| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||
ABSTRACT |
Thursday, May 4, 10:00 AM-12:00 PM
Abstracts 218-226
Moderators: Ihab R. Kamel, MD and Fergus M. Coakley, MD
10:00 AM
Keynote Address: New Liver Techniques
Ihab R. Kamel, MD, Johns Hopkins Hospital, Baltimore, MD
10:30 AM
218. Heat Shock Protein Expression by Percutaneous Radiofrequency Ablation of Hepatocellular Carcinoma In Vivo
Schueller G.*; Kettenbach J.; Lammer J.; 1. Radiology, Medical University of Vienna, Vienna, Austria.
Address correspondence to G. Schueller (gerd.schueller{at}meduniwien.ac.at)
Objective: Tumor cells respond to stress such as heat with the synthesis of heat shock proteins (HSPs). HSPs are crucially important in anti-tumoral host reactions. Study purpose was to evaluate the expression of HSPs before and after treatment with percutaneous radiofrequency (RF) ablation in hepatocellular carcinoma (HCC).
Materials and Methods: Human HCCs were inoculated in 41 athymic rats. RF ablation was performed applying 100-500 Ws energy, and tumors were excised 6-24 h later. Furthermore, in one patient with unresectable HCC, CT-guided biopsies were obtained from the tumor before and 24 h after RF ablation (ablation time, 10 min; total RF energy, 19.000 Ws). Immunohistochemistry of the specimens determined the expression of HSP 70 and 90.
Results: In the animal model, HSP 70 expression was 0% and HSP 90 expression was 30% before RF ablation. After treatment, the HSP expression was significantly increased, with a maximum after 12 h (HSP 70, 60%; HSP 90, 90%). The expression of HSPs correlated with the energy applied. The specimens of the one patient treated showed an eight-fold increase in HSP 70 expression and a 1.2-fold increase in HSP 90 expression. No adverse side effects were observed.
Conclusion: In both animal model and treatment of one patient we demonstrated the increase of HSP expression subsequent to RF ablation in HCC. Our results may be of relevance in further therapeutic anti-tumor strategies.
219. Radiologic-Pathologic Correlation of Multi-Slice CT in Determining the Volume of Necrosis in Hepatic Radiofrequency Ablation
Ko C.C.; Ho S.; Radiology, Vancouver General Hospital, Vancouver, BC, Canada.
Address correspondence to C.C. Ko (koc.canada{at}gmail.com)
Objective: We investigate the correlation of the low attenuation change in the followup CT post hepatic radiofrequency ablation with the pathologic evidence of parenchymal necrosis in pigs.
Materials and Methods: Radiofrequency ablations were performed in the normal livers of six pigs using three different thermal ablation systems: Boston Scientific, Radionics and RITA. Each ablation was carried out following the standard treatmentprotocols outlined by the individual manufacturers. Each pig received three to four ablations in different part of the liver,. A total of 21 ablations were performed, equally divided among the three ablation systems. Triple phase CT scans were acquired immediately following the ablations. The pigs were then sacrificed. The livers were then removed and the zones of necrosis were determined by both microscopic and macroscopic pathologic correlation. The volume of the ablated parenchyma in each specimen was measured with the water-displacement method. Radiologically, the post-ablation low attenuation change in the follow up CT was interpreted to represent parenchymal necrosis. The product of the surface area of the necrotic zone with the slice thickness calculated the volume of necrosis in the slice. The total ablated volume was then calculated by summating the volume of necrosis in individual slices demonstrating the necrosis. Comparisons were made between the radiologically calculated volumes of necrosis (Radiological Ablation Volume, RAV) with the corresponding measurements obtained by the pathologist (Pathological Ablation Volume, PAV).
Results: Using the PAV as the baseline, the RAV derived from CT consistently over-estimated the ablation volumes from all three thermal ablation systems. The percentage of the over-estimation in relation to the size of the ablation (%OE) was calculated using the formula: (RAV-PAV)/PAV x 100%. The average %OE was found to be around 140%. We hypothesize that the process of cell death takes days to fully evolve. The CT may be detecting the irreversible cellular injury at cellular enzymes level that would lead to cell death but not become pathologically apparent until days later. This process of cellular necrosis was arrested when the pigs were sacrificed straight after the ablations.
Conclusion: CT over-estimated the extent of post-ablation necrosis by a large margin. Further study is required to show that this may be partly due to CT detecting irreversible cellular injury. This zone of tissue may not be pathologically visible until a later time.
10:50 AM
220. Percutaneous Radiofrequency Ablation of Recurrent Liver Tumors after Hepatic Resection
Schindera S.T.1*; Nelson R.C.1; DeLong D.1; Clary B.2; 1. Department of Radiology, Duke University Medical Center, Durham, NC; 2. Department of General Surgery, Duke University Medical Center, Durham, NC.
Address correspondence to S.T. Schindera (sebastian.schindera{at}duke.edu)
Objective: To assess the therapeutic effectiveness, safety and long-term outcome of percutaneous radiofrequency ablation (RFA) of recurrent hepatic tumors after prior hepatic resection.
Materials and Methods: Retrospective review of our RFA database from June 1999 to July 2005 yielded 36 patients (22 men, 14 women; median age, 62 years; age range, 28 - 83 years) who have undergone prior hepatectomy for malignant (primary, n = 9; metastatic, n = 26) and benign (n = 1) hepatic tumor. A total of 69 recurrent lesions (mean diameter, 1.7 cm; range, 0.5 - 5.3 cm) were treated in 49 sessions (mean number, 1.4; range, 1 - 7) under general anesthesia using CT-fluoroscopic and/or US guidance. CT and MR imaging were used for evaluation of the therapeutic effectiveness. Overall survival rate of the patients with a malignant hepatic tumor was calculated using the Kaplan-Meier method.
Results: Complete ablation was accomplished in 57 (82.6%) out of 69 hepatic tumors. During a mean follow-up of 19.1 months (range, 1.1 - 65.2 months), 12 (17.4%) hepatic tumors in 9 patients showed either incomplete tumor destruction (n = 8) or local tumor recurrence (n = 4). Seven (36.8%) out of 19 ablated lesions guided by US alone showed local failure whereas only 5 (10%) out of 50 lesions guided by the combination of US and CT recurred. The difference between these two procedure guidances was statistically significant (p < 0.01). Comparing the size of tumors between the group of recurrent lesions (mean 2.6 cm; range, 0.8 - 5.0 cm) and the group of completely ablated lesions (mean 1.8 cm; range, 0.5 - 5.3 cm) was also significantly different (p = 0.02). Distant intrahepatic recurrence appeared in 22 (61.1 %) out of 36 patients. One death occurred following RFA due to fulminant liver failure. There were 9 (13.0%) other complications: 8 minor and one major (abscess formation). The overall survival rates for all patients with malignant hepatic tumors at 1-, 2- and 3-years were 76%, 68% and 45%, respectively; for patients with metastases from colorectal cancer (n = 14) the survival rates were 72%, 60% and 60%, respectively; and for patients with hepatocellular carcinoma (n = 8) the survival rates were 72%, 58% and 44%, respectively.
Conclusion: The size of the hepatic lesion and the type of imaging guidance has a significant influence on the therapeutic response of RFA. Percutaneous radiofrequency ablation offers a safe and effective treatment option for recurrent hepatic tumors after previous hepatectomy.
221. The Ability of CT to Determine the Degree of Hepatic Steatosis
Kodama Y.1*; Ng C.S.1; Wu T.T.2; Johnson V.E.3; Ayers G.D.3; Curley S.A.4; Abdalla E.K.4; Vauthey J.N.4; Charnsangavej C.1; 1. Diagnostic Radiology, MD Anderson Cancer Center, Houston, TX; 2. Pathology, MD Anderson Cancer Center, Houston, TX; 3. Biostatistics and Applied Mathematics, MD Anderson Cancer Center, Houston, TX; 4. Surgical Oncology, MD Anderson Cancer Center, Houston, TX.
Address correspondence to Y. Kodama (ykodama{at}mud.biglobe.ne.jp)
Objective: Significant hepatic steatosis increases the difficulty and risk associated with major hepatic resection and impacts graft function after partial-liver transplantation. The purpose of this study was to assess whether liver Hounsfield density measurements by CT could be used to predict hepatic fat content.
Materials and Methods: Eighty-eight consecutive patients who underwent liver resections for metastatic disease formed the basis of this retrospective study. Mean age 56 years old (range 18 to 81 years); 42 females and 46 males. Hounsfield density measurements were obtained from 12 defined, representative, liver locations in non-contrast, and corresponding portal phase contrast enhanced, pre-operative hepatic CTs. ROIs were 1.0 +/- 0.1 cm-square and located in liver parenchyma, avoiding focal liver lesions and vessels. The mean interval between the pre-operative CTs and resection was 16.4 day (range 1-36 days). Average Hounsfield densities of the resected portion of liver (3-9 ROIs) were compared to the percent fat content of the resected liver specimen, assessed histopathologically. Hounsfield densities were also obtained from the spleen to see if this could be used to improve the overall predictions. Linear regression analysis was undertaken employing a log-log scale, the latter to account for non-negative fat contents.
Results: There was a significant association between non-contrast CT Hounsfield density (mean 56.4, range -10.8 to 77.9 Hounsfield Units) and histological percent fat content (mean 12.9%, range 0% to 70%) [R-square = 0.65, p < 0.001], with the linear regression model on the log-log scale. The predictive equation for percent fat content was F% = 100 * exp (-exp (-1.915 + 0.0514 * NC), where F%= percent fat content, exp = exponential function (i.e, 2.718), and NC = non-contrast Hounsfield density. The association between CT Hounsfield densities obtained from contrast enhanced CTs and histological percent fat content was less strong [R-square = 0.52, p < 0.001]. No additional benefit was derived from including splenic CT densities in the regression function.
Conclusion: Non-contrast CT Hounsfield density predicts hepatic fat content with reasonable accuracy. Contrast enhanced CT is less reliable, and spleen Hounsfield density measurements are non-contributory.
222. Comparison of Contrast Material-enhanced Ultrasound versus Baseline Ultrasound and Contrast Material-enhanced Computed Tomography in Liver Metastases Diagnosis
Quaia E.1*; D'Onofrio M.2; Degobbis F.,; Rossi S.1; Cova M.1; 1. Radiology, University of Trieste, Trieste, Italy; 2. Radiology, University of Verona, Verona, Italy.
Address correspondence to E. Quaia (equaia{at}yahoo.com)
Objective: The accurate assessment of liver metastatic disease is fundamental for the treatment planning of primary malignancies. The aim of this study was to compare contrast material-enhanced Ultrasound (CEUS) to baseline US and contrast material-enhanced Computed Tomography (CT) in liver metastases diagnosis.
Materials and Methods: Two hundred-fifty-three patients with 1-5 proven or suspected liver metastases at baseline US were included. All patients underwent US before and after microbubble injection, and multiphase contrast material-enhanced CT. Independent panels of readers reviewed US and CT scans, and recorded liver metastases according to a 5-grade diagnostic confidence scale (1: absolutely benign; 2: probably benign; 3: indeterminate; 4: probably metastasis; 5: absolutely metastasis). Reference standards included CT with (n = 46 patients) or without transcutaneous US-guided lesions biopsy (n = 95) combined to follow-up data, Gd-BOPTA-enhanced MR imaging (n = 63), or intraoperative US (n = 49).
Results: A total number of 551 metastases and 55 benign lesions were finally diagnosed. According to a patient-by-patient analysis reference standards revealed no metastases in 57/253 (22%), > 5 metastases in 59/253 (23%), and 1-5 metastases in 137/253 patients (55%). In patients with 1-5 metastases CEUS vs baseline US revealed more metastases in 64/137 (46%), and the same number in 73/137 patients (54%),while CEUS vs CT revealed more metastases in 10/137 (7%), the same number in 99/137 (73%), and lower number in 28/137 patients (20%). According to a lesion-by-lesion analysis, readers changed diagnostic confidence score in 320 metastatic lesions after the additional review of CEUS in comparison to baseline US. In 262/320 metastases CEUS allowed readers to propose a correct diagnosis by shifting the diagnostic score from 1-3 to 4-5, while in the remaining 58/320 metastases readers were more confident for the correct characterization by shifting the diagnostic score from 4 to 5. The mean number of metastases per patient, the sensitivity, specificity, and area under ROC curve in CEUS (1.82 ± 1.79, 461/551 [83%], 50/55 [91%], 0.929) differed significantly from baseline US (0.88 ± 0.61, 225/551 [40%], 35/55 [63%], 0.579; p < 0.01), while did not differ from contrast material-enhanced CT (1.93 +/- 1.91, 490/551 [89%], 51/55 [93%], 0.945; p > 0.05).
Conclusion: CEUS improved liver metastases diagnosis in comparison to baseline US, while it revealed similar diagnostic performance and confidence to contrast material-enhanced CT.
223. A Novel Method for Analysis of Postoperative Liver Regeneration in the Adult Live Liver Recipient
Wald C.1*; Bourquain H.2; Kankanala N.3; Scheirey C.1; Verbesey J.4; Peitgen H.O.2; Pomfret E.A.4; 1. Department of Diagnostic Radiology, Lahey Clinic Medical Center, Burlington, MA; 2. Liver Imaging, MeVis, Bremen, Germany; 3. Medical School, Tufts University, Boston, MA; 4. Department of Hepatobiliary Surgery and Liver Transplantation, Lahey Clinic Medical Center, Burlington, MA.
Address correspondence to C. Wald (christoph.wald{at}lahey.org)
Objective: Novel image-analysis software conveys better understanding of the determinants and dynamics of liver regeneration in recipients of right lobe liver grafts which could lead to optimization of surgical technique and improve prediction of expected regeneration in an individual patient.
Materials and Methods: Dynamic contrast enhanced spiral CT was performed preoperatively in live liver donors, and in the matching recipients 1week, 1, 3, 6 and 12 months postoperatively after right lobe liver transplantation. 120 datasets were analyzed with Hepavision2 (© MeVis, Bremen, Germany). Portal vein and hepatic veins casts were electronically extracted from the CT data. Based on a hierarchical mathematical model the volumes of individual segmental vascular territories were calculated for all patients. The ratio of right anterior graft volume was divided by the volume of the right posterior graft (RAG/RPG ratio) and normalized to the segmental ratio calculated in the donor before surgery.
Results: The mean relative graft growth was 54% (+-44%, range -1% to 139%) in male and 48% (+- 35%, range 8% to 130%) in female patients. The mean growth of the anterior graft of all recipients was 47% (male 48%, female 46%) while growth in the posterior graft was calculated at 63% (male 71%, female 54%). 15 of 20 patients demonstrated preferential growth in the posterior graft. In 3 male recipients and 2 female recipients the relative growth of the anterior graft exceeded the relative growth in the posterior graft, indicating an asymmetric preferential regeneration of the anterior graft segments.
Conclusion: In most recipients, regeneration in posterior right lobe graft segments exceeded that in the anterior segments. Impairment of regeneration in the anterior graft segments may be due to compromise of venous outflow secondary to disruption of middle hepatic vein side branches during surgery. Correlation of volume and distribution of potential venous congestion in the postoperative graft with the regeneration pattern observed in this study may influence surgical decision making regarding necessity of venous reconstruction in the individual patient.
224. Comparison of FDG-PET to MRI for Identification of Malignancy in the Liver
Parker R.A.*; Mavi A.; Alavi A.; Radiology, University of Pennsylvania, Philadelphia, PA.
Address correspondence to R.A. Parker (rexp{at}mail.med.upenn.edu)
Objective: The goal of this study was to assess the relative performance of MRI of the abdomen and whole-body FDG PET for the detection of malignant lesions in the liver.
Materials and Methods: Imaging data on 109 patients who underwent both MRI of the abdomen and FDG-PET within 90 days (avg = 25.1d) of each other was retrospectively reviewed for the presence of malignant lesions within the liver. Histopathology (n = 16) or follow-up imaging (n = 93) served as the standard of reference. The sensitivity, specificity, positive predictive value and negative predictive value were calculated for each technique. A McNemar test was used to assess differences between the two methods. A p value of less than 0.05 was considered significant.
Results: Forty patients had hepatic malignancies and sixty-nine
had no malignant lesions according to the standard of reference. Based on a
per-patient analysis, the sensitivity, specificity, and positive and negative
predictive values on MRI were 87.5%, 84.0%, 81.4%, and 92.4%, respectively,
compared to 87.5%, 92.8%, 87.5%, and 92.8% on FDG-PET. The difference between
the two methods was not significant (
2 = 1.45, p = 0.228). When
both imaging modalities were used in combination the sensitivity and
specificity were 92.5% and 95.7 % respectively. FDG-PET identified 27
extrahepatic metastases while MRI identified only 20.
Conclusion: In patients with known or suspected malignancies, MRI of the abdomen and whole-body FDG-PET were comparable in identification of patients with malignant lesions within the liver. FDG-PET had the added advantage of detecting more extrahepatic metastases, a feature that could significantly impact future therapeutic choices. FDG-PET is, therefore, an excellent choice for first-line imaging of patients with suspected hepatic malignancy.
225. Effects of T1 and T2* Relaxation on Liver Fat Quantification Using GRE Sequences
Hughes F.; Bydder M.; Middleton M.S.; Chavez A.D.; Sitarz R.; Znamirowski R.; Hassanein T.; Bydder G.; Sirlin C.B.*; Radiology, University of California, San Diego, San Diego, CA.
Address correspondence to C.B. Sirlin (csirlin{at}ucsd.edu)
Objective: To assess the impact of T1 and T2* relaxation on liver fat fraction (FF) quantification using gradient-recalled echo (GRE) techniques, and to develop a fast GRE method that measures and takes into account T1 and T2* to accurately quantify FF.
Materials and Methods: The most common magnetic resonance (MR) imaging method to assess liver fat is 2-echo GRE imaging. This method attributes changes in signal amplitude at different echo times (TEs) to phase interference between fat and water. Typically, images are obtained at TEs when fat and water are assumed in phase (IP) and opposed phase (OP). Fat fraction (FF) can be quantified from measured signals as FF = (IP - OP)/(2*IP). However, 2-echo GRE ignores relaxation effects, which may cause errors. We modeled the fatty liver as a 2-component system made of fat and water and derived mathematical expressions showing the dependence of estimated FF on flip angle (FA), TE, T1, and T2*. To assess T2* effects, we implemented a rapid 6-echo GRE sequence in 51 patients with liver disease. Data obtained at multiple TEs were fitted to our model to obtain T2*-corrected FF estimations and generate fat and water T2* values. To assess T1 effects, the 6-echo GRE was repeated at many FAs (15-90°) in 15 patients. Data obtained at different FAs were fitted to obtain T1-corrected FF estimations and generate fat and water T1 values. FFs estimated by standard 2-echo GRE sequences were compared to T1- and T2*-corrected FF estimations.
Results: T1 and T2* effects caused the 2-echo GRE to make errors in FF estimation. T1 relaxation led to overestimation. The percent overestimation increased from 10-20% at FA of 15° to 70-110% at FA of 90°. T2* relaxation led to underestimation. For FFs > 0.14, mean percent underestimation was 34%. For FFs < 0.14, 2-echo GREs failed to detect any hepatic fat. The multi-FA, 6-echo GRE sequence accounted for T1 and T2* relaxation and permitted accurate FF estimations. Small amounts of fat (FF < 0.10) could be detected reliably and quantified. Fitted fat and water T1 values were 200-300 and 800-1,000 ms, respectively. Fitted fat and water T2* values were 6±2 ms and 26±6 ms, respectively.
Conclusion: Relaxation causes errors in FF estimation using standard 2-echo GRE methods. T1 causes overestimation and T2 causes underestimation. A rapid multi-FA, 6-echo GRE sequence accounts for and measures relaxation effects and calculates T1- and T2*-corrected FF. This method permits detection of small amounts of fat that were undetectable with 2-echo GRE methods.
226. Helical CT Diagnosis of Liver Injury after Roux-en-Y Gastric Bypass Surgery
Yu J.*; Turner M.A.; Fulcher A.S.; Lai E.; Carucci L.R.; Halvorsen R.; Diagnostic Radiology, MCV/VCU Medical Center, Richmond, VA.
Address correspondence to J. Yu (jyu1{at}vcu.edu)
Objective: Purpose: To identify the cause, CT features and appearance of liver injury following Roux-en-Y gastric bypass surgery.
Materials and Methods: From January 2001 to July 2005, a total of 1,210 patients had Roux-en-Y gastric bypass surgery; 220 with open gastric bypass and 990 with laparoscopic gastric bypass. Of 1,210, 137 patients had abdominal CTs. Abdominal CTscans were obtained in the portal venous phase following administration of oral and IV contrast at 5mm intervals. Two attending abdominal radiologists retrospectively reviewed the CT features in consensus for liver injury. Assessments were made for presence, location, shape and size of the liver attenuation abnormality. CT findings were compared with clinical and surgical findings.
Results: 11 of 1,210 patients had CT evidence of liver injury. All 11 patients had undergone laparoscopic gastric bypass using a Nathanson liver retractor. The extent and appearance of the liver injury corresponded to the shape and size of the retractor. None of the patients with open gastric bypass had liver injury. All injuries involved the lateral segment of the left lobe with areas of decreased attenuation and with a mean size of 5.5 x 2 x 2.5 cm. The liver injury was elongated extending from the inferior surface to the superior aspect of the lateral segment in 9 of 11 patients. One patient developed an abscess within the area of the liver injury and had surgical resection of the lateral segment of the left lobe. Ten patients required no additional treatment and recovered uneventfully.
Conclusion: Liver injury after laparoscopic Roux-en-Y gastric bypass surgery is attributable to usage of the Nathanson liver retractor. Knowledge of the cause, CT features and distribution of the liver injury may be helpful in preventing occurrence of the injury and in aiding diagnosis at CT.
* Will present paper
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
