Abstracts 258 to 521
Liver Cancer Program, New York, NY Purpose: To determine the patient and tumor characteristics associated with survival and recurrence after resection of HCC. Methods: We analyzed data on 300 consecutive patients undergoing resection for HCC at a single institution between 1/1/1990 and 1/1/2005. Selection criteria were in 2001refined to include only patients with bilirubinՅ 1.0 and platelets Ն 100,000 (era 1 vs. 2). Uni-and multivariate analyses of survival and recurrence were conducted with
... rank test and Cox proportional hazard analysis. Predictors of perioperative mortality were determined with logistic regression. Results: Noncirrhotic patients (Metavir 0-2, nϭ106) had significantly larger and less well-differentiated tumors than severe fibrotic/cirrhotics (Metavir 3-4, nϭ194). Overall 1, 3 and 5 yr survivals were 76%, 58% and 50% (median 61 mo) with 5 yr survival being similar for cirrhotics (47%) and noncirrhotics (52%). For cirrhotics, independent predictors of survival were transfusion, macroscopic vascular invasion (macro VI), albumin Ͻ 3.5, and era. Based on preoperatively available variables, the best cirrhotic candidates (albumin Ͼ3.5 & no macro VI) resected in era 2 achieved 4 yr survival of 86%. Macro VI and satellites were independent predictors of recurrence in cirrhotics. Looking only at cirrhotics without macro VI (nϭ123), era and transfusion predicted survival while satellites and non-HBV liver disease predicted recurrence on multivariate analysis. For noncirrhotics, tumor sizeϾ10cm, any vascular invasion, and non-HBV liver disease were independent predictors of survival. The best noncirrhotic candidates selected based on preoperative variables, achieved 5 yr survival of 80%. Satellites, mod-poor differentiation, macro VI and AFPϾ100 predicted recurrence on multivariate analysis in noncirrhotics. Perioperative mortality was 7% and dropped from 11% in era 1 to 3% in era 2 (pϭ0.006). Predictors of perioperative mortality included era, portal thrombectomy, and Child's B class. Conclusions: Despite similar long-term outcomes, predictors of survival and recurrence vary significantly based on the presence of concomitant cirrhosis. In cirrhotics, macro VI and liver function are important predictors of survival, while only tumor-related factors predict survival in noncirrhotics. Patient selection (era 1 vs. 2) and conduct of the surgery (transfusion) also play a significant role in the outcome of cirrhotics. Properly selected patients, with or without cirrhosis, can achieve 4 yr survival over 80%. Disclosures: The following authors have indicated they have no relationships to disclose: Sasan Roayaie, Josep Llovet, Parissa Tabrizian, Francesco D'Amico, Maria Fiel, Myron Schwartz IMMUNE, METABOLIC AND CIRCULATORY RESPONSES TO LIVER SURGERY masashi ishikawa, tokushima redcross hospital, komatsushima, Japan Recent studies have shown that the type 1/2 T-helper (Th1/2) cell balance is shifted toward a Th 2-type immune response not only by malignancy but also by surgical stress. In the present study, immune, metabolic and circulatory responses to liver surgery were measured for determination of surgical stress in comparison with those in other major abdominal surgeries. Patients and Methods: Ninety patients who underwent abdominal surgeries were divided into three groups: hepatic resection (nϭ20), gastric resection (nϭ39), colorectal resection (nϭ31). Blood sampling was performed before surgery, and on postoperative days (POD) 2 and 14 to determine Th 1/2 cell balance. The proportion of CD4 positive lymphocytes producing IFN-r, IL-2, IL-4 and IL-6 were measured by flow cytometry as described by Openshaw et al. Results were calculated as a ratio of the percent IFN-r-producing (Th1) cells to IL-4-producing (Th2) cells. Energy expenditure was measured by indirect calorimetry, and hemodynamics was studied using pulse dye densitometry by DDG analyzer until POD 14. Results: The postoperative ratio of Th1 to Th2 was decreased significantly in all subjects to 4.6ᮀ }3.1 on POD 2 from 6.8ᮀ }4.1(before operation). The ratio of Th1 to Th2 in patients who underwent hepatic resection markedly decreased to 2.7ᮀ }1.5 on POD 2, significantly different from those in patients undergoing gastric resection or colorectal resection. However, on POD 14 there were no significant differences among the three groups. Resting energy expenditure and cardiac index on POD 1 and 3 in patients with hepatic resection increased significantly above levels in the other surgical groups. However, blood volume in the hepatic resection patients was significantly lower than that of other patients until POD 3. Ten patients who developed postoperative complications had significantly lower Th 1/2 ratios and more hypermetabolism. Conclusions: This study reveals that hepatic resection induces a more marked shift toward a Th 2 helper T cell response and significantly more hypermetabolism than other major surgeries. A distinct pattern of Th 1/2 ratio changes during the early phase of the postoperative course in hepatic resection may be related to changes in metabolism and circulation. Therefore, determination of Th 1/2 balance may be of help in evaluating different surgical procedures, and the differences may relate to excessive surgical stress, synthesis of acute phase proteins and underlying liver disease. This finding would have significant implications for the development of prophylactic strategies in patients undergoing hepatic resection. Introduction: Liver resection for HCC often involves patients with chronic HBV infection. The incidence, outcomes and risk factors of HBV reactivation after liver resection in these patients are not well described. Methods: To address this, we performed a retrospective review of consecutive HepBsAgϩ patients who underwent liver resection (nϭ77) for hepatocellular carcinoma in our centre between January 2002 to December 2004. Data were systematically collected to determine: i) baseline demographic, biochemical, virological and surgical characteristics, ii) postoperative hepatitis and hepatitis B flare, iii) outcomes -hospital stay, mortality, tumour recurrence and iv) risk factor(s) for hepatitis B flare. Definitions: i) Postoperative hepatitis -ALT Ͼ2x baseline or ALT Ͼ200 IU/L between 2 and 24 weeks post-resection (to distinguish from post-resection transient elevation of ALT); ii) HBV flare: postoperative hepatitis associated with detectable HBV DNA. All results are expressed as median unless otherwise stated. Results: Baseline characteristics: 77 patients underwent 82 liver resections (70% segmentectomy, 30% hemi-hepatectomy ); mean age 58.0 ϩ 12.1 years; 87% male; 20% HBeAg ϩ; HBV DNA tested in 21%. Transient elevation of ALT in the first week after resection occurred in 92% of cases -peak ALT 222.0 IU/L (14-656) and declined by week 2 post-resection. The incidence of postoperative hepatitis was 26% (nϭ21) -peak ALT 231.0 IU/L observed at day 85 post-resection. The incidence of HBV flare was 9% (nϭ7) -peak ALT 312 IU/L (147-1400) at day 84 post-resection. Hepatic decompensation was more frequent and more severe in postoperative hepatitis arising from HBV flare (incidence 88% vs. 39%, pϭ0.03; peak bilirubin 322.0 vs. 253.0 umol/L, pϭ0.5; peak PT 19.5 vs. 17.3 seconds, pϭ0.5). 83% of decompensated patients required readmission. Overall mortality rate was significantly higher for those with HBV flare compared with those without postoperative hepatitis (67% vs 21%, pϭ0.03). HBV flare resulted in 2 deaths despite HEPATOLOGY, Vol. 42, No. 4, Suppl. 1, 2005 AASLD ABSTRACTS 299A institution of antiviral treatment. Neither postoperative hepatitis nor HBV flare resulted in higher HCC recurrence. Preoperative ALT Ͼ100 IU/L was found to be the only risk factor for postoperative hepatitis but not for hepatitis B reactivation flare. Conclusion: In HBsAgϩ patients undergoing liver resection, HBV flare is seen 9%, is associated with pre-operative ALT Ͼ 100 IU/L and associated with poorer outcomes. Whether HBVDNA level is a risk factor for HBV flare cannot be answered by this study and should be explored in future studies. There is a role in identifying patients at-risk for HBV flare for pre-emptive treatment. According to a French multicentric study, adenomas can be divided into 4 phenotype/genotype groups. Group A (50%) is steatotic mutated for the HNF1a gene (90% somatic, 10% constitutional); group B (10%) presents cytological atypia, mutated for b catenin and prones to develop hepatocellular carcinoma. No mutation has been observed in the 2 other groups. Contrary to group C (no particular morphological characteristics), group D differs from the others because it presents features that include criteria observed in so-called TFNH. The aim of our study was to review our cases of nodules previously described as TFNH (certain, possible, question mark). These cases cannot be classified in the FNH group on the following criteria: soft nodule, often hemorrhagic, ill defined border, absence of fibrotic bands but were also questionable for classification in the adenoma group because they often contain one or several particular features which are usually not present in adenoma (biliary structures, thick-walled arteries and obvious inflammation). Our cases of TFNH were characterized according to the number of nodules (unique, or multiple); the easiness or not to recognize the particular features; the similarities or differences between nodules from the same case. Clonality, HNF1a and b catenin mutations were looked for in each case. 146 patients had a hepatic resection for benign nodules : 60 cases for FNH; 86 cases for adenomas: 70 were typical adenomas and 16 presented those particular features in the entire or part of the nodule, such as pseudo portal tracts containing inflammatory cells, thick-walled arteries, and a more or less obvious ductular reaction, associated or not with a centrolobular sinusoidal dilatation. All these 16 cases were women (all on OC); 3 presented acute clinical symptoms due to bleeding. Six out of 16 were unique; 5 were multiple and in both groups, 5 and 4 cases respectively presented thick arteries, inflammation and ductular reaction. In 5 additional cases, 3 were diffuse with some nodules corresponding either to TFNH, or to typical adenomas and 2 were recurrent cases. Association with FNH was observed in 5 out of 16 cases. All 16 cases were monoclonal and did not exhibit HNF1 a nor b catenin mutations. Several arguments tend to unify, under a same entity, nodules sharing in common some features: monoclonality, no known gene mutation, risk of hemorrhage and particular morphological features. At present it seems reasonable to consider these nodules as a subgroup of inflammatory, HNF1a non mutated adenomas, and accordingly, the terms TFNH and mixed tumors (adenoma/FNH) should not be used to characterize this type of nodules. Disclosures: The following authors have indicated they have no relationships to disclose: Hepatocellular carcinoma (HCC) is a global medical concern. This investigation compares HCC treatment practices of quaternary care medical centers located in Asia (East) and North America (West). Demographic, physiologic, and pathologic information from all referrals was obtained and treatment categorized as liver transplantation (LTX), surgical resection (RES), or loco-regional therapy (LCL). Outcome variables were tumor recurrence, patient survival, and occurrence of a complication. From 1999 through 2004 (median follow-up: 2.5yr), 1357 East and 243 West patients were evaluated (Table I) . Underlying liver disease was principally hepatitis B (68%) in EAST and hepatitis C (37%) in West. At referral, EAST patients were typically diagnosed with HCC and had received initial treatment from the referring facility while WEST patients were referred for management of underlying liver disease or suggestive imaging without fulfilling diagnostic criteria. Surgical resection (Ͼ2 segments) was applied more frequently among EAST patients with a higher estimated MELD (8.8 Ϯ 1.9) and AJCC stage. RES in WEST patients was limited to non-cirrhotics or well-compensated cirrhotics (MELD: 7.9 Ϯ 1.7) and patients not a candidate for LTX. One-year recurrence was EAST 30% versus WEST 21% (pϽ0.05). Liver transplantation in EAST was principally (42%) applied to resection failures, multifocality, or advanced disease (AJCC Ն II) and was achieved through living donation (68%). WEST LTX recipients had AJCC Stage I or II disease at LTX had received LCL as interim therapy with only 8% of WEST LTX recipients RES failures. Living donor accounted for 13% of allografts. LTX one-year recurrence / patient survival were EAST 25% / 78% and WEST 17% / 81%, respectively. Technical complications were comparable between groups. LCL was applied at each center to advanced tumor (ϾAJCC IIIA), significant underlying liver disease (MELD: 11 Ϯ 0.8) or medical co-morbidities including age that precluded RES or LTX. Overall HCC recurrence at one year at both centers exceeded 35%. In conclusion, significant differences were observed in the management of HCC between the Asian and North American centers studied. These include the initial diagnosis and management of HCC by the referring physicians, the underlying etiology and severity of liver disease, and the applicability of surgical resection. While cadaver donation limits LTX application in Asia, satisfactory outcomes with surgical resection suggest expansion of these techniques to select North American patients could reduce current demand for liver transplantation.