| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ultrasonography |
1 From the Departments of Radiology (H.J.P., P.E.B.), Pathology (H.P.W.K.), and Research Computing and Biostatistics (D.Z.), and the Division of Plastic Surgery (J.B.M.), Children's Hospital and Harvard Medical School, 300 Longwood Ave, Boston, MA 02115. From the 1994 RSNA scientific assembly. Received February 11, 1999; revision requested April 2; revision received June 29; accepted July 20. Address reprint requests to H.J.P. (e-mail: paltiel@a1.tch.harvard.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Eighty-seven vascular anomalies were evaluated by means of US. Lesions were assessed for the presence of solid tissue and abnormal arteries, veins, or cysts. Vessel density, peak flow velocities, and resistive indexes were compared.
RESULTS: There were 49 hemangiomas and 38 vascular malformations. A significantly greater proportion of hemangiomas (48 of 49) compared with vascular malformations (zero of 38) consisted of a solid-tissue mass (P < .001). Vessel density was comparable for hemangioma and arteriovenous malformation (AVM) but significantly greater compared with the other vascular malformations (P < .001 in each case). No differences in mean arterial peak velocity were detected between hemangiomas and malformations. Mean venous peak velocity was significantly higher for AVM than for other vascular malformations and hemangioma. Mean resistive index was greater for lymphatic malformation than for hemangioma or AVM. Abnormal veins, arteries and veins, or cysts were univariate predictors for distinguishing between venous, arteriovenous, and lymphatic malformations (P < .001 in all cases). Solid-tissue mass was the only multivariate predictor for differentiating hemangioma from vascular malformation (likelihood ratio test = 109.8, P < .001).
CONCLUSION: US can be used to distinguish hemangioma from vascular malformation and detect arterial flow. These distinctions are critical for subsequent management and assessing prognosis.
Index terms: Angioma, 9*.312, 9*.83 • Arteries, abnormalities, 9*.141 • Arteriovenous malformations, 9*.14 • Lymphangioma, 9*.83 • Soft tissues, abnormalities, 9*.14 • Ultrasound (US), Doppler studies, 9*.12983 • Ultrasound (US), in infants and children, 9*.12983 • Veins, abnormalities, 9*.142
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Vascular anomalies of infancy and childhood can be divided into two major categories—hemangiomas and malformations—on the basis of physical findings, clinical behavior, histologic findings, and cellular kinetics (3).
Hemangiomas are the most common soft-tissue tumors of infancy and consist of small blood vessels of capillary morphology and size. They occur with a prevalence of 1%–2% in neonates (4) and of 12% in infants by the age of 1 year (5). The natural history of hemangioma is characterized clinically by rapid postnatal growth for 8–12 months, followed by slow but inevitable regression for the next 1–5 years (1).
Vascular malformations are errors of morphogenesis and consist of abnormal channels lined by normal endothelium that has a low turnover. These anomalies are subclassified according to the predominant channel type, namely, arterial, venous, capillary, lymphatic, or combinations thereof (3). Vascular malformations presumably are present at birth, although they may not become evident until adolescence or adulthood, and they persist throughout life.
Our hospital is a referral center for the diagnosis and treatment of vascular anomalies. Between 400 and 500 children and adults are seen by members of the vascular anomalies team per year. In some patients, the precise nature of a vascular anomaly is unclear, despite a thorough review of their history and careful physical examination. Such patients often are referred for US evaluation, especially if the abnormality is localized and superficial.
Gray-scale ultrasonography (US) coupled with color Doppler flow imaging has the advantage of providing a rapid, relatively inexpensive, and noninvasive assessment of lesion morphology and vascular components. Pulsed Doppler US permits spectral analysis of arterial and venous flow and measurement of flow velocities.
The aim of our study was to determine which gray-scale, color Doppler, and spectral Doppler US features of localized and previously untreated superficial soft-tissue vascular anomalies could be used to differentiate hemangiomas from vascular malformations and one type of malformation from another.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
All patients were examined by members of the vascular anomalies team that included physician specialists in general and plastic surgery, hematology, radiology, otorhinolaryngology, and dermatology. For each patient, a retrospective diagnosis was assigned by consensus, on the basis of a review of the medical record, including clinical history, physical examination results, and evolution of the lesion; computed tomographic (CT), magnetic resonance (MR) imaging, and/or angiographic findings; and histologic features.
Lesions were categorized as hemangiomas or vascular malformations. Malformations were subclassified into capillary, venous, lymphatic, arteriovenous, and combined types, according to the classification proposed by Mulliken and Glowacki in 1982 (3). Patients with a history of prior pharmacologic, radiologic, or surgical intervention, as well as individuals whose lesions were posttraumatic, were excluded from the study. The number and type of additional imaging studies, subsequent biopsy, and/or surgery were recorded for each patient.
US Evaluation
US was performed by the first author (H.J.P.) in 75 patients (87%). The 11 other patients (13%) were examined by one of several examiners, all of whom are experienced sonologists. Studies were performed without sedation with a 128 XP unit (Acuson, Mountain View, Calif) and linear-array transducers with imaging and Doppler frequencies of 7 MHz (79 patients), 5 MHz (one patient), or both (one patient) or with a Logiq 700 unit (GE Medical Systems, Milwaukee, Wis) and linear-array transducers with frequencies of 5–10 MHz (three patients) and 6–13 MHz (two patients) and a Doppler frequency of 6.25 MHz. The transducer power output was set below 100 mW/cm2, in accordance with the U.S. Food and Drug Administration's acoustic output guidelines for US studies.
Gray-scale images of the soft-tissue abnormalities were obtained in transverse and longitudinal planes. Lesions with well-defined margins were measured in three dimensions—length, width, and thickness. Low-flow color Doppler settings were used in all cases to permit optimal visualization of small vessels and detection of low-velocity arterial and venous flow. Color Doppler flow imaging was used to guide placement of the Doppler gate for pulsed Doppler analysis. Spectral Doppler analysis was attempted in all patients. An effort was made to obtain spectra from multiple sites within every lesion. Each study was recorded on film hard copy.
Image Analysis
A retrospective evaluation of all US images was performed by the first author (H.J.P.) as follows: A determination was made regarding the presence of either a well-circumscribed or an infiltrating soft-tissue lesion on gray-scale images. When a lesion with well-defined borders was identified, its largest diameter was recorded, and the presence or absence of a solid-tissue mass was noted. If the lesion was too large to be accurately measured with US, the largest diameter was determined by means of another imaging modality, physical examination, or intraoperative assessment.
When a parenchymatous mass was identified, its echogenicity relative to the adjacent normal soft tissues was determined and classified as hypoechoic, hyperechoic, or heterogeneous if it contained both hypo- and hyperechoic components and/or calcification. When no solid-tissue mass was depicted, the presence or absence of cysts or abnormal channels within the soft tissues was recorded.
Color Doppler flow images were analyzed for the presence and location of vessels within the lesion. Vessel location was categorized as central, peripheral, or both central and peripheral. By using a modification of the method described by Dubois et al (6) to assess vessel density, the area of greatest vascularity was identified and the number of color flow signals within an area 1 cm2 were counted. A vessel was defined as a linear or punctate structure that was not associated with adjacent, scattered, randomly distributed color pixels (color noise). Vessel density was defined semiquantitatively as low, fewer than two vessels per square centimeter; moderate, two to four vessels per square centimeter; or high, five or more vessels per square centimeter.
Spectral tracings were reviewed to document arterial flow, venous flow, or both. A study was deemed adequate if there was a minimum of two consecutive, equivalent waveforms from at least one artery or vein. This is the smallest number of waveforms that will ensure that a change in the Doppler interrogation angle or movement of the sample volume has not occurred.
Peak arterial and venous velocities and the lowest arterial resistive index values were recorded and used for statistical analysis. In the majority of lesions, macroscopic vessels could not be identified with gray-scale US alone. When flow was detected with color Doppler US, vessels usually were depicted as short vascular segments having a variable orientation to the scanning plane. Therefore, reliable angle correction was not possible for most examinations.
Since the Doppler display on both the Acuson and GE Medical Systems US equipment is calibrated in centimeters per second rather than in kilohertz, velocities were recorded in centimeters per second, and a vessel-Doppler US beam angle of 0° was assumed (7). Arterial resistive indexes were determined as follows: peak systolic Doppler shift - end diastolic Doppler shift/peak systolic Doppler shift.
Statistical Analysis
The Fisher exact test was used to determine significant differences between hemangiomas and vascular malformations with respect to the proportion of patients with (a) a solid-tissue mass, (b) multiple arteries and veins, (c) multiple venous channels, and (d) multiple cysts.
Vessel density was compared by means of the Pearson 
2 test with Yates correction for continuity. Arterial and venous peak flow velocities and resistive index measurements were tested for normality by means of the Kolmogorov-Smirnov statistic (8), and no serious departures were detected. Therefore, continuous variables were compared between groups with analysis of variance, or ANOVA, and the Student t test and expressed in terms of the mean and SEM. Stepwise logistic regression with the likelihood ratio test (9) was used to determine independent multivariate predictors to differentiate hemangiomas from vascular malformations. A two-tailed P value less than .05 was considered to indicate a statistically significant difference. Statistical analysis was performed with a software package (SPSS version 8.0; SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Results of the Fisher exact test demonstrated significant differences between hemangiomas and all vascular malformations with respect to the presence of a solid-tissue mass; for AVM versus the other vascular malformations and hemangioma regarding infiltrating arteries and veins; for venous malformation versus the other vascular malformations and hemangioma regarding well-circumscribed or infiltrating collections of veins; and for lymphatic malformation versus the other vascular malformations and hemangioma with respect to the presence of multiple cysts (all values of P < .001).
Logistic regression indicated that the presence or absence of a solid-tissue mass, infiltrating arteries and veins, well-circumscribed or infiltrating venous channels, and multiple cysts were each significant univariate predictors in differentiating between hemangioma and vascular malformation (all values of P < .001). However, stepwise logistic regression revealed that a solid-tissue mass was the only significant multivariate predictor that distinguished hemangioma from vascular malformation (likelihood ratio test = 109.8, P < .001).
There was no statistically significant difference in vessel density between hemangioma containing arteries and AVM (P = .34). However, vessel density for these hemangiomas versus the low-flow malformations (ie, venous, lymphatic, capillary, and capillary-lymphatic malformations) and for AVM versus the low-flow malformations was significantly higher (P < .001 in each case).
No statistically significant differences were detected between hemangioma versus vascular malformation with respect to mean arterial peak velocity (P = .71). However, mean venous peak velocity was significantly greater for AVM versus hemangioma (P = .001), venous malformation (P = .04), and lymphatic malformation (P = .03).
No statistically significant differences were detected in mean resistive index values for hemangioma versus AVM (P = .78). However, patients with lymphatic malformation had a mean resistive index value significantly greater than those with hemangioma or AVM (P < .001 in both cases).
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Surprisingly little has been written regarding the role of US in the diagnosis of vascular anomalies, despite its wide availability, ease of use, noninvasiveness, and relatively low cost (6,11–16). To our knowledge, no prior publication has attempted to characterize the differential US imaging features of the various soft-tissue vascular anomalies.
Hemangioma with arterial flow was distinguished from AVM in every patient in our series by the presence of solid parenchymal tissue. This finding is in keeping with published descriptions of these lesions at MR imaging and angiography (11,17–21). As previously noted by Dubois et al (6), hemangioma could not be differentiated from AVM on the basis of mean vessel density, mean arterial peak flow velocity, or mean resistive index values.
In our series, the mean venous peak velocity was significantly higher for AVM than for hemangioma. This may be because arteriovenous shunting is a prominent feature of AVM, whereas it is variably present in association with hemangioma, depending on its stage within the life cycle, but is not seen in lymphatic malformation (11,20,21). The distinction between hemangioma and AVM is critical, since their prognosis and treatment are so different.
For the majority of hemangiomas, no treatment is required because of their spontaneous involution. In rare instances, hemangiomas may divert a large volume of blood and cause congestive heart failure. Drug therapy is indicated for those hemangiomas that distort anatomic structures, impair function, or are life threatening (approximately 1%).
An AVM is a high-flow anomaly consisting of multiple abnormal communications between arteries and veins. AVMs usually are detected initially in infancy and remain quiescent during childhood. Puberty, pregnancy, hormonal treatment, or trauma often precipitate evolution. Arteriovenous shunts bypass the capillary bed and result in cutaneous ischemia, ulceration, and hemorrhage. The lesions may be disfiguring, impair function, or result in increased cardiac output and congestive heart failure. When progression occurs, arterial embolization is the preferred treatment, with or without radical surgical excision (22).
Three hemangiomas in our series had no detectable arterial or venous flow. Two of them were clinically in the involuting phase. They were distinguished from vascular malformations owing to the presence of solid tissue. The third was an involuted hemangioma with no gray-scale or Doppler US abnormality. Therefore, it could not be differentiated from the low-flow vascular anomalies that revealed either no US abnormality or only diffuse thickening of the skin and subcutaneous tissues (ie, four capillary, two venous, and one capillary-lymphatic malformation).
Clinically, there is unlikely to be confusion regarding the distinction between an involuted hemangioma that has already undergone growth and regression and a more indolent low-flow vascular malformation. This is reflected in the composition of our series, wherein patients with hemangiomas were referred to our program with tumors almost exclusively in the proliferating phase while they still demonstrated high-velocity, low-resistance arterial flow. In one patient with a proliferating hemangioma, only arterial spectra were documented. This particular study was not performed by the first author (H.J.P.), and the absence of venous spectra is, in retrospect, likely because of incomplete sampling at the time of examination.
Either well-circumscribed, spongelike vascular spaces or poorly marginated collections of veins are characteristic of venous malformation, whereas multiple cysts are characteristic of macrocystic lymphatic malformation. Five of the eight macrocystic lymphatic malformations in our series demonstrated arterial and venous flow within the septa or within the stromal tissue. The presence of flow correlates with previous angiographic descriptions of rim enhancement and capillary blush (21). Histologically, small arteries and veins frequently are seen surrounding the fibrous walls of the lymphatic channels. The high mean resistive index values and relatively low mean venous peak velocities detected in five of the lymphatic malformations is reflective of their slowly progressive clinical behavior as compared with high-flow hemangioma and AVM.
Low-flow malformations pursue a more indolent course than high-flow lesions. Treatment is administered on the basis of symptoms and is supportive in nature. A capillary malformation is usually flat and sharply demarcated and grows proportionately with the child. In older patients, facial capillary malformations often are associated with soft-tissue overgrowth, skeletal overgrowth, or both.
Venous malformations are present from birth and initially appear as either a faint blue patch or a soft mass. They often enlarge slowly during childhood and adolescence and to a lesser degree during adulthood. They are easily compressed and swell when dependent. Episodic thromboses occur, and large venous malformations may be associated with chronic consumption of fibrinogen and release of fibrin split products.
Lymphatic malformations are present from birth and usually are detected by 2 years of age. They may be macrocystic, microcystic, or combined (11). Macrocystic lymphatic malformations are cool, soft, smooth, translucent masses that occur beneath normal or bluish skin. Microcystic lymphatic malformations permeate the skin and muscles and often are associated with tiny, clear cutaneous vesicles. Associated soft-tissue and skeletal overgrowth is common. Recurrent infection and intralesional hemorrhage are frequent complications.
Angiokeratoma is a localized vascular malformation of the dermis consisting of small vessels believed to be of capillary and lymphatic origin and with associated hyperkeratosis (23).
We made no attempt to assess the reliability of gray-scale US and color Doppler flow imaging in distinguishing vascular anomalies from other soft-tissue masses. There is an overlap of imaging features between vascular anomalies and other soft-tissue lesions, both benign and malignant (15). The results of imaging studies must be integrated with the history and physical examination results. We always perform biopsy of any atypical lesion when the clinical features, imaging results, or both are equivocal.
The other main limitation of this study is the problem of performing accurate measurements of vessel density and flow velocity. Retrospective color Doppler US analysis is difficult because artifacts may be difficult to distinguish from true color flow signals. Furthermore, owing to the tortuous course of many of the vessels within hemangiomas and vascular malformations, it may be impossible to differentiate two or more separate channels from a single channel that has multiple entry and exit points within a single imaging plane. Dubois et al (6) never specifically dealt with this issue when describing their technique for measuring vessel density. Our results are comparable with those of Dubois et al (6) in that hemangiomas with arterial flow and AVMs had high or moderate vessel density.
Dubois and co-workers (6) avoided the problem of angle correction by reporting Doppler shifts in kilohertz instead of velocity measurements in centimeters per second. However, most current US equipment automatically computes and displays velocity measurements. An angle of 0° is assumed if angle correction is not performed by the operator. Lack of angle correction can result in serious underestimation of velocity measurements.
We found that mean arterial peak velocity measurements were not useful in differentiating hemangioma from AVM and were not statistically different from the mean arterial peak velocity measurements for lymphatic malformation. Although the mean resistive indexes for hemangioma and AVM were significantly lower than the mean resistive indexes for lymphatic malformation, and the mean venous peak velocity for AVM was higher than that for hemangioma, venous malformation, and lymphatic malformation, there was substantial variation in values within each group of abnormalities. These findings and inherent problems with methodology cause us to question the value of vessel density, peak arterial and venous velocities, and resistive index measurements as a basis for differentiating between the various vascular anomalies.
Limitations of US include the small field of view; restricted depth of penetration, especially with high-frequency transducers; difficulty in depicting flat superficial lesions; and detecting tiny vessels with low flow. Thus, this technique cannot always substitute for CT, MR imaging, or angiography when determination of the full extent of an anomaly is necessary or when a more precise diagnosis of a low-flow malformation is required and the US findings are not characteristic. In the future, the use of power Doppler imaging, as well as of Doppler contrast agents, undoubtedly will improve the detection of low velocity flow.
In conclusion, gray-scale US coupled with color Doppler flow imaging and spectral analysis is a very useful initial screening procedure in patients with soft-tissue lesions of presumed vascular origin owing to its ability to distinguish hemangioma, a tumor of solid tissue, from vascular malformations composed of abnormal vessels or cysts, and lesions containing arteries from those in which arterial flow is absent. These distinctions are of great value in directing subsequent investigation and management, since prognosis and appropriate treatment vary substantially for each type of anomaly.
|
Footnotes |
|---|
Abbreviation: AVM = arteriovenous malformation
Author contributions: Guarantor of integrity of entire study, H.J.P.; study concepts, H.J.P., P.E.B., J.B.M.; study design, H.J.P.; definition of intellectual content, H.J.P.; literature research, H.J.P.; clinical studies, H.J.P., H.P.W.K.; data acquisition and analysis, H.J.P.; statistical analysis, D.Z.; manuscript preparation, H.J.P.; manuscript editing and review, all authors.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Mazoyer, O. Enjolras, A. Bisdorff, J. Perdu, M. Wassef, and L. Drouet Coagulation Disorders in Patients With Venous Malformation of the Limbs and Trunk: A Case Series of 118 Patients Arch Dermatol, July 1, 2008; 144(7): 861 - 867. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W.-H. Tan, H. N Baris, P. E Burrows, C. D Robson, A. I Alomari, J. B Mulliken, S. J Fishman, and M. B Irons The spectrum of vascular anomalies in patients with PTEN mutations: implications for diagnosis and management J. Med. Genet., September 1, 2007; 44(9): 594 - 602. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. R. Stewart, S. Sriprasad, S. Pomplun, K. Walsh, and P. S. Sidhu Sonographic Features of a Spermatic Cord Capillary Hemangioma J. Ultrasound Med., January 1, 2007; 26(1): 139 - 142. [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Ohgiya, T. Hashimoto, T. Gokan, S. Watanabe, M. Kuroda, M. Hirose, S. Matsui, H. Nobusawa, T. Kitanosono, and H. Munechika Dynamic MRI for Distinguishing High-Flow from Low-Flow Peripheral Vascular Malformations Am. J. Roentgenol., November 1, 2005; 185(5): 1131 - 1137. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. M. Boon, J. B. Mulliken, O. Enjolras, and M. Vikkula Glomuvenous Malformation (Glomangioma) and Venous Malformation: Distinct Clinicopathologic and Genetic Entities Arch Dermatol, August 1, 2004; 140(8): 971 - 976. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. I. Olsen, G. S. Stacy, and A. Montag Soft-Tissue Cavernous Hemangioma RadioGraphics, May 1, 2004; 24(3): 849 - 854. [Full Text] [PDF]
|
|
|
|
 |
|
 I Vargel, B E Cil, P Kiratli, D Akinci, and Y Erk Hereditary intraosseous vascular malformation of the craniofacial region: imaging findings Br. J. Radiol., March 1, 2004; 77(915): 197 - 203. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I Beggs Ultrasound of soft tissue masses Imaging, June 1, 2002; 14(3): 202 - 208. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Dubois, L. Garel, M. David, and J. Powell Vascular Soft-Tissue Tumors in Infancy: Distinguishing Features on Doppler Sonography Am. J. Roentgenol., June 1, 2002; 178(6): 1541 - 1545. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Bodner, M. F. H. Schocke, F. Rachbauer, K. Seppi, S. Peer, A. Fierlinger, T. Sununu, and W. R. Jaschke Differentiation of Malignant and Benign Musculoskeletal Tumors: Combined Color and Power Doppler US and Spectral Wave Analysis Radiology, May 1, 2002; 223(2): 410 - 416. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. S. P. van Rijswijk, E. van der Linden, H.-J. van der Woude, J. M. van Baalen, and J. L. Bloem Value of Dynamic Contrast-Enhanced MR Imaging in Diagnosing and Classifying Peripheral Vascular Malformations Am. J. Roentgenol., May 1, 2002; 178(5): 1181 - 1187. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Marrocco-Trischitta, E. M. Nicodemi, C. Nater, and F. Stillo Management of Congenital Venous Malformations of the Vulva Obstet. Gynecol., November 1, 2001; 98(5): 789 - 793. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Dubois, G. Soulez, V. L. Oliva, M.-J. Berthiaume, C. Lapierre, and E. Therasse Soft-Tissue Venous Malformations in Adult Patients: Imaging and Therapeutic Issues RadioGraphics, November 1, 2001; 21(6): 1519 - 1531. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||
