CardioVascular and Intervenlional Radiology

9 Springer-Verlag New York, Inc. 1999

Cardiovasc Intervent Radiol (1999) 22:169-179

CIRSE '98 S U M M A R Y REPORT

Proceedings of the Annual Meeting of ClRSE Venice, Italy, September 27-October 1, 1998 Mario Bezzi, Rome, Italy Chairman of the Program Committee The scientific program of the 1998 CIRSE Meeting included 326 oral presentations and 111 posters. All aspects of vascular and non-vascular interventional radiology were covered. The presentations related to the most prominent topics are summarized here. The abstract numbers refer to the Main Programme of the Meeting published in CardioVascular and lnterventional Radiology 20 [Suppl 1], 1998.

Vena Cava Filters, Pulmonary Embolism, and Venous Intervention Patrizio Capasso (Lausanne, Switzerland) The 1998 annual meeting included one scientific session with seven, six, and eight oral presentations made in each of the aforementioned subjects, respectively. In addition, one round table discussion dealt with diagnostic modalities of acute pulmonary embolism and eleven posters covered subjects involving the venous system. In the vena cava filters session, Abstract 270 confirmed that the long-term results of the Bird's Nest filter were comparable to those of other filters without any increased inferior vena cava (IVC) obstruction rate. This was demonstrated in a single center long-term follow-up of 78 patients. Recurrent pulmonary embolism occurred in 1.3%. Although wire prolapse was observed in 87% and IVC wall perforation in 85% of patients, the former had no apparent effect on the filtering capacity of the device while the latter produced no symptoms. Abstracts 271 and 272 presented the results from the team in Dortmund who have created an in vitro model of the IVC with a computer-controlled flow system mimicking human venous return. This unique and extremely interesting model can be easily constructed and its design is available from the authors. With this system, measures in filtration capacity were demonstrated to be especially important with multiple-shot testing. Most filters were capable of capturing single emboli but conical filters quickly saturated after receiving multiple emboli and thus became less efficient. Once the apex of the filters was filled, the periphery allowed the passage of clots of substantial size. Filters that segmented the IVC like the Bird's Nest and, to a lesser extent, the Giinther tulip, had the highest capture rates with this test. When

dealing with temporary filters, the instability of the anchoring devices rendered most models precarious, a fact which was proven clinically in Abstract 273 with the atrial migration of the Tempofilter. The Gtinther temporary filter had the most stable metallic anchoring system although, under a heavy load of thrombus, this too demonstrated some degree of displacement. In order to avoid most of these potential complications of temporary filters, Abstract 274 described how the GUnther tulip filter was used successfully for the perioperative, peripartum, and immediate postembolic protection of eight patients. These implanted permanent filters were all successfully removed via the right internal jugular vein route after 2 to 12 days. Abstract 275 demonstrated that filter placement had a tendency to protect patients during thrombolysis in individuals with pelvic and caval thrombotic extension of lower-limb deep venous thrombosis (DVT), although the results were not significant. Finally, Abstract 276 demonstrated that in certain situations, embolic protection, complication rates, and proper positioning of IVC filters remains suboptimal. Today, the armamentarium for the diagnosis of pulmonary embolism (PE) includes spiral CT. The first portion of the session dealing with PE confirmed this. Although there was no angiographic correlation in the majority of cases, Abstract 138 demonstrated that spiral CT could become the method of choice in the initial diagnostic evaluation of patients with suspected PE. Although clots were missed on CT in 8% of the patients who were also evaluated with pulmonary angiography, spiral CT enabled this center to accurately diagnose different pathologic entities in the great majority of cases. In fact, other thoracic pathologies like infectious infiltrates and pleural effusions were observed in

170

42% of cases. Spiral CT was also useful in the follow-up of patients with PE. In Abstract 139, it was demonstrated that a decreasing adapted CT Miller score was observed in all treated patients, correlating well with the improvements in their clinical status. Another method for spiral CT quantification of the severity of PE was described in Abstract 140. This study dealt with the obstruction index, defined as the product of the anatomic site of the clot and the degree of its associated obstruction. The obstruction index was then compared to the Miller index on pulmonary angiography. Linear correlation with the regression model was observed between both techniques and the evaluation of obstruction and perfusion parameters on spiral CT. Massive pulmonary embolism can be treated with mechanical thrombus fragmentation using a pigtail rotation catheter as described in Abstracts 141 and 142. This mechanical disruption of the central clots was demonstrated to require additional locoregional thrombolysis in a significant number of cases (Abstract 141). These studies demonstrated that these relatively inexpensive materials can also recanalize the main pulmonary arteries in about 30% of cases with immediate and dramatic pressure reductions in over 50%. In cases presenting minor PE, systemic heparin therapy usually provides satisfactory restoration of the pulmonary circulation in short-term follow-up without additional locoregional thrombolysis. The presentations in the venous intervention session were varied. In Abstract 181, replacement of tunneled hemodialysis catheters through their existing tract was shown to be safe and easy, without any increased infection rate. With a technical success rate of 100% in 129 catheter exchanges, demonstrating 120-day primary and secondary patency rates of 19% and 81%, respectively, and an infection rate of 0.67 per 1000 catheter days, this technique seems to be an option which may be considered in some situations. Indirect measurements of D-dimers may be performed at the patient's bedside with the agglutination of the patient's red blood cells in the presence of elevated levels of cross-linked fibrin derivatives using the simpliRED test described in Abstract 182. Obviously, this test is rather specific for the presence of DVT when other clinical conditions, like infectious processes, which may elevate circulating fibrin sprit products and thus cause false positive agglutinations, are not present. Acute DVT responded well to locoregional transcatheter thrombolytic therapy applied 30 min after the creation of the DVT although factor Xa antagonist, thrombin antagonist, or iloprost offered no significant benefit to heparin in the porcine model. These results were presented by the team from the University of Oslo (Abstracts 183 and 184). Superior vena cava (SVC) syndrome was successfully treated with metallic stents in benign and malignant disease with immediate symptomatic relief (Abstracts 185, 186, 405). The results in these presentations compared well with those already described in the literature. In one study (Abstract 185), the Memotherm nitinol stent was used in 20 patients, demonstrating good results with an endoprosthesis that has not before been widely used for this indication.

CIRSE '98 Summary Report Similar favorable results were observed in arteriovenous fistulae-related central venous stenoses and obstructions in 42 patients. Although the technical success rate was much lower when dealing with venous occlusions, the primary and assisted 6-month patency rates were similar with values of 54% and 92% respectively (Abstract 187). Finally, it was interesting to observe that venous pseudoaneurysms were successfully treated with uncovered Wallstents in six of nine patients (Abstract 188). The occlusion of the pseudoaneurysms could be associated with endothelial extension along the stent or simple thrombosis from changes in the flow pattern occurring up to 8 weeks after the placement of the stents. The endoprostheses demonstrated secondary patency rates of 86% at 3 months. These lesions were located either in the polytetrafluoroethylene (PTFE) graft or within the venous outflow tract of eight hemodialysis shunts and one transjugular intrahepatic portosystemic shunt (TIPS). Posters also presented favorable results using metallic endoprostheses in the treatment of benign central venous lesions (Abstracts 398, 400, 403). As mentioned before in the literature, PTA alone was found to be suboptimal in the treatment of these lesions in a significant number of patients, thus requiring the placement of stents. With central outflow obstructions of hemodialysis shunts, this treatment gave primary patency rates of 55% to 67%.

Central Venous Access Anthony F. Watkinson (London, Great Britain) The placement of central venous access catheters and ports under radiological guidance is increasing, and is a routine procedure in most departments. This expanding area of work was reflected in an excellent session dedicated to this topic at the CIRSE conference. The main topic centered around preferred access routes (subclavian versus jugular vein), guidance techniques (fluoroscopy, CO 2 or conventional contrast venography, or ultrasound) and the trend away from tunneled catheters towards implantable subcutaneous ports. The internal jugular vein approach was demonstrated to have the highest success rate with lower complication rates (Abstract 42) when compared to the subclavian route, both routes being accessed using ultrasound guidance. The technical success was 100% for the internal jugular route versus 97% for the subclavian route, with the former having fewer passes to achieve access, reduced time, (23 min versus 43 min) and reduced pneumothorax rates (5% versus 0%). In addition, the subclavian vein approach had increased line sepsis, symptomatic thrombosis, and an increased early removal rate. Abstracts 40 and 47 also supported the preferential use of access via the jugular vein using ultrasound guidance. Abstract 46, however, presented findings, albeit in a small group of patients (10) with central venous port access using fluoroscopy and the subclavian vein approach, of no complications related to the puncture or in the early study period. However, in a larger group (Abstract 38) using

CIRSE '98 Summary Report

fluoroscopic guidance for the port implantation in the subclavian vein, 140 successful implantations were performed although the pneumothorax rate was greater than 4% and catheter fragmentation (pinch off peculiar to the subclavian approach?) was reported in two cases. In general, particularly with inexperienced operators, the jugular approach appears to be emerging as the preferred route using ultrasound guidance, with the subclavian route reserved for patients with difficult jugular access. Subcutaneous port implantation in the radiology suite (Abstract 38) appears to have comparable success rate and complications with surgical series and depending on clinician and patient preference may gradually replace tunneled catheter placement. The use of C O 2 venography in the antecubitical vein (Abstract 37) was performed in 2005 patients to guide access via the brachial, basilic, cephalic, or axillary veins. Venous access devices were placed successfully in 99.85% without complications, with iodine contrast medium being required in only 1.7% of patients. Nine patients (0.45%) complained of pain with two (0.1%) procedures being terminated due to severe pain. One patient died (0.05%) from cardiac arrest although it was felt that the patient had cardiac problems prior to the procedttre. This latter statistic raised a question mark in this otherwise excellent series which suggests CO 2 is a good and cheaper alternative to conventional contrast medium if venography is being used as the guidance technique. Fibrin sheath stripping for maintenance of function of central venous infusion ports was supported in one study (Abstract 43). There have been suggestions in the literature that the benefit is limited and patency is difficult to maintain, however this study in 37 patients demonstrated a 90% success in restoring function in obstructed central venous infusion ports after a mean time to malfunction of 8 months. The mean added patency was 183 days (26-598 days). This paper, in addition, suggests improved longevity of central venous infusion ports with reduced infection when compared with tunneled catheters. An interesting paper (Abstract 44) presented the findings in a small group of patients with predominately refractory malignant ascites treated with a Denver peritoneal venous shunt. In a relatively short, simple procedure the shunt was performed radiologically and all shunts functioned well until patient death (maximum 6 months). The theoretical complications of anemia, disseminated intravascular coagulation, pulmonary embolus, and cardiac failure were not observed. This is an interesting procedure which would be easily performed in the radiology suite and may have wider applications particularly in benign patients.

171

ogy. Radiologic study of permanently implanted devices for the management of dialysis patients is being done. Treatment of shunts and grafts was included in the program of the 1998 CIRSE Meeting with one scientific session including nine scientific presentations, some related papers in the central venous access session, one workshop, and eight posters. Abstract 27 reported on percutaneous declotting of native fistulas using aspiration techniques by use of an 8 Fr aspiration catheter. The author included 49 lower- or upper-arm fistulas and reported a technical success of 89% in the forearm and 75% in the upper arm. Stent implantation became necessary in 8 patients. Primary patency after 3 months was 86% in forearm and 75% in upper-arm fistulas. The same authors (Abstract 28) evaluated their technique in the treatment of bridge-grafts achieving a technical success of 100%, and a primary patency of 70% after 3 months. Hagspiel et al. (Abstract 29) reported on directional atherectomy in lesions resistant to percutaneous transluminal angioplasty (PTA). All 16 procedures were successful without procedure-related complications. While Abstracts 30 and 31 dealt with thrombolysis, Abstract 32 reported on the silent pulmonary embolism (PE) rate discovered by pulmonary perfusion scintigraphy after pulsed-spray lysis, comparing the agents of urokinase and heparinized saline. Surprising findings were that baseline scans were already abnormal in 56% of cases, although the authors did not exclude patients with previous embolectomy procedures. Additionally, the silent PE rate was higher if saline was used (50% versus 14%); the difference was, however, not significant and no event became clinically relevant. Abstracts 33 and 34 reported on results of treating central venous obstructions with PTA and stents. Abstract 35 evaluated the use of stents in peripheral veins of arteriovenous shunts in the case of selective indications such as rupture. Poster 396 reported on translumbar access for permanent catheter placement in 28 cases; there were no technical complications but late problems with catheter removal occurred in six cases. Posters 407 and 408 dealt with thrombolysis of grafts and fistulas showing an excellent long-term patency after thrombolysis in native fistulas. Poster 412 described the first and encouraging experience with the new Arrow thrombectomy device in hemodialysis fistulas and grafts. In conclusion the topic of percutaneous management of dialysis shunts focused more on thrombectomy techniques than PTA, showing that there is more interest in a widespread and complete concept in the management of the problematic hemodialysis access.

Hemodialysis Shunts and Grafts

Arterial Interventions: Angioplasty and Related Techniques

Dierk Vorwerk (Ingolstadt, Germany)

Jim A. Reekers (Amsterdam, The Netherlands)

Interventional procedures in failing hemodialysis grafts and shunts are of growing importance in interventional radiol-

During the meeting, five scientific sessions featuring 45 papers covered the topic of angioplasty and vascular stents.

172

Wagner et al. (Abstract 225) addressed the very important topic of bacteremia after a percutaneous intervention. To my knowledge there has never been such an (extensive) report about this much underestimated or even neglected subject. In this well-conducted prospective evaluation of 100 patients it was shown that temporal bacteremia occurs frequently with all invasive radiological procedures, even with simple diagnostic angiography (16%). Although a wide range of bacteria were isolated, there were no infectious complications, despite one patient with cholangitis after a biliary procedure. The important conclusion is that strictly aseptic procedures and antibiotic prophylaxis are necessary in the case of stent placement. With the increase of more complicated endovascular procedures, especially with stent-grafts, this paper could be a guideline for peri- and post-procedural management. Schmitz et al. (Abstract 226) investigated the belief that after percutaneous transluminal angioplasty (PTA) for critical ischemia, there is a deterioration of the microcirculation of the skin on the foot. In this well-conducted study they showed a significant increase in skin cpO 2 after PTA. This increase was not only seen immediately after the procedure but was even more significant after 6 weeks, which again supports PTA for critical ischemia. While most papers reported on the clinical use of several types of stents in different vascular territories, few papers dealt with the problem of restenosis. Link et al. (Abstract 49) reported on their initial experience with the REDHA-CUT atherectomy device. The device was used in 16 patients with arterial restenosis after PTA, located mainly in the superficial femoral artery. After treatment the degree of restenosis was reduced from 89% to 43%. At a follow-up of 6 months there was a substantial clinical improvement, since the walking distance increased from a mean of 114 m to a mean of 785 m. No serious complications were observed. Minaret al. (Abstract 230) discussed the hot topic of brachytherapy to prevent intimal hyperplasia in their study of more than a 100 patients. Although their conclusion was that brachytherapy showed a significant improvement in patency, their results were disputed because the design of their study raised many questions. There was no stratification of patients, clinical category, lesions lengths, or outflow. So finally this presentation did not contribute much to the discussion on brachytherapy. Both session chairmen expressed their feeling about a missed opportunity. Interventional techniques are extremely useful in cases of post-surgical arterial restenosis. Hoksbergen et al. from Amsterdam showed that, within a bypass surveillance program, PTA of a (duplex-detected) restenosis has a 1-year primary patency of 60% (Abstract 228). They also emphasized that bypass stenosis is an important problem which occurs mainly within the first year after surgery. Contrary to other reports they showed that body graft stenoses have a better patency than stenosis at one of the anastomosis sites; 86% versus 40%-48% respectively. Their conclusion was that PTA of bypass stenosis is a worthwhile procedure.

CIRSE '98 Summary Report Acute lower limb ischemia occurring below the knee may safely be approached with endovascular techniques. Degranges et al. (Abstract 232) reported on the treatment of popliteal and/or tibial occlusions with a combination of several techniques: aspiration, thrombolysis, and PTA in a small series. They showed that endovascular treatment of these lesions is safe and has a good (66%) primary cumulative patency with a limb-salvage rate of more than 90%. This presentation shows again that a prospective randomized trial of endovascular techniques versus bypass surgery for limb salvage has a high priority. Covered stents are now part of the armamentarium of the interventional radiologists, and the follow-up confirms the clinical usefulness of these devices. Delcour et al. (Abstract 303) reported on a long-term follow-up of 27 Cragg Endopro systems placed in 25 patients for the treatment of arterial aneurysms, stenoses, and occlusions. Technical success was achieved in 96% of cases. The patients were anticoagulated with heparin during the procedure and received 150 mg of aspirin daily after the procedure. At a follow-up of 2 years, the primary and secondary patency rates were 85% and 96% respectively.

Renal Artery Intervention Anthony F. Watkinson (London, Great Britain) Thirteen abstracts were presented in two sessions on renal artery intervention. Previous literature has suggested that the most challenging renal artery intervention is performed on the solitary kidney. An interesting abstract (Abstract 174) presented findings in 21 patients with solitary kidneys in whom renal artery stenting was performed. Mean follow-up was performed at 15 months with angiography in 76% of the patients. There was a technical success rate of 100% with a clinical benefit in 70% of patients. Interestingly, dialysis was stopped in four patients who were dialysis-dependent. There were no immediate complications associated with the procedure, and no significant stenoses were noted at follow up at 15 months. Abstract 180 provided further data on the role of renal artery stents with a high initial, technical, and clinical success rate, although there was only 3-month follow-up. Abstract number 266 demonstrated high initial technical success with the AVE balloon-expandable bridge stent, although, again, long-term follow-up with this new stent is not yet available. Abstract 269 demonstrated clinical follow-up from 1 to 4 years in 53 patients who underwent angioplasty and renal stent placement. Interestingly after high initial technical success, percutaneous transluminal renal angioplasty (PTRA) and stent placement was most clinically useful when used in patients with progressive renal failure whose creatinine was still above 400 /zmls/L. In addition there was sustained improvement in blood pressure in 55% of patients treated.

CIRSE '98 Summary Report

Several authors (Abstracts 175, 177, 178, 264) addressed the difficult problem of transplant artery stenosis. Abstract 264 reviewed the results of 83 angioplasties performed in 55 patients in a cohort of 1448 patients who had undergone renal transplantation over a period of 16 years. There was a high technical failure rate due to difficult access or failure to cross the stenosis. In those in whom the stenosis was traversed there was approximately 60% radiographic success, with only 20% gaining clinical improvement. In addition there was a complication rate of 16% with either arterial dissection or occlusion. This very honest presentation demonstrates the difficulties of renal transplant angioplasty, however they emphasized that although this may be a high-risk procedure, it remains the first-line treatment prior to surgical reconstruction. In view of the length of the series the results are skewed by the early cases, however renal artery transplant stenosis often has difficult access and lesions may not dilate despite high inflation pressures. The technical failures did not appear to relate to early restenosis or dissection which would have been salvagable by stent placement. Abstracts 175 and 178, however, both present a small series of patients in whom angioplasty failure in the renal transplant anastomotic stenosis was successfully treated by stent placement. Both demonstrated good clinical results with greater than 90% technical success rate, 90% improvement in blood pressure, and 60% in whom renal function was stabilized. In both series the restenosis rate at mean follow-up of 2 years in one series and 34 months in the other series, demonstrated only one restenosis out of 13 patients. This success may be related to the high degree of immunosuppressive medication the patients were receiving. Stent implantation in patients with renal artery transplant stenosis in whom angioplasty has failed may become a good option if the stent is kept short to give the option of surgical reconstruction should the stent fail. Abstract 177 presented the results of angioplasty in 14 patients with renal graft stenosis. It documents some improvement in hypertension but no significant effect on renal function. The complication rate appeared to increase in patients with multiple renal arteries. An interesting aside to this paper was the suggested use of Gadolinium or carbon dioxide as alternative contrast agents in patients with deteriorating renal function. Abstract 179 assessed the results of 18 patients who underwent simultaneous bilateral renal artery intervention, which appeared to be successful with no patients undergoing complications. Historically, this procedure has not been performed as a single procedure but as a two-stage procedure due to the risk of renal artery damage as a complication of renal artery angioplasty. Results from a larger study group are required. However, if this procedure can be proven safe then it would certainly save the patient from having two procedures. Two papers (Abstracts 265,267) indicate a role for intravascular ultrasound (IVUS) in both assessing the significance of a stenosis pre-percutaneous intervention in the renal artery in which other imaging modalities have proved inconclusive and also to assess the outcome follow-

173

ing stent placement to see if stent expansion is satisfactory. A severe limitation of both these studies is that no correlation with pressure measurements was performed. This is unfortunate as most clinicians in the United Kingdom still consider IVUS a research tool which is too expensive to consider as a viable option on every patient. Abstract 265 suggests additional intervention on patients in whom stent expansion appeared incomplete on the basis of IVUS, with no correlation as to whether there was a significant pressure drop to indicate if percutaneous intervention was indicated. In addition, abstract 267 supported the use of IVUS as 20 significant stenoses were identified in 30 patients in whom other imaging modalities were inconclusive. However, it appears that simple catheter pressure measurements which are much cheaper would also performed the task.

Arterial and Venous Thrombolysis Hans-Joachim Wagner (Marburg, Germany) Two scientific sessions were dedicated to this topic, one to arterial thrombolysis and thrombectomy (Abstracts 189, 197), one to venous thrombolysis (Abstracts 277, 283). The session on arterial thrombolysis this year was dominated by pharmacological fibrinolytic techniques, which represented eight of the nine communications. Only one paper dealt with a solely mechanical approach to thrombectomy to achieve clearing of the occlusive material. Ronnen et at. (Abstract 189) showed the value of a balloon-tipped sheath to prevent downstream embolization during local mechanical thrombectomy procedures in acute and subacute iliac artery thrombosis. Distal embolization could be prevented in all treated cases. The technique seems to be promising in dealing with thromboembolic occlusions above the inguinal ligament, a field that currently is often managed by open surgery. The presentations on the fibrinolytic strategy to dissolve thrombus dealt with established techniques and known fibrinolytic agents. Tsetis et al. (Abstract 190) used high boluses of r-tPA to lyse subacute arterial thromboses with a success rate of >80%. Only two minor complications were noted. Siablis et al. (Abstract 197) evaluated the effect of pulse-spray lysis in acute, subacute, and chronic ischemia. The median time of lysis was shorter with r-tPA than with urokinase. The overall results gave a limb salvage rate of 79% at one year. A retrospective analysis of urokinase lysis for acute graft thromboses in 77 patients by Geschwind et al. (Abstract 192) achieved a success rate of 69% in reopening the graft and unmasking an underlying lesion. In all these cases a successful outcome was achieved, whereas in the cases with unsuccessful reopening of the graft, the eventual outcome was unsuccessful and resulted in amputation. This study again gave evidence for the predictive value of a successful lysis for the overall outcome. Two papers dealt with the combination of a pharmacological and a mechanical technique for the aim of thrombus

174

dissolution. Pickel et al. (Abstract 191) used mechanical thrombolysis together with local pharmacological thrombolysis and retrospectively analyzed their results on more than 100 patients. They could demonstrate that a combination of mechanical and pharmacological techniques could shorten procedure time. A combination of fibrinolysis and mechanical devices to aspirate and fragment thrombus was also used by Bulvas et al. (Abstract 194) in over 200 patients. The recanalization rate was 74%. The group of Krause et al. (Abstract 193) stratified their analysis of low-dose urokinase treatment according to patient age (above and below 80 years). The results demonstrated a worse prognosis for patients over 80 years with regard to patency rate, limb salvage, and survival. The outcome was mainly influenced by the possibility of restoring outflow vessels at the end of fibrinolysis. Two papers addressed the thrombolytic treatment of acute occlusions of popliteal artery aneurysms. Holtzmann et al. (Abstract 195) and Greenberg et al. (Abstract 196) showed that an initial treatment with fibrinolysis to reopen the occluded aneurysmatic segment in conjunction with immediate vascular surgery after recanalization allowed limb salvage in 80% and 100%, respectively. During the session on venous thrombolysis several excellent experimental papers were presented. Roy et al. (Abstracts 277, 278, 281) focused on local pharmacological fibrinolysis in deep vein thrombosis. The group had created a porcine model of acute deep vein thrombosis. In one trial (Abstract 277), different application devices were tested. A nipple balloon catheter used to infuse t-PA achieved superior results to a sequestration technique using a multi-sided port infusion wire. In another trial (Abstract 278), the influence of the route of delivery of the fibrinolytic agent was investigated. A superior efficacy was demonstrated for partitioning the total dose of t-PA for intravenous and intraarterial administration together versus the intraarterial application or the intravenous application alone. In a third trial (Abstract 281), urokinase versus t-PA or t-PA plus saruplase to achieve venous thrombolysis was studied. There were no significant differences for the different fibrinolytic agents with regard to thrombus reduction, residual thrombus mass, or residual thrombosis in tributaries. Another experimental paper was presented by Wildberger et al. (Abstract 283). The ability to retrieve venous thrombi through an 18 Fr sheath with a balloon catheter and a rotating pigtail catheter aimed to fragment the thrombus was evaluated in an in vitro flow model. Within 10 min all thrombi could be removed without any downstream embolization. An interesting clinical paper was presented by Raynaud et al. (Abstract 280). The group treated three patients with complete renal transplant vein thrombosis with percutaneous thrombus aspiration only. In all three cases the vein could be reopened and the transplant function was restored. Although the experience is limited, this seems to be a new, valuable technique to preserve the renal transplant and avoid invasive measures. A larger patient population of 92 patients with iliofemoral deep vein thrombosis was treated with fibrinolysis in combination with

CIRSE '98 Summary Report

PTA and stenting by Bjarnason et al. (Abstract 282). The authors placed 98 stents in 45 patients (>50%). The bleeding complication rate was 16%. The immediate success (87%) was much better for fresh (<4 weeks old) than for older thrombi. Two-year primary and secondary patency rates were 60% and 80% respectively. It seems that the role of local thrombolysis for the management of DVT still has to be defined. In conclusion, the scientific sessions on arterial and venous thrombolysis presented interesting new experimental data. In the clinical field we are awaiting prospective randomized trials to better define the role of pharmacological as well as mechanical thrombolysis in both arterial and venous thrombotic occlusions.

Aortic Stent-Grafts Herv( Rousseau (Toulouse, France) Three scientific sessions, incorporating 27 papers (Abstracts 83-91, 92-102, 292-296, 299-300), five posters, and one state-of-the-art lecture, were dedicated to this topic. Two national multi-center studies from Austria and the UK, the Eurostar and European registry, and one state-ofthe-art were dedicated to the use of stent-grafts in the treatment of abdominal aortic aneurysms (AAA). All concluded that endovascular repair in good risk patients is associated with a low morbidity and a satisfactory success rate, with a technical success rate of 97% and less than 3% surgical conversion for the Austria study (Abstract 087). Major complications were observed in 10.6% in this study. Leaks remain the main drawback of this new investigational method. Primary leaks are observed in 2%-9% and originate in the limbs with inappropriate diameter of the iliac artery or from reperfusion via collaterals (i.e., lumbar or inferior mesenteric arteries). Secondary failure may occur by de novo reperfusion via collaterals, distal migration, or stentgraft kinking. These latter complications could be devicerelated or secondary to morphologic changes of the AAA sizes after complete exclusion. In fact, after exclusion the transverse and longitudinal retraction of the aorta can change the position of the different components of the stent-graft with a risk of limb thrombosis, leaks, or stent fracture. For the group from Austria (Abstract 087), the mean primary success, defined as AAA exclusion, was 84%. The secondary success rates obtained by spontaneous thrombosis, embolization, or stent-graft insertion, were 97% and 100%, respectively, at 6 and 24 months. Edwards et al. (Abstract 091), on behalf of the Eurostar steering committee, presented the results of this registry and compared two patient groups (before and after 1996). Evolution of the procedure was reflected by a shorter procedure time, lower perfusion requirement, and shorter hospital stay, but, surprisingly, the intraoperative rate of procedural complications was higher in the later group. This may be explained by a higher acceptance rate of more complex AAA

CIRSE '98 Summary Report

patients for endovascular repair. Finally, all these authors stress the importance of good patient selection to reduce the complication rate and the need for close surveillance. The Malmo group presented three different abstracts (Abstract 90, 98, 100). By comparing the proximal aortic diameter at the level of renal arteries by CT before and after AAA repair by open surgery or endovascular methods, they found an increase of aortic diameter in both groups. They concluded that stent-grafts will benefit from oversizing, suprarenal insertion, and improved mechanisms of stent-graft fixation to prevent migrations and endoleaks. Several other authors reported on their results with various types of stent-grafts for AAA. They confirmed the feasibility for clinical application, but, generally, numbers were too small and follow-up time too limited to draw definite conclusions. Four abstracts demonstrated that stent-grafting of the thoracic aorta is emerging as a new technique in aneurysms of various etiologies, in traumatic ruptures of the aortic isthmus, and in complicated aortic dissections. The Jikei University of Tokyo and the Stanford team Abstract 85) reported a series of 47 thoracic aneurysms combined with a history of prior or simultaneous surgical repair of AAA treated by transluminal stent-graft placement. Their results, with a 98% technical success and a mortality and paraplegia rate of 8.5%, supported the hypothesis that endovascular stent-graft placement may be a safe and durable treatment for selected patients with aneurysmal disease of the descending thoracic aorta. The group from Malm6 (Abstract 86) reported a critical reappraisal of their experience with thoracic grafts. They observed mortality rates of 23.5% and 37.5% for elective-repaired aneurysms and ruptured aneurysms, respectively. They showed that endovascular exclusion is technically feasible but associated with significant morbidity. They showed also that rigidity of the stent-graft could be a problem because the inability to adhere securely to the aortic arch might cause a perigraft leak. Endovascular stent-grafts, in combination with aggressive medical management to reduce cardiac shear forces, were also investigated for elective therapy of thoracic aortic trauma (Abstract 84). The authors demonstrated that endovascular treatment of acute rupture has multiple theoretical advantages: a simple arteriotomy is performed without thoracotomy or aortic clamping which, theoretically, lessens the risk of medullar ischemia and, as stent-graft insertion can be done with a mild anticoagulation, the major bleeding complications observed with a full surgical heparinization could theoretically be avoided. The aneurysm frequently originates just beyond the left subclavian artery with an insufficient proximal anchoring length to have safe support of the stentgraft on healthy aortic wall; this was resolved by different means: (1) by intentionally putting the non-covered part of the stent over the ostium of the left subclavian artery, allowing the polyester covering just after this ostium; (2) by creating a small pre-implant window in the polyester in front of this ostium during manufacturing. By these techniques no

175

arm ischemia or neurologic complications were observed by the authors and it was not necessary to create a carotid-tosubclavian arterial bypass graft before or after stent-graft insertion. Kato et al. from Japan (Abstract 102) demonstrated that in dissections involving the descending thoracic aorta, stentgrafting also offers an interesting alternative for occluding the primary entry tear or the residual false lumen at the level of the descending aorta, in order to avoid aortic dilation and rupture. In conclusion, aortic stent-grafts were a major subject of CIRSE in Venice. More consolidated data from multi-center trials were presented, but no randomized study to compare this new investigational method to conventional surgical therapy were published.

TIPS and Portal Vein Interventions Josd I. Bilbao (Pamplona, Spain) On the Meeting programme there were 27 presentations on the topic "TIPS and Portal Vein Interventions". There were two workshops, one guest lecture, four posters, and twenty oral presentations. A large number of the presentations on TIPS discussed various technical aspects. Most of the authors did not use any kind of complementary guidance during the transhepatic advancement of the needle. Others (Abstract 113) recommended wedged injections of CO 2 in the hepatic vein, which yield excellent images of the portal vein; their quality can be regarded as excellent, and the results are comparable to those obtained by direct portography. Others used ultrasound (Abstract 104) to reach the portal vein. Rose et al. discussed the use of an innovative 3-D ultrasound technique by means of which, before the intervention is performed, one can obtain a sonographic volume of data which can be used to optimize the selection of the hepatic vein, the direction and angle of the needle, etc. (Abstract 200). Several papers dealt with the use of different designs of prosthesis for the portohepatic communication. Krajina et al. (Abstract 205) recommended the use of the spiral Z-stent in cases with straight paths between the two veins (71% of their patients); for curved paths which may provoke kinking, flexible prostheses were recommended. In these authors' view, there are no differences in clinical results, as far as patency and efficiency are concerned, but the spiral Z-stent is generally much less expensive than other self-expanding designs. The Turin group (Abstract 104) performed 224 procedures (66% of their series) using the Memotherm prosthesis. Primary patency at 12 months was 71%, while assisted primary patency over the same period was 95%. There was a 14% incidence of encephalopathy, and there was no case of migration of the stent. Pampana et al. (Abstract 204) presented a series of 43 consecutive cases in which an Ultraflex prosthesis had been inserted, with results which, according to the authors, resembled those in other series with different

176 prostheses as far as patency and encephalopathy were concerned. However, in two cases there was stent dislodgement during the procedure, probably due to the low radial force of this prosthesis. Of particular interest was the study presented by Bustamante et al. (Abstract 394) comparing the technical, morphological, and clinical results obtained in two groups of similar patients treated with the Strecker (22 patients) and the Memotherm (21 patients) prostheses. Although no concrete figures were given, the authors concluded that their experience had not brought to light any differences between the two stents. Rosenblum et al. (Abstract 110) presented a series of pediatric patients (aged 2.5-16.8 years) on whom TIPS had been performed. They stated that this technique, used as a bridge to liver transplantation, is just as feasible in this population group, and has very low morbidity. Zajko et al. (Abstract 199) presented the application of TIPS in another interesting subgroup of patients, namely liver transplant recipients with recurrence of symptoms (gastrointestinal hemorrhage or refractory ascites). In the 13 cases in their study, no particular technical difficulties were noted, and the results were good for control of hemorrhage from varices, but not for the treatment of refractory ascites. In the latter case, generally Child C patients, survival was poor; retransplantation was recommended. Without focusing particularly on the group of patients who were given the TIPS or the type of prosthesis used, there were two presentations in which various concrete technical recommendations were made for carrying out the procedure. Although dissections of the portal vein are unusual, their consequences can be very serious, and so Lazar et al. (Abstract 395) recommends that particular attention should be paid to their early detection, and once they have been detected, the damaged portal area should be covered immediately with a second prosthesis. Tesdal et al. (Abstract 202) carried out an interesting classification of his 84 patients into five groups according to the morphology and characteristics of the gastroesophageal collateral varices, and came to several interesting conclusions. First, there is no correlation between the portosystemic pressure gradient and the degree (size, morphology, etc.) of the varices. Second, the patients with intrathoracic varices (groups II and IV in their classification) presented more severe varicose bleeding than the others. Lastly, the patients whose varices were embolized (using Ethibloc and coils) during the TIPS procedure had a lower incidence of bleeding than those in whom no embolization was performed. In the technical section, it is interesting to recall the efforts which are being made in research to find new means of percutaneous portosystemic decompression. Kaminou et al. (Abstract 109) presented an interesting study in which they described the percutaneous retroperitoneal splenorenal shunt (PRESS) as an alternative technique to TIPS. Regarding the clinical results, there were three series which were focused particularly on this issue. The Lille group (Abstract 201) carried out a lengthy study of their first

CIRSE '98 Summary Report

101 patients (90% Child B and Child C, and 80% with alcoholic cirrhosis), in which they reported a 35% incidence of obstructions, with primary patency of 55% after 1 year and 32% after 4 years; the secondary patency was 85% after 1 year and 82% after 5 years. Three-year survival for the Child C cases was 33%, significantly lower than for the rest of the series (72% for the same follow-up). Recurrent hemorrhage occurred in 10% of cases, recurrent ascites in 18% and encephalopathy in 20%. Rousseau et al. (Abstract 103) presented the data from a multi-center randomized study comparing the results of sclerotherapy and TIPS in patients with Child C who presented with digestive hemorrhage. Rebleeding rates were 21% for TIPS and 48% for sclerotherapy. The mortality rates (46% and 49%) were similar. For Child B and C with active hemorrhage, the first-line treatment recommended is sclerotherapy, with TIPS reserved for cases with rebleeding. Finally, Textor et al. (Abstract 114) presented the results of using TIPS in the difficult management of patients with hepato-renal syndrome (16 patients); in their view, these patients improve significantly, with hemodialysis being suspended in some cases, and a significant improvement in analytical data (creatinine, urea, etc.) occurs. They also observed an improvement in survival, although it was only 56% at 1 year. One problem in the area of TIPS which awaits solution is the high rate of restenosis and the need for reintervention which is still essential to ensure good secondary patency. Hausegger et al. (Abstract 106) presented a study in which they analyzed the frequency and site of stenosis. They observed 78% stenosis, which was more frequent in the parenchymal tract than in the hepatic vein; stenotic relapse tends to appear again at the same site as the first occurrence. As one third of stenoses had clinical repercussions, the authors recommend check-ups at least every 6 months, at which time portography should always be performed, with measurement of the portosystemic gradient. Although stenosis can occur after 24 months, its frequency declines after this point. Another study (Abstract 105) analyzed the incidence of restenosis in the long-term follow-up of patients with TIPS; no differences were found between the frequency or time of occurrence of intraparenchymal and hepatic vein stenoses. It would seem that there is a clear (significant?) drop in the appearance of restenoses in the follow-up after 24 months, with a reduced need for further operations after this point. The histological substrate which causes recurrence of stenosis is still a subject of great importance. The data in the series by Rossi et al. (Abstract 107) are interesting in that they discuss the histological specimens from nine patients with TIPS who were never treated with new prostheses or balloon dilatation. In no case did they observe bile contamination, and the morphology (amount, thickness, etc.) of the connective tissue was directly related to the age of the TIPS. The prosthetic wires were always surrounded by giant multinuclear cells.

CIRSE '98 Summary Report

If the cause of recurrent restenosis or premature occlusion is bile contamination, one study (Abstract 108) analyzed the levels of circulating bile salts as a possible method for early detection of this complication; the results were not satisfactory since there was no difference between either the place at which the sample was taken or the moment at which this procedure was carried out. Two groups (Abstracts 108, 111) presented the results they obtained with metal prostheses covered with Dacron (stent-grafts) in TIPS, both as the initial prosthesis and as a means of delaying overgrowth of the neointima. Although it was demonstrated both angiographically and histologically that overgrowth of the intima is clearly delayed, signs were also found that bile was passing inside the shunt. Even though the series were not large, the primary patency was low and there was no evidence (Abstract 111) that these prostheses offer an improvement on the clinical results obtained with non-covered prostheses. In patients who have previously undergone abdominal surgery, gastrointestinal bleeding may be related to the presence of segmentary obstruction of the portal or superior mesenteric vein which result in pre-hepatic portal hypertension (Abstract 198). According to these authors, management of bleeding from these varices is difficult and the prognosis in such cases is poor. As Escalante et al. indicate, in segmentary stenosis and obstructions of the portal vein, the insertion of endoprostheses was technically feasible, offering good results for patency in the medium and longterm follow-up (Abstract 339). Together with the insertion of an endoprosthesis, Marita et al. (Abstract 206) recommend the embolization of the collaterals to ensure hepatopetal flow and patency of the prosthesis.

B nterventions i land i aGastrointestinal ry Wojciech Cwikiel (Lund, Sweden) One state-of-the-art lecture, 15 abstracts for oral presentation, and eight posters illustrated the progress of interventions in the gastrointestinal tract. Biliary interventions were mentioned in 12 abstracts for oral presentation and eight posters. Palliation of dysphagia depends on the etiology of the esophageal disease. Good results have been achieved with dilation using large balloons in patients with achalasia of the esophagus (Abstract 245). About 25% of patients required a repeat procedure, but there were no serious complications. Epithelial hyperplasia impairs the results of stent-treatment in patients with benign esophageal disease and is more frequent at the edges of the stents. In the case of benign esophageal disease stenting should be avoided, and in selected cases should be performed preferably with uncovered stents (State-of-the-Art IV, Abstract 235). Ingrowth of tumor through the stent mesh is a well-known disadvantage of uncoated stents. For palliation of dysphagia in patients with malignant esophageal disease covered stents

177

are preferred to avoid ingrowth of the tumor. However, the major drawback of covered stents is migration, a problem that has not been completely solved (State-of-the-Art IV). An attempt at preventing tumor ingrowth by coating of the stent wires with Paclitaxel, a drug that inhibits neoangiogenesis and therefore tumor growth, did not have significant effect in the presented preliminary study (Abstract 234). In tumors near the cardia, metallic stents protruding into the stomach may cause gastroesophageal reflux. A simple antireflux valve to prevent this was presented by Kocher et al. (Abstract 238). In 18 patients, they were able to prevent reflux in all cases as shown by follow-up barium swallow studies. Percutaneous gastrostomy performed under fluoroscopic guidance, was successfully established in 99% of 500 patients, and had, as expected, a very low rate of complications (Abstract 244). Ultrasonography-guided percutaneous jejunostomy, recommended by authors for patients with nutritional problems (Abstract 240) had an 81% success rate, a 19% complication rate, and no procedure-related deaths. Obstruction of the gastric outlet and duodenum may be treated with the insertion of self-expanding stents. Experience with such treatment is still limited, but the preliminary results presented in small series (Abstracts 237, 239) are promising. A new interesting method for creation of an anastomosis between two parts of the bowel, using magnets, was evaluated in an experimental study (Abstract 137). Survival of patients with malignant biliary disease is generally limited to about 4-5 months (Abstracts 17, 145). Although many authors prefer metallic stents for palliating malignant biliary obstruction, Orlacchio et al. reported that the mean patency of a percutaneously inserted Carey-Coons endoprosthesis was as long as the mean survival of the patient (Abstract 17). The authors did not report any serious complications secondary to insertion of this plastic endoprosthesis. Covered stents may represent an alternative to conventional metallic stents in the management of malignant jaundice. Out of 15 patients with malignant biliary obstruction treated by insertion of covered stents only two stents (13.3%) become occluded (Abstract 14). The stent patency rate was 86% at 6 and 12 months. No stent migration was reported. The extremely high cost of a covered metallic stent should be discussed on the basis of results in comparison to the low cost of a plastic endoprosthesis. The results of uncovered Wallstents were analyzed in a large series of 320 patients with malignant biliary obstruction treated in a single center over a period of seven years. Only 13% of the uncovered stents became occluded (Abstract 145). In this study performed in a group of unselected patients, however, the 30-day mortality was 27%, higher than in other previously published studies, and the mean survival was 138 days. The short survival may have increased the patency rate, since patients probably have not lived long enough to have their stent occlude. The use of metallic stents is usually not indicated in patients with benign strictures. However, a group of patients that may benefit from metallic stent im-

178 plantation is represented by patients with biliary strictures after liver transplantation. Sergent et al. (Abstract 148), in a group of 12 patients treated with metallic biliary stents after transplantation, reported primary and secondary patency rates of 60% and 90%, respectively.

Interventional Procedures in Liver Neoplasms Paolo Ricci (Rome, Italy) During the CIRSE meeting two sessions, including 16 papers (Abstracts 115-125, 247-251), covered the topic of liver tumors, by presenting experimental studies and clinical applications. Experimental papers explored the use of new agents for chemoembolization and the possibilities of isolated liver perfusion. Treatment of hepatic tumors by combining a nitrogen oxide inhibitor with a vasoconstrictor agent, prior to selective embolization, was presented in a murine model (Abstract 115). The results indicated that the combination of these agents attenuates the tumoral growth by effective obstruction of microcirculation. Two groups from Sweden reported their experimental evaluation on total isolated liver perfusion in animal models. Murata et al. (Abstract 116) demonstrated that a percutaneous approach is technically feasible and that the hemodynamic changes may be helpful in improving the treatment of hepatic malignancies. Harnek et al. (Abstract 247) showed the effectiveness of a new double-balloon catheter inserted with a transjugular approach in the portal and inferior caval veins. Four papers dealt with the clinical results, the possible complications, and the follow-up of chemoembolization of hepatocellular carcinoma (HCC). Abstract 117 reported the efficacy and safety of transarterial chemotherapy with degradable starch microspheres (DSM), in comparison with intraarterial chemotherapy (IAC). In a group of 33 patients the tumor response rates were as follow: 3 partial remission (3 DSM vs 0 IAC), 20 stable disease (12 DSM vs 8 IAC), 10 tumor progression under therapy (2 DSM vs 8 IAC) and 2 tumor progression in DSM group after therapy was stopped. Three patients died in each group due to causes not directly related to the therapy. The real advantage of this new embolic device needs to be validated by further studies. The efficacy and safety of intraarterial hepatic I-131 Lipiodol for the treatment of patients with HCC and portal vein thrombosis was also investigated (Abstract 118). In a group of 23 patients the response to treatment was partial in 3 patients, the disease was stable in 12, and progressive in 8. Survival rates were respectively 70%, 33%, 12%, and 0% at 3, 6, 9, and 12 months. The conclusion was that this kind of treatment is, at present, not valid for primary liver tumors in the presence of portal vein thrombosis. The use of contrast-enhanced Doppler sonography as a valid method in the follow-up of chemoembolization was reported by Catalano et al. (Abstract 120). Another study

CIRSE '98 Summary Report (Abstract 123), from the Universities of Rome and Pisa, assessed the efficacy of contrast-enhanced US in monitoring the response of HCC treated by percutaneous ethanol injection (PEI). A new follow-up protocol was proposed, including contrast-enhanced US after each PEI session, retreatment until complete disappearance of color flow signals, and spiral CT as a final control. The general idea was that the availability of US contrast media can really improve the value of color Doppler US in assessing the outcome of therapies. Finally Matsura et al. (Abstract 119) stressed the incidence of hepatic bilomas after chemoembolization in 190 patients (4.2% in their series). Three papers were devoted to tumor ablation, both in liver and pancreas. Gillams et al. (Abstract 121) reported their experience in over 200 patients, with an average liver lesion size of 4 cm, with both interstitial laser photocoagulation and radiofrequency ablation. Their conclusion was that there is a linear relationship between the ablated volume and the total energy deposited for laser, that is approximately 1 cc of ablated volume per 1000 Joules. The effectiveness of treatment is enhanced by preliminary use of ethanol; this is particularly true for the most vascular lesion where PEI determines thrombosis of the tumor vasculature. The same group dealt with the debulking of small pancreatic tumors by means of percutaneous needle photodynamic therapy (Abstract 124). Seven patients were treated after sensitization with the agent MtI-IPC. The effectiveness of the treatment was assessed by dynamic contrast-enhanced CT and the results were that the area of tumoral necrosis was 13-18 mm in size and that no clinically relevant complications occurred. The impact of this procedure on survival should emerge rapidly from a larger series of patients with longer follow-up. Roche et al. (Abstract 122) presented the feasibility and safety of percutaneous radiofrequency ablation with cooled-tip electrodes for the treatment of hepatic metastases in 10 patients with 13 lesions. In this series all lesions were completely avascular on CT scans 2 months after the procedure and no treated lesion recurred during a 60-330 day follow-up. During this period distant metastases, located in the lung and the liver, occurred in five patients. Mancini et al. (Abstract 125) reported on 100 patients with liver metastases or HCC in whom a catheter was percutaneously implanted, via an axillary approach, in the hepatic artery for continuous intraarterial chemotherapy. The implantation procedure was successfully performed in all patients and serious complications occurred in nine cases (3 pseudoaneurysms of the axillary artery and 6 of the hepatic artery). Two other papers dealt with intraarterial chemotherapy in the treatment of metastatic liver disease. Cyjon et al. (Abstract 248) stressed the value of regional intraarterial treatment in patients with metastases from colorectal cancer in case of failure of conventional treatment. Grasso et at. (Abstract 250) reported their experience in locoregional high-dose intraarterial chemotherapy via a transaxillary approach, without implantation of a subcutaneous port and without embolization procedures. The conclusion was that

CIRSE '98 Summary Report

implantable devices represent the best option for intraarterial treatment of liver metastases. The last paper of the second session on liver tumors, by Goffette et al. (Abstract 251), demonstrated in 11 patients a significant increase in the volume of the left hepatic lobe

179

after percutaneous occlusion of the right portal vein with cyanoacrylate mixed with Lipiodol. The procedure allowed the performance of extended right hepatectomy in cases where the future remnant liver volume was initially of insufficient volume.