| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pediatric Imaging |
1 From the Department of Medical Imaging/Neuroradiology, Hospital for Sick Children, 555 University Ave, Toronto, Ontario, Canada M5G 1X8. From the 2000 RSNA scientific assembly. Received June 4, 2001; revision requested January 17; final revision received October 1; accepted October 16. Address correspondence to D.A. (e-mail: derek.armstrong@sickkids.on.ca).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: MR findings and clinical results of percutaneous alcohol sclerotherapy in 59 pediatric patients with VVMs were retrospectively reviewed. Before treatment, lesions were graded with MR imaging on the basis of margins and size, respectively: grade 1, well defined, less than or equal to 5 cm; grade 2A, well defined, greater than 5 cm; grade 2B, ill defined, less than or equal to 5 cm; and grade 3, ill defined, greater than 5 cm. Regression models were used to test trends in therapy across the MR classification grades, including the repeat sclerotherapies, volumes of ethanol and metrizamide administered for each lesion, and number of access sites. Clinical response to sclerotherapy, which was evaluated with consensus by a multidisciplinary team, was graded as poor, good, or excellent. Association between MR imaging grade and clinical assessment was tested with the Fisher exact test.
RESULTS: There were 14 grade 1 lesions, nine grade 2A, 15 grade 2B, and 21 grade 3. Twenty-four patients had a poor response to sclerotherapy; 19, good; and 16, excellent. Ten of 14 (71%) grade 1 lesions had an excellent response; none, a poor response. Twelve of 21 (57%) grade 3 lesions had a poor response; none, an excellent result. Grade 2 lesions were relatively equally distributed among the three categories, with the exception of nine of 15 (60%) grade 2B lesions that had a poor response (P < .001). There was a trend with increasing lesion grade for increasing numbers of sclerotherapy sessions, volumes of ethanol and metrizamide for each lesion, and numbers of access sites.
CONCLUSION: There is a strong association between this proposed MR imaging classification and the results of percutaneous sclerotherapy.
© RSNA, 2002
Index terms: Alcohol, 9*.12642, 9*.759 • Arteriovenous malformations, 9*.1264, 9*.759 • Interventional procedures, in infants and children, 9*.1264, 9*.759 • Veins, malformations, 9*.1264, 9*.759 • Veins, therapeutic embolization, 9*.1264, 9*.759
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
According to the classification system of Mulliken and Glowacki (1), vascular malformations are structural abnormalities that result from abnormal blood or lymphatic vessel morphogenesis and display normal endothelial cell mitotic activity. Vascular malformations may possess any combination of capillary, venous, arterial, and lymphatic components. Venous and lymphovenous malformations are congenital, but they may not become clinically apparent until late infancy or early childhood. They are rare and occur with equal frequency in male and female patients. VVMs do not regress or involute but rather grow commensurately with the child (1). Rapid enlargement may occur during puberty or pregnancy or with trauma (3). Potential complications, some of which have the potential to be life threatening, include pain, compression or invasion of adjacent structures, decreased range of motion, bleeding, consumptive coagulopathy, and cosmetic deformity (4–6). Vascular malformations are distinct from hemangiomas, which are the most common tumors in infancy. Hemangiomas may or may not be present at birth, and they occur predominantly in female patients. They demonstrate a proliferative phase of increased endothelial cell mitotic activity and an involuting phase, which is usually complete by age 5–7 years (1).
In the classification system of Kaban and Mulliken (2), which is based on vascular flow dynamics, arteriovenous malformations are categorized as either low flow (eg, capillary, venous, and lymphatic malformations) or high flow. Management of low-flow VVMs is difficult. Surgical resection may be hazardous owing to major blood loss and incomplete resection. In addition, recurrence and poor cosmetic results are common (7). Medical therapies, including steroid administration, laser photocoagulation, and transcatheter arterial embolization, have only limited success (8–12). Percutaneous sclerotherapy, most commonly with ethanol, with or without surgical excision, is now advocated as the treatment of choice (4–6,13). In contrast, arteriovenous malformations are usually treated with selective arterial embolization before surgical resection (4–6,12–15).
Diagnosis of VVMs is usually based on medical history and physical examination (16). Magnetic resonance (MR) imaging features for VVMs have been previously established (17–19), including high signal intensity on T2-weighted images, intermediate signal intensity on T1-weighted images, multicystic spaces in lymphatic components with or without associated hemorrhage, contrast enhancement of venous components, lack of signal voids, presence of phleboliths, and the general appearances of superficial well-defined lesions or deeply infiltrating lesions along fascial planes. Therefore, MR imaging can help characterization and diagnosis of the lesion, but its most important role is in the display of the extent and relationship of the lesion to surrounding structures (17,18,20,21). It is this capability of MR imaging that helps therapeutic planning, and we postulate that MR imaging has the potential to help predict treatment outcome.
At our institution, a multidisciplinary team, including pediatric plastic surgeons and interventional radiologists, undertakes the management of VVMs. A carefully reasoned treatment plan that can be comprehended by the patient and family, including realistic therapeutic expectations, is crucial. Treatment strategies are based on a comprehensive clinical assessment and MR imaging. To our knowledge, there are no data published for the MR features of VVMs that help predict outcome of these lesions after sclerotherapy.
The purpose of this study was to compare the clinical results of percutaneous sclerotherapy of VVMs in children with our proposed MR classification system.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Two radiologists (M.G., P.C.), who were blinded to the treatment outcome, retrospectively and independently reviewed the preprocedural MR images. Any differences in opinion were resolved with consensus. Imaging was performed with a 1.5-T superconducting magnet. Spin-echo T1-weighted (repetition time msec/echo time msec = 600/20) and spin-echo T2-weighted (2,500/20) images were obtained of all lesions in multiple planes. Additional imaging was performed when the clinical scenario dictated, including MR angiography if there was suspicion of an arteriovenous malformation and gadolinium-enhanced (Magnevist [0.1 mmol per kilogram of body weight]; Berlex Laboratories, Wayne, NJ) MR imaging if there was suspicion of a neoplasm. Lesions were categorized on the basis of our proposed MR classification system according to VVM morphology, including size and margins, which were well defined, ill defined, or infiltrative (Table; Figs 1–4). A well-defined or circumscribed margin has a sharp abrupt transition from the surrounding tissue, which remains confined to its tissue and fascial plane. An ill-defined margin has an irregular interface with the surrounding tissue because the lesion may cross tissue and fascial planes. Differentiation between a venous and a lymphovenous malformation was based on MR characteristics, including nonenhancing cystic spaces with or without hemorrhage in lymphatic portions and enhancing venous components. The diagnosis of a lymphovenous malformation was confirmed at venography on the basis of aspiration of lymph and blood before the injection of contrast material.
|
|
|
|
|
|
|
|
|
When possible, tourniquets were placed proximal to the lesion to slow venous drainage and maximize the time for alcohol to sclerose the lesion. During the procedure, the injection was fluoroscopically evaluated to ensure intralesional administration of ethanol without extravasation into the surrounding soft tissues or draining veins. A maximum volume of 1 mL/kg was injected at one sitting. The injection was stopped before this volume was reached when the lesion was completely opacified, resistance to injection was encountered, or opacification of the venous drainage was seen. In 23 lesions, multiple injections were required at different sittings (range, two to eight sittings).
Patients were initially followed up at our vascular malformation clinic approximately 12 weeks after the procedure. This time allowed resolution of swelling and changes directly related to sclerotherapy and proper assessment of the treatment results. Recommendations for repeat sclerosis or clinical follow-up were based on this initial assessment. For the purpose of this study, the minimum 12-week follow-up was required.
Results of sclerosis therapy were graded with subjective clinical assessment, including interview and examination. MR imaging follow-up was available for some patients but was not included in the evaluation of clinical results. The multidisciplinary team at our vascular malformation clinic, including a pediatric interventional neuroradiologist (D.A.) and a plastic surgeon, performed the assessments. Both lesion size and patient symptoms were assessed, including pain, limitation of movement and function, airway compression, and cosmetic deformity. Patients were followed up from 1 to 5 years. Clinical results were graded as follows: poor, little or no improvement; good, significant decrease in size and symptoms; and excellent, clinical obliteration.
The Fisher exact test was used to test the association between the proposed MR grades and the clinical results of sclerotherapy. Interobserver agreement about lesion margins on MR images was evaluated with the 
statistic. Regression models were used to test for a trend in clinical outcomes across the proposed MR classification groups. Logarithmic transformations of volume and the number of puncture sites were used to ensure approximately constant variances in the corresponding regression analyses.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Table shows the association between the proposed MR classification of VVMs and the clinical assessment in the 59 patients. The association between MR grade and clinical assessment was statistically significant (P < .001, Fisher exact test). Owing to the small number of lesions in the lymphovenous group, subgroup statistical analysis was not performed.
Interobserver agreement was measured for the MR features evaluated in the grading system. There were no discrepancies between the two independent reviewers in the assessment of lesion size. The observers had substantial agreement about lesion margins ( = 0.79).
During the 1–5-year clinical follow-up, 23 (39%) of the 59 patients required repeat sclerosis. The decision for repeat therapy was determined at either the time of follow-up, on the basis of a lack of clinical improvement, or at the time of therapy, if the lesion was large and multiloculated and therefore required volumes of ethanol that were too large to be administered at one sitting. The mean numbers of repeat sclerotherapy sessions required for each patient for each lesion grade are as follows: grade 1, n = 1, median = 1.0 patients; grade 2A, n = 1.4, range = 1.0–3.0 patients, median = 1.0 patients; grade 2B, n = 2, median = 2.0 patients, range = 1.0–4.0 patients; grade 3, n = 3, median = 2.0 patients, range = 1.0–8.0 patients. With Poisson regression analysis, a trend was found for an increasing number of sessions with increasing lesion grade (P = .024).
Mean volumes of ethanol and metrizamide mixture injected for each lesion for each sitting were the following: grade 1 lesions, 8.4 mL (median, 8.5 mL; range, 3.0–18.0 mL); grade 2A lesions, 17.1 mL (median, 15.0 mL; range, 7.0–30.0 mL); grade 2B lesions, 6.8 mL (median, 12.9 mL; range, 2.0–60.0 mL); and grade 3 lesions, 26.4 mL (median, 20.0 mL; range 3.0–70.0 mL). Linear regression analysis showed that there is an increasing trend in the logarithm of volumes of ethanol and metrizamide with increasing lesion grade (P < .001). Mean total volumes of ethanol and metrizamide mixture injected for each lesion over multiple sittings were the following: grade 1 lesions, 8.4 mL (median, 8.5 mL; range, 3.0–18.0 mL); grade 2A lesions, 24.8 mL (median, 20.0 mL; range, 9.0–60.0 mL); grade 2B lesions, 32.4 mL (median, 29.5 mL; range, 2.0–76.0 mL); grade 3 lesions, 105.4 mL (median, 38 mL; range, 3.0–566.0 mL). With a similar analysis, there was a trend of increasing total lesional volumes of injected ethanol and metrizamide with increasing grade (P < .001).
In addition, 42 (71%) of the 59 lesions required multiple access sites to opacify the malformation. The mean number of access sites for each lesion averaged for the total number of sessions were as follows: grade 1 lesions, 1.8 sites (median, 2.0 sites; range, 1.0–4.0 sites); grade 2A lesions, 3.4 sites (median, 4.0 sites; range, 1.0–6.0 sites); grade 2B lesions, 3.9 sites (median, 3.0 sites; range, 1.0–12.0 sites); and grade 3 lesions, 6.5 sites (median, 5.0 sites; range, 1.0–17.0 sites). A linear regression on the logarithm of the number of puncture sites showed an increasing trend across grade (P < .001). No lesion recurred after initial clinical improvement or resolution during follow-up.
Complications were self-limiting in all cases and included skin necrosis in one patient and skin blistering in three patients. No definite association between complications and lesion grade was identified, and results of statistical evaluation would be inconclusive due to the small numbers. In our experience, very few patients experienced mild transient postprocedural pain and one patient required admission to the hospital for pain control with intravenous analgesics for less than 24 hours.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although an excellent response was not seen in large infiltrating VVMs (grade 3), ethanol sclerotherapy proved beneficial in nine (43%) of the 21 patients with improved symptoms. Therefore, we believe ethanol sclerotherapy is a worthwhile treatment option in such patients. In addition, poor definition of a lesion is a poorer prognostic factor than is lesion size, as suggested by the results in our patients with grade 2 VVMs. However, the number of lesions in this category is too small to predict statistical significance.
Our results show a trend for the necessity to increase the number of repeat sclerotherapies, the mean volumes of ethanol and metrizamide mixture injected in each lesion during each sitting and over multiple sittings, and the number of access puncture sites with increasing lesion grade. These data may prove useful in procedure planning.
We studied an exclusively pediatric population. Various patient populations have been investigated in previous studies, including a similar population (4,6) and a mixed population of children and adults (22–24). No substantial difference in outcomes is evident between these investigations.
Of the 59 patients in our study, four (7%) experienced self-limiting complications, including skin necrosis and skin blistering. In previous studies, complication rates ranged from 4% to 25%, including skin necrosis that was reversible or required skin grafting, blistering, and reversible neuropathy (4,22–24). The complication rate in our study is within the range of rates in previous studies.
At our institution, patients with VVMs undergo evaluation in a multidisciplinary clinic that specializes in the treatment of vascular lesions. As part of their diagnostic work-up, most patients undergo MR imaging to help determine the diagnosis but, more important, to help determine the treatment plan. On the basis of our clinical experience and judgment, we question the potential usefulness of MR imaging to help predict outcome of VVMs treated with percutaneous ethanol sclerotherapy. VVMs are difficult lesions to treat; therefore, it is important for treating physicians to be familiar with the potential successes and limitations of the proposed treatment when they determine appropriate candidates for such therapy and when they counsel patients and families about expected outcomes (7). Our proposed MR classification is based on clinical experience. The results in our study are preliminary and were obtained retrospectively in a relatively small number of patients. Further validation is needed.
In conclusion, our results demonstrate a strong association between the proposed MR classification and the results of percutaneous ethanol sclerotherapy. An excellent response was seen in most small well-defined VVMs, and a poor result was seen in most large infiltrating lesions. Although poor results were seen in the majority of grade 3 lesions and none resolved with treatment, improvement in symptoms and lesion size was seen in some patients; therefore, we believe this is a potentially beneficial therapy.
These data may prove useful in the choice of appropriate patients for treatment and in the counseling of patients and families regarding realistic expectations of ethanol sclerosis therapy.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviation: VVM = venous vascular malformation
Author contributions: Guarantor of integrity of entire study, M.G.; study concepts and design, M.G.; literature research, P.A.C.; clinical studies, D.A.; data acquisition, P.A.C., M.G.; data analysis/interpretation, M.G.; statistical analysis, M.G.; manuscript preparation, definition of intellectual content and editing, P.A.C.; manuscript revision/review, P.A.C., M.G.; manuscript final version approval, P.A.C., M.G., D.A.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Hyodoh, M. Hori, H. Akiba, M. Tamakawa, K. Hyodoh, and M. Hareyama Peripheral Vascular Malformations: Imaging, Treatment Approaches, and Therapeutic Issues RadioGraphics, October 1, 2005; 25(suppl_1): S159 - S171. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. O. Arat, M. E. Mawad, and M. Boniuk Orbital Venous Malformations: Current Multidisciplinary Treatment Approach Arch Ophthalmol, August 1, 2004; 122(8): 1151 - 1158. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. U. Herborn, M. Goyen, T. C. Lauenstein, J. F. Debatin, S. G. Ruehm, and K. Kroger Comprehensive Time-Resolved MRI of Peripheral Vascular Malformations Am. J. Roentgenol., September 1, 2003; 181(3): 729 - 735. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
