- Email: [email protected]
Efficacy of Venous Sac Embolization for Pulmonary Arteriovenous Malformations: Comparison with Feeding Artery Embolization
CLINICAL STUDY
Efficacy of Venous Sac Embolization for Pulmonary Arteriovenous Malformations: Comparison with Feeding Artery Embolization Sadao Hayashi, MD, PhD, Yasutaka Baba, MD, PhD, Terutoshi Senokuchi, MD, and Masayuki Nakajo, MD, PhD
ABSTRACT Purpose: To examine the efficacy of venous sac embolization (VSE) in comparison with transcatheter feeding artery embolization (FAE) for treatment of pulmonary arteriovenous malformations (PAVMs). Materials and Methods: From 1989–2009, 21 patients underwent embolization of 37 PAVMs. Safety and long-term efficacy of VSE were evaluated retrospectively and compared with FAE. Results: FAE was performed in 22 (18 simple and 4 complex type) PAVMs, and VSE was performed in 15 (14 simple and 1 complex type) PAVMs. There were significant differences between FAE and VSE in treated periods, PAVM location, coil type used, number of coil combinations and coils per PAVM, coil position, and reperfusion; there were no significant differences in most PAVM characteristics, follow-up durations (58 mo ⫾ 54 vs 42 mo ⫾ 42; P ¼ .32), and minor complications (pleurisy [2 vs 2]). Reperfusion occurred in 11 (50%) of 22 PAVMs in the FAE group and no PAVMs in the VSE group (P o .01). Of 22 PAVMs in the FAE group, 17 (77%) were treated with 0.035-inch coils alone; of 15 PAVMs in the VSE group, 14 (93%) were treated with 0.018-inch interlocking detachable coils (IDCs), 0.018-inch pushable fibered coils, or IDCs and pushable fibered coils combined (P o .01). The number of coils used was 8 ⫾ 4 in the VSE group and 4 ⫾ 4 in the FAE group (P ¼ .002). Conclusions: The high reperfusion rate in the FAE group was mainly due to the use of large 0.035-inch coils alone. Although more coils are needed, VSE can be used to treat PAVMs with a venous sac safely and achieve long-term efficacy.
ABBREVIATIONS: FAE = feeding artery embolization, HHT = hereditary hemorrhagic telangiectasia, IDC = interlocking detachable coil, PAVM = pulmonary arteriovenous malformation, VSE = venous sac embolization
Pulmonary arteriovenous malformations (PAVMs) are rare pulmonary vascular anomalies and direct pulmonary artery-to-vein connections without intervening capillary beds. PAVMs are most commonly congenital in nature; can occur as either an isolated abnormality or part of hereditary hemorrhagic telangiectasia (HHT); and are associated with a risk of paradoxical embolization, resulting in events such as stroke, transient ischemic attack, and brain abscess, with the strong potential for severe From the Department of Radiology, Kagoshima University Graduate School of Medical and Dental Sciences, 8-35-1 Sakuragaoka, Kagoshima-shi, Kagoshima 890-8520, Japan. Received May 10, 2012; final revision received September 6, 2012; accepted September 9, 2012. Address correspondence to S.H.; E-mail: [email protected] None of the authors have identified a conflict of interest. & SIR, 2012 J Vasc Interv Radiol 2012; 23:1566–1577 http://dx.doi.org/10.1016/j.jvir.2012.09.008
neurologic deficits (1–4). Treatment is indicated to obliterate or reduce the shunt and prevent neurologic complications. Embolization was introduced as the treatment of choice for PAVMs by Porstmann in 1977 (5). Since then, several published series have documented the efficacy of embolization with metallic coils or detachable balloons in the treatment of PAVMs in patients with hypoxemia, patients with neurologic complications, or asymptomatic patients with a feeding artery Z 3 mm in diameter, obviating the need for surgery in most cases; the longterm efficacy of embolization for PAVMs has also been documented (6–14). More recently, the AMPLATZER Vascular Plug (St. Jude Medical, St. Paul, Minnesota) has also been used for embolization of PAVMs (15–18). Embolization for PAVM can be performed in the outpatient setting (19,20). Although most PAVMS are successfully treated, recanalization or reperfusion can occur after the first embolization in some patients during longterm follow-up.
Volume 23
’
Number 12
’
December
’
2012
The standard percutaneous technique is transcatheter feeding artery embolization (FAE), involving the deployment of coils, balloons, or plugs within the feeding artery as close to the neck of the PAVM as possible, to avoid occlusion of vessels to the normal lung (3,11,20). Transcatheter venous sac embolization (VSE) using interlocking detachable coils (IDCs) has also been used when the target artery of the PAVM is too short to avoid sacrifice of large normal pulmonary artery branches or when the artery is a high-flow type with a higher risk of paradoxical embolization of a coil or balloon (21–23). In addition, VSE using detachable coils may be useful to resolve the risk of migration of embolic materials in embolization of PAVMs (22,23). We reviewed embolization procedures for PAVMs that were performed in our department during the past 21 years to examine the safety and long-term efficacy of VSE in comparison with FAE.
MATERIALS AND METHODS This retrospective study had institutional review board approval, and informed consent was waived.
Patients We reviewed clinical charts and angiographic and computed tomography (CT) images of patients with PAVMs who underwent embolization in our department and identified 21 patients treated between October 1989 and May 2009. No patient was excluded, and the study population comprised these 21 patients.
Angiographic and Embolization Techniques After written informed consent was obtained from all patients, all pulmonary angiographic and embolization procedures were performed by the Seldinger method. Because the procedures were done over the course of 21 years, the operators also changed in this period. These
1567
procedures were performed by at least two operators at each therapeutic session, and one operator had a minimum of 4 5 years of experience in endovascular intervention. To prevent clot formation during angiographic procedures, a dose of 5,000 units (1,000 units/mL) of heparin was administered intravenously, and an additional 1,000 units of heparin was added every hour when necessary. Pulmonary angiography was performed with a 5-F pigtail catheter (Medikit Co, Ltd, Tokyo, Japan) through the femoral vein under local anesthesia. Selective right and left pulmonary angiograms were obtained in frontal and lateral projections using a 5-F JB1 catheter (Create Medic Co, Ltd, Tokyo, Japan). The type, size, and number of available coils and their deployment positions at the feeding artery or venous sac were determined by the operators from the findings on selective angiography. When the feeding artery of the PAVM was visualized, the JB1 catheter was inserted into the feeding artery and advanced as close to the venous sac as possible. FAE was performed using 0.035-inch Gianturco-Wallace stainless steel coils (coil size range, 3–8 mm) (Cook, Inc, Bloomington, Indiana) or 0.035-inch VortX 35 platinum coils (coil size range, 3–8 mm) (Boston Scientific, Natick, Massachusetts); the diameter of the coils used was about 20% larger than the diameter of the feeding artery undergoing embolization. For VSE, a 2.4-F Renegade microcatheter (Boston Scientific) was advanced coaxially through the JB1 catheter into the venous sac via the feeding artery. In the embolization, we used 0.018inch IDCs (Boston Scientific) about 20% larger than the drainage vein at first, which have been available since 1997. We introduced the larger sized IDCs (coil size range, 4–20 mm) in the venous sac first, detaching the coils after confirming the size and position fit for the sac. A wire mesh of IDCs was created to prevent coil migration of the subsequent pushable small coils (coil size range, 3–8 mm) (ie, 0.018-inch fibered platinum coils; [Boston Scientific] for embolization of the venous sac (Fig 1a–c). Complete occlusion of the feeding artery or venous sac was confirmed with pulmonary angiography performed
Figure 1. Patient 18. A 15-year-old boy with complex-type PAVM in right segment 4 treated by VSE. (a) Pulmonary arteriogram shows a complex-type PAVM with two feeding artery branches and a venous sac. (b) Radiograph shows a venous sac filled with 17 IDCs, of which the largest one is 18 mm in maximum diameter. (c) Pulmonary arteriogram reveals no contrast material in the PAVM and draining vein after VSE.
1568
’
Venous Sac Embolization of PAVMs
Hayashi et al
’
JVIR
Figure 2. Patient 13. A 67-year-old woman with two simple-type PAVMs treated by embolization. (a) Pulmonary arteriogram shows one PAVM in right segment 4 treated by VSE and the other in right segment 10 treated by FAE. (b) Pulmonary arteriogram demonstrates complete packing of the PAVM in right segment 4 treated by VSE and recanalization of the PAVM in right segment 10 treated by FAE 68 months later. (c) Contrast-enhanced CT image shows the contrast-enhanced draining vein (arrow) of the recanalized segment 10 PAVM.
after embolization in all PAVMs. After IDCs became available, VSE was deemed preferable when the PAVM had a venous sac; it was also thought that VSE was beneficial in PAVMs with large outflow vessels or short feeding arteries (22) because the microcatheter enabled us to avoid injuring the vessel wall and carry the smaller coils safely into the venous sac. Also, the 5-Fr JB1catheter occasionally could not be advanced enough to reach the proximal site of the sac with or without spasms of the feeding artery. However, VSE was not performed when the microcatheter could not be put stably in the venous sac.
Follow-up Studies for Reperfusion Chest radiography and non–contrast-enhanced and contrastenhanced CT were planned principally for follow-up studies at 3–6 months, 6–12 months, and every 1–2 years thereafter. When the venous sac and draining vein in the case of FAE and the draining vein in the case of VSE showed contrast enhancement on CT, which indicated reperfusion, angiography was planned. When these structures were visible without contrast enhancement, they were followed up by chest radiography and non–contrast-enhanced and contrastenhanced CT. When complete resolution (ie, disappearance or scarlike structure formation of the venous sac or draining vein) was noted without contrast enhancement on CT scan, chest radiography and non–contrast-enhanced CT were principally planned for the next follow-up examination. Non–contrast-enhanced and contrast-enhanced CT scan of the chest was performed using a single-detector CT scanner (X-Vigor; Toshiba, Tokyo, Japan) or 16-detector CT scanner (Aquilion 16; Toshiba). Using the singledetector CT scanner, all scans began from the apex of the lung to the diaphragm. Additional section at selected levels, synchronized scanning with longitudinal table motion at a speed of one section thickness per second (10-mm collimation, 120 kVp, 200 mA), and two subsequent sequences were used to evaluate contrast enhancement of the PAVM
(5-mm collimation, 120 kVp, 250 mA; 2-mm collimation, 120 kVp, 230 mA) after a bolus injection of 100 mL of nonionic iodinated contrast material. Using the 16-detector CT scanner, all scans began from the apex to the diaphragm with tube voltage of 120 kVp and gantry rotation speed of 0.5 second/rotation. Maximum allowable tube current was 400 mA, detector row configuration was 16 1 mm, and table increment was 15 mm/rotation. Contrast-enhanced CT scans were performed with the following image parameters: maximum allowable tube current of 400 mA, detector row configuration of 16 0.5 mm, and table increment of 7.5 mm/rotation in the longitudinal direction. Bolus tracking technique (Sure Start software; Toshiba) was used to time the start of the scanning after contrast material injection. Dynamic contrast-enhanced CT scan was performed after a bolus injection of 100 mL of nonionic iodinated contrast material. Reperfusion of the PAVM was confirmed with pulmonary arteriography and bronchial arteriography (Figs 2a–c and 3a and b). Bronchial angiography was performed using a 5-F shepherd’s hook catheter (Hanako Medical, Tokyo, Japan) through the femoral artery under local anesthesia. We defined reperfusion as sac or draining vein blood flow regardless of the inflow route and recanalization as sac blood flow from the recanalized artery that had previous embolization with coils.
Comparative Studies of FAE and VSE We reviewed the patients’ age and sex; the year when embolization was performed; type and location of the PAVM; the initial findings of the venous sac, feeding artery, and its branches and draining vein; the follow-up findings of the venous sac and draining vein; the types and number of coils used; coil position; follow-up periods; presence and absence of reperfusion; and complications related to embolization. We classified PAVMs into simple type with a segmental feeding artery or its branch and complex type with two or more segmental arteries or branches according to the classification by White et al in
Volume 23
’
Number 12
’
December
’
2012
1569
classified as minor or major according to Society of Interventional Radiology (SIR) guidelines (27).
Morphologic Measurement Methods The initial maximum diameter of the venous sac, feeding artery, and draining vein was measured on contrastenhanced CT or the angiogram obtained before embolization when contrast-enhanced CT was unavailable. The length of the feeding artery or its branch was measured from the bifurcated point of the segmental artery or its branch to the proximal point of the venous sac on the angiogram obtained before embolization. Coil position was measured from the distal point of coils in the feeding artery to the proximal point of the venous sac on the angiogram obtained after embolization. The follow-up maximum diameter of the venous sac or draining vein was measured on the follow-up pulmonary window CT images.
Statistical Methods Comparison between groups were made by using the binominal, w2, or Fisher exact test for categorical variables. The Mann-Whitney U test was used to analyze the differences of continuous variables in venous sac diameter, diameter and length of the feeding artery, diameter of the draining vein, and coil placement position between FAE and VSE groups or between FAE reperfusion and nonreperfusion groups. Data are expressed as mean ⫾ standard deviation and range for continuous variables. P values o .05 were considered significant. Statistical analyses were performed on a computer using StatView version 5.0 software (SAS Institute, Cary, North Carolina).
RESULTS
Figure 3. Patient 8. A 51-year-old woman with simple-type PAVM in right segment 8 treated by FAE. (a) Pulmonary arteriogram shows recanalization of the PAVM (arrow). (b) Bronchial arteriogram shows reperfusion of the PAVM (arrow) from the bronchial artery at the time of additional FAE.
1983 (24) for our analysis, although the newly defined complex type needs two or more segmental arteries (2). HHT was diagnosed by the Curac- ao criteria (25). In each case, the diagnosis of PAVM was established based on findings from thoracic CT and pulmonary angiography (26). The follow-up months were from the date of embolization to the follow-up date of the last contrastenhanced CT scan in patients without reperfusion and to the date of reperfusion in patients with reperfusion noted on contrast-enhanced CT scan. Complications were
Follow-up results were available for all 21 patients. The last follow-up examination was performed in May 2011. A PAVM was initially suspected because of a nodule with vessel-like structures on chest radiographs or unenhanced CT and subsequently confirmed with contrast-enhanced CT and pulmonary angiography to identify the feeding artery or arteries, venous sac or shunt, and draining vein. Detailed data of FAE and VSE groups are presented in Tables 1 and 2.
Overall Results Four patients received both FAE and VSE. Age range was 15–67 years (mean, 43 years ⫾ 16). Female patients constituted 81% (17 of 21) of the study group. Patient and PAVM breakdown was as follows: 3 patients with HHT and 18 patients without HHT; 12 patients with single PAVM and 9 patients with multiple PAVMs; and 37 PAVMs (32 simple and 5 complex types; 15 treated before 1996 and 22 treated in 1997 and thereafter; 7 upper, 9 middle, and 21 lower lobe PAVMs; and 24 right and 13 left ones). One of five complex types was diffuse type, which was located at the left lower lobe supplied by left A8, A9,
1570
Table 1 . Characteristics and Therapeutic Results of Pulmonary Arteriovenous Malformations Initially Treated by Feeding Artery Embolization
’
Initial
Follow-up
Number of Coils Used
Patient No./Age/Sex Year No.
Type
Feeder
Dsac 15
3
1/19/F
1989
1
Complex
Lt. A9
2/63/F 3/17/F
1991 1991
2 3
Simple Simple
Lt. A3 Rt. A6
20 20
4/62/M*
1992
4
Simple
Rt. A3
1993 1992
5 6
Simple Simple
1993
7
Complex
5
36
8 8
10 10
20
5
Rt. A10 Rt. A9
50 5
Lt. A8
20
Lt. A9
8
35 coil
18 coil
Coil
Follow-up
Position
(mo)
Reperfused Artery
Management of Reperfusion Additional FAE
4
þ
1
32
66
Rec. and bronchial
30 50
CR 15
CR 7
þ
1 1
15 32
175 36
Rec. and accessory
Operation
8
47
5
3
þ
3
18
24
Rec.
Additional FAE
10 3
15 5
15 36
5 3
3 3
þ þ
8 3
6 15
22 24
Bronchial Rec.
Follow-up Follow-up
5
12
50
5
3
þ
22
Bronchial
Follow-up
23
Rec.
Additional FAE
Rec. and bronchial Rec. and bronchial
Additional FAE Additional FAE
22
5 3
30 30
Complex
Lt. A10
10
5
8
40
5
8
þ
8
30
3 2
17 24
4
15
1994
9
Simple
Lt. A10 Lt. A3
40
3 9
20
32 10
CR
CR
7
5/48/F
1994
10
Simple
Lt. A8
13
3
4
36
CR
CR
1
6/48/F 7/43/F*
1994 1994
11 12
Simple Simple
Lt. A5 Rt. A1
20 10
5 3
7 4
30 18
CR CR
CR CR
1994
13
Simple
Rt. A2
55
7
14
30
CR
CR
8/51/F
1994 1995
14 15
Simple Simple
Lt. A1þ2 Rt. A8
15 10
3 7
2 9
20 47
8 7
2 2
9/15/F
1997
16
Complex
Lt. A8
Diffuse
8
15
30
CR
CR
30 20
CR CR
Lt. A9 Lt. A4
IDC
7
2
Lt. A8 Lt. A8 1992
Dfa Ddv Lfa Dsac Ddv CE
7 3
1
27
18
160
3 1
24 5
159 15
2
22
15
þ þ
1 3
3 12
4 45
5
5
137
8 2
5 10
10/37/M*
1997
17
Simple
Lt. A10
40
7
9
36
CR
3
11/55/F
1997 1997
18 19
Simple Simple
Rt. A8 Rt. A8
15 20
3 4
3 5
38 46
CR CR
CR CR
12/50/F
1999
20
Simple
Rt. A5
15
3
3
28
5
2
þ
13/67/F 14/43/F
2003 2008
21 22
Simple Simple
Rt. A10 Rt. A3
20 3
5 3
5 5
32 13
7 CR
3 CR
þ
1
1 3 3 2 3 3
5 1
5
26
30 5
27 109
7
76
Rec.
Additional FAE
3 3
68 19
Rec.
Additional FAE Hayashi et al ’
JVIR
Key to headings and abbreviations, in order of appearance in table. PAVM ¼ pulmonary arteriovenous malformations; Year ¼ year of FAE; Dsac ¼ diameter of sac; Dfa ¼ diameter of feeding artery; Ddv ¼ diameter of draining vein; Lfa ¼ length of feeding artery; CE ¼ contrast enhancement of venous sac and drainage vein; 35 coil ¼ 0.035-inch pushable fibered coil; IDC ¼ interlocking detachable coil; 18 coil ¼ 0.018-inch pushable fibered coil; Coil Position ¼ distance from sac (mm); Lt. ¼ left; Rt. ¼ right; A ¼ segmental pulmonary artery; CR ¼ complete resolution, disappearance or scarlike structure formation of sac and draining vein on CT; Rec. ¼ recanalized; FAE ¼ feeding artery embolization; Diffuse ¼ diffuse telangiectatic communications. n Patient with hereditary hemorrhagic telangiectasia. Reperfusion was noted in 11 of 22 PAVMs.
Venous Sac Embolization of PAVMs
Findings of PAVM (unit; mm)
Volume 23 ’
Number 12
Table 2 . Characteristics and Therapeutic Results of Pulmonary Arteriovenous Malformations Initially Treated by Venous Sac Embolization
(mo)
Number of Coils Used
Patient
35 Year
No.
Type
Feeder
Dsac
Dfa
Ddv
Lfa
Ddv
CE
10/37/M*
1997
1
Simple
Rt. A9
25
7
10
26
2
1997
2
Simple
Lt. A4
8
3
3
28
1999 2003
3 4
Simple Simple
Rt. A4 Rt. A4
20 15
6 5
7 5
40 15
coil
IDC
18 coil
1
6
CR
2
CR CR
3 4
2
150 68
27 26
14/43/F
2007
5
Simple
Rt. A8
15
4
5
15
CR
6
2
41
6 7
Simple Simple
Rt. A5 Rt. A7
4 15
3 3
4 5
27 24
CR CR
4
15/45/F
2008 1998
3 11
19 82
16/45/F
2002
8
Simple
Rt. A8
30
4
6
35
2
17/45/F 18/15/M
2005 2006
9 10
Simple Complex
Lt. A4 Rt. A4
15 15
4 4
5 5
65 17
CR CR
3 17
2
2
15
CR
0
19/51/F
2006 2006
11 12
Simple Simple
Rt. A8 Rt. A8
15 17
3 3
3 3
38 36
CR CR
2006
13
Simple
Rt. A9
7
4
4
24
CR
20/46/F 21/27/M
2007 2009
14 15
Simple Simple
Rt. A8 Lt. A8
8 15
3 3
4 4
15 16
CR CR CR
Rt. A4
2
2
8
2012
No./Age/Sex
12/50/F 13/67/F
Follow-up
’
Follow-up
December
Initial
’
Findings of PAVM (unit; mm)
100 4
12 25
3 3
4 5
4 4
3
3
4
6 7
2 6
45 17
Key to headings and abbreviations, in order of appearance in table. PAVM ¼ pulmonary arteriovenous malformations; Year ¼ year of venous sac embolization; Dsac ¼ diameter of sac; Dfa ¼ diameter of feeding artery; Ddv ¼ diameter of draining vein; Lfa ¼ length of feeding artery; CE ¼ contrast enhancement of drainage vein; 35 coil, 0.035-inch pushable fibered coil; IDC ¼ interlocking detachable coil; 18 coil ¼ 0.018-inch pushable fibered coil; Rt. ¼ right; Lt. ¼ left; A ¼ segmental pulmonary artery; CR ¼ complete resolution, disappearance or scarlike structure formation of sac and draining vein on CT. n Patient with hereditary hemorrhagic telangiectasia. No reperfusion was noted in all 15 PAVMs.
1571
1572
’
Venous Sac Embolization of PAVMs
Hayashi et al
and A4 segmental arteries with multiple small arteriovenous shunts. Embolization was performed on average in one patient (range, 0–4) and 2 PAVMs (range, 0–6) per year during 21 years. The diameters of the venous sac, feeding artery, and draining vein were 18 mm ⫾ 12 (range, 3–55 mm), 5 mm ⫾ 2 (range, 2–10 mm), and 7 mm ⫾ 4 (range, 2–20 mm). The length of the feeding artery was 30 mm ⫾ 12 (range, 10–65 mm). For 37 PAVMs, 209 coils (72 0.035-inch coils, 61 0.018 pushable fibered coils, and 76 IDCs) were used. There were 17 PAVMs treated with 0.035-inch coils alone and 20 PAVMs treated with other coils or a combination of coils. The number of coils used was 6 ⫾ 4 (range, 1–17) per PAVM. Reperfusion occurred in 11 (30%) of 37 PAVMs. The follow-up period was 51 months ⫾ 50 (range, 4–175 months). There were no major procedure-related complications (ie, paradoxical embolization owing to embolization device or blood clot and vessel perforation). A minor complication of chest pain secondary to pleurisy was noted in four patients and disappeared within a few days without medication.
Results of Comparison between FAE and VSE There were no significant differences between FAE and VSE in the number of patients (14 vs 11) and PAVMs (22
’
JVIR
vs 15), age (44 years ⫾ 17 vs 43 years ⫾ 13), male-tofemale ratio (2:12 vs 3:8), ratio of HHT to non-HHT (3:11 vs 1:10), ratio of single to multiple PAVMs (7:7 vs 6:5), ratio of simple to complex PAVMs (18:4 vs 14:1), and minor complications (pleurisy) related to embolization (2 vs 2). The other clinical features of FAE and VSE groups are summarized in Table 3. Significant differences between FAE and VSE groups were as follows. FAE was performed in all 15 PAVMs between 1989 and 1996, and VSE was performed in 15 (68%) of 22 PAVMs in 1997 and thereafter, when IDCs became available (P o .01). Of 22 PAVMs treated by FAE, 15 (68%) were located in the middle and lower lobes, whereas all 15 (100%) PAVMs treated by VSE were located in these lobes (P o .05). The diameter of the draining vein was larger in the FAE group (8 mm ⫾ 5) than the VSE group (5 mm ⫾ 2) with marginal significance (P ¼ .045). In the FAE group, 70 (79%) of 89 coils used were 0.035-inch coils, whereas in the VSE group, 70 (58%) and 48 (40%) of 120 coils used were IDCs (70 coils) and 0.018-inch coils (48 coils) (P o .01). In the FAE group, 17 (77%) of 22 PAVMs were treated with 0.035-inch coils alone, whereas in the VSE group, 14 (93%) of 15 PAVMs were treated with IDCs, 0.018-inch coils, or their combination, and only 2 0.035-inch coils were used in a patient combined with other coils (P o .01). VSE used twice as
Table 3 . Comparison of Clinical Features of Pulmonary Arteriovenous Malformations between Feeding Artery Embolization and Venous Sac Embolization Groups Group Clinical Feature
FAE
VSE
P Value
No. PAVMs
22
15
1989–1996:1997–2009
15:7
0:15
o .01
NS
Location (upper/middle/lower) Diameter (mm)
22 (7/2/13)
15 (0/6/9)
o .05
Venous sac
21 ⫾ 14 (3–55)
15 ⫾ 7 (4–30)
NS
Feeding artery Draining vein
5 ⫾ 2 (2–10) 8 ⫾ 5 (2–20)
4 ⫾ 1 (3–7) 5 ⫾ 2 (2–10)
NS .045
32 ⫾ 11 (10–50)
28 ⫾ 14 (15–65)
NS
Length Feeding artery (mm) Coil type used
o .01
35 coils
70
2
18 coils IDC
13 6
48 70
35 coils alone Others
17 5
0 15
No. coils used/PAVM
4 ⫾ 4 (1–15)
8 ⫾ 4 (2–17)
Coil position (mm)
13 ⫾ 9 (þ: 22)
In venous sac (: 11)
o .01*
Reperfusion
11
0
o .01
Follow-up (mo)
58 ⫾ 54 (4–175)
42 ⫾ 42 (4–150)
o .01
No. coil combinations
.002
NS
P value between FAE and VSE; NS ¼ not significant FAE ¼ feeding artery embolization; IDC ¼ interlocking detachable coil; PAVM ¼ pulmonary arteriovenous malformation; VSE ¼ venous sac embolization. n When estimated by coil position (þ) or () from the venous sac.
Volume 23
’
Number 12
’
December
’
2012
1573
Table 4 . Comparison between Reperfusion and Nonreperfusion Groups in the Feeding Artery Embolization Group Group Clinical Feature No. patients No. PAVMs
Reperfusion
Nonreperfusion
7
7
P Value
11
11
9:2 11 (2/1/8)
6:5 11 (5/1/5)
NS NS
Venous sac Feeding artery
18 ⫾ 11 (5–20) 5 ⫾ 2 (3–8)
24 ⫾ 16 (3–55) 5 ⫾ 2 (3–10)
NS NS
Draining vein
7 ⫾ 4 (2–10)
9 ⫾ 6 (3–20)
NS
34 ⫾ 11 (18–50)
28 ⫾11 (10–46)
NS
35 coils 18 coils
39 5
31 8
IDC
3
3
No. coil combinations 35 coils alone
10
7
1989–1996:1997–2009 Location (upper/middle/lower) Diameter (mm)
Length Feeding artery (mm) No. coil types used
Others
NS 1
4
No. coils used/PAVM Coil position (mm)
4 ⫾ 4 (1–13) 13 ⫾ 9 (1–30)
4 ⫾ 4 (1–15) 14 ⫾ 10 (3–32)
NS NS
Follow-up (mo)
37 ⫾ 23 (4–76)
79 ⫾ 68 (15–175)
NS
P value between reperfusion and nonreperfusion groups; NS ¼ not significant. IDC ¼ interlocking detachable coil; PAVM ¼ pulmonary arteriovenous malformation.
many coils as FAE on average (8 ⫾ 4 vs 4 ⫾ 4; P ¼ .002). Reperfusion occurred in 11 (50%) of 22 PAVMs treated by FAE but did not occur in 15 PAVMs treated by VSE (P o .01). The follow-up period did not differ between FAE (58 months ⫾ 54 [range, 4–175 months]) and VSE (42 months ⫾ 42 [range, 4–150 months]) groups (P ¼ .32).
Results of Comparison between FAE Reperfusion and Nonreperfusion Groups Reperfusion occurred in 7 (50%) of 14 patients and 11 (50%) of 22 PAVMs in the FAE group and consisted of five recanalized arteries alone, three recanalized and bronchial arteries, one recanalized and one accessory artery, and two bronchial arteries alone. Of 11 reperfusions, 6 were noted between 4 months and 2 years, and 5 were noted between 2 years and 6.3 years. There were no significant differences between FAE reperfusion and nonreperfusion patients or PAVMs in age (39 y ⫾ 21 vs 47 y ⫾ 15), male-to-female ratio (1:6 vs 1:6), ratio of HHT to non-HHT (1:6 vs 1:6), ratio of single to multiple PAVMs (3:4 vs 4:3), and ration of simple to complex PAVMs (8:3 vs 10:1). None of the 7 patients with reperfusion showed symptoms attributable to reperfusion. There were no significant differences between the groups in the proportion of the number of PAVMs between the treated periods (4 0.1) and among upper, middle, and lower lobes; diameter of the venous sac (P ¼ .66), feeding artery (P ¼ .81), and draining vein (P ¼ .18); length of the feeding
artery (P ¼ .11); number of coils used; coil combination and coils used per PAVM (P ¼ .38); coil position (P ¼ .31); and follow-up months (P ¼ .11) (Table 4). However, reperfusion occurred more frequently before compared with during and after 1997 (60% [9 of 15] vs 29% [2 of 7]) and in the PAVMs treated with 0.035-inch coils alone compared with PAVMs treated with other coils and combinations of coils (59% [10 of 17] vs 20% [1 of 5]). Concerning management after reperfusion, of the 9 of 22 PAVMs with recanalization (41%), 7 underwent additional FAE, one was treated surgically, and one could not be treated because the patient refused additional therapy and died of pneumonia 6 months after the last follow-up examination. Two patients with reperfusion from the bronchial artery alone were followed without treatment.
Changes in CT Findings of Reperfused and Nonreperfused PAVMs after Embolization All 11 FAE reperfused PAVMs showed venous sac contrast enhancement with 7 mm ⫾ 3 (range, 3–15 mm) venous sac diameter and change in diameter of 44% ⫾ 20% (range, 10%–75%) (diameter at follow-up 100%/pretherapeutic diameter) and draining vein contrast enhancement with 4 mm ⫾ 2 (range, 2–8 mm) diameter and change in diameter of 58% ⫾ 29% (range, 22%–100%) (Table 5; Fig 2c). All 11 FAE nonreperfused PAVMs showed no contrast enhancement in the venous sacs and draining veins (Fig 4a–d). The venous sac changed to complete
1574
’
Venous Sac Embolization of PAVMs
Hayashi et al
’
JVIR
Table 5 . Changes in CT Findings of Reperfused and Nonreperfused Pulmonary Arteriovenous Malformations after Embolization Feeding Artery Embolization Venous Sac Embolization Finding
Reperfusion (n ¼ 11)
Venous sac diameter
7 mm ⫾ 3 (3–15 mm)
% change
44% ⫾ 20% (10%–75%)
Enhancement Draining vein
Nonreperfusion (n ¼ 11) CR (n ¼ 11)
Nonreperfusion (n ¼ 15) NA NA
Yes
No
Diameter
4 mm ⫾ 2 (2–8 mm)
CR (n ¼ 10); 3 mm (n ¼ 1)
2 mm (n ¼ 2)
% change Enhancement
58% ⫾ 29% (22%–100%) Yes
33% (n ¼ 1)
20% (n ¼ 1); 33% (n ¼ 1) No
No
% change ¼ pretherapeutic diameter 100/diameter at follow-up; Enhancement ¼ contrast enhancement. CR ¼ complete resolution, disappearance or scarlike structure formation of sac and draining vein on CT; NA ¼ not available owing to metallic artifacts.
Figure 4. Patient 6. A 48-year-old woman with simple-type PAVM in left segment 5 treated by FAE. (a, b) Pretreatment contrast-enhanced mediastinal (a) and lung (b) window CT images reveal a large PAVM in left segment 5 (arrows). (c) Follow-up mediastinal window CT image reveals disappearance of the PAVM. (d) Follow-up lung window CT image reveals scarlike structures of the PAVM (arrow).
Figure 5. Patient 12. A 50-year-old woman with simple-type PAVM in right segment 4 treated by VSE. (a) Pretreatment contrastenhanced CT image reveals the PAVM and its draining vein (arrow). (b, c) Follow-up contrast-enhanced mediastinal (b) and lung (c) window CT images reveal a tiny scarlike structure of the draining vein without contrast enhancement (arrows).
resolution in 11 FAE nonreperfused PAVMs, whereas the draining vein changed to complete resolution in 10 FAE nonreperfused PAVMs and to 3 mm with the change in diameter of 33% in 1 FAE nonreperfused PAVM. All 15 VSE nonreperfused PAVMs showed no draining vein contrast enhancement (Fig 5a–c). The draining vein changed to complete resolution in 13 VSE nonreperfused
PAVMs and to 2 mm each with the change in diameter of 20% and 33% in 2 VSE nonreperfused PAVMs.
DISCUSSION The present study showed that reperfusion occurred solely in the FAE group, in which most patients were treated before
Volume 23
’
Number 12
’
December
’
2012
1996, at a frequency of 50% (11 of 22). No reperfusion was noted in the 15 PAVMs of the more recent VSE group. The reperfusion or recanalization rate of PAVMs with FAE using coils or balloons ranged from 0%–57% after long-term follow-up, and the wide range in reperfusion rate may be mainly related to the follow-up methods and definition of reperfusion or recanalization (7–14,28). Haitjema et al (7) used chest radiography and arterial blood gas measurement and found recanalization in 2 (2%) of 92 PAVMs. Lee et al (8) used chest radiography or non– contrast-enhanced CT and arterial blood gas measurement and found reperfusion in 8 (15%) of 53 PAVMs. Prasad et al (9) used principally non–contrast-enhanced CT and found persistence of PAVM in 18 (7%) of 267 PAVMs treated with stainless steel coils and 4 (10%) of 39 PAVMs treated with platinum coils. Mager et al (10) used chest radiography and blood gas measurement and found recanalization in 25 (8%) of 296 PAVMs. Pollak et al (11) used chest radiography and non–contrast-enhanced CT and found reperfusion in 11 (3%) of 393 PAVMs. Remy-Jardin et al (12) used non–contrast-enhanced and contrast-enhanced CT and found recanalization requiring embolization in 12 (19%) of 64 PAVMs. However, they classified treatment outcomes for four categories: 30 successful (47%), 18 partially successful (28%), 2 partially failed (3%), and 14 failed (22%) PAVMs. Anderson and Kjeldsen (13,14) used chest radiography, contrast echocardiography, and measurement of arterial blood gases and found recanalization in 8 (8%) of 106 PAVMs treated with coils and balloons (13) and in none of 13 PAVMs treated with balloons (14). Sagara et al (28) used contrastenhanced CT and found reperfusion in 8 (57%) of 14 PAVMs. However, they detected reperfusion in 1 (7%) of 14 PAVMs when they used chest radiography alone and in 2 (14%) of them when non–contrast-enhanced CT alone was used. We believe that the most reliable method to detect reperfusion is to confirm the blood flow of the venous sac and draining vein, which can be achieved by contrastenhanced CT followed by pulmonary angiography. In the present study, reperfusion was represented as measurable venous sacs 3–15 mm in diameter and draining veins 2–8 mm, all of which showed contrast enhancement on CT examinations, and a mean reduction in diameter (posttherapy/pretherapy 100%) was 44% ⫾ 20% (range, 10%– 75%) and 58% ⫾ 29% (range, 22%–100%). Mechanisms underlying reperfusion of PAVMs after embolization were summarized by Pollak et al (11) as follows: (i) recanalization of the vessel after embolization, (ii) growth of a missed or previously small accessory artery, (iii) bronchial and other systemic artery collateral flow into the pulmonary artery beyond the level of the embolization (creating a left-to-left shunt), and (iv) pulmonary artery-to-pulmonary artery collateral flow around the occlusion. In the present study, recanalization of the artery after embolization occurred in 41% (9 of 22) in the FAE group, in 23% (5 of 22) with recanalization of the
1575
artery after embolization alone, in 14% (3 of 22) also accompanied by reperfusion from the bronchial artery, and in 4.5% (1 of 22) accompanied by an accessory artery. Reperfusion from the bronchial artery alone occurred in 9% (2 of 22). Reperfusion occurred secondary to mechanisms (i)–(iii), with recanalization as the mechanism of reperfusion in 82% (9 of 11), similar to the 88% incidence reported by Milic et al (29). The last mechanism, (iv), has been observed only in young children and presumably results from the ability for continued lung growth (11). Besides reperfusion, significant differences between the FAE and VSE groups were noted in other clinical features. PAVMs significantly differed in location between FAE and VSE groups owing to distribution of 7 FAE PAVMs and no VSE PAVMs in upper lobes. However, no difference was observed in reperfusion rate between 2 (29%) of 7 FAE upper lobe PAVMs and 9 (60%) of 15 FAE middle or lower lobe PAVMs (P 4 .1). The draining vein was significantly larger in diameter in the FAE group (8 mm ⫾ 5) than the VSE group (5 mm ⫾ 2; P ¼ .045). This difference might be related to the fact that PAVMs were larger in the sac size in the FAE group (21 mm ⫾ 14) than the VSE group (15 mm ⫾ 7), although not statistically significant. Before 1996, operators performed FAE in all 15 PAVMs, and reperfusion occurred in 9 (60%) of them. Operators thereafter performed FAE in 7 (32%) and VSE in 15 (68%) of 22 PAVMs, and reperfusion occurred in 2 (9%) of them. This difference (P o .01) coincided with differences in the types and number of coils used and their combination (ie, the 0.035-inch and 0.018-inch coils alone were available before 1996, and IDCs were available after 1996). The types and number of coils used differed significantly between both groups; 0.035-inch coils were used predominantly in the FAE group (79% [70 of 89]). The IDCs were used most frequently (58% [70 of 120]) followed by 0.018inch coils (40% [48 of 120]), and only two 0.035-inch coils were used in the VSE group. In the FAE group, 0.035-inch coils alone were used in 17 of 22 PAVMs, but 0.035-inch coils alone were not used in the VSE group. Reperfusion occurred significantly more frequently in the group that used 0.035-inch coils alone (56% [10 of 18]) than in the other group (5% [1 of 19]; P o .01). Comparison of reperfused and nonreperfused PAVM groups in FAE showed no significant differences in all characteristics of PAVMs, including the diameter of the venous sac, feeding artery, and draining vein and length of the feeding artery. Characteristics of PAVMs in the present study might not be responsible for reperfusion. Although other parameters, including treated periods, the number of coils used, and coil combination, were also not significantly different between reperfused and nonreperfused groups, reperfused PAVMs occurred predominantly from 1989–1996 (82% [9 of 11]) before 0.018-inch IDCs became available and were more frequent in the group that used 0.035-inch coils alone (59% [10 of 17]) than in the other coil combinations (20% [1 of 5]). Causes of
1576
’
Venous Sac Embolization of PAVMs
reperfusion might be related to insufficient distal embolization because of the use of larger 0.035-inch coils, including inappropriate size selection and deployment position of the coils through a 5-F JB1 catheter. No reperfused PAVMs were encountered in the VSE group. No significant differences in characteristics of PAVMs were evident between the FAE and VSE groups except for a marginally significant larger diameter of the draining veins in the FAE group as mentioned previously. These results suggest that differences in reperfusion rates between FAE and VSE groups might be due to embolization methods using the different types of coils and number of coils used. FAE was performed using larger coils through a larger catheter. It might be more difficult to select appropriate coils in size and number and appropriate position for deployment than with VSE. VSE was performed through a microcatheter to achieve successful deployment of smaller coils in the sac using more appropriate coils in size and number. Coley and Jackson (21) argued theoretically against the routine placement of coils in the venous sac of PAVMs. First, rupture might occur during coil packing. However, rupture did not occur in our series or in other cases in which VSE was performed (20–23). VSE avoids risky tension that may induce rupture of the venous sac, using IDCs with soft, compliant characteristics that enhance fit. Second, thrombus forms on the coils during embolization, and dislodgment during further packing of the sac is perhaps more likely to occur than when the arterial neck is being occluded. This procedure needs many more coils for VSE compared with FAE, prolonging procedure time. However, we did not encounter thrombus migration into the draining vein during VSE. To avoid paradoxical embolization of the thrombus, we used heparin before and during the procedure in all cases. If thrombus dislodgment is anticipated, we recommend flow control with balloon occlusion of the feeding artery in large sac cases (30). Third, coils protruding from the venous sac into normal pulmonary veins might impair pulmonary venous drainage from the normal lung. Accurate deployment of IDCs with soft and compliant nature by the detachment system can prevent coils protruding into the normal pulmonary vein. Fourth, if platinum microcoils are used, the reduced radial force compared with conventional steel coils might increase the risk of migration. However, to avoid this risk, we used IDCs of a suitable size to fit the venous sac and deployed these carefully in the venous sac. Even if conventional steel coils were used, coils of an unsuitable size would result in systemic migration. We encountered one case (patient 16 in Table 2) in which a 0.035-inch pushable coil was dislocated from the feeding artery into the venous sac. This coil did not migrate into the systemic circulation and remained within the venous sac. VSE was performed in this case without complications. Selection of coils larger than the drainage vein can reduce the risk of coil migration in VSE, despite the slight radial force of the IDC. However, pushable steel coils may be
Hayashi et al
’
JVIR
associated with a risk of injury to the sac walls and cause rupture of the venous sac. Fifth, the venous sac cannot decompress after this form of treatment. However, after successful conventional embolization of the neck of a PAVM, the sac can decompress. Sac decompression in FAE is useful to evaluate treatment success. However, in VSE, success can be evaluated from complete resolution and no enhancement of the draining vein using CT. VSE may not be always indicated for PAVMs. FAE may be solely indicated for the following PAVMs: (i) venous sac filled with clot, in which FAE is recommended to prevent paradoxical embolization of clot in the venous sac (31); (ii) diffuse or complex type, in which VSE is difficult to perform when such PAVMs have no venous sacs and are composed of multiple or small vessel shunts (24,32) similar to the diffuse type (patient 9) in the present study; (iii) giant venous sac 4 3 cm in diameter, in which VSE requires much more coils, and packing the venous sac is more expensive than FAE; and (iv) PAVM with venous sac in which the microcatheter cannot be put stably. As shown in Table 2, the type and number of coils used were not parallel with the venous sac size in VSE; this suggests that the sac size alone was not the determinant of the type and number of coils used for VSE. For example, the venous sac was 15 mm in the PAVMs in segment 10 (Fig 1a–c) and segment 4 (Fig 2a–c) in Table 2; for the former PAVM, 17 IDCs were used for embolization, and for the latter, 4 IDCs and 2 0.018-inch coils were used. Embolization was successful in both cases. A key issue may be how to perform embolization of the proximal portion of the large venous sac completely without coil migration to reduce the number of appropriate coils to as few as possible to achieve successful embolization. For FAE, the AMPLATZER Vascular Plug (St. Jude Medical, St. Paul, Minnesota) (15–18,33) and 0.035-inch Interlock fibered IDC (Boston Scientific) are available in the United States and Europe; these may offer sufficient radial force for FAE and allow selection of devices with an appropriate size and deployment in an appropriate distal position of the feeding artery. Vascular plugs with or without coils shorten the procedure time (15) and decrease reperfusion rates, which were reported from 0–7% (16– 18,33). More recently, another promising result was reported for embolization of complex PAVMs with large outflow vessels and short feeding arteries using VSE with detachable coils combined with FAE with AMPLATZER Vascular Plugs (34). Although we have not yet performed embolization for PAVMs with feeding arteries o 3 mm, the need to treat such small PAVMs has been recognized when they are accessible (19,35). VSE would be one of the therapy options for such small PAVMs, if IDCs with soft and compliant characteristics can be safely put in the venous sac through a microcatheter via the small feeding artery. The following issues must be considered when interpreting the results of the present study. First, the small sample size and retrospective design are key limitations,
Volume 23
’
Number 12
’
December
’
2012
and the procedure time was not recorded, although about 2 hours was needed to complete one therapeutic session. Second, the high reperfusion rate of FAE was observed in the older period when 0.035-inch coils were used and small IDCs were unavailable. Third, although no significant difference was evident, follow-up periods tended to be shorter with VSE than with FAE. Further follow-up is necessary for VSE, given later (beyond 2 years to 76 months) occurrence of about half of reperfusions in FAE in the present study. Fourth, we measured the feeding artery, venous sac, and draining vein in diameter. Volumetric measurement of these structures using modern multidetector IVR-CT/angiosystem can serve our selection of more appropriate type and number of coils for embolization. In conclusion, although more coils are needed, VSE can be used to treat PAVMs with a venous sac safely and achieve long-term efficacy.
REFERENCES 1. Bosher LH Jr, Blake DA, Byrd BR. An analysis of the pathologic anatomy of pulmonary arteriovenous aneurysms with particular reference to the applicability of local excision. Surgery 1959; 45:91–104. 2. White RI Jr, Pollak JS, Wirth JA. Pulmonary arteriovenous malformations: diagnosis and transcatheter embolotherapy. J Vasc Interv Radiol 1996; 7:787–804. 3. Khurshid I, Downie GH. Pulmonary arteriovenous malformation. Postgrad Med J 2002; 78:191–197. 4. Govani FS, Shovlin CL. Hereditary haemorrhagic telangiectasia: a clinical and scientific review. Eur J Hum Genet 2009; 17:860–871. 5. Porstmann W. Therapeutic embolization of arteriovenous fistula by catheter technique. In: Kelop O, editor. Current Concepts in Pediatric Radiology. Berlin, Germany: Springer; 1977. p. 23–31. 6. Dutton JA, Jackson JE, Hughes JM, et al. Pulmonary arteriovenous malformations: results of treatment with coil embolization in 53 patients. AJR Am J Roentgenol 1995; 165:1119–1125. 7. Haitjema TJ, Overtoom TT, Westermann CJ, et al. Embolisation of pulmonary arteriovenous malformations: results and follow up in 32 patients. Thorax 1995; 50:719–723. 8. Lee DW, White RI Jr, Egglin TK, et al. Embolotherapy of large pulmonary arteriovenous malformations: long-term results. Ann Thorac Surg 1997; 64:930–940. 9. Prasad V, Chan RP, Faughnan ME. Embolotherapy of pulmonary arteriovenous malformations: efficacy of platinum versus stainless steel coils. J Vasc Interv Radiol 2004; 15:153–160. 10. Mager JJ, Overtoom TT, Blauw H, et al. Embolotherapy of pulmonary arteriovenous malformations: long-term results in 112 patients. J Vasc Interv Radiol 2004; 15:451–456. 11. Pollak JS, Saluja S, Thabet A, et al. Clinical and anatomic outcomes after embolotherapy of pulmonary arteriovenous malformations. J Vasc Interv Radiol 2006; 17:35–44. 12. Remy-Jardin M, Dumont P, Brillet PY, et al. Pulmonary arteriovenous malformations treated with embolotherapy: helical CT evaluation of longterm effectiveness after 2-21-yearfollow-up. Radiology 2006; 239: 576–585. 13. Andersen PE, Kjeldsen AD. Clinical and radiological long-term follow-up after embolization of pulmonary arteriovenous malformations. Cardiovasc Intervent Radiol 2006; 29:70–74. 14. Andersen PE, Kjeldsen AD. Long-term follow-up after embolization of pulmonary arteriovenous malformations with detachable silicone balloons. Cardiovasc Intervent Radiol 2008; 31:569–574.
1577
15. Abdel Aal AK, Hamed MF, Biosca RF, et al. Occlusion time for Amplatzer vascular plug in the management of pulmonary arteriovenous malformations. AJR Am J Roentgenol 2009; 192:793–799. 16. Hart JL, Aldin Z, Braude P, et al. Embolization of pulmonary arteriovenous malformations using the Amplatzer vascular plug: successful treatment of 69 consecutive patients. Eur Radiol 2010; 20:2663–2670. 17. Letourneau-Guillon L, Faughnan ME, Soulez G, et al. Embolization of pulmonary arteriovenous malformations with Amplatzer vascular plugs: safety and midterm effectiveness. J Vasc Interv Radiol 2010; 21: 649–656. 18. Tapping CR, Ettles DF, Robinson GJ. Long-term follow-up of treatment of pulmonary arteriovenous malformations with AMPLATZER Vascular Plug and AMPLATZER Vascular Plug II devices. J Vasc Interv Radiol 2011; 22:1740–1746. 19. Trerotola SO, Pyeritz RE, Bernhardt BA. Outpatient single-session pulmonary arteriovenous malformation embolization. J Vasc Interv Radiol 2009; 20:1287–1291. 20. White RI Jr. Pulmonary arteriovenous malformations: how do I embolize? Tech Vasc Interv Radiol 2007; 10:283–290 21. Coley SC, Jackson JE. Venous sac embolization of pulmonary arteriovenous malformations in two patients. AJR Am J Roentgenol 1996; 167:452–454. 22. Takahashi K, Tanimura K, Honda M, et al. Venous sac embolization of pulmonary arteriovenous malformation: preliminary experience using interlocking detachable coils. Cardiovasc Intervent Radiol 1999; 22: 210–213. 23. Dinkel HP, Triller J. Pulmonary arteriovenous malformations: embolotherapy with superselective coaxial catheter placement and filling of venous sac with Guglielmi detachable coils. Radiology 2002; 223: 709–714. 24. White RI Jr, Mitchell SE, Barth KH, et al. Angioarchitecture of pulmonary arteriovenous malformations: an important consideration before embolotherapy. AJR Am J Roentgenol 1983; 140:681–686. 25. Shovlin CL, Guttmacher AE, Buscarini E, et al. Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber syndrome). Am J Med Genet 2000; 91:66–67. 26. Remy J, Remy-Jardin M, Wattinne L, et al. Pulmonary arteriovenous malformations: evaluation with CT of the chest before and after treatment. Radiology 1992; 182:809–816. 27. Sacks D, McClenny TE, Cardella JF, et al. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14(9 Pt 2): S199–S202. 28. Sagara K, Miyazono N, Inoue H, et al. Recanalization after coil embolotherapy of pulmonary arteriovenous malformations: study of long-term outcome and mechanism for recanalization. AJR Am J Roentgenol 1998; 170:727–730. 29. Milic A, Chan RP, Cohen JH, et al. Reperfusion of pulmonary arteriovenous malformations after embolotherapy. J Vasc Interv Radiol 2005; 16:1675–1683. 30. Mori K, Shiigai M, Saida T, et al. A modified metallic coil embolization technique for pulmonary arteriovenous malformations using coil anchors and occlusion balloon catheters. Cardiovasc Intervent Radiol 2008; 31: 638–642. 31. Graves AD, Gregorius JC, Smith DC. Management of a patient with a clot-filled pulmonary arteriovenous malformation. J Vasc Interv Radiol 2009; 20:652–655. 32. Wei CW, Faughnan ME, Menard A, et al. Lobar embolization for treatment of diffuse pulmonary arteriovenous malformations in hereditary hemorrhagic telangiectasia: a case report. J Vasc Interv Radiol 2010; 21:1105–1108. 33. Trerotola SO, Pyeritz RE. Does use of coils in addition to Amplatzer vascular plugs prevent recanalization? AJR Am J Roentgenol 2010; 195: 766–771 34. Hundt W, Kalinowski M, Kiessling A, et al. Novel approach to complex pulmonary arteriovenous malformation embolization using detachable coils and Amplatzer vascular plugs. Eur J Radiol 2012; 81:e732–e738. 35. Trerotola SO, Pyeritz RE. PAVM embolization: an update. AJR Am J Roentgenol 2010; 195:837–845.
Efficacy of Venous Sac Embolization for Pulmonary Arteriovenous Malformations: Comparison with Feeding Artery Embolization Sadao Hayashi, MD, PhD, Yasutaka Baba, MD, PhD, Terutoshi Senokuchi, MD, and Masayuki Nakajo, MD, PhD
ABSTRACT Purpose: To examine the efficacy of venous sac embolization (VSE) in comparison with transcatheter feeding artery embolization (FAE) for treatment of pulmonary arteriovenous malformations (PAVMs). Materials and Methods: From 1989–2009, 21 patients underwent embolization of 37 PAVMs. Safety and long-term efficacy of VSE were evaluated retrospectively and compared with FAE. Results: FAE was performed in 22 (18 simple and 4 complex type) PAVMs, and VSE was performed in 15 (14 simple and 1 complex type) PAVMs. There were significant differences between FAE and VSE in treated periods, PAVM location, coil type used, number of coil combinations and coils per PAVM, coil position, and reperfusion; there were no significant differences in most PAVM characteristics, follow-up durations (58 mo ⫾ 54 vs 42 mo ⫾ 42; P ¼ .32), and minor complications (pleurisy [2 vs 2]). Reperfusion occurred in 11 (50%) of 22 PAVMs in the FAE group and no PAVMs in the VSE group (P o .01). Of 22 PAVMs in the FAE group, 17 (77%) were treated with 0.035-inch coils alone; of 15 PAVMs in the VSE group, 14 (93%) were treated with 0.018-inch interlocking detachable coils (IDCs), 0.018-inch pushable fibered coils, or IDCs and pushable fibered coils combined (P o .01). The number of coils used was 8 ⫾ 4 in the VSE group and 4 ⫾ 4 in the FAE group (P ¼ .002). Conclusions: The high reperfusion rate in the FAE group was mainly due to the use of large 0.035-inch coils alone. Although more coils are needed, VSE can be used to treat PAVMs with a venous sac safely and achieve long-term efficacy.
ABBREVIATIONS: FAE = feeding artery embolization, HHT = hereditary hemorrhagic telangiectasia, IDC = interlocking detachable coil, PAVM = pulmonary arteriovenous malformation, VSE = venous sac embolization
Pulmonary arteriovenous malformations (PAVMs) are rare pulmonary vascular anomalies and direct pulmonary artery-to-vein connections without intervening capillary beds. PAVMs are most commonly congenital in nature; can occur as either an isolated abnormality or part of hereditary hemorrhagic telangiectasia (HHT); and are associated with a risk of paradoxical embolization, resulting in events such as stroke, transient ischemic attack, and brain abscess, with the strong potential for severe From the Department of Radiology, Kagoshima University Graduate School of Medical and Dental Sciences, 8-35-1 Sakuragaoka, Kagoshima-shi, Kagoshima 890-8520, Japan. Received May 10, 2012; final revision received September 6, 2012; accepted September 9, 2012. Address correspondence to S.H.; E-mail: [email protected] None of the authors have identified a conflict of interest. & SIR, 2012 J Vasc Interv Radiol 2012; 23:1566–1577 http://dx.doi.org/10.1016/j.jvir.2012.09.008
neurologic deficits (1–4). Treatment is indicated to obliterate or reduce the shunt and prevent neurologic complications. Embolization was introduced as the treatment of choice for PAVMs by Porstmann in 1977 (5). Since then, several published series have documented the efficacy of embolization with metallic coils or detachable balloons in the treatment of PAVMs in patients with hypoxemia, patients with neurologic complications, or asymptomatic patients with a feeding artery Z 3 mm in diameter, obviating the need for surgery in most cases; the longterm efficacy of embolization for PAVMs has also been documented (6–14). More recently, the AMPLATZER Vascular Plug (St. Jude Medical, St. Paul, Minnesota) has also been used for embolization of PAVMs (15–18). Embolization for PAVM can be performed in the outpatient setting (19,20). Although most PAVMS are successfully treated, recanalization or reperfusion can occur after the first embolization in some patients during longterm follow-up.
Volume 23
’
Number 12
’
December
’
2012
The standard percutaneous technique is transcatheter feeding artery embolization (FAE), involving the deployment of coils, balloons, or plugs within the feeding artery as close to the neck of the PAVM as possible, to avoid occlusion of vessels to the normal lung (3,11,20). Transcatheter venous sac embolization (VSE) using interlocking detachable coils (IDCs) has also been used when the target artery of the PAVM is too short to avoid sacrifice of large normal pulmonary artery branches or when the artery is a high-flow type with a higher risk of paradoxical embolization of a coil or balloon (21–23). In addition, VSE using detachable coils may be useful to resolve the risk of migration of embolic materials in embolization of PAVMs (22,23). We reviewed embolization procedures for PAVMs that were performed in our department during the past 21 years to examine the safety and long-term efficacy of VSE in comparison with FAE.
MATERIALS AND METHODS This retrospective study had institutional review board approval, and informed consent was waived.
Patients We reviewed clinical charts and angiographic and computed tomography (CT) images of patients with PAVMs who underwent embolization in our department and identified 21 patients treated between October 1989 and May 2009. No patient was excluded, and the study population comprised these 21 patients.
Angiographic and Embolization Techniques After written informed consent was obtained from all patients, all pulmonary angiographic and embolization procedures were performed by the Seldinger method. Because the procedures were done over the course of 21 years, the operators also changed in this period. These
1567
procedures were performed by at least two operators at each therapeutic session, and one operator had a minimum of 4 5 years of experience in endovascular intervention. To prevent clot formation during angiographic procedures, a dose of 5,000 units (1,000 units/mL) of heparin was administered intravenously, and an additional 1,000 units of heparin was added every hour when necessary. Pulmonary angiography was performed with a 5-F pigtail catheter (Medikit Co, Ltd, Tokyo, Japan) through the femoral vein under local anesthesia. Selective right and left pulmonary angiograms were obtained in frontal and lateral projections using a 5-F JB1 catheter (Create Medic Co, Ltd, Tokyo, Japan). The type, size, and number of available coils and their deployment positions at the feeding artery or venous sac were determined by the operators from the findings on selective angiography. When the feeding artery of the PAVM was visualized, the JB1 catheter was inserted into the feeding artery and advanced as close to the venous sac as possible. FAE was performed using 0.035-inch Gianturco-Wallace stainless steel coils (coil size range, 3–8 mm) (Cook, Inc, Bloomington, Indiana) or 0.035-inch VortX 35 platinum coils (coil size range, 3–8 mm) (Boston Scientific, Natick, Massachusetts); the diameter of the coils used was about 20% larger than the diameter of the feeding artery undergoing embolization. For VSE, a 2.4-F Renegade microcatheter (Boston Scientific) was advanced coaxially through the JB1 catheter into the venous sac via the feeding artery. In the embolization, we used 0.018inch IDCs (Boston Scientific) about 20% larger than the drainage vein at first, which have been available since 1997. We introduced the larger sized IDCs (coil size range, 4–20 mm) in the venous sac first, detaching the coils after confirming the size and position fit for the sac. A wire mesh of IDCs was created to prevent coil migration of the subsequent pushable small coils (coil size range, 3–8 mm) (ie, 0.018-inch fibered platinum coils; [Boston Scientific] for embolization of the venous sac (Fig 1a–c). Complete occlusion of the feeding artery or venous sac was confirmed with pulmonary angiography performed
Figure 1. Patient 18. A 15-year-old boy with complex-type PAVM in right segment 4 treated by VSE. (a) Pulmonary arteriogram shows a complex-type PAVM with two feeding artery branches and a venous sac. (b) Radiograph shows a venous sac filled with 17 IDCs, of which the largest one is 18 mm in maximum diameter. (c) Pulmonary arteriogram reveals no contrast material in the PAVM and draining vein after VSE.
1568
’
Venous Sac Embolization of PAVMs
Hayashi et al
’
JVIR
Figure 2. Patient 13. A 67-year-old woman with two simple-type PAVMs treated by embolization. (a) Pulmonary arteriogram shows one PAVM in right segment 4 treated by VSE and the other in right segment 10 treated by FAE. (b) Pulmonary arteriogram demonstrates complete packing of the PAVM in right segment 4 treated by VSE and recanalization of the PAVM in right segment 10 treated by FAE 68 months later. (c) Contrast-enhanced CT image shows the contrast-enhanced draining vein (arrow) of the recanalized segment 10 PAVM.
after embolization in all PAVMs. After IDCs became available, VSE was deemed preferable when the PAVM had a venous sac; it was also thought that VSE was beneficial in PAVMs with large outflow vessels or short feeding arteries (22) because the microcatheter enabled us to avoid injuring the vessel wall and carry the smaller coils safely into the venous sac. Also, the 5-Fr JB1catheter occasionally could not be advanced enough to reach the proximal site of the sac with or without spasms of the feeding artery. However, VSE was not performed when the microcatheter could not be put stably in the venous sac.
Follow-up Studies for Reperfusion Chest radiography and non–contrast-enhanced and contrastenhanced CT were planned principally for follow-up studies at 3–6 months, 6–12 months, and every 1–2 years thereafter. When the venous sac and draining vein in the case of FAE and the draining vein in the case of VSE showed contrast enhancement on CT, which indicated reperfusion, angiography was planned. When these structures were visible without contrast enhancement, they were followed up by chest radiography and non–contrast-enhanced and contrastenhanced CT. When complete resolution (ie, disappearance or scarlike structure formation of the venous sac or draining vein) was noted without contrast enhancement on CT scan, chest radiography and non–contrast-enhanced CT were principally planned for the next follow-up examination. Non–contrast-enhanced and contrast-enhanced CT scan of the chest was performed using a single-detector CT scanner (X-Vigor; Toshiba, Tokyo, Japan) or 16-detector CT scanner (Aquilion 16; Toshiba). Using the singledetector CT scanner, all scans began from the apex of the lung to the diaphragm. Additional section at selected levels, synchronized scanning with longitudinal table motion at a speed of one section thickness per second (10-mm collimation, 120 kVp, 200 mA), and two subsequent sequences were used to evaluate contrast enhancement of the PAVM
(5-mm collimation, 120 kVp, 250 mA; 2-mm collimation, 120 kVp, 230 mA) after a bolus injection of 100 mL of nonionic iodinated contrast material. Using the 16-detector CT scanner, all scans began from the apex to the diaphragm with tube voltage of 120 kVp and gantry rotation speed of 0.5 second/rotation. Maximum allowable tube current was 400 mA, detector row configuration was 16 1 mm, and table increment was 15 mm/rotation. Contrast-enhanced CT scans were performed with the following image parameters: maximum allowable tube current of 400 mA, detector row configuration of 16 0.5 mm, and table increment of 7.5 mm/rotation in the longitudinal direction. Bolus tracking technique (Sure Start software; Toshiba) was used to time the start of the scanning after contrast material injection. Dynamic contrast-enhanced CT scan was performed after a bolus injection of 100 mL of nonionic iodinated contrast material. Reperfusion of the PAVM was confirmed with pulmonary arteriography and bronchial arteriography (Figs 2a–c and 3a and b). Bronchial angiography was performed using a 5-F shepherd’s hook catheter (Hanako Medical, Tokyo, Japan) through the femoral artery under local anesthesia. We defined reperfusion as sac or draining vein blood flow regardless of the inflow route and recanalization as sac blood flow from the recanalized artery that had previous embolization with coils.
Comparative Studies of FAE and VSE We reviewed the patients’ age and sex; the year when embolization was performed; type and location of the PAVM; the initial findings of the venous sac, feeding artery, and its branches and draining vein; the follow-up findings of the venous sac and draining vein; the types and number of coils used; coil position; follow-up periods; presence and absence of reperfusion; and complications related to embolization. We classified PAVMs into simple type with a segmental feeding artery or its branch and complex type with two or more segmental arteries or branches according to the classification by White et al in
Volume 23
’
Number 12
’
December
’
2012
1569
classified as minor or major according to Society of Interventional Radiology (SIR) guidelines (27).
Morphologic Measurement Methods The initial maximum diameter of the venous sac, feeding artery, and draining vein was measured on contrastenhanced CT or the angiogram obtained before embolization when contrast-enhanced CT was unavailable. The length of the feeding artery or its branch was measured from the bifurcated point of the segmental artery or its branch to the proximal point of the venous sac on the angiogram obtained before embolization. Coil position was measured from the distal point of coils in the feeding artery to the proximal point of the venous sac on the angiogram obtained after embolization. The follow-up maximum diameter of the venous sac or draining vein was measured on the follow-up pulmonary window CT images.
Statistical Methods Comparison between groups were made by using the binominal, w2, or Fisher exact test for categorical variables. The Mann-Whitney U test was used to analyze the differences of continuous variables in venous sac diameter, diameter and length of the feeding artery, diameter of the draining vein, and coil placement position between FAE and VSE groups or between FAE reperfusion and nonreperfusion groups. Data are expressed as mean ⫾ standard deviation and range for continuous variables. P values o .05 were considered significant. Statistical analyses were performed on a computer using StatView version 5.0 software (SAS Institute, Cary, North Carolina).
RESULTS
Figure 3. Patient 8. A 51-year-old woman with simple-type PAVM in right segment 8 treated by FAE. (a) Pulmonary arteriogram shows recanalization of the PAVM (arrow). (b) Bronchial arteriogram shows reperfusion of the PAVM (arrow) from the bronchial artery at the time of additional FAE.
1983 (24) for our analysis, although the newly defined complex type needs two or more segmental arteries (2). HHT was diagnosed by the Curac- ao criteria (25). In each case, the diagnosis of PAVM was established based on findings from thoracic CT and pulmonary angiography (26). The follow-up months were from the date of embolization to the follow-up date of the last contrastenhanced CT scan in patients without reperfusion and to the date of reperfusion in patients with reperfusion noted on contrast-enhanced CT scan. Complications were
Follow-up results were available for all 21 patients. The last follow-up examination was performed in May 2011. A PAVM was initially suspected because of a nodule with vessel-like structures on chest radiographs or unenhanced CT and subsequently confirmed with contrast-enhanced CT and pulmonary angiography to identify the feeding artery or arteries, venous sac or shunt, and draining vein. Detailed data of FAE and VSE groups are presented in Tables 1 and 2.
Overall Results Four patients received both FAE and VSE. Age range was 15–67 years (mean, 43 years ⫾ 16). Female patients constituted 81% (17 of 21) of the study group. Patient and PAVM breakdown was as follows: 3 patients with HHT and 18 patients without HHT; 12 patients with single PAVM and 9 patients with multiple PAVMs; and 37 PAVMs (32 simple and 5 complex types; 15 treated before 1996 and 22 treated in 1997 and thereafter; 7 upper, 9 middle, and 21 lower lobe PAVMs; and 24 right and 13 left ones). One of five complex types was diffuse type, which was located at the left lower lobe supplied by left A8, A9,
1570
Table 1 . Characteristics and Therapeutic Results of Pulmonary Arteriovenous Malformations Initially Treated by Feeding Artery Embolization
’
Initial
Follow-up
Number of Coils Used
Patient No./Age/Sex Year No.
Type
Feeder
Dsac 15
3
1/19/F
1989
1
Complex
Lt. A9
2/63/F 3/17/F
1991 1991
2 3
Simple Simple
Lt. A3 Rt. A6
20 20
4/62/M*
1992
4
Simple
Rt. A3
1993 1992
5 6
Simple Simple
1993
7
Complex
5
36
8 8
10 10
20
5
Rt. A10 Rt. A9
50 5
Lt. A8
20
Lt. A9
8
35 coil
18 coil
Coil
Follow-up
Position
(mo)
Reperfused Artery
Management of Reperfusion Additional FAE
4
þ
1
32
66
Rec. and bronchial
30 50
CR 15
CR 7
þ
1 1
15 32
175 36
Rec. and accessory
Operation
8
47
5
3
þ
3
18
24
Rec.
Additional FAE
10 3
15 5
15 36
5 3
3 3
þ þ
8 3
6 15
22 24
Bronchial Rec.
Follow-up Follow-up
5
12
50
5
3
þ
22
Bronchial
Follow-up
23
Rec.
Additional FAE
Rec. and bronchial Rec. and bronchial
Additional FAE Additional FAE
22
5 3
30 30
Complex
Lt. A10
10
5
8
40
5
8
þ
8
30
3 2
17 24
4
15
1994
9
Simple
Lt. A10 Lt. A3
40
3 9
20
32 10
CR
CR
7
5/48/F
1994
10
Simple
Lt. A8
13
3
4
36
CR
CR
1
6/48/F 7/43/F*
1994 1994
11 12
Simple Simple
Lt. A5 Rt. A1
20 10
5 3
7 4
30 18
CR CR
CR CR
1994
13
Simple
Rt. A2
55
7
14
30
CR
CR
8/51/F
1994 1995
14 15
Simple Simple
Lt. A1þ2 Rt. A8
15 10
3 7
2 9
20 47
8 7
2 2
9/15/F
1997
16
Complex
Lt. A8
Diffuse
8
15
30
CR
CR
30 20
CR CR
Lt. A9 Lt. A4
IDC
7
2
Lt. A8 Lt. A8 1992
Dfa Ddv Lfa Dsac Ddv CE
7 3
1
27
18
160
3 1
24 5
159 15
2
22
15
þ þ
1 3
3 12
4 45
5
5
137
8 2
5 10
10/37/M*
1997
17
Simple
Lt. A10
40
7
9
36
CR
3
11/55/F
1997 1997
18 19
Simple Simple
Rt. A8 Rt. A8
15 20
3 4
3 5
38 46
CR CR
CR CR
12/50/F
1999
20
Simple
Rt. A5
15
3
3
28
5
2
þ
13/67/F 14/43/F
2003 2008
21 22
Simple Simple
Rt. A10 Rt. A3
20 3
5 3
5 5
32 13
7 CR
3 CR
þ
1
1 3 3 2 3 3
5 1
5
26
30 5
27 109
7
76
Rec.
Additional FAE
3 3
68 19
Rec.
Additional FAE Hayashi et al ’
JVIR
Key to headings and abbreviations, in order of appearance in table. PAVM ¼ pulmonary arteriovenous malformations; Year ¼ year of FAE; Dsac ¼ diameter of sac; Dfa ¼ diameter of feeding artery; Ddv ¼ diameter of draining vein; Lfa ¼ length of feeding artery; CE ¼ contrast enhancement of venous sac and drainage vein; 35 coil ¼ 0.035-inch pushable fibered coil; IDC ¼ interlocking detachable coil; 18 coil ¼ 0.018-inch pushable fibered coil; Coil Position ¼ distance from sac (mm); Lt. ¼ left; Rt. ¼ right; A ¼ segmental pulmonary artery; CR ¼ complete resolution, disappearance or scarlike structure formation of sac and draining vein on CT; Rec. ¼ recanalized; FAE ¼ feeding artery embolization; Diffuse ¼ diffuse telangiectatic communications. n Patient with hereditary hemorrhagic telangiectasia. Reperfusion was noted in 11 of 22 PAVMs.
Venous Sac Embolization of PAVMs
Findings of PAVM (unit; mm)
Volume 23 ’
Number 12
Table 2 . Characteristics and Therapeutic Results of Pulmonary Arteriovenous Malformations Initially Treated by Venous Sac Embolization
(mo)
Number of Coils Used
Patient
35 Year
No.
Type
Feeder
Dsac
Dfa
Ddv
Lfa
Ddv
CE
10/37/M*
1997
1
Simple
Rt. A9
25
7
10
26
2
1997
2
Simple
Lt. A4
8
3
3
28
1999 2003
3 4
Simple Simple
Rt. A4 Rt. A4
20 15
6 5
7 5
40 15
coil
IDC
18 coil
1
6
CR
2
CR CR
3 4
2
150 68
27 26
14/43/F
2007
5
Simple
Rt. A8
15
4
5
15
CR
6
2
41
6 7
Simple Simple
Rt. A5 Rt. A7
4 15
3 3
4 5
27 24
CR CR
4
15/45/F
2008 1998
3 11
19 82
16/45/F
2002
8
Simple
Rt. A8
30
4
6
35
2
17/45/F 18/15/M
2005 2006
9 10
Simple Complex
Lt. A4 Rt. A4
15 15
4 4
5 5
65 17
CR CR
3 17
2
2
15
CR
0
19/51/F
2006 2006
11 12
Simple Simple
Rt. A8 Rt. A8
15 17
3 3
3 3
38 36
CR CR
2006
13
Simple
Rt. A9
7
4
4
24
CR
20/46/F 21/27/M
2007 2009
14 15
Simple Simple
Rt. A8 Lt. A8
8 15
3 3
4 4
15 16
CR CR CR
Rt. A4
2
2
8
2012
No./Age/Sex
12/50/F 13/67/F
Follow-up
’
Follow-up
December
Initial
’
Findings of PAVM (unit; mm)
100 4
12 25
3 3
4 5
4 4
3
3
4
6 7
2 6
45 17
Key to headings and abbreviations, in order of appearance in table. PAVM ¼ pulmonary arteriovenous malformations; Year ¼ year of venous sac embolization; Dsac ¼ diameter of sac; Dfa ¼ diameter of feeding artery; Ddv ¼ diameter of draining vein; Lfa ¼ length of feeding artery; CE ¼ contrast enhancement of drainage vein; 35 coil, 0.035-inch pushable fibered coil; IDC ¼ interlocking detachable coil; 18 coil ¼ 0.018-inch pushable fibered coil; Rt. ¼ right; Lt. ¼ left; A ¼ segmental pulmonary artery; CR ¼ complete resolution, disappearance or scarlike structure formation of sac and draining vein on CT. n Patient with hereditary hemorrhagic telangiectasia. No reperfusion was noted in all 15 PAVMs.
1571
1572
’
Venous Sac Embolization of PAVMs
Hayashi et al
and A4 segmental arteries with multiple small arteriovenous shunts. Embolization was performed on average in one patient (range, 0–4) and 2 PAVMs (range, 0–6) per year during 21 years. The diameters of the venous sac, feeding artery, and draining vein were 18 mm ⫾ 12 (range, 3–55 mm), 5 mm ⫾ 2 (range, 2–10 mm), and 7 mm ⫾ 4 (range, 2–20 mm). The length of the feeding artery was 30 mm ⫾ 12 (range, 10–65 mm). For 37 PAVMs, 209 coils (72 0.035-inch coils, 61 0.018 pushable fibered coils, and 76 IDCs) were used. There were 17 PAVMs treated with 0.035-inch coils alone and 20 PAVMs treated with other coils or a combination of coils. The number of coils used was 6 ⫾ 4 (range, 1–17) per PAVM. Reperfusion occurred in 11 (30%) of 37 PAVMs. The follow-up period was 51 months ⫾ 50 (range, 4–175 months). There were no major procedure-related complications (ie, paradoxical embolization owing to embolization device or blood clot and vessel perforation). A minor complication of chest pain secondary to pleurisy was noted in four patients and disappeared within a few days without medication.
Results of Comparison between FAE and VSE There were no significant differences between FAE and VSE in the number of patients (14 vs 11) and PAVMs (22
’
JVIR
vs 15), age (44 years ⫾ 17 vs 43 years ⫾ 13), male-tofemale ratio (2:12 vs 3:8), ratio of HHT to non-HHT (3:11 vs 1:10), ratio of single to multiple PAVMs (7:7 vs 6:5), ratio of simple to complex PAVMs (18:4 vs 14:1), and minor complications (pleurisy) related to embolization (2 vs 2). The other clinical features of FAE and VSE groups are summarized in Table 3. Significant differences between FAE and VSE groups were as follows. FAE was performed in all 15 PAVMs between 1989 and 1996, and VSE was performed in 15 (68%) of 22 PAVMs in 1997 and thereafter, when IDCs became available (P o .01). Of 22 PAVMs treated by FAE, 15 (68%) were located in the middle and lower lobes, whereas all 15 (100%) PAVMs treated by VSE were located in these lobes (P o .05). The diameter of the draining vein was larger in the FAE group (8 mm ⫾ 5) than the VSE group (5 mm ⫾ 2) with marginal significance (P ¼ .045). In the FAE group, 70 (79%) of 89 coils used were 0.035-inch coils, whereas in the VSE group, 70 (58%) and 48 (40%) of 120 coils used were IDCs (70 coils) and 0.018-inch coils (48 coils) (P o .01). In the FAE group, 17 (77%) of 22 PAVMs were treated with 0.035-inch coils alone, whereas in the VSE group, 14 (93%) of 15 PAVMs were treated with IDCs, 0.018-inch coils, or their combination, and only 2 0.035-inch coils were used in a patient combined with other coils (P o .01). VSE used twice as
Table 3 . Comparison of Clinical Features of Pulmonary Arteriovenous Malformations between Feeding Artery Embolization and Venous Sac Embolization Groups Group Clinical Feature
FAE
VSE
P Value
No. PAVMs
22
15
1989–1996:1997–2009
15:7
0:15
o .01
NS
Location (upper/middle/lower) Diameter (mm)
22 (7/2/13)
15 (0/6/9)
o .05
Venous sac
21 ⫾ 14 (3–55)
15 ⫾ 7 (4–30)
NS
Feeding artery Draining vein
5 ⫾ 2 (2–10) 8 ⫾ 5 (2–20)
4 ⫾ 1 (3–7) 5 ⫾ 2 (2–10)
NS .045
32 ⫾ 11 (10–50)
28 ⫾ 14 (15–65)
NS
Length Feeding artery (mm) Coil type used
o .01
35 coils
70
2
18 coils IDC
13 6
48 70
35 coils alone Others
17 5
0 15
No. coils used/PAVM
4 ⫾ 4 (1–15)
8 ⫾ 4 (2–17)
Coil position (mm)
13 ⫾ 9 (þ: 22)
In venous sac (: 11)
o .01*
Reperfusion
11
0
o .01
Follow-up (mo)
58 ⫾ 54 (4–175)
42 ⫾ 42 (4–150)
o .01
No. coil combinations
.002
NS
P value between FAE and VSE; NS ¼ not significant FAE ¼ feeding artery embolization; IDC ¼ interlocking detachable coil; PAVM ¼ pulmonary arteriovenous malformation; VSE ¼ venous sac embolization. n When estimated by coil position (þ) or () from the venous sac.
Volume 23
’
Number 12
’
December
’
2012
1573
Table 4 . Comparison between Reperfusion and Nonreperfusion Groups in the Feeding Artery Embolization Group Group Clinical Feature No. patients No. PAVMs
Reperfusion
Nonreperfusion
7
7
P Value
11
11
9:2 11 (2/1/8)
6:5 11 (5/1/5)
NS NS
Venous sac Feeding artery
18 ⫾ 11 (5–20) 5 ⫾ 2 (3–8)
24 ⫾ 16 (3–55) 5 ⫾ 2 (3–10)
NS NS
Draining vein
7 ⫾ 4 (2–10)
9 ⫾ 6 (3–20)
NS
34 ⫾ 11 (18–50)
28 ⫾11 (10–46)
NS
35 coils 18 coils
39 5
31 8
IDC
3
3
No. coil combinations 35 coils alone
10
7
1989–1996:1997–2009 Location (upper/middle/lower) Diameter (mm)
Length Feeding artery (mm) No. coil types used
Others
NS 1
4
No. coils used/PAVM Coil position (mm)
4 ⫾ 4 (1–13) 13 ⫾ 9 (1–30)
4 ⫾ 4 (1–15) 14 ⫾ 10 (3–32)
NS NS
Follow-up (mo)
37 ⫾ 23 (4–76)
79 ⫾ 68 (15–175)
NS
P value between reperfusion and nonreperfusion groups; NS ¼ not significant. IDC ¼ interlocking detachable coil; PAVM ¼ pulmonary arteriovenous malformation.
many coils as FAE on average (8 ⫾ 4 vs 4 ⫾ 4; P ¼ .002). Reperfusion occurred in 11 (50%) of 22 PAVMs treated by FAE but did not occur in 15 PAVMs treated by VSE (P o .01). The follow-up period did not differ between FAE (58 months ⫾ 54 [range, 4–175 months]) and VSE (42 months ⫾ 42 [range, 4–150 months]) groups (P ¼ .32).
Results of Comparison between FAE Reperfusion and Nonreperfusion Groups Reperfusion occurred in 7 (50%) of 14 patients and 11 (50%) of 22 PAVMs in the FAE group and consisted of five recanalized arteries alone, three recanalized and bronchial arteries, one recanalized and one accessory artery, and two bronchial arteries alone. Of 11 reperfusions, 6 were noted between 4 months and 2 years, and 5 were noted between 2 years and 6.3 years. There were no significant differences between FAE reperfusion and nonreperfusion patients or PAVMs in age (39 y ⫾ 21 vs 47 y ⫾ 15), male-to-female ratio (1:6 vs 1:6), ratio of HHT to non-HHT (1:6 vs 1:6), ratio of single to multiple PAVMs (3:4 vs 4:3), and ration of simple to complex PAVMs (8:3 vs 10:1). None of the 7 patients with reperfusion showed symptoms attributable to reperfusion. There were no significant differences between the groups in the proportion of the number of PAVMs between the treated periods (4 0.1) and among upper, middle, and lower lobes; diameter of the venous sac (P ¼ .66), feeding artery (P ¼ .81), and draining vein (P ¼ .18); length of the feeding
artery (P ¼ .11); number of coils used; coil combination and coils used per PAVM (P ¼ .38); coil position (P ¼ .31); and follow-up months (P ¼ .11) (Table 4). However, reperfusion occurred more frequently before compared with during and after 1997 (60% [9 of 15] vs 29% [2 of 7]) and in the PAVMs treated with 0.035-inch coils alone compared with PAVMs treated with other coils and combinations of coils (59% [10 of 17] vs 20% [1 of 5]). Concerning management after reperfusion, of the 9 of 22 PAVMs with recanalization (41%), 7 underwent additional FAE, one was treated surgically, and one could not be treated because the patient refused additional therapy and died of pneumonia 6 months after the last follow-up examination. Two patients with reperfusion from the bronchial artery alone were followed without treatment.
Changes in CT Findings of Reperfused and Nonreperfused PAVMs after Embolization All 11 FAE reperfused PAVMs showed venous sac contrast enhancement with 7 mm ⫾ 3 (range, 3–15 mm) venous sac diameter and change in diameter of 44% ⫾ 20% (range, 10%–75%) (diameter at follow-up 100%/pretherapeutic diameter) and draining vein contrast enhancement with 4 mm ⫾ 2 (range, 2–8 mm) diameter and change in diameter of 58% ⫾ 29% (range, 22%–100%) (Table 5; Fig 2c). All 11 FAE nonreperfused PAVMs showed no contrast enhancement in the venous sacs and draining veins (Fig 4a–d). The venous sac changed to complete
1574
’
Venous Sac Embolization of PAVMs
Hayashi et al
’
JVIR
Table 5 . Changes in CT Findings of Reperfused and Nonreperfused Pulmonary Arteriovenous Malformations after Embolization Feeding Artery Embolization Venous Sac Embolization Finding
Reperfusion (n ¼ 11)
Venous sac diameter
7 mm ⫾ 3 (3–15 mm)
% change
44% ⫾ 20% (10%–75%)
Enhancement Draining vein
Nonreperfusion (n ¼ 11) CR (n ¼ 11)
Nonreperfusion (n ¼ 15) NA NA
Yes
No
Diameter
4 mm ⫾ 2 (2–8 mm)
CR (n ¼ 10); 3 mm (n ¼ 1)
2 mm (n ¼ 2)
% change Enhancement
58% ⫾ 29% (22%–100%) Yes
33% (n ¼ 1)
20% (n ¼ 1); 33% (n ¼ 1) No
No
% change ¼ pretherapeutic diameter 100/diameter at follow-up; Enhancement ¼ contrast enhancement. CR ¼ complete resolution, disappearance or scarlike structure formation of sac and draining vein on CT; NA ¼ not available owing to metallic artifacts.
Figure 4. Patient 6. A 48-year-old woman with simple-type PAVM in left segment 5 treated by FAE. (a, b) Pretreatment contrast-enhanced mediastinal (a) and lung (b) window CT images reveal a large PAVM in left segment 5 (arrows). (c) Follow-up mediastinal window CT image reveals disappearance of the PAVM. (d) Follow-up lung window CT image reveals scarlike structures of the PAVM (arrow).
Figure 5. Patient 12. A 50-year-old woman with simple-type PAVM in right segment 4 treated by VSE. (a) Pretreatment contrastenhanced CT image reveals the PAVM and its draining vein (arrow). (b, c) Follow-up contrast-enhanced mediastinal (b) and lung (c) window CT images reveal a tiny scarlike structure of the draining vein without contrast enhancement (arrows).
resolution in 11 FAE nonreperfused PAVMs, whereas the draining vein changed to complete resolution in 10 FAE nonreperfused PAVMs and to 3 mm with the change in diameter of 33% in 1 FAE nonreperfused PAVM. All 15 VSE nonreperfused PAVMs showed no draining vein contrast enhancement (Fig 5a–c). The draining vein changed to complete resolution in 13 VSE nonreperfused
PAVMs and to 2 mm each with the change in diameter of 20% and 33% in 2 VSE nonreperfused PAVMs.
DISCUSSION The present study showed that reperfusion occurred solely in the FAE group, in which most patients were treated before
Volume 23
’
Number 12
’
December
’
2012
1996, at a frequency of 50% (11 of 22). No reperfusion was noted in the 15 PAVMs of the more recent VSE group. The reperfusion or recanalization rate of PAVMs with FAE using coils or balloons ranged from 0%–57% after long-term follow-up, and the wide range in reperfusion rate may be mainly related to the follow-up methods and definition of reperfusion or recanalization (7–14,28). Haitjema et al (7) used chest radiography and arterial blood gas measurement and found recanalization in 2 (2%) of 92 PAVMs. Lee et al (8) used chest radiography or non– contrast-enhanced CT and arterial blood gas measurement and found reperfusion in 8 (15%) of 53 PAVMs. Prasad et al (9) used principally non–contrast-enhanced CT and found persistence of PAVM in 18 (7%) of 267 PAVMs treated with stainless steel coils and 4 (10%) of 39 PAVMs treated with platinum coils. Mager et al (10) used chest radiography and blood gas measurement and found recanalization in 25 (8%) of 296 PAVMs. Pollak et al (11) used chest radiography and non–contrast-enhanced CT and found reperfusion in 11 (3%) of 393 PAVMs. Remy-Jardin et al (12) used non–contrast-enhanced and contrast-enhanced CT and found recanalization requiring embolization in 12 (19%) of 64 PAVMs. However, they classified treatment outcomes for four categories: 30 successful (47%), 18 partially successful (28%), 2 partially failed (3%), and 14 failed (22%) PAVMs. Anderson and Kjeldsen (13,14) used chest radiography, contrast echocardiography, and measurement of arterial blood gases and found recanalization in 8 (8%) of 106 PAVMs treated with coils and balloons (13) and in none of 13 PAVMs treated with balloons (14). Sagara et al (28) used contrastenhanced CT and found reperfusion in 8 (57%) of 14 PAVMs. However, they detected reperfusion in 1 (7%) of 14 PAVMs when they used chest radiography alone and in 2 (14%) of them when non–contrast-enhanced CT alone was used. We believe that the most reliable method to detect reperfusion is to confirm the blood flow of the venous sac and draining vein, which can be achieved by contrastenhanced CT followed by pulmonary angiography. In the present study, reperfusion was represented as measurable venous sacs 3–15 mm in diameter and draining veins 2–8 mm, all of which showed contrast enhancement on CT examinations, and a mean reduction in diameter (posttherapy/pretherapy 100%) was 44% ⫾ 20% (range, 10%– 75%) and 58% ⫾ 29% (range, 22%–100%). Mechanisms underlying reperfusion of PAVMs after embolization were summarized by Pollak et al (11) as follows: (i) recanalization of the vessel after embolization, (ii) growth of a missed or previously small accessory artery, (iii) bronchial and other systemic artery collateral flow into the pulmonary artery beyond the level of the embolization (creating a left-to-left shunt), and (iv) pulmonary artery-to-pulmonary artery collateral flow around the occlusion. In the present study, recanalization of the artery after embolization occurred in 41% (9 of 22) in the FAE group, in 23% (5 of 22) with recanalization of the
1575
artery after embolization alone, in 14% (3 of 22) also accompanied by reperfusion from the bronchial artery, and in 4.5% (1 of 22) accompanied by an accessory artery. Reperfusion from the bronchial artery alone occurred in 9% (2 of 22). Reperfusion occurred secondary to mechanisms (i)–(iii), with recanalization as the mechanism of reperfusion in 82% (9 of 11), similar to the 88% incidence reported by Milic et al (29). The last mechanism, (iv), has been observed only in young children and presumably results from the ability for continued lung growth (11). Besides reperfusion, significant differences between the FAE and VSE groups were noted in other clinical features. PAVMs significantly differed in location between FAE and VSE groups owing to distribution of 7 FAE PAVMs and no VSE PAVMs in upper lobes. However, no difference was observed in reperfusion rate between 2 (29%) of 7 FAE upper lobe PAVMs and 9 (60%) of 15 FAE middle or lower lobe PAVMs (P 4 .1). The draining vein was significantly larger in diameter in the FAE group (8 mm ⫾ 5) than the VSE group (5 mm ⫾ 2; P ¼ .045). This difference might be related to the fact that PAVMs were larger in the sac size in the FAE group (21 mm ⫾ 14) than the VSE group (15 mm ⫾ 7), although not statistically significant. Before 1996, operators performed FAE in all 15 PAVMs, and reperfusion occurred in 9 (60%) of them. Operators thereafter performed FAE in 7 (32%) and VSE in 15 (68%) of 22 PAVMs, and reperfusion occurred in 2 (9%) of them. This difference (P o .01) coincided with differences in the types and number of coils used and their combination (ie, the 0.035-inch and 0.018-inch coils alone were available before 1996, and IDCs were available after 1996). The types and number of coils used differed significantly between both groups; 0.035-inch coils were used predominantly in the FAE group (79% [70 of 89]). The IDCs were used most frequently (58% [70 of 120]) followed by 0.018inch coils (40% [48 of 120]), and only two 0.035-inch coils were used in the VSE group. In the FAE group, 0.035-inch coils alone were used in 17 of 22 PAVMs, but 0.035-inch coils alone were not used in the VSE group. Reperfusion occurred significantly more frequently in the group that used 0.035-inch coils alone (56% [10 of 18]) than in the other group (5% [1 of 19]; P o .01). Comparison of reperfused and nonreperfused PAVM groups in FAE showed no significant differences in all characteristics of PAVMs, including the diameter of the venous sac, feeding artery, and draining vein and length of the feeding artery. Characteristics of PAVMs in the present study might not be responsible for reperfusion. Although other parameters, including treated periods, the number of coils used, and coil combination, were also not significantly different between reperfused and nonreperfused groups, reperfused PAVMs occurred predominantly from 1989–1996 (82% [9 of 11]) before 0.018-inch IDCs became available and were more frequent in the group that used 0.035-inch coils alone (59% [10 of 17]) than in the other coil combinations (20% [1 of 5]). Causes of
1576
’
Venous Sac Embolization of PAVMs
reperfusion might be related to insufficient distal embolization because of the use of larger 0.035-inch coils, including inappropriate size selection and deployment position of the coils through a 5-F JB1 catheter. No reperfused PAVMs were encountered in the VSE group. No significant differences in characteristics of PAVMs were evident between the FAE and VSE groups except for a marginally significant larger diameter of the draining veins in the FAE group as mentioned previously. These results suggest that differences in reperfusion rates between FAE and VSE groups might be due to embolization methods using the different types of coils and number of coils used. FAE was performed using larger coils through a larger catheter. It might be more difficult to select appropriate coils in size and number and appropriate position for deployment than with VSE. VSE was performed through a microcatheter to achieve successful deployment of smaller coils in the sac using more appropriate coils in size and number. Coley and Jackson (21) argued theoretically against the routine placement of coils in the venous sac of PAVMs. First, rupture might occur during coil packing. However, rupture did not occur in our series or in other cases in which VSE was performed (20–23). VSE avoids risky tension that may induce rupture of the venous sac, using IDCs with soft, compliant characteristics that enhance fit. Second, thrombus forms on the coils during embolization, and dislodgment during further packing of the sac is perhaps more likely to occur than when the arterial neck is being occluded. This procedure needs many more coils for VSE compared with FAE, prolonging procedure time. However, we did not encounter thrombus migration into the draining vein during VSE. To avoid paradoxical embolization of the thrombus, we used heparin before and during the procedure in all cases. If thrombus dislodgment is anticipated, we recommend flow control with balloon occlusion of the feeding artery in large sac cases (30). Third, coils protruding from the venous sac into normal pulmonary veins might impair pulmonary venous drainage from the normal lung. Accurate deployment of IDCs with soft and compliant nature by the detachment system can prevent coils protruding into the normal pulmonary vein. Fourth, if platinum microcoils are used, the reduced radial force compared with conventional steel coils might increase the risk of migration. However, to avoid this risk, we used IDCs of a suitable size to fit the venous sac and deployed these carefully in the venous sac. Even if conventional steel coils were used, coils of an unsuitable size would result in systemic migration. We encountered one case (patient 16 in Table 2) in which a 0.035-inch pushable coil was dislocated from the feeding artery into the venous sac. This coil did not migrate into the systemic circulation and remained within the venous sac. VSE was performed in this case without complications. Selection of coils larger than the drainage vein can reduce the risk of coil migration in VSE, despite the slight radial force of the IDC. However, pushable steel coils may be
Hayashi et al
’
JVIR
associated with a risk of injury to the sac walls and cause rupture of the venous sac. Fifth, the venous sac cannot decompress after this form of treatment. However, after successful conventional embolization of the neck of a PAVM, the sac can decompress. Sac decompression in FAE is useful to evaluate treatment success. However, in VSE, success can be evaluated from complete resolution and no enhancement of the draining vein using CT. VSE may not be always indicated for PAVMs. FAE may be solely indicated for the following PAVMs: (i) venous sac filled with clot, in which FAE is recommended to prevent paradoxical embolization of clot in the venous sac (31); (ii) diffuse or complex type, in which VSE is difficult to perform when such PAVMs have no venous sacs and are composed of multiple or small vessel shunts (24,32) similar to the diffuse type (patient 9) in the present study; (iii) giant venous sac 4 3 cm in diameter, in which VSE requires much more coils, and packing the venous sac is more expensive than FAE; and (iv) PAVM with venous sac in which the microcatheter cannot be put stably. As shown in Table 2, the type and number of coils used were not parallel with the venous sac size in VSE; this suggests that the sac size alone was not the determinant of the type and number of coils used for VSE. For example, the venous sac was 15 mm in the PAVMs in segment 10 (Fig 1a–c) and segment 4 (Fig 2a–c) in Table 2; for the former PAVM, 17 IDCs were used for embolization, and for the latter, 4 IDCs and 2 0.018-inch coils were used. Embolization was successful in both cases. A key issue may be how to perform embolization of the proximal portion of the large venous sac completely without coil migration to reduce the number of appropriate coils to as few as possible to achieve successful embolization. For FAE, the AMPLATZER Vascular Plug (St. Jude Medical, St. Paul, Minnesota) (15–18,33) and 0.035-inch Interlock fibered IDC (Boston Scientific) are available in the United States and Europe; these may offer sufficient radial force for FAE and allow selection of devices with an appropriate size and deployment in an appropriate distal position of the feeding artery. Vascular plugs with or without coils shorten the procedure time (15) and decrease reperfusion rates, which were reported from 0–7% (16– 18,33). More recently, another promising result was reported for embolization of complex PAVMs with large outflow vessels and short feeding arteries using VSE with detachable coils combined with FAE with AMPLATZER Vascular Plugs (34). Although we have not yet performed embolization for PAVMs with feeding arteries o 3 mm, the need to treat such small PAVMs has been recognized when they are accessible (19,35). VSE would be one of the therapy options for such small PAVMs, if IDCs with soft and compliant characteristics can be safely put in the venous sac through a microcatheter via the small feeding artery. The following issues must be considered when interpreting the results of the present study. First, the small sample size and retrospective design are key limitations,
Volume 23
’
Number 12
’
December
’
2012
and the procedure time was not recorded, although about 2 hours was needed to complete one therapeutic session. Second, the high reperfusion rate of FAE was observed in the older period when 0.035-inch coils were used and small IDCs were unavailable. Third, although no significant difference was evident, follow-up periods tended to be shorter with VSE than with FAE. Further follow-up is necessary for VSE, given later (beyond 2 years to 76 months) occurrence of about half of reperfusions in FAE in the present study. Fourth, we measured the feeding artery, venous sac, and draining vein in diameter. Volumetric measurement of these structures using modern multidetector IVR-CT/angiosystem can serve our selection of more appropriate type and number of coils for embolization. In conclusion, although more coils are needed, VSE can be used to treat PAVMs with a venous sac safely and achieve long-term efficacy.
REFERENCES 1. Bosher LH Jr, Blake DA, Byrd BR. An analysis of the pathologic anatomy of pulmonary arteriovenous aneurysms with particular reference to the applicability of local excision. Surgery 1959; 45:91–104. 2. White RI Jr, Pollak JS, Wirth JA. Pulmonary arteriovenous malformations: diagnosis and transcatheter embolotherapy. J Vasc Interv Radiol 1996; 7:787–804. 3. Khurshid I, Downie GH. Pulmonary arteriovenous malformation. Postgrad Med J 2002; 78:191–197. 4. Govani FS, Shovlin CL. Hereditary haemorrhagic telangiectasia: a clinical and scientific review. Eur J Hum Genet 2009; 17:860–871. 5. Porstmann W. Therapeutic embolization of arteriovenous fistula by catheter technique. In: Kelop O, editor. Current Concepts in Pediatric Radiology. Berlin, Germany: Springer; 1977. p. 23–31. 6. Dutton JA, Jackson JE, Hughes JM, et al. Pulmonary arteriovenous malformations: results of treatment with coil embolization in 53 patients. AJR Am J Roentgenol 1995; 165:1119–1125. 7. Haitjema TJ, Overtoom TT, Westermann CJ, et al. Embolisation of pulmonary arteriovenous malformations: results and follow up in 32 patients. Thorax 1995; 50:719–723. 8. Lee DW, White RI Jr, Egglin TK, et al. Embolotherapy of large pulmonary arteriovenous malformations: long-term results. Ann Thorac Surg 1997; 64:930–940. 9. Prasad V, Chan RP, Faughnan ME. Embolotherapy of pulmonary arteriovenous malformations: efficacy of platinum versus stainless steel coils. J Vasc Interv Radiol 2004; 15:153–160. 10. Mager JJ, Overtoom TT, Blauw H, et al. Embolotherapy of pulmonary arteriovenous malformations: long-term results in 112 patients. J Vasc Interv Radiol 2004; 15:451–456. 11. Pollak JS, Saluja S, Thabet A, et al. Clinical and anatomic outcomes after embolotherapy of pulmonary arteriovenous malformations. J Vasc Interv Radiol 2006; 17:35–44. 12. Remy-Jardin M, Dumont P, Brillet PY, et al. Pulmonary arteriovenous malformations treated with embolotherapy: helical CT evaluation of longterm effectiveness after 2-21-yearfollow-up. Radiology 2006; 239: 576–585. 13. Andersen PE, Kjeldsen AD. Clinical and radiological long-term follow-up after embolization of pulmonary arteriovenous malformations. Cardiovasc Intervent Radiol 2006; 29:70–74. 14. Andersen PE, Kjeldsen AD. Long-term follow-up after embolization of pulmonary arteriovenous malformations with detachable silicone balloons. Cardiovasc Intervent Radiol 2008; 31:569–574.
1577
15. Abdel Aal AK, Hamed MF, Biosca RF, et al. Occlusion time for Amplatzer vascular plug in the management of pulmonary arteriovenous malformations. AJR Am J Roentgenol 2009; 192:793–799. 16. Hart JL, Aldin Z, Braude P, et al. Embolization of pulmonary arteriovenous malformations using the Amplatzer vascular plug: successful treatment of 69 consecutive patients. Eur Radiol 2010; 20:2663–2670. 17. Letourneau-Guillon L, Faughnan ME, Soulez G, et al. Embolization of pulmonary arteriovenous malformations with Amplatzer vascular plugs: safety and midterm effectiveness. J Vasc Interv Radiol 2010; 21: 649–656. 18. Tapping CR, Ettles DF, Robinson GJ. Long-term follow-up of treatment of pulmonary arteriovenous malformations with AMPLATZER Vascular Plug and AMPLATZER Vascular Plug II devices. J Vasc Interv Radiol 2011; 22:1740–1746. 19. Trerotola SO, Pyeritz RE, Bernhardt BA. Outpatient single-session pulmonary arteriovenous malformation embolization. J Vasc Interv Radiol 2009; 20:1287–1291. 20. White RI Jr. Pulmonary arteriovenous malformations: how do I embolize? Tech Vasc Interv Radiol 2007; 10:283–290 21. Coley SC, Jackson JE. Venous sac embolization of pulmonary arteriovenous malformations in two patients. AJR Am J Roentgenol 1996; 167:452–454. 22. Takahashi K, Tanimura K, Honda M, et al. Venous sac embolization of pulmonary arteriovenous malformation: preliminary experience using interlocking detachable coils. Cardiovasc Intervent Radiol 1999; 22: 210–213. 23. Dinkel HP, Triller J. Pulmonary arteriovenous malformations: embolotherapy with superselective coaxial catheter placement and filling of venous sac with Guglielmi detachable coils. Radiology 2002; 223: 709–714. 24. White RI Jr, Mitchell SE, Barth KH, et al. Angioarchitecture of pulmonary arteriovenous malformations: an important consideration before embolotherapy. AJR Am J Roentgenol 1983; 140:681–686. 25. Shovlin CL, Guttmacher AE, Buscarini E, et al. Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber syndrome). Am J Med Genet 2000; 91:66–67. 26. Remy J, Remy-Jardin M, Wattinne L, et al. Pulmonary arteriovenous malformations: evaluation with CT of the chest before and after treatment. Radiology 1992; 182:809–816. 27. Sacks D, McClenny TE, Cardella JF, et al. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14(9 Pt 2): S199–S202. 28. Sagara K, Miyazono N, Inoue H, et al. Recanalization after coil embolotherapy of pulmonary arteriovenous malformations: study of long-term outcome and mechanism for recanalization. AJR Am J Roentgenol 1998; 170:727–730. 29. Milic A, Chan RP, Cohen JH, et al. Reperfusion of pulmonary arteriovenous malformations after embolotherapy. J Vasc Interv Radiol 2005; 16:1675–1683. 30. Mori K, Shiigai M, Saida T, et al. A modified metallic coil embolization technique for pulmonary arteriovenous malformations using coil anchors and occlusion balloon catheters. Cardiovasc Intervent Radiol 2008; 31: 638–642. 31. Graves AD, Gregorius JC, Smith DC. Management of a patient with a clot-filled pulmonary arteriovenous malformation. J Vasc Interv Radiol 2009; 20:652–655. 32. Wei CW, Faughnan ME, Menard A, et al. Lobar embolization for treatment of diffuse pulmonary arteriovenous malformations in hereditary hemorrhagic telangiectasia: a case report. J Vasc Interv Radiol 2010; 21:1105–1108. 33. Trerotola SO, Pyeritz RE. Does use of coils in addition to Amplatzer vascular plugs prevent recanalization? AJR Am J Roentgenol 2010; 195: 766–771 34. Hundt W, Kalinowski M, Kiessling A, et al. Novel approach to complex pulmonary arteriovenous malformation embolization using detachable coils and Amplatzer vascular plugs. Eur J Radiol 2012; 81:e732–e738. 35. Trerotola SO, Pyeritz RE. PAVM embolization: an update. AJR Am J Roentgenol 2010; 195:837–845.