Outcomes analysis of laparoscopic resection of pancreatic neoplasms.

BACKGROUND: Experience with laparoscopic resection of pancreatic neoplasms remains limited. The purpose of this study is to critically analyze the indications for and outcomes after laparoscopic resection of pancreatic neoplasms. METHODS: The medical records of all patients undergoing laparoscopic resection of pancreatic neoplasms from July 2000 to February 2006 were reviewed. Data are expressed as mean +/- standard deviation. RESULTS: Laparoscopic pancreatic resection was performed in 22 patients (M:F, 8:14) with a mean age of 56.3 +/- 15.1 years and mean body mass index (BMI) of 26.3 +/- 4.5 kg/m2. Nine patients had undergone previous intra-abdominal surgery. Indications for pancreatic resection were cyst (1), glucagonoma (1), gastrinoma (2), insulinoma (3), metastatic tumor (2), IPMT (4), nonfunctioning neuroendocrine tumor (3), and mucinous/serous cystadenoma (6). Mean tumor size was 2.4 +/- 1.6 cm. Laparoscopic distal pancreatectomy was attempted in 18 patients and completed in 17, and enucleation was performed in 4 patients. Laparoscopic ultrasound (n = 10) and a hand-assisted technique (n = 4) were utilized selectively. Mean operative time was 236 +/- 60 min and mean blood loss was 244 +/- 516 ml. There was one conversion to an open procedure because of bleeding from the splenic vein. The mean postoperative LOS was 4.5 +/- 2.0 days. Seven patients experienced a total of ten postoperative complications, including a urinary tract infection (UTI) (1), lower-extremity deep venous thrombosis (DVT) and pulmonary embolus (1), infected peripancreatic fluid collection (1), pancreatic pseudocyst (1), and pancreatic fistula (6). Five pancreatic fistulas were managed by percutaneous drainage. The reoperation rate was 4.5% and the overall pancreatic-related complication rate was 36.4%. One patient developed pancreatitis and a pseudocyst 5 months postoperatively, which was managed successfully with a pancreatic duct stent. There was no 30-day mortality. CONCLUSIONS: Laparoscopic pancreatic resection is safe and feasible in selected patients with pancreatic neoplasms. With a pancreatic duct leak rate of 27%, this problem remains an area of development for the minimally invasive technique.

Influence of biologic factors and anatomic site in completely resected liposarcoma.

PURPOSE: Soft tissue sarcoma (STS) encompasses a group of neoplasms that are anatomically and biologically diverse. Retroperitoneal/visceral (RP/V) tumors have a poorer prognosis than extremity/trunk (E/T) lesions, and this has been attributed to frequent presentation with tumors of large size and multiorgan involvement that precludes complete resection. The worse prognosis that is associated with RP/V tumors has also been thought to be histopathologically dependent and not necessarily related to anatomic site. The aim of this study was to determine the role of anatomic site and biologic features in prognosis and outcome in patients after complete resection by examining a large cohort of STS patients with a single histopathology, ie, liposarcoma. METHODS: All patients who were treated for liposarcoma from July 1, 1982, through July 1, 1998, were included. Univariate analyses were performed using log-rank test and Kaplan-Meier estimates, and multivariate analyses were performed using Cox regression. The three end points examined were local recurrence (LR), distant recurrence, and disease-specific survival (DSS). RESULTS: Seven hundred twenty patients with liposarcoma were evaluated, and of these, 460 had completely resected primary or completely resected locally recurrent disease. Breakdown of anatomic site was 65% E/T (n = 301) and 35% RP/V (n = 159). The median follow-up period for patients who underwent complete resection was 42 months (range, 1 to 194 months). We found that RP/V site is a poor prognosticator that is independent of patient sex and age; tumor size, grade, and margin; and recurrent presentation. Sixty-nine percent of patients with RP/V tumors who died had local disease only and no distant metastasis at the time of death. CONCLUSION: In liposarcoma, tumor location exerts as strong an influence on prognosis as biology. In contrast to extremity liposarcoma, LR without distant metastasis often results in death for patients with RP/V tumors. For these patients, local control accomplished by complete surgical resection +/- adjuvant radiation therapy should impact strongly on DSS.

Immunoregulatory effects of CD4+ T helper subsets in human melanoma.

BACKGROUND: The elucidation of CD4+ T helper (Th) cell traits is important for the understanding of immunoregulatory mechanisms in patients with cancer, in particular the Th-cell effect on cytotoxic CD8+ tumor-specific lymphocytes (CTL). METHODS: Sixty-six T-cell receptor alpha beta+/CD4+ clones were generated from tumor-infiltrating lymphocytes of five patients with melanoma and classified into subsets by cytokine production. Transwell experiments were performed to test how the soluble factors of each Th-clone subset affected the cytotoxicity of the tumor-specific CTL against autologous tumor. RESULTS: Th0 clones enhanced cytotoxicity of the CD8+ CTL compared with control CTL cultured in cytokine-free medium. Th1-clone supernatant also enhanced cytotoxicity by CD8+ CTL. In contrast, Th2 clones decreased killing compared with control CTL. Replacement of the Th clones by exogenous interleukin (IL)-2 in concentrations similar to that produced by Th0 and Th1 clones enhanced cytotoxicity. However, suppression of cytotoxicity was observed when similar concentrations of IL-4 were added instead. The helper effect of Th0-soluble factors could be inhibited by anti-IL-2 antibody, whereas anti-IL-4 antibody did not show a significant enhancement. CONCLUSIONS: The majority of the CD4+ tumor-infiltrating lymphocytes (Th0) in patients with melanoma enhance the CTL response to autologous tumor by their soluble factors, whereas Th2 cells suppress the CTL response.

Sentinel lymph node biopsy in breast cancer: unfiltered radioisotope is superior to filtered.

BACKGROUND: The combination of gamma-probe radiolocalization and blue-dye mapping of sentinel lymph nodes (SLNs) has been advocated as the most accurate method for staging the clinically negative axilla in breast cancer patients, but the technical aspects of these procedures are not fully characterized in the literature. In this study, we compared the success of SLN localization in 134 consecutive breast cancer patients using blue dye plus two different preparations of radiocolloid. STUDY DESIGN: A retrospective analysis of a prospectively maintained data base was performed to assess SLN localization in two cohorts of patients. Unfiltered technetium-99m (Tc-99m) sulfur colloid (in 77 patients; group I) was compared with filtered Tc-99m sulfur colloid (in 57 patients; group II). All patients had a peritumoral injection of blue dye and isotope, followed immediately by lymphoscintigraphy to confirm radioactivity at the injection site and to image the SLN. Statistical analysis was performed using the Pearson chi-square test. RESULTS: Unfiltered Tc-99m sulfur colloid was superior to the filtered radiocolloid in localizing the SLN (88% versus 73%; p = 0.03). SLN imaging by lymphoscintigraphy was also more successful in the unfiltered group. Using the combination of blue dye and radiolocalization, SLNs were identified in 94% of patients. CONCLUSIONS: For optimal localization of the SLN in breast cancer patients, surgeons should use the combined technique of blue-dye mapping and gamma-probe localization using unfiltered Tc-99m sulfur colloid.

In vitro stimulation of ovarian tumour-associated lymphocytes with a peptide derived from HER2/neu induces cytotoxicity against autologous tumour.

The proto-oncogene HER2/neu encodes for a 185 kDa transmembrane protein with extensive homology to the epidermal growth factor (EGF) receptor. We have previously shown a correlation between HER2/neu expression and the level of in vitro cytotoxicity of tumour-associated lymphocytes (TAL) versus autologous tumour. In addition, we have recently demonstrated that tumour-associated cytotoxic T-lymphocytes (CTL) from ovarian and breast cancer patients can recognize a HER2/neu derived peptide epitope when presented in the context of HLA-A2. Since repeated tumour stimulation of CTL enhances both proliferation and cytotoxicity against autologous tumour, we hypothesized that repeated peptide antigen stimulation would have a similar effect. To be therapeutically useful, the peptide antigen must meet the following conditions: (1) the peptide must be immunogenic and cause a proliferation of CTL to adequate therapeutic numbers, and (2) the peptide-specific CTL which are generated must be cytotoxic against autologous tumour. To test our hypothesis, T-lymphocytes isolated from the ascites of four consecutive HER2/neu+ ovarian cancer patients were initially stimulated with solid phase anti-CD3 antibody and divided into three groups: (1) treatment with recombinant interleukin-2 (IL-2) alone, (2) IL-2 plus weekly stimulation with irradiated autologous tumour cells, and (3) IL-2 plus weekly stimulation with a HER2/neu derived peptide. Peptide-stimulated and tumour-stimulated CTL showed similar increases in proliferation with both groups consistently reaching therapeutic numbers. Peptide-stimulated CTL demonstrated significantly enhanced cytotoxicity against autologous tumour in 4-h chromium release assays as compared to the IL-2 alone group.(ABSTRACT TRUNCATED AT 250 WORDS)

Myeloid-derived suppressor cells: general characteristics and relevance to clinical management of pancreatic cancer.

Recent studies describe a heterogeneous population of cells of the myeloid lineage, termed myeloid derived suppressor cells (MDSC), which are observed with increased prevalence in the peripheral blood and tumor microenvironment of cancer patients, including pancreatic cancer. Accumulation of MDSC in the peripheral circulation has been related to extent of disease, and correlates with stage. MDSC have primarily been implicated in promoting tumor growth by suppressing antitumor immunity. There is also compelling evidence MDSC are also involved in angiogenesis and metastatic spread. Two main subsets of MDSC have been identified in cancer patients: a monocytic subset, characterized by expression of CD14, and a granulocytic subset characterized by expression of CD15. Both subsets of MDSC actively suppress host immunity through a variety of mechanisms including production of reactive oxygen species and arginase. Just as in humans, accumulation of monocytic and granulocytic MDSC has been noted in the bone marrow, spleen, peripheral circulation, and tumors of tumor bearing mice. Successful targeting of MDSC in mice is associated with improved immune responses, delayed tumor growth, improved survival, and increased efficacy of vaccine therapy. By further elucidating mechanisms of MDSC recruitment and maintenance in the tumor environment, strategies could be developed to reverse immune tolerance to tumor. We discuss here what is currently known about MDSC as well as some potential strategies targeting MDSC in the context of our work on pancreatic cancer and recent literature. Due to the number of new reports on MDSC, the most pertinent ones have been selected.

Vaccine therapy for cancer.

BACKGROUND: Tumor-specific cytotoxic T-lymphocytes (CTLs) can be isolated from the solid tumors, draining lymph nodes, metastatic effusions, and peripheral blood of cancer patients. Despite this evidence for a cell-mediated immune response to cancer, attempts at active specific immunotherapy using cancer vaccines have met with little success in clinical trials. METHODS: We have reviewed the immunobiology of the cell-mediated immune response to cancer by focusing on what is known about the major histocompatibility complex (MHC)-restricted interaction between tumor cells and CD8+ or CD4+ T-cells. In addition, we review the recent advances in the identification of tumor-associated antigens (TAAs) that are recognized by tumor-specific CTLs in melanoma and other cancers. In discussing these antigens, we highlight the recent identification of several MHC-restricted antigenic peptides that are recognized by CTLs from patients with melanoma and those with ovarian and breast cancer. We examine the implications that the discovery of these TAAs and peptides will have on the development of new anticancer vaccines. We review the most recent vaccine trials in melanoma and other cancers and focus on current concepts aimed at improving the therapeutic efficacy of future vaccines, including genetically engineered tumor cell vaccines. CONCLUSIONS: With the recent identification of several TAAs and antigenic peptide epitopes in melanoma and other cancers, immunotherapy researchers are now focusing on new strategies for the development of anticancer vaccines. As the repertoire of known TAAs increases and our understanding of the immunobiology of cell-mediated immunity to cancer improves, immunotherapists remain cautiously optimistic in their quest for effective cancer vaccines.

Immunotherapeutic approaches to sarcoma.

In the last decade, the most important factor in the rekindled interest in immune therapy for cancer is the development of new methods to identify tumor antigens that can be recognized by T-cells and other immune effectors. In addition, greater knowledge about tolerance and mechanisms of tumor cell evasion from immune effectors has made the prospect of developing clinically effective immune therapies for cancer seem promising. Research in immune therapies for sarcoma has been limited, mainly because of the previous lack of defined tumor antigens in this disease and the low prevalence of sarcoma in the general population. We will review the fundamental concepts of tumor immunobiology, both cellular and humoral, and highlight the new, powerful methods for identifying novel tumor antigens. Further, we will focus on the unique situation presented by sarcoma as the only solid tumor in which many cytogenetic abnormalities have been characterized which encode for unique, tumor-specific fusion proteins that are ideal targets for immune-based therapy. We will review the specifics of vaccine therapy approach to this disease, with emphasis on strategies to improve the immunogenicity of newly defined tumor antigens in sarcoma.

The HER2/neu-derived peptide p654-662 is a tumor-associated antigen in human pancreatic cancer recognized by cytotoxic T lymphocytes.

The protooncogene HER2/neu encodes a 185-kDa transmembrane protein with extensive homology to the epidermal growth factor receptor. It is overexpressed in several human cancers of epithelial origin, such as pancreatic cancer. Previously, we demonstrated that cytotoxic T lymphocytes (CTL) derived from breast, ovarian, and non-small cell lung cancer recognized a peptide derived from HER2/neu. To evaluate whether this HLA-A2-binding peptide is a tumor-associated antigen (TAA) in pancreatic cancer, the ability of HER2/neu-reactive CTL to lyse human pancreatic carcinoma cells was tested. CTL were generated from tumor-associated T lymphocytes from HLA-A2+ HER2/neu+ breast and ovarian cancer patients. All CTL recognized autologous and allogeneic HER2/ neu+ tumor cells in an HLA-A2-restricted fashion. Furthermore, all CTL recognized p654-662 (GP2) derived from HER2/neu. These CTL also recognized HER2/neu+ pancreatic cancer cells in an HLA-A2-restricted fashion. HER2/neu+ HLA-A2- pancreatic cancer were not or only poorly lysed. Repeated stimulation of HLA-A2+ PBL from pancreatic cancer patients using the HER2/neu-derived peptide resulted in specific recognition of this peptide and, more importantly, HER2/neu+ pancreatic tumors in an HLA-A2-restricted fashion. Autologous HLA-A2+ fibroblasts or HLA-A2+ malignant melanoma cells were not recognized. HLA-A2- peptide-stimulated T lymphocytes showed no significant cytotoxicity. These results demonstrate that this HER2/neu-derived peptide is a shared TAA among several adenocarcinomas including pancreatic carcinoma, suggesting a common mechanism of recognition of these human tumors by T lymphocytes. The identification of the HER2/neu-derived peptide GP2 as a TAA in pancreatic cancer provides an opportunity for the design of novel immunotherapy and vaccine strategies.

Shared T cell epitopes in epithelial tumors.

We have previously shown the importance of human leukocyte antigen (HLA)-A2 and the proto-oncogene HER2/neu in the T cell recognition of ovarian cancer. Since these proteins are ubiquitously expressed in epithelial-derived tumors, we have acid-eluted HLA-bound peptides from ovarian cancers, fractionated the peptides, and reconstituted T cell epitopes on the HLA-A2+ T2 cell line to determine if common tumor-associated antigens exist among HLA-A2+, HER2/neu+ epithelial cancers. We demonstrate that tumor-specific cytotoxic T lymphocytes (CTL) generated from tumor-infiltrating lymphocytes isolated from three ovarian, two breast, and two non-small-cell lung cancers recognize at least three of the same peptide fractions from multiple elutions. One of these peptide fractions coelutes with a HER2/neu-derived peptide which has been shown recently to be recognized by these same CTL. These findings demonstrate that a common peptide-based tumor vaccine is theoretically possible for many different epithelial-derived cancers.

Tumor-specific and HLA-A2-restricted cytolysis by tumor-associated lymphocytes in human metastatic breast cancer.

A tumor-specific cytotoxic T lymphocyte (CTL) immune response has been well documented in melanoma, renal cell carcinoma, and ovarian cancer. Conflicting evidence exists regarding the existence of tumor-specific CTL populations in breast cancer. Tumor cells and tumor-associated lymphocytes (TAL) were isolated from the pleural effusions of six consecutive patients with metastatic breast cancer. After solid-phase anti-CD3 stimulation, TAL cultures were expanded with weekly autologous tumor stimulation and low-dose IL-2 for 3 wk. T cell populations were characterized using flow cytometric analysis and ranged from 49 to 91% CD8+, > 98% CD3+, and < 3% CD16+. Functionally, tumor-stimulated TAL showed tumor-specific recognition of autologous tumor cells (241 +/- 142 LU20/10(7)) and no detectable lysis of autologous fibroblasts, Daudi or K562. Cytotoxicity of TAL against HLA-A2+ allogeneic targets was significantly higher when compared with HLA-A2- tumor cell lines (127 +/- 76 vs 6 +/- 18 LU, p = 0.0001). This cytotoxicity against autologous and allogeneic tumor cells was blocked by anti-HLA-A2 mAb and cold HLA-A2+ targets in cold-target inhibition assays. TAL from all HLA-A2+ patients recognized GP2, a known, HER2/neu-derived tumor-associated peptide Ag that is HLA-A2 restricted. We have shown that TAL obtained from metastatic effusions of breast cancer patients contain lymphocytes that can recognize and lyse autologous and allogeneic tumor cells in a tumor-specific, HLA-A2-restricted fashion. In addition, tumor-specific TAL derived from breast cancer patients can selectively lyse HLA-A2+ pancreatic and ovarian tumor cell targets, suggesting a common HLA-A2-restricted tumor-associated Ag between these distinct epithelial cancers. Further elucidation of the cell-mediated immune response to breast cancer and the identification of shared TAA could result in the development of broadly applicable vaccine therapies for many cancers.

Magnetic resonance imaging facilitates breast conservation for occult breast cancer.

INTRODUCTION: Occult primary breast cancer, i.e., isolated axillary adenocarcinoma without detectable tumor in the breast by either physical exam or mammography, represents up to 1% of operable breast cancer. Modified radical mastectomy (MRM) is generally the accepted treatment for this condition although tumor is identified in only two-thirds of mastectomy specimens. Breast magnetic resonance imaging (MRI) can identify occult breast carcinoma and may direct therapy. This study examined the ability of breast MRI to detect occult breast cancer and to facilitate breast conservation therapy. METHODS: Forty women with biopsy-proven metastatic adenocarcinoma to an axillary lymph node and no evidence of primary cancer were studied. All patients had a physical examination, mammography, and MRI of the breast. Using a dedicated breast coil, MRI imaging was performed with and without gadolinium enhancement. Positive MRI scans were compared with histopathologic findings at the time of operation (n = 21). RESULTS: MRI identified the primary breast lesion in 28 of 40 women (70%). Of these 28 patients, 11 had MRM, 11 had lumpectomy/axillary lymph node dissection (ALND)/radiotherapy (XRT), 2 had ALND/XRT alone, and 4 had no local treatment secondary to stage IV disease. Two women initially treated with lumpectomy/ALND subsequently had mastectomy for positive margins. Of the women with positive MRI who had breast surgery, 21 of 22 (95%) had tumor within the surgical specimen. Twelve women had negative MRI of the breast. Five of these 12 underwent MRM, of whom 4 had no tumor in the mastectomy specimen. The remaining 7 patients had ALND and whole breast radiation (ALND/XRT) (n = 5), or were observed (n = 2). Overall, 18 of 34 women surgically treated had MRM, while 16 (47%) preserved their breast. Tumor yield for patients having breast surgery was 81%. CONCLUSIONS: MRI of the breast can identify occult breast cancer in many patients and may facilitate breast conservation in select women. Negative breast MRI predicts low tumor yield at mastectomy.

Breast and ovarian cancer-specific cytotoxic T lymphocytes recognize the same HER2/neu-derived peptide.

The identification of antigenic peptides presented on the tumor cell surface by HLA class I molecules and recognized by tumor-specific cytotoxic T lymphocytes may lead to a peptide vaccine capable of inducing protective cellular immunity. We demonstrate that both HLA-A2-restricted breast and ovarian tumor-specific cytotoxic T lymphocytes recognize shared antigenic peptides. At least one of these peptides is derived from the oncogene product of HER2/neu, which is overexpressed in 30-40% of all breast and ovarian cancers. T cells sensitized against this nine-amino acid sequence demonstrate significant recognition of HLA-A2+, HER2/neu+ tumors. Since 50% of the tumor-cell population is HLA-A2+ and many different tumors express HER2/neu, this peptide may be widely recognized and have many clinical applications.

Simultaneous production of T helper-1-like cytokines and cytolytic activity by tumor-specific T cells in ovarian and breast cancer.

Cytotoxic T-cell (CTL) cultures were generated from five ovarian cancer patients (OvCTL) and from three breast cancer patients (BrCTL). All CTL lines were T-cell receptor (TcR) alphabeta+ and predominantly CD8+ (73 +/- 13%). These CTL lines preferentially recognized autologous tumor cells in an HLA class I-restricted, and in part HLA-A2-restricted, manner. In addition, the CTL lines recognized allogeneic HLA-A2+ ovarian and breast tumor cells. Specific recognition was determined by T-cell-mediated cytotoxicity as well as cytokine release. Coculture of irradiated autologous tumor cells with OvCTL induced secretion of IFN-gamma, GM-CSF and TNF-alpha, but not IL-4, indicating a T helper-1-type response. Similar results were obtained when OvCTL and BrCTL were stimulated with histologically matched HLA-A2+ tumor cells. Also, BrCTL stimulated with HLA-A2+ but not HLA-A2- ovarian tumor cells produced significant levels of GM-CSF and TNF-alpha. Finally, the Her2/neu peptide p654-662, earlier identified as a tumor antigen in both ovarian and breast cancer, induced cytotoxicity as well as the specific release of IFN-gamma and TNF-alpha but not IL-4 by OvCTL and BrCTL. Thus, tumor-specific recognition by CTL was verified by cytotoxicity and cytokine release. The secretion of Th1-like cytokines as opposed to Th2-like cytokines suggest that therapeutically OvCTL and BrCTL could potentially enhance the endogenous immune response to tumor.

Intradermal radiocolloid and intraparenchymal blue dye injection optimize sentinel node identification in breast cancer patients.

BACKGROUND: Radiotracer and blue dye mapping of sentinel lymph nodes (SLN) have been advocated as accurate methods to stage the clinically negative axilla in breast cancer patients. The technical aspects of SLN biopsy are not fully characterized. In this study we compare the results of intraparenchymal (IP) and intradermal (ID) injection of Tc-99m sulfur colloid, to establish an optimal method for SLN localization. METHODS: 200 consecutive patients had SLN biopsy performed by a single surgeon. Of these, 100 (Group I) had IP injection and 100 (Group II) had ID injection of Tc-99m sulfur colloid. All patients had IP injection of blue dye as well. Endpoints included (1) successful SLN localization by lymphoscintigraphy, (2) successful SLN localization at surgery, and (3) blue dye-isotope concordance (uptake of dye and isotope by the same SLN). RESULTS: Isotope SLN localization was successful in 78% of Group I and 97% of group II patients (P < .001). When isotope was combined with blue dye, SLN were found in 92% of group I and 100% of Group II (P < .01). In cases where both dye and isotope were found in the axilla, dye mapped the same SLN as radiotracer in 97% of Group I and 95% of Group II patients. CONCLUSIONS: The dermal and parenchymal lymphatics of the breast drain to the same SLN in most patients. Because ID injection is easier to perform and more effective, this technique may simplify and optimize SLN localization.

Surgical margin in hepatic resection for colorectal metastasis: a critical and improvable determinant of outcome.

OBJECTIVE: To update the analysis of technical and biologic factors related to hepatic resection for colorectal metastasis in a large single-institution series to identify important prognostic indicators and patterns of failure. SUMMARY BACKGROUND DATA: Surgical therapy for colorectal carcinoma metastatic to the liver is the only potentially curable treatment. Careful patient selection of those with resectable liver-only metastatic disease is crucial to the success of surgical therapy. METHODS: Two hundred forty-four consecutive patients undergoing curative hepatic resection for metastatic colorectal carcinoma were analyzed retrospectively. Variables examined included sex, stage of primary lesion, size of liver lesion(s), number of lesions, disease-free interval, ploidy, differentiation, preoperative carcinoembryonic antigen level, and operative factors such as resection margin, use of cryotherapy, intraoperative ultrasound, and blood loss. RESULTS: Surgical margin, number of lesions, and carcinoembryonic antigen (CEA) levels significantly control prognosis. Patients with only one or two liver lesions, a 1-cm surgical margin, and low CEA levels have a 5-year disease-free survival rate of more than 30%. Disease-free interval, original stage, bilobar involvement, size of metastasis, differentiation, and ploidy were not significant predictors of recurrence. The pattern of failure correlates with surgical margin. Routine use of intraoperative ultrasound resulted in an increased incidence of negative surgical margin during the period examined. CONCLUSIONS: Surgical resection or cryotherapy of hepatic metastasis from colorectal cancer is safe and curable in appropriately selected patients. Biologic factors, such as number of lesions and carcinoembryonic antigen levels, determine potential curability, and surgical margin governs the patterns of failure and outcome in potentially curable patients. Optimization of selection criteria and surgical resection margins will improve outcome.