Large-Bore Nitinol Stents for Malignant Superior Vena Cava Syndrome: Factors Influencing Outcome

Transcription

1 Vascular and Interventional Radiology Original Research Maleux et al. Treatment of Malignant Superior Vena Cava Syndrome With Large-Bore Nitinol Stents Vascular and Interventional Radiology Original Research Geert Maleux Patrick Gillardin Steffen Fieuws 2 Sam Heye Johan Vaninbroukx Kristiaan Nackaerts 3 Maleux G, Gillardin P, Fieuws S, Heye S, Vaninbroukx J, Nackaerts K Keywords: angioplasty, malignant, stenosis, venous stent DOI:.224/AJR Received July, 22; accepted after revision January 7, 23. Department of Radiology, University Hospitals Leuven, Herestraat 49, 3 Leuven, Belgium. Address correspondence to G. Maleux (geert.maleux@uzleuven.be). 2 Interuniversity Institute for Biostatistics and Statistical Bioinformatics, Katholieke Universiteit Leuven and Universiteit Hasselt, Leuven, Belgium. 3 Department of Pneumology, University Hospitals Leuven, Leuven, Belgium. AJR 23; 2: X/3/ American Roentgen Ray Society Large-Bore Nitinol Stents for Malignant Superior Vena Cava Syndrome: Factors Influencing Outcome OBJECTIVE. The purpose of this article is to retrospectively evaluate the technical and clinical outcomes of large-bore nitinol stents for treating malignant superior vena cava syndrome. In addition, we analyzed factors potentially influencing the outcome. Materials and methods. Over a 7-year period, 78 consecutive patients presented with superior vena cava syndrome related to primary lung tumor (n = 62) or malignant lymphadenopathies (n = 6). The factors analyzed were Kishi score at admission, tumor type, and need for an additional balloon-expandable stent. RESULTS. Technical success was obtained in all but one patient (99%), who presented with a stent migration immediately after insertion. In 7 patients (22%), an additional balloon-expandable stent was needed for complete expansion of the nitinol stent. For patients with symptomatic malignant lymphadenopathies or primary lung tumor, overall survival rates were 5% (n = 8) and 54% (n = 34), respectively, at 6 months and 9% (n = 3) and 34% (n = 2), respectively, at 2 months (p =.376). There was no difference in survival as a function of the Kishi score (p =.8) or of the placement of an additional balloon-expandable stent (p =.35). Finally, reocclusion events were noted in patients both with (n = ) and without (n = 7) a balloon-expandable stent. CONCLUSION. Large-bore nitinol stents are highly effective for malignant superior vena cava syndrome. The survival rates of patients with caval vein stenosis due to either the primary tumor or secondary enlarged adenopathies were equal. An additional balloon-expandable stent was required in 22% of cases owing to incomplete expansion of the nitinol stent but was not associated with higher thrombosis rate. S uperior vena cava (SVC) syndrome is caused by the obstruction of the venous blood flow from the upper body (including the head) into the right atrium. In most cases, the underlying cause is a malignant tumoral mass compressing the SVC or one or both brachiocephalic veins. However, a benign origin of the SVC syndrome is becoming more and more frequent, mainly because of the increased number of patients with indwelling hemodialysis catheters, port catheters, pacemaker leads, and other devices [, 2]. Typical symptoms are associated with venous hypertension of the upper body and head, resulting in symptoms of congestion and edema of the face and upper thorax. Other symptoms may include dyspnea, dysphagia, severe headache, and even cognitive dysfunction. For many years, palliative treatment of SVC syndrome consisted of radiation therapy or chemotherapy. However, true symptom relief can be obtained only several weeks after the start of chemoradiotherapy. In addition, in the case of chemotherapy with platinum-based drugs, sufficient hydration during and after the chemotherapy infusion is required for safe drug administration. However, this hydration may aggravate the SVC syndrome and finally contraindicate continued chemotherapy [3]. Endovascular stent placement for the palliation of malignant SVC syndrome is a minimally invasive alternative to radiotherapy and chemotherapy and is increasingly being regarded as the first-line approach, because of the rapid and efficient relief it achieves of all venous flow obstruction symptoms [3 6]. In most of the publications dealing with endovascular management of malignant SVC syndrome, stainless steel stents, such as the Gianturco Z stent (Cook Medical), the Wallstent (Boston Scientific), or the balloon-expandable Palmaz stent (Cordis), have been used [7]. However, each type of these vascular stents may have some drawbacks: the Gianturco Z stent is rigid, with wide gaps between the stent wires, which allow tumor AJR:2, September

2 Maleux et al. TABLE : Type of Tumors Associated With Superior Vena Cava Occlusion TABLE 2: Kishi Classification Score of All Patients Included Type of Tumor ingrowth, and with a short length, which requires multiple insertions; the Wallstent has a substantial foreshortening, which may result in insufficient covering of the tumoral stenosis and may also present with partial wall contact; the Palmaz stent is a balloon-expandable stent [8] and may crush when inserted into peripheral [9] or central thoracic veins []. Nitinol stents do not have these disadvantages and may therefore be more suitable for safe and efficient endovascular management of malignant SVC syndrome. Experience, however, is limited, and outcome data are scarce and mostly mixed with the aforementioned types of steel stents for the treatment of SVC syndrome [3, 4]. In this study, we evaluated the performance of the flexible large-bore diameter nitinol stent (Zilver, Cook Medical) for the treatment of malignant SVC syndrome, and we analyzed several factors potentially affecting the overall outcome. Materials and Methods Patient and Tumor Characteristics Between December 24 and May 2, 78 consecutive patients (54 men and 24 women) with a mean age of 64. years (median, 64.7 years; range, years) were referred to the interventional radiology unit for endovascular palliative treatment of a malignant SVC syndrome. The endovascular procedure was performed as the first step in the overall management of malignant SVC syndrome, which is in accordance with a study by Lanciego et al. [5]. Patients with a benign stenosis No. (%) of Tumors (n = 78) Primary lung tumor 62 (79.5) Non small cell lung cancer 46 (58.9) Squamous carcinoma 6 Adenocarcinoma 3 Undifferentiated 7 Small cell lung cancer 6 (2.5) Tumors associated with malignant lymphadenopathies 6 (2.5) Non small cell lung cancer 6 Small cell lung cancer 4 Prostate carcinoma Esophageal carcinoma Renal cell carcinoma 2 Colon carcinoma Urothelial carcinoma of the superior vena cava superior were excluded from this study. All patients had undergone contrast-enhanced thoracic CT for tumor staging and evaluation of the SVC syndrome using MDCT. The Kishi classification score [] was used to grade the clinical degree of SVC syndrome. This retrospective study was approved by the local ethical committee (approval S-54458). Angiographic Stent Procedure All angiographic stent procedures were performed after bilateral upper limb venography, including opacification of the central thoracic veins and using nonionic iodixanol contrast media (Visipaque 27, GE Healthcare). The degree of malignant stenosis was assessed according to the venographic images; the degree of stenosis was defined in relation to the percentage of endovascular occlusion and was classified as follows: high-grade stenosis, 8 99%; moderate stenosis, 5 8%; and low-grade stenosis, 3 5%. In the event of a thrombotic occlusion of the SVC, catheter-directed thrombolysis was started with the use of urokinase (Actosolv, Eumedica) infused through a multi side hole catheter, which was placed in the clot as described elsewhere [2]. Daily control venography was performed, and, if thrombotic material was lysed, the procedure for stent placement was prepared. Venous access was made in the right groin under local anesthesia by placing a 7-French sheath. After IV injection of 5 IU of heparin, catheterization of the stenosis was performed with a conventional diagnostic 4-French catheter (Vertebral Optitorque, Terumo Europe) and hydrophilic guidewire (Glidewire, Terumo Europe). Kishi Score Primary Lung Tumor Mediastinal Adenopathy Total a 55 5 Note Kishi scores were available for further analysis for only 7 of the 78 patients. The mean (± SD) Kishi scores were 3. ±.67 for the 55 patients with primary lung tumors, 3. ±.5 for the 5 patients with mediastinal adenopathy, and 3. ±.63 for all 7 patients (p =.66). a p =.6679 for primary lung tumor versus mediastinal adenopathy. After the hydrophilic guidewire was exchanged for a stiff guidewire (Amplatz, Cook Medical), a selfexpandable nitinol stent (Zilver, Cook Medical) with a diameter of 2 or 4 mm and a length of 4, 6, or 8 mm was inserted after predilatation of the malignant stenosis with an 8-mm angioplasty balloon. The stent diameter depended on the diameter of the SVC below and above the stenosis: all stents were oversized by 2 mm in comparison with the nominal diameter of the superior vena cava. If the length of the stented segment was more than 8 mm, an additional nitinol stent was inserted. Afterward, the stents were postdilated with a -, 2-, or 4-mm-diameter angioplasty balloon, depending on the nominal diameter of the treated vessel. If the nitinol stent still presented with a more than 5% residual stenosis, due to recoil phenomena, and after repeated (three inflations of 3 seconds each) inflation of the balloon was performed, then an additional cobalt-chromium balloon-expandable stent (Scuba, Medtronic-Invatec) was inserted and postdilated up to 2 mm diameter. Postinterventional antithrombotic therapy consisted of a therapeutic dose of low-molecular-weight heparin for the first month after the procedure, combined with 8 mg of aspirin daily. Later, only aspirin was prescribed, instead of coumarin, to avoid late bleeding complications in this population of frail patients. Statistical Analysis and Follow-Up Variables were compared between the groups using Mann-Whitney U and Fisher exact tests. Kaplan-Meier estimates were used to construct curves for overall survival, which were compared 668 AJR:2, September 23

3 Treatment of Malignant Superior Vena Cava Syndrome With Large-Bore Nitinol Stents TABLE 3: Description of Kishi Scoring System for Symptoms of Superior Vena Cava Syndrome Signs and Symptoms Neurologic symptoms Stupor, coma, or blackout 4 Blurry vision, headache, dizziness, or amnesia 3 Changes in mentation 2 Uneasiness Laryngopharyngeal or thoracic symptoms Orthopnea or laryngeal edema 3 Stridor, hoarseness, dysphagia, glossal edema, or shortness of breath 2 Cough or pleural effusion Nasal and facial signs or symptoms Lip edema, nasal stiffness, epistaxis, or rhinorrhea Facial swelling Venous dilatation Neck vein or arm vein distention, upper extremity swelling, or upper body plethora Note Scoring system was published by Kishi et al. []. The total score for signs and symptoms was calculated as the sum of the highest grade in each class. using a log-rank test. p values less than.5 were considered statistically significant. No corrections for multiple testing were performed. All analyses were performed using SAS software (version 9.2 for Windows, SAS Institute). The follow-up period was defined from stent insertion date up to the patient s death or to the end of the follow-up period, which was January 22. Results Demographic Data In 62 patients (79.5%), a primary lung tumor invaded the SVC, whereas in the remaining 6 patients (2.5%), the SVC occlusion was related to malignant mediastinal lymphadenopathies of a primary tumor located in either the thorax or the abdomen. The type of tumor and the location of the primary malignancy are summarized in Table. In the group of primary lung tumors (n = 62), 55 Kishi scores were recorded and available for further analysis; in the group of mediastinal adenopathies (n = 6), 5 scores were recorded (Table 2). The Kishi scoring system [] is described in Table 3. Score Radiologic Outcome Upper limb venography revealed a highgrade stenosis in 86% (n = 67) of cases, a moderate stenosis in 6% (n = 5) of cases, and a low-grade stenosis in % (n = ) of cases. In 6% (n = 5) of cases, a complete thrombotic occlusion was seen and subsequently treated by in situ thrombolysis and stenting afterward. The malignant stenosis was located in the SVC in 88% (n = 69) of cases (Fig. ), in the SVC with extension into the left brachiocephalic vein in 4% (n = 3) of cases, and in the SVC with extension into the right brachiocephalic vein in 8% (n = 6) of cases. In total, 8 Zilver stents (average,.49 stents per procedure; minimum, one stent; maximum, three stents) were placed and postdilated; in 7 patients (22%), placement of an additional balloon-expandable stent was required to manage insufficient expansion of the Zilver stent (Fig. 2). Of these 7 patients, 5 presented with a malignant SVC syndrome owing to primary lung tumor invasion. The remaining two patients presented with occluding mediastinal lymphadenopathies (p =.49). In all but one patient (99%), the nitinol stents were placed correctly, completely covering the malignant stenosis. However, in one patient who presented with a high-grade SVC stenosis owing to an invading non small cell lung cancer, the Zilver stent with 4 mm diameter and 6 mm length migrated into the right atrium immediately after deployment and with the stiff Amplatz guidewire still in place. A snare catheter (Goose neck, EV3) with a 2-mm diameter loop was introduced over the stiff wire to capture the distal end of the migrated Zilver stent. The snared stent was safely pulled down and deployed in the right external iliac vein without visible stent or vein wall damage (Figs. 3A 3D). In a second session, a large-bore Wallstent (Boston Scientific) with a nominal diameter of 6 mm and a nominal length of 4 mm was correctly and safely deployed over the malignant SVC stenosis (Fig. 3E). Other procedural complications were minor and included tachypnea (n = 2) and an axillary hematoma (n = ); these complications did not require additional treatment or a longer hospital stay. Clinical Outcome In all patients, the clinical symptoms related to the SVC occlusion disappeared within 72 hours after the procedure. There were symptomatic reocclusion events of the stented vein in eight (%) patients; one patient underwent three procedures of restenting because of in-stent restenosis at 47, 279, and 283 days, respectively, after the initial stent procedure. Symptomatic reocclusion occurred with a median interval of 62 days (minimum, 4 days; maximum, 279 days) after successful SVC stenting. Reocclusion events were noted both in a patient with an additional bare stent (n = /7 [6%]) and in patients without a balloon-expandable stent (n = 7/6 [.5%]) (p =.6). Finally, thrombotic occlusion of the stented segments was noted in one patient who was previously treated by catheter-directed thrombolysis followed by Zilver stent implantation. Management of the reocclusion consisted of catheter-directed thrombolysis and stenting (n = ) or recanalization and primary stenting (n = 8). One patient presented with a reocclusion of the stented SVC days after the Zilver stent procedure, but because of the patient s poor general status, reintervention was not performed. Restenting was performed with use of the Zilver nitinol stent in seven procedures, with the use of a Wallstent in one procedure; in one other patient, in-stent restenosis was successfully treated with the insertion of a covered stent (Viabahn, W. L. Gore). The Kaplan-Meier survival curve of all the patients included in the study (n = 78) revealed estimated survival rates of 53.8%, 34.6%, 6.%, and 2.2% at 6, 2, 24, and 36 months, respectively (Fig. 4). Factors Potentially Influencing Outcome Kishi classification There was no evidence for a difference in overall survival as a function of Kishi classification, whether we considered the classification as a continuous predictor (p =.8) or when we divided the Kishi classification into three groups: Kishi 2, Kishi 3 4, and Kishi 5 7 (p =.42) (Fig. 5). Type of tumor invading the SVC The group of patients with a primary lung tumor invad- AJR:2, September

4 Maleux et al. ing the SVC had estimated survival rates of 54.8%, 37.7%, and 8.3% at 6, 2, and 24 months, respectively. The group of patients with malignant mediastinal lymphadenopathies had estimated survival rates of 5.%, 2.4%, and 7.% at 6, 2, and 36 months, respectively. No statistical evidence for a difference in overall survival (p =.38) between the two groups was observed (Fig. 6). Need for an additional balloon-expandable stent In the group of patients for whom an additional balloon-expandable stent was deployed into the nitinol stent because of substantial recoil and subsequent insufficient expansion, estimated survival rates of 64.7%, 43.%, and 23.% were found at 6, 2, and C A 24 months, respectively. In the group of patients who did not need an additional balloonexpandable stent, the estimated survival rates were 5.8%, 32.2%, and 4.4% at 6, 2, and 24 months, respectively (p =.35). Discussion In this retrospective study, we found that the Zilver nitinol stent can be used to treat malignant SVC stenoses with very high technical and immediate clinical success rates (99%). This observed effectiveness is at least as high as the effectiveness rates obtained with other types of stents, mainly self-expanding steel stents. Several series dealing with the Wallstent [5, 3 5] or Gianturco Z stent [6] also B D Fig. 7-year-old man. A, Axial CT reveals symptomatic large malignant mass lesion (arrows) corresponding to non small cell lung cancer and invading superior vena cava. B, Corresponding cavography image shows tumorous invasion (arrows) with nondeployed Zilver stent (Cook Medical) positioned over malignant stenosis. C, Control cavography image obtained after stent deployment and dilation shows correct positioning of stent (arrows) completely covering malignant stenosis and restoring patency of superior vena cava. D, Control CT obtained 2 months later shows completely patent Zilver stent (arrows) within tumorous mass, which has shrunk as result of radiation therapy. found effectiveness rates of 9% or more. In two other series dealing with the mixed use of both Wallstent and Memotherm nitinol stents (CR Bard), the immediate clinical success rate was 8% and 95%, respectively, but no distinction of technical or clinical outcome was made between the two types of stents, nor was there a clear indication of which cases involved the use of a Wallstent and which cases involved the use of a nitinol stent [3, 4]. In addition, in contrast to most other authors using self-expanding stents, we decided to place an additional balloon-expandable stent in cases involving more than 5% residual stenosis after Zilver stent placement and postdilation to achieve better immediate clinical outcome and 67 AJR:2, September 23

5 Treatment of Malignant Superior Vena Cava Syndrome With Large-Bore Nitinol Stents to prevent early stent thrombosis. Additional balloon-expandable stent placement within the Zilver stent was performed in 22% of the cases. Although we do not have real arguments that placement of an additional balloonexpandable stent will result in higher patency rates and better clinical outcome, we are more comfortable completing the stent procedure with a fully expanded stent, an approach that is in agreement with a study by Kee et al. [7]. Those authors used a Palmaz stent in the event of insufficient expansion of a Wallstent. They did not clearly mention, however, in how many cases additional Palmaz stenting was required. Finally, we did not observe a higher incidence of early or late stent occlusion in cases A D Fig year-old woman. A, Axial CT reveals large mediastinal mass lesion (arrows), identified as small cell lung cancer, with clear invasion of superior vena cava. B and C, Corresponding cavography images show malignant stenosis (arrows) before (B) and after (C) repeat dilation. D, Radiograph of stent shows residual recoil of stent (arrows), resulting in more than 5% residual stenosis. E, Placement of additional balloon-expandable stent (arrows) resulted in complete expansion of stented segment. F, Completion cavography image shows completely patent stent (white arrows) and stented superior vena cava (black arrows). in which additional balloon-expandable stenting was initially performed, despite the larger amount of metallic material in the vessel wall. The technical complication rate in this series is also acceptably low: in one of 78 cases (.3%) we had to manage a periprocedural, potentially fatal, stent migration into the right atrium. In agreement with Anand et al. [8] and Bagul et al. [9], we successfully snared, pulled down, and redeployed the nitinol stent into the right iliac vein. Another potential reason for the very low rate of technical complications may be the fact that we never used nitinol stents with a diameter exceeding 4 mm, which is in agreement with Fagedet et al. [4], who found that stents more than 6 B E mm in diameter are a predictive factor for complications (p =.8), potentially because very-large-diameter stents may cause vein rupture with hemorrhage or pulmonary edema because of access flow. The overall outcome data are also favorable for the Zilver nitinol stent. The restenosis and rethrombosis rates after successful SVC stent placement are also relatively low, with an overall reocclusion rate of %, which is lower than or in the same range as those in other series using other types of stents (Table 4). The low restenosis rate may be associated with the fact that we always performed a unilateral stent placement and never performed a bilateral or kissing stent placement, even for stenoses at C F AJR:2, September 23 67

6 Maleux et al. the venous confluence. These findings are in agreement with those of Dinkel et al. [3], who used the Wallstent in a nonrandomized study and found a significantly lower reocclusion rate (p <.5) in the unilateral group. The overall survival rates in our study were 53.8% and 34.6% at 6 and 2 months of follow-up, respectively. However, comparison with historical series is difficult and probably not meaningful because the patient selection group may be different, the timing C Fig year-old man. A, Axial CT shows large tumoral mass lesion (asterisk) invading superior vena cava (arrow). B, Corresponding venography image confirms high-grade stenosis (arrows) of superior vena cava. C, Immediately after Zilver stent (Cook Medical) placement, migration of stent (arrows) into right atrium occurred. D, After snaring, stent (arrows) was pulled down and repositioned into right external iliac vein. E, In second session, large Wallstent (Boston Scientific) (arrows) was deployed correctly over superior vena cava stenosis. of endovascular treatment of SVC syndrome before or after radiotherapy and chemotherapy is not always defined, and, finally, the medical management of these patients with cancer has also dramatically improved over the past decades. However, on the basis of these study results, it is clear that both patients with primary and secondary tumors invading the SVC can be considered good candidates for stenting without significant differences in survival A D and complication rates for the two groups. Finally, the degree of severity of symptoms of SVC syndrome, defined as a Kishi score, is not a prognostic factor for outcome (p =.8). This study also has some limitations. First, it is a retrospective study over a long study period with some missing data, although the technique of stenting has not changed over the years. Second, the outcome data regarding the stent reocclusion rate might be underestimated, because some patients in very B E 672 AJR:2, September 23

7 Treatment of Malignant Superior Vena Cava Syndrome With Large-Bore Nitinol Stents Overall Survival Rate (%) No. of Patients at Risk Time Since Surgery (mo) Overall Survival Rate (%) No. of Patients at Risk Primary lung tumor Mediastinal adenopathy Overall Survival Rate (%) No. of Patients at Risk Kishi score 2 Kishi score 3 4 Kishi score Primary lung tumor Mediastinal adenopathy Kishi score 2 Kishi score 3 4 Kishi score 5 7 Time Since Surgery (mo) Log-rank, p = Log-rank, p = Time Since Surgery (mo) Fig. 4 Kaplan-Meier survival curve (solid black line) of all included patients reveals estimated survival rate at 6 and 2 months of 53.8% and 34.6%, respectively. Dashed gray lines correspond to 95% CIs. Fig. 5 Kaplan-Meier survival curve, comparing group of patients with primary lung tumor (solid black line) with group of patients with mediastinal adenopathy (dashed gray line), does not show significant difference in overall survival (p =.376). Fig. 6 Kaplan-Meier survival curves, analyzing overall survival as function of Kishi score, cannot show significant difference in overall survival between these groups (p =.4228) AJR:2, September

8 Maleux et al. TABLE 4: Restenosis Rates for Different Stent Types Type of Stent (Manufacturer) [Reference] Restenosis Rate (%) Wallstent (Boston Scientific) [8, 9] 3 9 Gianturco Z stent (Cook Medical) [9] 25 Wallstent (Boston Scientific) and Palmaz stent (Cordis) [5] Memotherm (CR Bard) and Wallstent (Boston Scientific) [] 22 Zilver stent (Cook Medical) [current study] poor clinical condition owing to widespread cancer may have preferred further palliative care at home instead of hospital readmission. Finally, no direct comparison was made with other types of SVC stents, and no more than three clinical factors were analyzed for their potential influence on global outcome. In conclusion, this study shows that largebore Zilver nitinol stents are safe and very effective for the palliative treatment of malignant SVC stenosis, with encouraging global survival data and low reocclusion rates. References. de Gregorio Ariza M, Gamboa P, Gimeno M, et al. Percutaneous treatment of superior vena cava syndrome using metallic stents. Eur Radiol 23; 3: Rajan DK, Saluja JS. Use of nitinol stents following recanalization of central venous occlusions in hemodialysis patients. Cardiovasc Intervent Radiol 27; 3: Urruticoechea A, Mesia R, Dominguez J, et al. Treatment of malignant superior vena cava superior syndrome by endovascular stent insertion: experience on 52 patients with lung cancer. Lung Cancer 24; 43: Fagedet D, Thony F, Timsit JF, et al. Endovascular treatment of malignant superior vena cava syndrome: results and predictive factors of clinical efficacy. Cardiovasc Intervent Radiol 23; 36: Lanciego C, Pangua C, Chacon J, et al. Endovascular stenting as the first step in the overall management of malignant superior vena cava syndrome. AJR 29; 93: Da Ines D, Chabrot P, Cassagnes L, et al. Endovascular treatment of SVC syndrome from neoplastic origin: a review of 34 cases. J Radiol 28; 89: Nguyen NP, Borok TL, Welch J, Vinh-Hung V. Safety and effectiveness of vascular endoprosthesis for malignant superior vena cava syndrome. Thorax 29; 64: Elson JD, Becker GJ, Wholey MH, Ehrman KO. Vena caval and central venous stenoses: management with Palmaz balloon-expandable intraluminal stents. J Vasc Interv Radiol 99; 2: Maleux G, Rousseau H, Otal P, Joffre F. Collapsed Palmaz stent after deployment for hemodialysis access-related venous stenosis. J Vasc Interv Radiol 998; 9:69 7. Hammer F, Becker D, Goffette P, Mathurin P. Crushed stents in benign left brachiocephalic vein stenosis. J Vasc Surg 2; 32: Kishi K, Sonomura T, Mitsuzane K, et al. Selfexpandable metallic stent therapy for superior vena cava syndrome: clinical observations. Radiology 993; 89: Maleux G, Marchal P, Palmers M, et al. Catheterdirected thrombolytic therapy for thoracic deep vein thrombosis is safe and effective in selected patients with and without cancer. Eur Radiol 2; 2: Dinkel HP, Mettke B, Schmind F, Baumgartner I, Triller J, Do DD. Endovascular treatment of malignant superior vena cava syndrome: is bilateral Wallstent placement superior to unilateral placement? J Endovasc Ther 23; : Thony F, Moro D, Witmeyer P, et al. Endovascular treatment of superior vena cava obstruction in patients with malignancies. Eur Radiol 999; 9: Watkinson AF, Hansell DM. Expandable Wallstent for the treatment of obstruction of the superior vena cava. Thorax 993; 48: Gaines PA, Belli AM, Anderson PB, McBride K, Hemingway AP. Superior vena caval obstruction managed by the Gianturco Z stent. Clin Radiol 994; 49:22 26; discussion, Kee ST, Kinoshita L, Razavi MK, Nyman UR, Semba CP, Dake MD. Superior vena cava syndrome: treatment with catheter-directed thrombolysis and endovascular stent placement. Radiology 998; 26: Anand G, Lewanski C, Cowman S, Jackson J. Superior vena cava stent migration into the pulmonary artery causing fatal pulmonary infarction. Cardiovasc Intervent Radiol 2; 34(suppl 2): S98 S2 9. Bagul NB, Moth P, Menon NJ, Myint F, Hamilton G. Migration of the superior vena cava stent. (letter to the editor) J Cardiothorac Surg 28; 3: AJR:2, September 23