Color doppler ultrasonography of soft-tissue masses

Transcription

1 Acta Radiologica ISSN: (Print) (Online) Journal homepage: Color doppler ultrasonography of soft-tissue masses Roberto Lagalla, A. Iovane, G. Caruso, M. Lo Bello & L. E. Derchi To cite this article: Roberto Lagalla, A. Iovane, G. Caruso, M. Lo Bello & L. E. Derchi (1998) Color doppler ultrasonography of soft-tissue masses, Acta Radiologica, 39:4, To link to this article: Published online: 04 Jan Submit your article to this journal Article views: 1039 Full Terms & Conditions of access and use can be found at

2 Acta Radiologica 39 (1998) Printed in Denmark - All rights reserved Copyright 0 Acta Radiologica 1998 ACTA R A DI 0 LOG I C A ISSN COLOR DOPPLER ULTRASONOGRAPHY OF SOFT-TISSUE MASSES R. LAGALLA', A. IOVANE', G. CARUSO', M. Lo BELLO' and L. E. DERCHI~ Departments of Radiology, 'University of Palermo, Palermo, and *University of Genoa, Genoa, Italy. Abstract Purpose: To evaluate the capability of color Doppler ultrasonography to differentiate between benign and malignant soft-tissue tumors. Material and Methods: We reviewed the ultrasonographic (US) and color Doppler (CD) findings in 46 consecutive patients with a palpable periskeletal mass. The presence of 3 or more vascular hila and of tortuous and irregular internal vessels within the lesions was considered an indication of malignancy. The CD diagnosis was compared with that obtained at US alone. Results: The sensitivity and specificity of CD were respectively 85% and 92%; these values were higher than those obtained at US alone, respectively 75% and 50%. Arteriovenous malformations presented as lesions with large internal vessels that had low vascular impedance and were easily diagnosed. The waveform patterns within solid tumors were not specific. Conclusion: At present, US is commonly employed to confirm the presence of a suspected soft-tissue mass, to locate it accurately, and to indicate its nature. CD findings enhance the role of the US technique in such lesions. The combined use of US and CD can allow the differentiation of benign from malignant lesions, and thus provide a better basis for treatment. Key words: Soft-tissue tumors, ultrasonography; tissue characterization, color Doppler. Correspondence: Roberto Lagalla, Department of Radiology, University of Palermo, Via del Vespro 127, Palermo, Italy. FAX Accepted for publication 8 October Ultrasonography (US) is one of the preferred diagnostic modalities in the study of lesions of the musculoskeletal system. In the evaluation of soft-tissue masses, US is employed: to confirm their presence; to identify their location and relationship to surrounding structures; and to analyze the inner texture and character of the margins for the purpose of differentiating between benign lesions and malignancies (1, 7, 9). However, US offers a relatively low level of accuracy as only malignant tumors can be diagnosed reliably by the detection of local invasion or by the presence of metastases, while the benign nature of a lesion cannot be ascertained. Analyses of flow by color Doppler (CD) techniques have recently been used to provide additional parameters for differentiation. Good results have been reported in masses that affect both superficial and deep-lying organs (4, 6, 8, 10, 11, 17). However, little attention has been paid to the use of CD in the study of lesions affecting the musculoskeletal system (3, 13). The aim of this paper was to use this technique as an aid to the differential diagnosis of benign and malignant space-occupying lesions in the periskeletal soft tissues. Material and Methods We reviewed the US and CD findings in 46 consecutive patients with a palpable mass affecting the periskeletal soft tissues. They comprised 21 male and 25 female patients, aged 3-67 years. The maximum diameter, location, and histological diagnosis of the lesion in each patient are presented in Table 1. Patients with subcutaneous lipomas were excluded as these lipomas present no diagnostic problem and they are virtually avascular. Both US imaging and the CD evaluation of the lesions were obtained on real-time equipment with 421

3 R. LAGALLA ET AL. a broadband linear-array probe that operates at a frequency of 5-10 MHz for imaging and 6 MHz for Doppler (Ultramark 9 HDI, Advanced Technology Laboratories (ATL), Bothell, WA, USA). Setting parameters for CD included a pulse-recurrence frequency (PRF) within the Hz range with a wall filter at 100 Hz. The color gain was adjusted to a level at which noise began to appear on the screen and then slightly lowered to achieve a clean image. The region of interest was scanned with minimum pressure from the transducer. In some cases, even slight compression could obliterate the flow signals from small intratumoral vessels, probably owing to low perfusion pressure. Pulsed Doppler spectral analysis was obtained in all lesions with vascular signals by means of a 2-mm sample volume and a PRF of 2000 Hz; waveforms were collected at a minimum of 5 dif- Fig. 1. Benign intramuscular myxoma. a) US demonstrates a hypo-echoic mass with clear edges that is well demarcated from the adjacent structures. b) CD shows an absence of intralesional flow with color signals visible only at the periphery of the mass. c) Histology shows mucoid material mixed with muscle cells and an absence of vascular structures. Fig. 2. Myxoid liposarcoma. a) US shows a large hypo-echoic mass with lobulated margins that is well demarcated from surrounding tissues. b) CD identifies one linear vessel at the periphery of the mass (4). c) Histology shows tissue in which stroma and anaplastic cells are scarce; some of those that are present have fat vacuoli. 422

4 COLOR DOPPLER US OF SOFT-TISSUE MASSES Fig. 3. Malignant schwannoma. Heterogeneous oval mass with irregular edges which, at CD, appears hypervascularized. Fig. 5. Cavernous hemangioma of the arm. CD shows a large vessel (+) surrounding the posterior part of the lesion; a fluidfluid level is clearly recognizable. ferent sites within each mass and a mean resistive index (peak systolic velocity minus end-diastolic velocity/peak systolic velocity) was calculated. We were unable to measure angle-corrected velocities in all patients: in many cases, the vascular signals could be seen only as small colored spots, and the direction of the vessel could not be determined, precluding angle correction. The US images were reviewed for dimension, echogenicity, textural pattern and lesion margins. The internal structure was graded as hypo-, hyperor iso-echogenic with respect to surrounding muscles; the textural pattern was considered to be homogeneous, heterogeneous, or complex (when both solid and liquid areas were visible within the mass). Margins were graded as clear and regular or as blurred and irregular. The CD results were classified according to criteria developed for the study of breast lesions (4, 6) with regard to: the presence or absence of flow signals; the number of vessels entering the mass (considered as vascular hila ) from the periphery; the distribution of color signals within the nodule (peripheral versus central); the shape of their course (linear, tortuous, or visible only as small spots); and the diameter of the vessels (smaller or larger than 2 mm). The vascular diameter was measured on the CD images. Confu-mation of the diagnosis was provided by Fig. 4. AV malformation of the thigh. a) US shows irregular muscular mass with mixed echo texture. b) CD demontrates large tortuous vessels with whirling flow, typical of AV malformation. c) MR angiography confirms the diagnosis. 423

5 R. LAGALLA ET AL. Table 1 Final diagnoses. lesion location, and lesion size in 46 patients Final diagnosis Size, cm Patients, n Benign lesions Intramuscular lipoma Schwannoma AV malformation Myxoma Cavernous hemangioma Fibrolipoma Giant-cell tumor Infected Baker cyst Cystic hematoma Malignant lesions Schwannoma Synovial sarcoma Liposarcoma Melanoma Rhabdomyosarcoma Myxoid chondrosarcoma Myxoid liposarcoma Chondrosarcoma Sarcoma Metastatic carcinoma Muscular lymphoma Fibrosarcoma Fibrous histiocytoma Table 2 US findings in 46 patients US findingdpattern Benign, Malignant, n=26 n=20 H ypo-echoic Hyper-echoic Iso-echoic Homogeneous Heterogeneous Complex Regular margins Blurred margins Table 3 Color Doppler findings in 46 patients Color Doppler findings Benign, Malignant, n=26 n=20 Presence of flow Absence of flow or more vascular hila 2 17 Fewer than 3 vascular hila 5 2 Peripheral flow signals 7 16 Central flow signals 6 19 Linear vessels 5 12 Tortuous vessels 4 12 Spot flow signals 3 9 Small vessels 8 13 Large vessels 5 8 histology either at percutaneous biopsy (17 cases) or at surgery (29 cases). Results Histology showed that 20 lesions were malignant and 26 were benign (Table 1). The US findings are presented in Table 2 and the CD findings in Table 3. The retrospective correlation of the US findings with the final diagnosis was based on the following criteria for malignancy: the presence of irregular and blurred margins and of a heterogeneous textural pattern. This correlation resulted in 5 falsenegative and 13 false-positive US results: sensitivity 75%, specificity 50%; negative predictive value 72%, positive predictive value 53%; accuracy 6 l YO. CD showed flow signals in 31/46 lesions. Of the 26 benign lesions, 14 proved to be avascular. The remaining 12 benign lesions comprised: 1 with a spotty peripheral flow (Fig. 1); 6 with a few internal vessels that had a tortuous course (4 cases) or only spotty internal flow signals (2 cases); and 5 with vessels both at the periphery and at the center that had a course that was either linear (4 cases) or tortuous (2 cases). One malignant tumor had only a small artery at its periphery and no internal vessels (Fig. 2). The others proved to be hypervascular, mostly with both linear and irregular internal signals (Fig. 3). When the indication of malignancy was based on the presence of 3 or more vascular hila and of tortuous and irregular internal vessels, we obtained 3 false-negative and 2 false-positive US results: sensitivity 85%, specificity 92%; negative predictive value 88%, positive predictive value 89%; accuracy 89%. Resistive index (RI) values did not show a differentiation between benign and malignant lesions. Benign masses had a mean value of 0.58 (range ) while malignancies had a mean value of 0.75 (range ); the difference was not statistically significant. A wide variability in the RI was also noted within the same lesions, with values ranging from 0.53 to In patients with vascular lesions, the CD findings suggested the benign nature of the disease. AV malformations presented as irregular masses of mixed echo texture that contained large vessels with lowimpedance arterial waveforms (Fig. 4). The large diameter of the vessels helped to identify these lesions as vascular, even when they presented at US as heterogeneous nodules with irregular outer margins. The patient with angioma of the posterior surface of the arm had a small mass containing hypo-echoic internal areas with a fluid-fluid level. No internal 424

6 COLOR DOPPLER US OF SOFT-TISSUE MASSES flow could be demonstrated, and only large peripheral vessels could be seen; these findings suggested a benign lesion (Fig. 5). Discussion The growth of malignant tumors depends heavily on the blood supply provided by new vessels which develop under the stimulus of tumor angiogenic factors. These new vessels penetrate the lesion from its periphery and are typically of small caliber with thin walls and a relative paucity of muscular fibers in the walls. Their course and distribution within the tumor are chaotic and irregular, and they present multiple anastomoses and shunts (5, 14). The irregular course of these vessels can be identified at CD while their anomalous flow, with high velocity and/or low impedance, can be detected by means of waveform analysis (2, 4, 6, 10-18). Anomalous flow has also been described in malignant tumors arising from soft tissues (3, 13); these reports correspond to the results obtained in our series. When the CD results were included as additional parameters in the differential diagnosis between benign lesions and malignancies, 11 falsepositive and 2 false-negative cases were corrected, reaching a sensitivity of 85% and a specificity of 92% for the combined US examination. Not only the presence of vessels but also their morphology, as imaged by CD, proved useful in differentiating the nature of the lesions. In 5 of the 12 benign vascular nodules, either internal or peripheral vascular images of linear course were obtained, suggesting either the displacement of a preexisting vessel or the presence of a regular network of intratumoral vessels. Most malignancies had either irregular and tortuous vascular images or only small spots of color scattered throughout the tumor. These spots of color indicated either tumor infiltration of a pre-existing vessel or an irregular network of neoplastic vascularity. The malignancies were imaged for only short segments and only when a good CD angle was obtained, and they were then visible only as spots. Analysis of the vascular diameter was particularly useful for identifying the nature of the AV malformation. Doppler instruments are highly sensitive to the presence of flow and have the capability to identify the presence of vascular signals either as spectra or as colors, even at sites where US imaging is unable to demonstrate a vessel. In CD images, the vessel diameter can be measured. The diameter is related not only to the diameter itself but also to flow velocity and signal intensity. Different equipment or different settings on the same machine can therefore give wide variations in vascular display. At present, it does not seem possible to compare the vascular diameters obtained in different subjects by different types of US unit. This type of comparison would be possible only in the same US laboratory, using the same piece of equipment, and following standardized scanning parameters. Our series of patients was widely heterogeneous, presenting a large variety of diseases with lesions of different diameters. Lesions with different histologies can have different degrees of vascularization and it was impossible to make a separate analysis of the Doppler characteristics of the different lesions in our series because of the relatively low number of each type. In addition to histology, tumor size can also be considered a factor that influences the degree of vascularization in a soft-tissue mass. Theoretically, the larger the lesion, the higher its metabolic needs, and the larger its vascular supply, although areas of tumor necrosis can frequently be encountered in large tumors particularly at their centers; this is possibly due to the overgrowth of the mass beyond the metabolic capabilities of the vascular network. The evaluation of intratumoral flow waveforms did not help in the differential diagnosis owing to the wide range of results, which showed both high and low impedance signals in both benign and malignant lesions. Furthermore, both high and low resistance flow patterns were observed within the same mass in many cases. This may be related to the presence in each lesion of either type of tissue or of both viable and necrotic areas with different flow characteristics. Conclusion: Both CT and MR imaging are needed for an accurate pre-operative evaluation of space-occupying lesions of soft tissues, and US cannot be a substitute for them. However, US can confirm the presence of a suspected mass, locate it accurately, and provide clues as to its nature. The combined use of US imaging and CD can indicate a differential diagnosis and distinguish between benign and malignant lesions, thereby providing a means of guiding the management of the patient. REFERENCES 1. BERNARDINO M. E., JING B.-S., THOMAS J. L., LINDELL M. M. & Z~RNOZA J.: The extremity soft-tissue lesion. A comparative study of ultrasound, computed tomography and xeroradiography. Radiology 139 (1981), BURNS l? N., HALLIWELL M. & WELLS l? N. T.: Ultrasonic Doppler studies of the breast. Ultrasound Med. Biol. 8 (1982), CAMPANI R., BOTTINELLI 0. & CALLIADA E: Tessuti molli perischeletrici. In: Eco-color-Doppler, p Edited by M. Bazzocchi. APC-SIRM, Milan

7 R. LAGALLA ET AL. 4. COSGROVE D. O., BAMBER J. C., DAVEY J. B., MCKINNA J. A. & SINNETT H. D.: Color Doppler signals from breast tumors. Radiology 176 (1990), FOLKMAN J. & KLAGSBURN M.: Angiogenic factors. Science 235 (1987), GIUSEPPETTI G. M., BALDASSARRE S. & ARGALIA G.: Evaluation of breast nodules with echo colour Doppler sonography. Preliminary findings. Eur. Radiol. 4 (1994), HARCKE H. T., GRISSOM L. E. & FINKELSTEIN M. S.: Evaluation of the musculoskeletal system with sonography. AJR 150 (1988), HORSTMAN W G., MELSON G. L., MIDDLETON W. D. & AN- DRIOLE G. L.: Testicular tumors. Findings with color Doppler US. Radiology 185 (1992), KAPLAN P. A., MATAMOROS A. & ANDERSON J. C.: Sonography of the musculoskeletal system. AJR 155 (1990), KELLY I. M. G., LEES W. R. & RICKARDS D.: Prostate cancer and the role of color Doppler US. Radiology 189 (1993), KUIJPERS D. & JASPE R.: Renal masses differential diagnosis with pulsed Doppler. Radiology 170 (1989), MARTINOLI C., DERCHI L. E., SOLBIATI L., RIZZATTO G., SILVESTRI E. & GIANNONI M.: Color Doppler sonography of the salivary glands. AJR 163 (1994), MENANTAU B., LADAM-MARCUS Y, SERVAGENT P., GAUS- SERAND E, PILVEN 0. & MARCUS C.: Color-Doppler sonography of soft tissue tumors. Eur. Radiol. 5 (1995), S SCHOR A. M. & SCHOR S. L.: Tumor angiogenesis. J. Pathol. 141 (1983), SOLBIATI L., CIOFFI V. & BALLARATI E.: Ultrasonography of the neck. Radiol. Clin. North Am. 30 (1992), TAYLOR K. J. W., RAMOS I., CARTER D., MORSE S. S., SNOW- ER D. & FORTUNE K.: Correlation of Doppler US tumor signals with neovascular morphologic features. Radiology 166 (1988), TAYLOR K. J. W., RAMOS I., MORSE S. S., FORTUNE K. L., HAMMERS L. & TAYLOR C. R.: Focal liver masses. Differential diagnosis with pulsed Doppler US. Radiology 164 (1987), VAN CAMPENHOUT I. & PATRIQUIN H.: Malignant microvasculature in abdominal tumors in children. Detection with Doppler US. Radiology 183 (1992),