Tumor Vascularity of Breast Lesions: Potentials and Limits of Contrast-Enhanced Doppler Sonography

Transcription

1 Markus Stuhrmann 1 Ramon Aronius Michael Schietzel Received June 28, 1999; accepted after revision May 16, All authors: Röntgenabteilung, Gemeinschaftskrankenhaus Herdecke, Universität Witten-Herdecke, Beckweg 4, Herdecke, Germany. Address correspondence to M. Stuhrmann, Radiologische Gemeinschaftspraxis, Alter Markt 10, D Wuppertal, Germany. AJR 2000;175: X/00/ American Roentgen Ray Society Tumor Vascularity of Breast Lesions: Potentials and Limits of Contrast-Enhanced Doppler Sonography OBJECTIVE. We investigated improving the evaluation of benignity in breast lesions using Doppler sonography with galactose palmitic acid coated microbubbles. SUBJECTS AND METHODS. In 77 patients with 84 breast tumors scheduled for surgical tumor removal, color-coded duplex sonography was performed before and after administration of Levovist. Of the 77 patients, 25 with 28 lesions had been treated for prior breast carcinoma. The parameters investigated were the following: degree of enhancement, number of tumor vessels, time to maximum enhancement, and the pattern of vascular morphology and course. RESULTS. Findings in malignant tumors (n = 53) showed a greater number of vessels and a faster stronger enhancement after Levovist administration, whereas a definite partial overlap with results from benign tumors (n = 31) was found. The best distinction was produced by vascular morphology and course, with a sensitivity of 90% and a specificity of 81%. In 23 of the 25 patients who previously underwent surgery, a clear distinction was possible between a postoperative scar (n = 11) and a tumor recurrence (n = 17). CONCLUSION. Although administration of the contrast agent clearly improved evaluation of benign features on Doppler sonography, absolute certainty cannot be achieved. The feasibility of making an otherwise difficult distinction between a scar and tumor recurrence on sonography and mammography appears to be promising, but further studies are necessary. I ncreasingly, tumor vascularization, particularly of breast tumors, is becoming the focus of scientific interest in prognostic, diagnostic, and possibly therapeutic terms [1, 2]. It plays an important role in the primary identification of a nonspecific breast lesion and in the diagnosis of a recurrence with MR imaging and duplex sonography. Using these additional examinations avoids unnecessary negative biopsies for nonspecific breast findings on mammography. The results of color-coded duplex sonography are disappointing according to earlier studies; in particular, there was a poor differentiation of hypovascular tumors [3 7]. Until now duplex sonography played only a minor role in the evaluation of a breast mass. Levovist (Schering, Berlin, Germany) is a sonographic contrast agent consisting of galactose microparticles coated with palmitic acid. The preliminary studies of this method with relatively low numbers of patients and different examination parameters indicated a high sensitivity [8 10]. Reports distinguishing between a tumor recurrence and a postoperative scar were not previously available. The aim of our study was to investigate which examination parameter is best suited to the evaluation of benign features and whether a clear distinction can be made between malignant and benign tumors and between a tumor recurrence and a postoperative scar. Subjects and Methods We conducted a prospective study of 84 breast tumors in 77 women who were years old (mean age, 47 years). In all patients, a surgical tumor removal was planned because of a suspicious finding on physical examination or mammography. Twenty-five patients with a history of ipsilateral breast carcinoma had 28 lesions suggestive of tumor recurrence. Of these 25, 19 patients had undergone lumpectomy and six had undergone mastectomy. The period between the first operation and the current examination ranged from 4 months to 17 years (average time, 39 months). Twenty-seven of the 28 lesions were palpable as localized masses. A coarse diffuse thickening of the scar was found in one patient after mastectomy. Mammography was available in 13 patients. Mammography in six patients showed suspicious malignant findings, four mammographic findings showed inconspicuous benign results, and three mammographic findings were indeterminate. Detailed information about the AJR:175, December

2 Stuhrmann et al. intervals between imaging studies and completion of breast therapy and about size, location, and histologic types of recurrences is given in Tables 1 and 2. Tumors ranged from 0.5 to 3.0 cm (mean size, 1.5 ± 0.4 cm). A precise histopathologic study was performed in all except four patients. In one patient with an invasive ductal carcinoma of the left breast and two TABLE 1 No. of Patients Findings in Patients with Tumor Recurrences Time Interval a (months) Location Size (cm) Carcinoma Degree of Enhanc. b Vascular Pattern c In scar 2.0 Invasive ductal 3 3 a + b Distant from scar 1.6 Invasive ductal 2 3 a Near scar 1.1 Tubular 3 3 a + b Distant from scar 1.6 Invasive ductal 2 3 a + b In scar 1.5 Invasive ductal 2 3 a In scar 2.8 Invasive ductal 2 3 a In scar 0.9 Invasive ductal In scar 1.0 Invasive ductal 2 3 a + b In scar 2.0 Invasive ductal 4 3 a + b In scar 0.9 Invasive lobular 4 3 a + b Distant from scar 1.4 Invasive lobular 2 3 a + b Near scar 1.4 Invasive ductal 3 3 a In scar 0.9 Invasive ductal 2 3 a + b In scar 1.3 Invasive ductal 4 3 a + b Near scar 1.6 Invasive ductal 4 3 a + b 76 9 In scar 1.3 Invasive ductal 4 3 a + b Near scar 1.1 Invasive lobular 4 3 a + b Note. Enhanc. = enhancement a Between imaging studies and completion of breast therapy. b 0 = no enhancement, 1 = just visible enhancement, 2 = low enhancement, 3 = medium enhancement, 4 = high increase in color Doppler signal intensity. c 1 = no vascularization, 2a = benign with same-size monomorphic vessels, 2b = benign with vessels running around the edge of the tumor, 3a = malignant with irregularly configured vessels varying in size, occasional fusing in color grading to pools, 3b = malignant type with irregularly running vessels that approach or penetrate the tumor (see Fig. 1). TABLE 2 No. of Patients Findings in Patients with Scars Time Interval a (months) Size (cm) Degree of Enhancement b Vascular Pattern c a a + b a a a a + b a Between imaging studies and completion of breast therapy. b 0 = no enhancement, 1 = just visible enhancement, 2 = low enhancement, 3 = medium enhancement. c 1 = no vascularization, 2a = benign with same-size monomorphic vessels, 2b = benign with vessels running around the edge of the tumor, 3a = malignant with irregularly configured vessels varying in size, occasional fusing in color grading to pools, 3b = malignant type with irregularly running vessels that approach or penetrate the tumor (see Fig. 1). sonographically evident lesions in the mastopathically altered right breast, only an aspiration cytology was performed. In three patients with a postoperative scar, no surgical excision was performed; follow-ups at 9, 13, and 25 months revealed no change in the findings. We conducted all examinations the day before surgery using an Ultrasound 9 HDI scanner (Advanced Technology Laboratories, Bothell, WA) with a 5-10 MHz linear transducer (frequency of color Doppler examination was 6 MHz). With B- mode sonography, the tumor was first located and imaged. Then imaging with color-coded duplex sonography was performed on a representative section of the tumor. Both the tumor itself and the surrounding tissue were included. The Doppler velocity scale was set at 0.02 m/sec, and to avoid artifact formation, we used a 65% gain. The parameters were not changed during the examination. We were careful not to exert any pressure on the transducer because small tumor vessels can be easily compressed. After the initial examination, Levovist was administered as a 4-g bolus of 300 mg/ml given over 15 sec. The entire examination was recorded with a video recorder; the contrast agent study was continuously depicted for the selected tumor section. Evaluation was based on the video recording. The degree of enhancement was studied and, as proposed by Kedar et al. [8], we used the following scale of five grades: 0, no enhancement; 1, just visible enhancement; 2, low enhancement; 3, medium enhancement; and 4, high increase in color Doppler signal intensity. Furthermore, the number of tumor vessels in total and per area of section were examined. We also evaluated the lag time between the start of injection and the maximum detectable enhancement. Further, the basic pattern of detectable tumor vascularization was analyzed and classified on the basis of the morphology, size, and course of the vessels (Fig. 1). Analysis was done retrospectively in the first 25 patients for whom the histologic results were already known. In subsequent patients evaluation of the vascularization was done prospectively according to the criteria seen in Figure 1. After we established the examination parameters, the sonographic examinations were evaluated independently by two researchers, and a consensus was reached in cases of discrepancy. Statistical comparison of the data was done just on examinations performed subsequent to the first 25 cases, which were used to develop examination parameters. We used the Student s t test (2nd and 3rd parameters) and Fisher s exact test (1st and 4th parameters). Results Of the 53 histopathologically proven cancers, 35 were invasive ductal carcinoma and 10, invasive lobular carcinoma. Furthermore, three tubular, one mucinous, one medullary carcinoma, and three cases of ductal carcinoma in situ were found. In the group of benign lesions (n = 31), the diagnosis included seven fibroadenomas, 12 fibrocystic changes, one abscess, and 11 postoperative scars. Six patients had multiple findings: four with a multicentric invasive ductal carcinoma had two tumor nodes in different breast quadrants, one patient with previous surgery had a postoperative scar and a tumor recurrence (Fig. 2), and finally the previously described patient 1586 AJR:175, December 2000

3 Contrast-Enhanced Doppler Sonography of Breast Lesions had a breast cancer on the left and two circumscribed mastopathic changes on the right. Before administering the contrast agent, we found visible vessels in 44 of 53 carcinomas and eight of 31 benign lesions. After the application of Levovist, vascularization was shown in almost all malignant tumors and in 21 benign lesions. No vascularization was seen in five postoperative scars, five circumscribed mastopathic lesions, and one diffuse tumor recurrence. Figure 3 shows the typical vascularization pattern of a carcinoma. The results of evaluation of the parameters are given in Tables 3 5. Although after Levovist administration carcinomas show a clear tendency toward faster and stronger enhancement with a greater number of vessels evident, some obvious overlap was found in the results. If the selected limit value (border value) is a degree of enhancement less than or equal to 2 and a relative number of tumor vessels less than or equal to 1.5/cm 2, then the result with respect to the prospectively reviewed cases (n = 59) is a sensitivity of 63% and 84% and a specificity of 90% and 71%, respectively. The pattern of vessel morphology and course (Table 5) afforded the best distinction between a malignant and a benign tumor. With regard to the prospectively reviewed patients, 34 of 38 carcinomas and 17 of 21 benign lesions were correctly judged and corresponded to a sensitivity of 90% and a specificity of 81%. In the group of the primarily diagnosed carcinomas, three false-negative results involved a medullary and two, invasive ductal carcinoma. Four benign lesions (1 abscess, 2 circumscribed lesions associated with a fibrocystic mastopathy, and a multiply septate fibroadenoma) showed false-positive results. Of the 28 lesions in patients who previously underwent surgery, 11 scars and 17 tumor recurrences were identified (Tables 1 and 2). In 10 postoperative scars only small single vessels could be seen after Levovist administration, if at all. A 63-year-old patient with excessive scar tissue showed an increased abnormal vascular- TABLE 3 Degree of Enhancement Finding No Malignant Benign Recurrence Scar Note. 0 = no enhancement, 1 = just visible enhancement, 2 = low enhancement, 3 = medium enhancement, 4 = high increase in color Doppler signal intensity. p value for malignant versus benign lesion: p = (calculated for the prospectively reviewed cases). Fig. 1. Drawings show basic patterns of tumor vascularization: type 1 (not shown) is avascular type with no obvious vascularization. Type 2 is benign type, with individual, roughly same size, monomorphic vessels (left) or vessels running around edge of tumor, which is regular course of vessels (right). Type 3 is malignant-type irregularly configured vessels varying in size, occasionally fusing in color grading to pools (left) or irregularly running vessels that approach or penetrate tumor (right). A C Fig year-old woman who underwent breast-sparing surgery 19 months earlier. Patient presented with suspected multifocal tumor recurrence because of two palpable lumps in scar from surgery and at separate location near scar. A and B, B-mode sonograms reveal two echo-poor lesions. Color-coded Doppler sonography (not shown) did not show any vessels. C, Contrast-enhanced sonogram shows that one lesion still has no obvious vascularization. Histopathology (not shown) confirmed scar. D, Contrast-enhanced sonogram of other lesion shows pathologic tumor vascularization with irregularly configured vessels varying in size. Histopathology (not shown) confirmed tumor recurrence (invasive ductal carcinoma). B D AJR:175, December

4 Stuhrmann et al. ization pattern. A 75-year-old patient with a diffuse tumor infiltration of the chest after total mastectomy showed no vascularization after Levovist administration. If the time between breast therapy and investigation is compared with the vascularization detectable after Levovist administration, a tendency is shown with respect to reduced circulation in the scar in relation to its age (Table 2). Discussion Previous studies using color-coded duplex sonography in the diagnosis of breast tumors revealed that the evaluation criteria of number of vessels, blood-flow velocity, and intratumoral vessel resistance did not permit clear differentiation between malignant and benign tumors [3 7]. Malignant tumors had a tendency to show increased vascularization, but this overlapped with the results from avascular, malignant, and hypervascular benign tumors, especially with fibroadenomas [3, 6]. The initial reports of using contrast agents with sonography to reveal the vascularization of breast tumors, primarily in small patient groups, are now available [8 10]. An exact Fig year-old woman with invasive ductal carcinoma. Doppler sonogram shows malignant type 3 vascularization pattern, with deep vessels penetrating tumor and following irregular course. TABLE 4 Finding comparison with our results is not possible because of the different methodologies. We could not confirm the sensitivity of 100% noted in previous studies [8 10]. Our results also revealed that the administration of Levovist improved the evaluation of vascularization of a breast tumor. Regarding the possibility of distinguishing between benign and malignant tumors, we found clear differences in the parameters investigated. Malignant tumors tended to be associated with a greater number of vessels and more intense enhancement after contrast agent administration, but some results clearly overlapped. The best diagnostic criterion proved to be the pattern of tumor perfusion. The vessel architecture of malignant tumors is known from pathologic studies. Tumor vessels are particularly characterized by caliber fluctuations, irregular course, formation of sinusoids, and arteriovenous shunts [11]. One focus of our simplified classification of the vascular pattern is the morphology and size of detectable vascularization. In benign processes we observed mostly individual vessels that were similar in size. Vascularization of different sizes resulted in the most malignant tumors, some of which were confluent, probably as a result of the presence of sinusoids and arteriovenous shunts. The course of vascularization formed the other focal point in the differentiation. In the cases of benign tumors, the vessels were gently curved and coursed along the margin of the masses, especially in fibroadenomas, similar to those observed by Kedar et al. [8]. Conversely, malignant tumors were chiefly distinguished by irregular or tortuous courses in which vessels penetrated the tumor. We did not evaluate the peripheral and central vascularization as a point for differentiation of benign from malignant disease. Kedar et al. [8] also indicated in their report that the tortuosity of the vessels and the number of intervascular shunts are important criteria in distinguishing malignant from benign tumors. We believe that these criteria are difficult to quantify; furthermore, neighboring vessels can be misinterpreted as intervascular Number of Tumor Vessels and Time to Peak Enhancement Unenhanced Enhanced Total Per cm 2 Note. Values are expressed as mean ± SD. p value was calculated for the prospectively reviewed cases. Time (sec) Malignant (n = 53) 1.8 ± ± ± ± 13 Benign (n = 31) 0.4 ± ± ± ± 18 p value Recurrence (n = 17) 1.9 ± ± ± ± 20 Scar (n = 11) 0.2 ± ± ± ± 15 shunts in the color-coded sonography. We propose that classification of the vascularization according to its underlying pattern is more practical in clinical practice. As our preliminary results indicate, examination with Doppler sonography after the administration of Levovist is associated with limitations in its specificity. In particular, inflammatory and mastopathic alterations can suggest a malignant-type vascularization pattern. Also, the vascularization pattern of a multiply septate fibroadenoma can resemble a malignant result. Unlike Kedar et al. [8], we found no significant difference in the time course of enhancement between malignant and benign tumors. One possible explanation could be that this parameter depends not just on the tumor s vascularization, but also on the patient s circulation parameters and on the rate of injection of the contrast agent. An interesting application arises from the difficult distinction between a postoperative scar and a tumor recurrence both clinically and on mammography. B-mode sonography does not permit a clear distinction between a scar and a tumor recurrence in patients with a similar morphology. Our preliminary results showed that Levovist administration improved the certainty of diagnosis. Previously, MR imaging was frequently performed as a noninvasive aid to diagnosis with varying success [12 14]. Contrast enhancement was the differentiating criterion here: at least 18 months after completion of the therapy, the scar tissue, in contrast to the tumor recurrence, usually showed no enhancement. Our study, taking into consideration the small number of patients, also showed the tendency of a reduced vascularization with the increasing age of the scar. The enhancement of scarring is less after long periods of time, when the scar is mature, compared with relatively new TABLE 5 Basic Patterns of Tumor Vascularization Finding No. Type 1 Type 2 Type 3 Malignant Benign Fibroadenoma Fibrocystic Abscess 1 1 Recurrence Scar Note. Classification scheme (see Fig. 1): type 1, avascular with no obvious vascularization; type 2, benign; type 3, malignant. p value for malignant versus benign lesion: (calculated for the prospectively reviewed cases) AJR:175, December 2000

5 Contrast-Enhanced Doppler Sonography of Breast Lesions scars. Our preliminary results also showed that proof or preclusion of vascularization is not enough, especially in the short period after breast therapy. Individual vessels are visible even in scars after administration of Levovist. Evaluations with respect to the degree of enhancement and the basic pattern of the tumor vascularization are necessary to enable further differentiation. In our study we observed a false-positive and a false-negative finding. No increase in tumor vascularization could be detected in the older patient with a diffuse chest wall recurrence. The findings in a patient with a localized coarse scar that developed from a postoperative hematoma showed increased vascularization with a pathologically classified basic pattern. Cases of similarly vascularized granulomas with a strong contrast enhancement were also found on MR imaging [15]. An exact histology was not available, however, in our patient. Our study and this method have certain limitations: first, only those lesions can be investigated that were visualized with B-mode sonography. This methodology excludes, therefore, patients with microcalcifications. In addition, in contrast to MR imaging, it is not possible to visualize the entire breast at once. Another limitation of this methodology is the uncertainty when determining the number of vessels: vessels lying close together can be visualized as one vessel because of fusing of the color pixels. The converse is also possible, that a tortuous vessel running through several sections could be counted twice. Another limitation involves the obvious subjectivity of some examination parameters. Because the best parameter in the differential diagnosis seems to be the classification by vascularization pattern, it is doubtful that extensive quantitative measurements [16, 17] would lead to an improved result. Finally, a wide variety of histopathologic results was seen in our patient group, whereas the number of patients was too small for a proper evaluation of the method. In summary, our preliminary results indicate that the use of Levovist improved the evaluation of benignity with Doppler sonography. Unlike the findings in earlier reports, however, we do not think that the evaluation will ever achieve absolute certainty. The characteristic pattern of vessel morphology and course has proven to be the best examination parameter. The otherwise difficult distinction between a postoperative scar and a tumor recurrence on sonography and mammography appears solvable with this method. Given the consequences associated with a diagnosis of breast carcinoma and the lack of an absolutely certain evaluation of a lesion on duplex sonography with a contrast agent, the current state of knowledge requires histopathologic confirmation. The exact value of the method (e.g., regarding the prognostic evaluation of a malignant lesion and the distinction between a scar and a recurrence) can be established only through further study. Acknowledgments We thank Christian Roefke, Thorsten Schwarz, Sina Delghandi, and Marianne Boesten for their help in this study and Michael Schulte for statistical analysis of the data. References 1. Passe TJ, Bluemke DA, Siegelmann SS. Tumor angiogenesis: tutorial on implications for imaging. Radiology 1997;203: Casparini G, Harris AL. Clinical importance of the determination of tumor angiogenesis in breast carcinoma: much more than a new prognostic tool. J Clin Oncol 1995;13: Cosgrove DO, Kedar RP, Bamber JC, et al. Breast diseases: color Doppler US in differential diagnosis. Radiology 1993;189: McNicholas MMJ, Mercer PM, Miller JC, Mc- Dermott EWM, O Higgins NJ, MacErlean DP. Color Doppler sonography in the evaluation of palpable breast masses. AJR 1993;161: Madjar H, Prömpeler HJ, Sauerbrei W, Wolfarth R, Pfleiderer A. Color Doppler flow criteria of breast lesions. Ultrasound Med Biol 1994;20: Fiedler V, Neubauer KD, Schneiders A, Herzig P. Ranking of color-coded duplex ultrasonography (CCDU) in the staging of breast tumors [in German]. Rofo Fortschr Geb Röntgenstr Neuen Bildgeb Verfahr 1996;165: Birdwell RL, Ikeda DM, Jeffrey SS, Jeffrey RB Jr. Preliminary experience with power Doppler imaging of solid breast masses. AJR 1997;169: Kedar RP, Cosgrove D, McCready VR, Bamber JC, Carter ER. Microbubble contrast agent for color Doppler US: effect on breast masses work in progress. Radiology 1996;198: Albrecht T, Patel N, Cosgrove DO, Jayaram V, Blomley MJK, Eckersley R. Enhancement of power Doppler signals from breast lesions with the ultrasound contrast agent EchoGen emulsion: subjective and quantitative assessment. Acad Radiol 1998;5[suppl 1]:S195 S Schröder RJ, Hadijuana J, Hidajat N, et al. Colorcoded signal-enhanced duplex ultrasonography of space-occupying intramammary processes [in German]. Rofo Fortschr Geb Rontgenstrahlen Neaen Bildgeb Verfahr 1998;168: Shubik P. Vascularization of tumors: a review. J Cancer Res Clin Oncol 1982;103: Cohen EK, Leonhardt CM, Shumak RS, et al. Magnetic resonance imaging in potential postsurgical recurrence of breast cancer: pitfalls and limitations. Can Assoc Radiol J 1996;47: Mumtaz H, Davidson T, Hall-Craggs MA, et al. Comparison of magnetic resonance imaging and conventional triple assessment in locally recurrent breast cancer. Br J Surg 1997;84: Rieber A, Merkle E, Zeitler H, et al. Value of MR mammography in the detection and exclusion of recurrent breast carcinoma. J Comput Assist Tomogr 1997;21: Heinig A, Heywang-Köbrunner SH, Viehweg P, et al. Value of contrast medium magnetic resonance tomography of the breast in breast reconstruction with implant [in German]. Radiologe 1997;37: Kedar RP, Cosgrove DO, Bamber JC, Bell DS. Automated quantification of color Doppler signals: a preliminary study in breast tumors. Radiology 1995;197: Huber S, Helbich T, Kettenbach J, Dock W, Zuna I, Delorme S. Effects of a microbubble contrast agent on breast tumors: computer-assisted quantitative assessment with color Doppler US early experience. Radiology 1998;208: AJR:175, December