| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Thoracic Imaging |
1 From the Department of Thoracic Imaging, Hospital Calmette, University Center of Lille, Boulevard Jules Leclerc, 59037 Lille CEDEX, France (M.R., P. Dumont, P.Y.B., P. Dupuis, J.R.); and Department of Medical Statistics, University of Lille, Lille, France (A.D.). Received March 2, 2005; revision requested April 27; revision received June 7; final version accepted June 21. Address correspondence to M.R. (e-mail: mremy-jardin{at}chru-lille.fr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Materials and Methods: Neither institutional review board approval nor patient consent was required for this retrospective study. Thirty-eight patients underwent follow-up helical chest CT 2–21 years after successful embolotherapy of 64 PAVMs. Four outcome categories were analyzed on the basis of the PAVM morphologic changes and perfusion findings seen on CT angiograms: successful treatment (marked reduction or disappearance of the aneurysmal sac), partially successful treatment (reduced size of the aneurysmal sac and pulmonary vessels, with feeding artery[ies] less than 3 mm in diameter, deemed too small to be occluded), partially failed treatment (reduced size of the aneurysmal sac and pulmonary vessels, with feeding artery[ies] larger than 3 mm and additional embolotherapy required), and failed treatment (similar size of or interim growth in the aneurysmal sac, with unchanged or enlarged pulmonary vessels). 
2 or Fisher exact tests were used to analyze categorical variables; Mann-Whitney rank tests were used to analyze continuous variables. P < .05 was considered to indicate statistical significance.
Results: Long-term follow-up of the 64 occluded PAVMs revealed successful treatment of 30 (47%), partially successful treatment of 18 (28%), partially failed treatment of two (3%), and failed treatment of 14 (22%) PAVMs. The overall treatment success rate was 75% (47% plus 28%). Delayed recanalization requiring repeat embolotherapy occurred in 12 (19%) cases. No relationship between failed treatment and number of coils deposited in the feeding arteries was found. The frequency of gastrointestinal tract and/or hepatic arteriovenous fistulas at initial diagnosis (P = .01) and/or the interim development of pulmonary hypertension with or without heart failure (P = .01) was significantly higher in patients with at least one PAVM for which embolotherapy failed (n = 9) than in patients who underwent successful or partially successful embolotherapy of all PAVMs (n = 29).
Conclusion: Long-term CT follow-up of initially successfully treated PAVMs revealed successful embolotherapy of 75% and partially or completely failed embolotherapy of 25% of PAVMs.
© RSNA, 2006
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Before the advent of computed tomography (CT), postprocedural outcome was mainly based on sequential analysis of chest radiographs, serial blood gas measurements, and right-to-left shunt calculations performed by using the 100% oxygen method or technetium 99m–labeled microspheres (2,6), with the recent introduction of contrast material–enhanced echocardiography in follow-up after transcatheter embolotherapy (7). Helical CT with three-dimensional reconstruction has become a reliable tool for detecting PAVMs and for preembolization mapping of the afferent vessels (8). The purpose of our investigation was to retrospectively evaluate the long-term results of transcatheter embolotherapy of PAVMs by using helical CT.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Our final study group therefore included 38 patients (21 female, 17 male) with a mean age of 34.35 years ± 13.66 (standard deviation) (age range, 11–65 years). The early follow-up findings for 13 of these 38 patients have been previously reported; their follow-up periods ranged from 3 months to 1 year (4). At initial presentation, 24 (63%) patients—16 referred for chest radiograph abnormalities and eight referred as part of a family screening program for hereditary hemorrhagic telangiectasia—were asymptomatic. Fourteen (37%) patients had cerebral (n = 10), pulmonary (n = 3), or abdominal (n = 1) symptoms that prompted them to seek medical attention. In 32 (84%) patients, the diagnosis of hereditary hemorrhagic telangiectasia was based on the presence of three of four features key to this diagnosis, as previously described (9). In our study group, 24 patients had a single PAVM and 14 had two or more PAVMs (range, 2–7).
The 38 patients underwent a total of 88 delayed follow-up CT examinations, with a mean of 2.31 examinations performed per patient (range, 1–6). Before the introduction of fast helical CT technology, all patients who were referred for follow-up of treated PAVMs at our institution underwent a nonenhanced CT follow-up examination, which was indicated every 2–5 years, as recommended in the literature (4,10). When it became possible to scan the entire thorax with thin collimation during a sufficient breath hold, nonenhanced CT scanning was replaced by a contrast-enhanced helical CT examination to obtain additional information regarding the outcome of occluded PAVMs. The inclusion of a CT angiographic examination was approved by the institutional review board and ethics committee of Hospital Calmette in the clinical context of treated PAVMs. For our retrospective study, the institutional review board required neither its approval nor the patients' informed consent.
Care was taken to avoid administering contrast material where there were contraindications, such as renal insufficiency (ie, creatinine levels > 150 µmol/L), iodine intolerance, and use of metformin or biguanides for treatment of diabetes mellitus. Owing to the presence of right-to-left shunts through the PAVMs, special care was taken to exclude residual air bubbles in the syringe of the power injector and in its connecting tube before the administration of contrast material. For 32 of the 38 patients included in the present study, the frequency of a systemic collateral supply to the PAVMs during the long-term follow-up of treated PAVMs has been previously reported (11).
Embolotherapy
The 38 patients had a total of 76 PAVMs with feeding arteries that were 3 mm in diameter or larger, for which embolotherapy was indicated (12). Twelve PAVMs were not treated because of the patients' reluctance to undergo multiple embolotherapy sessions. Thus, a total of 64 PAVMs were treated with embolotherapy, which was performed by two authors (J.R. and M.R., with 25 and 15 years of experience performing embolotherapy, respectively). For all PAVMs treated with embolotherapy, the pulmonary arterial circulation was occluded by using a similar technique, which involves the use of steel coils (Cook, Bloomington, Ind) that are deposited in the feeding artery(ies) after its superselective catheterization with a 7-F radiopaque nontapered catheter with an end hole. Once flow ceased, as confirmed by performing a test injection of contrast material or obtaining a control angiogram, the procedure was stopped. For the purposes of the present investigation, the most proximal level of occlusion of the pulmonary arterial bed that enabled cessation of the aneurysmal perfusion was retrospectively assessed on the follow-up angiograms, which were systematically obtained at the end of the embolotherapy procedures.
In patients who underwent several embolizations, the time between procedures did not exceed 1 year. The mean interval between embolotherapy and follow-up helical CT was 9.74 years ± 4.55 (standard deviation) (range, 2–21 years). When multiple endovascular procedures were performed to treat a given PAVM, the date of the first procedure was considered for this calculation. The mean number of occluded PAVMs per patient was 1.68 ± 1.16 (range, 1–6), and the mean number of embolotherapy sessions performed per patient was 1.08 ± 0.72 (range, 1–3). All treated PAVMs were initially successfully occluded by means of steel coil deposition, as confirmed by the cessation of aneurysmal perfusion at the end of the embolotherapy procedure.
Pretherapeutic Evaluation
The anatomic characteristics of the 64 PAVMs treated with embolotherapy in our study group were assessed on pulmonary angiograms (38 patients) and pretherapeutic CT images (37 patients). Owing to technologic improvements over time, the pretherapeutic CT evaluations were performed by using a sequential CT unit (Elscint 2400, Elscint, Haifa, Israel or DRH, Siemens, Erlangen, Germany) in three patients and by using a helical CT unit (Somatom Plus or Volume Zoom; Siemens) in 34 patients. All of these evaluations consisted of thick-collimation (8–10-mm), contiguous, nonenhanced CT examinations.
PAVM features were analyzed by consensus between two readers trained to read the CT images obtained in patients with PAVMs—a radiologist (P.Y.B.) with 5 years experience and a pulmonologist (P. Dumont) with 2 years experience—who were also involved in the interpretations of the postembolotherapy helical CT images. In cases of discordant interpretations, a final consensus was reached with a third reader (M.R.), a faculty radiologist with 15 years of experience with CT angiography of the chest.
The following features were analyzed: (a) size of the aneurysmal sac; (b) lobar location of the aneurysm (For comparative analysis purposes, aneurysms located in the right upper, right middle, and left upper lung lobes were considered to be located in the upper lung zones, and aneurysms located in the right and left lower lobes were considered to be located in the lower lung zones.); (c) central or peripheral location of the aneurysm within the lung parenchyma (Peripheral lung region was defined as the outer 4 cm of the lung parenchyma deep to the visceral pleura; central lung region was defined as the more centrally located portion of the lung parenchyma.); (d) angioarchitecture of the aneurysm, as assessed by using the CT criteria of Remy et al (4) (In a simple PAVM, a single artery feeds an aneurysmal communication with a single draining vein; in the complex type, one or more pulmonary artery branches communicate with an aneurysm that has two or more draining veins. In cases of complex PAVMs, the number of feeding arteries was noted.); and (e) diameter of the feeding arteries (ie, <3 mm, 3–5 mm, or >5 mm).
These parameters were analyzed on the pretherapeutic thoracic CT images by using measurements made with calipers. On diagnostic angiograms, this information was obtained from the analysis of unilateral pulmonary angiograms acquired with a conventional technique in each lung, completed by the analysis of the hyperselective angiograms obtained prior to the deposition of steel coils. Care was taken to measure the diameters of the aneurysmal sac and vascular pedicles on angiograms after correction of the geometric magnification. For the 37 patients with both helical CT images and pulmonary angiograms available, analysis of both sets of images led to the final determinations regarding the parameters listed above. In the case of the one patient who did not undergo pretherapeutic CT, the readers analyzed the above mentioned parameters on the unilateral and hyperselective angiograms obtained before the embolotherapy procedure.
Follow-up Helical CT
CT protocol.—The present investigation was focused on findings of the last helical CT examination performed for the long-term follow-up of each PAVM. The CT images were obtained by using a single–detector row scanner (Somatom Plus; Siemens) for seven PAVMs, a four–detector row scanner (Sensation 4; Siemens) for four PAVMs, and a 16–detector row scanner (Sensation 16; Siemens) for 53 PAVMs. The following scanning parameters were used: for the single–detector row scanner, 2–3-mm collimation, a pitch of 2.0, and a rotation time of 0.75 second; for the four–detector row scanner, 4 x 1-mm collimation, a pitch of 1.75, and a rotation time of 0.50 second; and for the 16–detector row scanner, 16 x 0.75-mm collimation, a pitch of 1.5, and a rotation time of 0.50 second.
The kilovoltage and milliamperage settings varied between 120 and 140 kV and between 80 and 100 mAs, respectively. The region scanned extended from the cervicothoracic junction to the upper part of the abdomen. Every examination was performed during deep inspiration. A radiation dose reduction technique based on online tube current control (Care Dose System; Siemens) was systematically applied for the multisection helical CT examinations.
Eleven long-term follow-up CT examinations were performed without contrast medium, and 53 examinations were performed after contrast medium administration. CT angiograms were obtained by injecting 80–100 mL of contrast material (300 mg of iodine per milliliter) at a rate of 4 mL/sec. The automatic bolus-triggering software program (Care Bolus) on our CT unit was systematically applied, a circular region of interest was positioned at the level of the ascending aorta, and an attenuation threshold for triggering the data acquisition was preset at 100 HU.
From each data set, contiguous 3–5-mm-thick transverse CT images were systematically reconstructed and viewed at mediastinal and lung window settings. Transverse imaging was completed with the generation of three-dimensional reconstructions as follows: From the nonenhanced CT examinations, we systematically reconstructed three-dimensional shaded-surface displays with attenuation threshold values ranging from –600 to –700 HU, which enabled the depiction of nonenhanced PAVMs while excluding peripheral vascular structures. Whenever a vessel with apparent discontinuities was seen in the vicinity of the aneurysmal part on the initial three-dimensional reconstruction, an additional display was systematically obtained with a lower threshold (–850 HU) to evaluate the relationship between the vessel and the aneurysmal sac (8).
From the contrast-enhanced CT examinations, we systematically obtained oblique coronal and sagittal maximum intensity projections and volume-rendered images of the thoracic vascular structures. Three-dimensional reconstructions were generated on commercially available workstations (Virtuoso and Leonardo consoles; Siemens) by technologists with 5–10 years of experience with helical CT. The transverse CT images and three-dimensional reconstructions were printed on laser film hard copies by using standardized lung and mediastinal window settings, with a photographic format of 20 images per sheet of film.
Image interpretation.—The same readers (P.Y.B., P. Dumont) retrospectively evaluated the long-term helical CT follow-up images in consensus; in cases of discordant interpretations, a final consensus with the same third reader (M.R.) was reached. Film hard-copy CT images were reviewed, with the option of using cine-mode display on the workstation for the multisection CT images. The readers were blinded to the patients' clinical histories.
Long-term effectiveness.—The long-term effectiveness of embolotherapy for treatment of the 64 occluded PAVMs was assessed by using the findings of the last helical CT examination performed for long-term follow-up and comparing the follow-up findings with the findings of the diagnostic examination (CT for 63 PAVMs, pulmonary angiography for one PAVM). Each PAVM was numbered for further follow-up. At follow-up helical CT, the size (enlarged, decreased, or unchanged) of and perfusion (present, absent, or not evaluated) in the aneurysm were evaluated. The size of the aneurysm was considered to be substantially modified if a greater than 30% change in diameter was observed. An aneurysm was considered to be no longer detectable on CT images when it was not detected on two successive sections or detectable as a micronodule.
The long-term effectiveness of embolotherapy was classified into four categories by using two series of criteria previously described in the literature (4), which enabled a comparison of the PAVM outcomes depicted on nonenhanced and contrast-enhanced helical CT images. With use of the nonenhanced helical CT images, the four outcome categories were as follows: (a) success, as indicated by a marked (>30%) shrinkage or the disappearance of the aneurysmal sac; (b) partial success, as indicated by a less than 30% reduction in the size of the aneurysmal sac and a reduction in the size of the feeding artery(ies) beyond the level of the deposited coils to less than 3 mm in diameter, which was deemed too small to be reoccluded; (c) partial failure, as indicated by a reduction in the size of the aneurysmal sac and a reduction in the size of the feeding artery(ies) beyond the level of deposited coils to larger than 3 mm in diameter, which was an indication for repeat embolotherapy; and (d) failure, as indicated by an unchanged or enlarged aneurysmal sac with an unchanged or enlarged feeding artery(ies) beyond the level of the deposited coils or the interim development of an additional feeding artery. Any aneurysm with an enlarged part, an undetectable feeding artery(ies), and a large draining vein was suspected of systemic perfusion and thus considered a result of failed embolotherapy.
With use of the contrast-enhanced helical CT images, the four outcome categories were as follows (11): (a) success, as indicated by a marked (>30%) reduction or the disappearance of the aneurysmal sac, with no pulmonary or systemic perfusion; (b) partial success, as indicated by a less than 30% reduction in the size of the aneurysmal sac, with persistent pulmonary perfusion and a patent feeding artery beyond the level of the deposited coils smaller than 3 mm in diameter, which was deemed too small to be reoccluded; (c) partial failure, as indicated by a substantial reduction in the size of the aneurysmal sac, persistent pulmonary perfusion, and a patent feeding artery(ies) beyond the level of the deposited coils larger than 3 mm in diameter, which was an indication for repeat embolotherapy; and (d) failure, as indicated by an unchanged or enlarged aneurysmal sac with persistent pulmonary perfusion. Any case in which the interim development of an additional feeding artery or systemic perfusion was observed at the aneurysm level was considered to represent failed embolotherapy. Use of the 3-mm threshold for feeding artery diameter was based on the results of natural history studies of PAVMs, which suggest that PAVMs that have feeding arteries with diameters of at least 3 mm yield the highest risk for embolic stroke and should be occluded (3,10).
Statistical Analyses
Statistical analyses were performed by using commercially available software (SAS; SAS Institute, Cary, NC). Results are expressed as means ± standard deviations for continuous variables. Comparisons between groups were performed by using the 2 or Fisher exact test for categorical variables and the Mann-Whitney or Kruskal-Wallis rank test for continuous variables. P < .05 was considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Pretherapeutic Characteristics of the Occluded PAVMs
Forty-seven (73%) of the 64 PAVMs had a peripheral location, and 17 (27%) had a central location. Fifty-three (83%) PAVMs were located in lower lung zones, and 11 (17%) were located in upper lung zones. Forty-one PAVMs were located in the right lung, and 23 were located in the left lung. The mean size of the aneurysmal sac was 18.5 mm ± 2.45 (range, 4–45 mm). Fifty-three (83%) PAVMs had a simple angioarchitecture, and 11 (17%) had a complex angioarchitecture. The 53 simple PAVMs had feeding arteries with diameters of 3–5 mm (n = 44) or 6–10 mm (n = 9). Of the 11 complex PAVMs, 10 had two feeding arteries each and one had three feeding arteries. Of the total 23 feeding arteries, 18 had diameters of 3–5 mm and five had diameters of 6–10 mm.
Therapeutic Modalities
For 54 (84%) PAVMs, successful occlusion was achieved in one session. For the remaining 10 (16%) PAVMs—eight complex and two simple malformations with technical difficulties in catheterizing the feeding arteries—two sessions were necessary to achieve complete cessation of blood flow within the aneurysmal sac. The mean interval between the two embolotherapy sessions was 6.5 months ± 0.4 (range, 2–8 months). The most proximal level of occlusion needed to achieve cessation of blood flow within the aneurysmal sac was segmental for 25 (39%) PAVMs, subsegmental (within the 4th-order branches) for 34 (53%) PAVMs, and within the 5th-order branches for five (8%) PAVMs. The mean number of coils deposited in feeding arteries was 3.92 ± 3.39 (range, 1–20) for the simple PAVMs and 7.09 ± 2.15 (range, 2–12) for the complex PAVMs.
Long-term Effectiveness of Embolotherapy Evaluated at Follow-up Helical CT
In terms of long-term effectiveness, embolotherapy was found to be successful for 30 (47%) of the 64 PAVMs (Figs 1, 2), partially successful for 18 (28%), partially failed for two (3%), and failed for 14 (22%) (Fig 3). The overall embolotherapy success rate was 75% (ie, complete occlusion of the malformations in 47% of cases, partial success in 28% of cases). The overall embolotherapy failure rate was 25% (ie, partial failure in 3% of cases, failure in 22% of cases). The PAVM outcomes assessed with CT protocols were as follows: (a) 30 successful results assessed on single-section (n = 3) and multisection (n = 27) CT images, including 24 CT angiograms; (b) 18 partially successful results assessed on single-section (n = 2) and multisection (n = 16) CT images, including 14 CT angiograms; (c) two partially failed results, both assessed on multisection CT angiograms; and (d) 14 failed results assessed on single-section (n = 2) and multisection (n = 12) CT images, including 13 CT angiograms. No significant differences in mean duration of CT follow-up were found among the four categories of postembolotherapy outcome of the 64 PAVMs (P = .09).
|
|
|
|
|
|
|
|
|
|
Systemic perfusion was suspected at a nonenhanced helical CT examination performed 3 years after the successful occlusion of a simple PAVM in the inferior segment of the lingula and was angiographically confirmed 3 weeks after helical CT; this case has been previously reported (8). In this case, numerous arterial branches originating from the left internal mammary and left inferior phrenic arteries led to surgical treatment of the malformation and the surrounding systemic collateral vessels. Apart from the patient in this case, none of the patients in whom embolotherapy of PAVMs failed underwent complementary procedures.
Risk Factors for Failed Embolotherapy
To analyze the factors of risk for failed embolotherapy, two groups of PAVMs were defined: 48 (75%) PAVMs (group 1) in which embolotherapy was completely (n = 30) or partially (n = 18) successful and 16 (25%) PAVMs (group 2) in which embolotherapy partially (n = 2) or completely (n = 14) failed.
Anatomic characteristics of PAVMs at time of embolotherapy.—There were no significant differences in angioarchitecture or mean aneurysmal sac size between PAVM groups 1 and 2 (Table 1). Apart from a significantly higher proportion of peripheral PAVMs in group 2, no significant differences in the locations of PAVMs in the upper or lower lung zones or in the right or left lungs were found.
|
|
|
As shown in Table 4, we observed no significant difference in mean age, sex ratio, or frequency of hereditary hemorrhagic telangiectasia at the time of embolotherapy between groups 1 and 2. However, the frequency of gastrointestinal tract and/or hepatic arteriovenous malformations was significantly higher in group 2 than in group 1. After embolotherapy, all of the group 1 patients had normal levels of pulmonary artery blood pressure, whereas three group 2 patients developed pulmonary hypertension. The hypertension was transient in one patient: Pulmonary artery catheterization revealed a pressure of 30/15 mm Hg (mean, 25 mm Hg). However, the pulmonary artery pressure was chronically elevated in the remaining two patients (P = .01): In one patient, pulmonary artery catheterization revealed a pressure of 40/20 mm Hg (mean, 29 mm Hg). In the other patient, chronically enlarged central and peripheral pulmonary arteries were seen on follow-up CT images for the first time 3 years after the last embolotherapy session, during which the pulmonary artery pressure was within the normal range. No further pulmonary artery catheterizations were performed. These three patients also had hepatic and/or gastrointestinal tract arteriovenous malformations in the clinical context of hereditary hemorrhagic telangiectasia at presentation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
To our knowledge, the long-term results of embolotherapy reported in the literature are limited to those reported in only two studies, both involving the use of steel coils to embolize the malformations (3,5). In one study (5), a cohort of 32 patients were treated with embolization, and 14 of them were reported to have been followed up for more than 2 years (mean, 40 months). The authors reported that 12 (85%) of the 14 patients had successful results. In the second study, Lee et al (3) reported that of 45 patients with 52 treated PAVMs, 38 (84%) patients with 44 (85%) PAVMs, who were followed up for a mean of 4.7 years, had successful long-term resolution of the PAVMs.
The lower embolotherapy success rate reported in the present study, compared with the above results, may be explained by differences in the tools used to assess the long-term results of embolotherapy. Although we stratified our analyses of morphologic criteria according to the level of each occluded malformation, Haitjema et al (5) used serial measurements of shunt fraction, with intravenous digital subtraction angiography of the pulmonary arteries indicated in cases of elevating shunt fraction. The criterion used by Lee et al (3) to determine successful embolization was resolution of the occluded PAVM, as assessed by using angiography, arterial blood gas partial pressure of oxygen levels, or a combination of arterial oxygen tension values and either chest radiography or CT. The longer period of follow-up after embolotherapy in our study may be another explanation for the discrepancy between our overall treatment success rate and those previously reported (3,5).
Delayed recanalization that required additional treatment was observed in a total of 12 (19%) PAVMs, including two PAVMs judged to be partially failed results of embolotherapy and 10 PAVMs judged to be completely failed results. Recanalization rates of 10%–14% at long-term follow-up have been reported (3,5). Sagara et al (14) investigated the outcome of 14 PAVMs in a smaller cohort of seven patients during follow-up periods that ranged from 4 to 66 months and documented recanalization of the feeding artery of six (43%) PAVMs that had been embolized with steel coils; two of these malformations were fed by pulmonary and bronchial arteries.
In the current study, additional feeding arteries that were not recognized at the time of initial management explained the failed treatment for three (5%) complex PAVMs. Similar findings have been reported in the literature (3,5,8). Three-dimensional CT reconstructions in a patient imaged 1 year after successful embolotherapy of a simple PAVM showed development of a previously undetectable pulmonary arterial branch connected to the aneurysmal sac (8). Three (6%) of the 52 occluded PAVMs reported by Lee et al (3) also exhibited interval growth of an accessory vessel, whereas Haitjema et al (5) reported the concurrent findings of recanalization of embolized vessels and new feeding vessels in one (7%) of 14 patients. As previously reported, recanalized PAVMs can be treated by means of secondary occlusion with successful and permanent results (3,5).
In one case, we observed the development of systemic perfusion of the aneurysmal sac 3 years after embolotherapy of a simple PAVM in the inferior segment of the lingula. A systemic blood supply from bronchial or nonbronchial systemic arteries also has been blamed for the persistent perfusion of treated PAVMs (14,15). However, this condition does not have the same clinical importance as persistent pulmonary perfusion because it represents a left-to-left shunt, with no chance for paradoxical embolization. Furthermore, the occurrence of systemic perfusion that leads to pulmonary hemorrhage is rare; to our knowledge, the only reported case of this condition occurred in a pregnant woman (15). Because of the anatomic complexity of the vascular lesion observed in the patient in our study, we performed surgery instead of repeat embolotherapy to ensure the definitive treatment of the malformation.
Because failed embolotherapy was observed in 25% of the occluded PAVMs, we attempted to identify the risk factors for treatment failure. No such factors were found among either the anatomic characteristics of the PAVMs at the time of embolotherapy or the technical characteristics of the embolization procedures performed at the time of initial presentation. However, our data suggest that failed embolotherapy could be linked to patient-related parameters—in particular, pulmonary hypertension and gastrointestinal and/or hepatic shunts. The deleterious effect of elevated cardiac output in patients with hepatic involvement should lead to the treatment of hepatic shunts before embolotherapy of PAVMs is performed. This sequence of treatment was followed for one of our patients; however, the extensive hepatic vascular malformations could not be cured despite successive embolization procedures. The rare association between pulmonary hypertension and PAVMs should also be taken into consideration when planning endovascular treatment.
Several patients have reportedly developed (16) or experienced increased (17) pulmonary hypertension after embolization—presumably secondary to a reduction in low resistance vascular circuits. This has led some authors to stress the importance of the pretherapeutic detection of pulmonary hypertension (18,19). Moreover, several investigators have recommended estimating the posttreatment pulmonary hemodynamics, which can be accomplished with transient occlusion of the blood supply to the PAVM with a balloon-tipped catheter (17,19). In addition, pregnancy has been described in the literature as a factor that triggers the development of PAVMs, particularly among patients with hereditary hemorrhagic telangiectasia (20–23). The factors linked to pregnancy include increased blood volume and estrogen-progestogen imbalance. The one patient in our study population who became pregnant during the time of this investigation did not develop such conditions.
Our study had limitations. First, 64 occluded PAVMs could be followed up for more than 2 years; thus, our results require further confirmation in a larger series of occluded malformations. Second, the CT follow-up did not systematically include the administration of contrast medium and thus did not systematically include PAVM perfusion analysis results in the outcomes of occluded PAVMs. However, this factor did not affect our recognition of successful results, which was based more on morphologic criteria than CT attenuation features. Third, we should mention the potential selection bias that was created owing to the patients who were lost to follow-up, the embolotherapy outcomes of whom are impossible to state. It should be emphasized too that none of the treated patients who did not undergo CT after 2 years received a diagnosis of failed results within the first 2 years after embolotherapy.
The statistical test results reported are based on the assumption that the result for any one PAVM was independent of the results for other PAVMs in the same patient. Because results such as the occlusion and location of PAVMs in the same patient were correlated, the reported P values may be inexact. In conclusion, our long-term helical CT follow-up of PAVMs that were initially successfully treated revealed successful results for 75% of PAVMs.
|
ADVANCES IN KNOWLEDGE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
|
FOOTNOTES |
|---|
Abbreviations: PAVM = pulmonary arteriovenous malformation
Author contributions: Guarantors of integrity of entire study, M.R., J.R.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, M.R., P. Dumont, P.Y.B., J.R.; clinical studies, M.R., P. Dumont, P.Y.B., P. Dupuis, J.R.; statistical analysis, A.D.; and manuscript editing, M.R., J.R.
Authors stated no financial relationship to disclose.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Schneider, M. Uder, M. Koehler, M. A. Kirchin, A. Massmann, A. Buecker, and U. Geisthoff MR Angiography for Detection of Pulmonary Arteriovenous Malformations in Patients with Hereditary Hemorrhagic Telangiectasia Am. J. Roentgenol., April 1, 2008; 190(4): 892 - 901. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. P. Chan, M. Faughnan, and R. White Pulmonary Arteriovenous Malformations Treated with Embolotherapy Radiology, September 1, 2007; 244(3): 932 - 932. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME |
= aneurysmal sac. (b) Findings on follow-up angiogram (left posterior oblique view) obtained after deposition of several steel coils (arrow) confirm cessation of blood flow within the malformation. (c) Multi–detector row helical CT angiogram (thin-slab maximum intensity projection; left posterior oblique view) obtained 11 years later shows lack of opacification beyond level of deposited coils and very thin diameter of occluded feeding artery (arrow) above the coils. These findings confirm successful embolotherapy.
