergoing upfront resection, inability to retain liver vascular supply, or synchronous colorectal primary presentation. Patients underwent baseline computed tomography, magnetic resonance imaging, and/or positron emission tomography (PET) for staging and received neoadjuvant CAPOX plus bevacizumab, with resectability assessed every four cycles. Primary end point was radiological objective response rate (ORR). RESULTS: Forty-six patients were recruited, of which 91% underwent PET to ensure metastases confined to liver. Following neoadjuvant CAPOX plus bevacizumab, the ORR was 78% (95% confidence interval 63% to 89%). This allowed 12 of 30 (40%) patients with initial nonsynchronous unresectable CLM to be converted to resectability. In addition, 10 of 15 (67%) patients with synchronous resectable CLM underwent liver resection, with four additional patients being observed alone due to excellent response to neoadjuvant therapy. No grade 3-4 perioperative complications were seen. CONCLUSION: Neoadjuvant CAPOX plus bevacizumab resulted in a high response rate for patients with CLMs with poor-risk features not selected for upfront resection and converted 40% of patients to resectability.

Defining patient outcomes in stage IV colorectal cancer: a prospective study with baseline stratification according to disease resectability status.

BACKGROUND: Stage IV colorectal cancer encompasses a broad patient population in which both curative and palliative management strategies may be used. In a phase II study primarily designed to assess the efficacy of capecitabine and oxaliplatin, we were able to prospectively examine the outcomes of patients with stage IV colorectal cancer according to the baseline resectability status. METHODS: At enrolment, patients were stratified into three subgroups according to the resectability of liver disease and treatment intent: palliative chemotherapy (subgroup A), conversion therapy (subgroup B) or neoadjuvant therapy (subgroup C). All patients received chemotherapy with capecitabine 2000 mg m(-2) on days 1-14 and oxaliplatin 130 mg m(-2) on day 1 repeated every 3 weeks. Imaging was repeated every four cycles where feasible liver resection was undertaken after four or eight cycles of chemotherapy. RESULTS: Of 128 enrolled patients, 74, 22 and 32 were stratified into subgroups A, B and C, respectively. Attempt at curative liver resection was undertaken in 10 (45%) patients in subgroup B and 19 (59%) in subgroup C. The median overall survival was 14.6, 24.5 and 52.9 months in subgroups A, B and C, respectively. For patients in subgroups B and C who underwent an attempt at curative resection, 3-year progression-free survival was 10% in subgroup B and 37% for subgroup C. CONCLUSIONS: This prospective study shows the wide variation in outcome according to baseline resectability status and highlights the potential clinical value of a modified staging system to distinguish between these patient subgroups.

Synchronous oesophagectomy and hepatic resection for metastatic oesophageal cancer: report of a case.

BACKGROUND: Oesophageal cancer with liver metastasis is rare and when diagnosed is usually advanced and surgical management is contraindicated.Method-Results: We report the case of a patient who presented with oesophageal cancer and liver metastasis. The patient received chemotherapy combined with RFA to liver tumour. Subsequently she was subjected to oesophagectomy and liver resection of segment 5 extended into segment 8. Patient underwent adjuvant chemotherapy post-operatively and remains disease-free until now, 29 months after operation. CONCLUSION: Oesophageal cancer with concomitant liver metastasis is a rare and lethal disease. Multimodal management including surgery may offer prolonged survival in highly selected patients.

Complete pathological response following down-staging chemoradiation in locally advanced pancreatic cancer: challenging the boundaries.

Pancreatic cancer is an aggressive malignancy, relatively resistant to chemotherapy and radiotherapy, which usually presents late. Disease specific mortality approaches unity despite advances in adjuvant therapy. We present the first reported case of complete pathological response following neoadjuvant therapy in a locally advanced pancreatic adenocarcinoma.

Refining the technique of hepatic parenchymal transection: combined saline-linked radiofrequency precoagulation and ultrasonic aspiration.

BACKGROUND/AIMS: With recognition of its benefits, there has been a trend towards minimizing blood loss during hepatic parenchymal transection but no one technique has been shown to be superior to another. We analyzed our experience with using a novel combined technique of saline-linked radiofrequency precoagulation and ultrasonic aspiration for hepatic parenchymal transection. METHODOLOGY: This combined technique was used in 12 patients for parenchymal transection for metastatic hepatic disease and data was collected prospectively. Total blood loss, bile leaks, parenchymal transection time, hepatic pedicle clamp requirement and 30-day mortality were used as outcome measures. RESULTS: Four minor and 8 major hepatic resections were performed in twelve patients of who two underwent a synchronous resection of the rectum. The median blood loss was 525 mL (IQR 312.5-1150) in these patients who had a median postoperative stay of 7 days (IQR 7-14). The median parenchymal transection time was 120 minutes (IQR 100-153.75). No patient required portal triad clamping at anytime and there was no mortality. CONCLUSIONS: Combined technique of saline-link radiofrequency ablation and ultrasonic aspiration appears to be comparable to other techniques and should be considered as an alternative.

Osteosarcoma metastasising to the duodenum and pancreas.

The incidence of metastatic osteosarcoma is increasing because of improved results following multi-agent chemotherapy and resection of the primary tumour. Metastases occur most commonly in the lungs, whereas bowel metastases are rare. We describe a 25-year-old female who presented with melaena six years after successful resection of an osteosarcoma of her right femur, and one year after resection of a solitary pulmonary metastasis. Imaging revealed a lesion arising within both the duodenum and the pancreas for which a Whipple's pancreatoduodenectomy was carried out, achieving complete resection. Histological examination confirmed the diagnosis of metastatic osteosarcoma. We believe this is only the second such case reported. At 11 months post-operatively she had no detectable disease. Although rare, osteosarcoma can metastasise to the intestine. The surgeon must be aware of this complication, and that bowel metastases are potentially resectable.

Pancreatic cancer in England and Wales 1975-2000: patterns and trends in incidence, survival and mortality.

BACKGROUND: Rates and time trends in mortality from pancreatic cancer vary considerably between countries. AIM: To examine trends and patterns in the incidence of, and the survival and mortality from, pancreatic cancer in England and Wales from 1975 to 2000; in particular, whether incidence and survival rates are related to socio-economic deprivation. METHODS: We calculated annual age-specific and overall age-standardized incidence and mortality rates by sex for pancreatic cancer in total, and by subsite. We also estimated survival by sex and age group and by subsite. RESULTS: In males, the age-standardized rate fluctuated in the late 1970s, to peak at 13.0 per 100,000 in 1979, declined steadily by an average of 1.3% per year to around 10.3 per 100,000 in the mid-1990s and then levelled off. For females, the rate peaked at 8.4 per 100,000 in the late 1980s before declining and fluctuating around 7.7 per 100,000 in the late 1990s. Patterns and trends in mortality rates were closely similar to those in incidence, due to the very low survival rates: only 2-3% at 5 years from diagnosis. Survival rates improved only minimally over the period 1971-99. Incidence and mortality rates were slightly higher in both males and females living in the most deprived areas, but survival was not consistently related to socio-economic deprivation. CONCLUSIONS: The incidence of, and mortality from, pancreatic cancer in England and Wales have fallen from peak levels observed in the 1970s and 1980s, and levelled off in the 1990s for both sexes; survival rates remain very low.

Computer analysis of upper gastrointestinal endoscope images.

Gastro-oesophageal reflux disease (GORD) occurs in up to 40% of adults in the West. Oesophagitis is a major determinant in the treatment of GORD but its current classification systems are subjective. In order to help to provide objective interpretation of upper gastro-intestinal (GI) endoscope examination and reduce inter-observer variability, we developed a computer image analysis system. Digital video recordings were made on patients with clinical evidence of reflux oesophagitis. Cross-sectional profiles of hue and saturation data were analysed on images from seven patients with grade B or C oesophagitis (LA grading). This analysis showed clear changes in hue (p = 0.01) and saturation (p = 0.001). These results suggest that quantification of upper GI endoscopic images is feasible and may help in objective assessment.

Aggressive combination treatment for invasive fungal sinusitis in immunocompromised patients.

Invasive sinonasal fungal disease is a potentially fatal complication of chemotherapy-induced immunosuppression and neutropenia. We reviewed the outcomes of seven cancer patients who had been diagnosed with invasive fungal sinusitis; six patients had hematologic malignancies and one had breast cancer. At the time of their sinus diagnosis, all patients had been hospitalized and were receiving combination chemotherapy for their underlying malignancy. Impairment of their immune function was characterized by an absolute neutrophil count of less than 1,000/mm3. Aggressive management of their sinonasal fungal disease consisted of surgical debridement and systemic amphotericin B for all patients, and treatment with granulocyte colony-stimulating factor for two patients. Invasive Aspergillus infection was identified in six patients and invasive Candida albicans infection in one. Although the prognosis for these patients was poor and two patients died of the fungal infection, the aggressive treatment strategy resulted in long-term survival for the remaining five patients.

Salvage surgery for patients with recurrent gastrointestinal sarcoma: prognostic factors to guide patient selection.

BACKGROUND: Gastrointestinal sarcomas are rare stromal tumors and most are classifiable as malignant gastrointestinal stromal tumors. They have a high propensity for intraabdominal recurrence. It is unclear whether there is a survival advantage from reoperation for recurrent disease or if surgery should be confined to symptom relief. The authors have attempted to identify features that may allow the selection of patients most likely to benefit from reoperation. METHODS: Retrospective univariate and multivariate analyses of 60 patients with recurrent gastrointestinal sarcoma, accrued from a prospective data base between July 1982 and September 1995, were performed. RESULTS: Initial recurrence was evident by a median of 20 months from primary resection. Most patients (85%) had an initial symptomatic recurrence. Local recurrence was seen in 76% of patients, but in half of these synchronous hepatic metastases were present. None had disease outside of the abdomen. Complete resection of recurrent disease was possible in only one-third of cases. Median survival after surgery for recurrence was 15 months, but was longest for patients whose recurrence took the form of liver metastases in the absence of disease elsewhere. In multivariate analysis of post-recurrence survival, the only significant determinant of survival was the length of the disease free interval between the initial operation and recurrence. CONCLUSIONS: Survival following recurrence of a gastrointestinal sarcoma is largely determined by the tumor biology, one manifestation of which is the disease free interval. Except for patients with a long disease free interval, surgery should be reserved largely for symptom control.