Differentiating Benign and Malignant Breast Lesions with T2*-Weighted First Pass Perfusion Imaging

Transcription

1 Acta Radiologica ISSN: (Print) (Online) Journal homepage: Differentiating Benign and Malignant Breast Lesions with T2*-Weighted First Pass Perfusion Imaging K. A. Kvistad, S. Lundgren, H. E. Fjøne, E. Smenes, H.-B. Smethurst & O. Haraldseth To cite this article: K. A. Kvistad, S. Lundgren, H. E. Fjøne, E. Smenes, H.-B. Smethurst & O. Haraldseth (1999) Differentiating Benign and Malignant Breast Lesions with T2*-Weighted First Pass Perfusion Imaging, Acta Radiologica, 40:1, To link to this article: Published online: 04 Jan Submit your article to this journal Article views: 150 Full Terms & Conditions of access and use can be found at

2 Acto Rudiologico 40 (1999) Printed in Dmmurk - All rights reserved Copyright 0 Actu Rudiologica 1999 ACTA RA DI OLO G I C A ISSN DIFFERENTIATING BENIGN AND MALIGNANT BREAST LESIONS WITH T2*-WEIGHTED FIRST PASS PERFUSION IMAGING K. A. KVISTAD', S. LUNDGREN', H. E. FJ0SNE3, E. SMENES', H.-B. smethurst4 and 0. HARALDSETH' 'MR-Center, 'Department of Oncology, 3Department of Surgery, 4Department of Pathology, University Hospital, Trondheim, Norway. Abstract Purpose: Invasive breast carcinomas and fibroadenomas are often difficult to differentiate in dynamic contrast-enhanced TI-weighted MR imaging of the breast, because both tumors can enhance strongly after contrast injection. The purpose of this study was to evaluate whether the addition of T2*-weighted first pass perfusion imaging can increase the differentiation of malignant from benign lesions. Material and Methods: Nine patients with invasive carcinomas and 10 patients with contrast enhancing fibroadenomas were examined by a dynamic contrast-enhanced T1-weighted 3D sequence immediately followed by a single slice T2*-weighted first pass perfusion sequence positioned in the contrast-enhancing lesion. Results: The carcinomas and the fibroadenomas were impossible to differentiate based on the contrast enhancement characteristics in the T1-weighted sequence. The signal loss in the T2*-weighted perfusion sequence was significantly stronger in the carcinomas than in the Conclusion: Addition of a T2*-weighted first pass perfusion sequence with a high temporal resolution can probably increase the differentiation of fibroadenomas from invasive carcinomas in contrast-enhanced MR imaging of the breast. Key words: Breast neoplasms, MR; contrast enhancement, perfusion imaging. Correspondence: Kjell Arne Kvistad, MR-Center, Medical Division, University Hospital, N-7006 Trondheim, Norway. FAX Accepted for publication 8 June One of the major problems in dynamic contrastenhanced T1-weighted MR imaging of the breast is that contrast enhancement is found not only in carcinomas, but also frequently in non-malignant lesions (3, 15, 16), and even in normal breast parenchyma (21). Fibroadenomas are the most commonly occurring benign breast tumors, and many fibroadenomas demonstrate strong and rapid contrast enhancement where the signal intensity-versus-time curves considerably overlap those of breast carcinomas (10, 12, 26). The contrast enhancement in fibroadenomas is most pronounced in premenopausal patients, and is correlated to the myxoid stromal component of the tumor (10, 26). Several larger studies have re- ported a high sensitivity (92,95%), but a rather low specificity (51-72%), for the detection of breast cancer with dynamic contrast-enhanced T1- weighted MR imaging (4, 11, 22). This poor specificity, which partly can be attributed to contrastenhancing fibroadenomas, is one of the main reasons for the considerable uncertainty that still exists about the indications for MR imaging of the breast (19). While contrast enhancement in TI -weighted imaging depends on several factors such as the precontrast TI value of the tumor, tumor perfusion, tumor interstitial matrix, capillary density and capillary permeability (14,20), rapid T2*-weighted first-pass perfusion imaging responds more specific 45

3 K. A. KVISTAD ET AL. to changes in microvascular perfusion (23, 27). While passing through the capillary network, a bolus injection of a paramagnetic contrast medium produces local field inhomogenities between the intra- and extravascular compartments, and this susceptibility difference can be detected by a series of rapid T2*-weighted gradient echo images (23). T2*-weighted first pass perfusion imaging has been implemented in studies of ischemic brain disease and in the grading of central nervous system tumors (1, 25). Recently, T2*-weighted perfusion MR imaging has been proposed as a method to help differentiate benign from malignant breast tumors (20). The aim of the present study was to evaluate if a protocol of combined T1- and T2*- weighted imaging in the same session could differentiate the contrast-enhancing fibroadenomas from invasive carcinomas, and thus be promising for improving the specificity of contrast-enhanced MR imaging of the breast. Material and Methods The study was approved by the local ethics committee, and all patients gave their informed consent before inclusion in the study. Nine patients with invasive breast carcinomas and 10 patients with contrast-enhancing fibroadenomas were studied. All patients with carcinomas were treated with mastectomy or breast-conserving surgery. At the histology examination, 7 invasive ductal carcinomas, 1 mucoid carcinoma and 1 carcinoma with neuroendocrine differentiation were found. Eight of the fibroadenomas were confirmed by open biopsy or percutaneous fine-needle biopsy, while biopsy was not perfomed in 2 of the patients with fibroadenomas because of young patient age combined with typical clinical, X-ray mammography and ultrasonography findings. The mean age was 57.6 years in patients with invasive carcinomas, and 38.4 years in patients with fibroadenomas. Only fibroadenomas with a signal intensity increase of at least 85% during the first 90 s after contrast injection were included in the study. All carcinomas enhanced above this threshold. The mean size of the carcinomas was 24 mm (range 8-36 mm), and the mean size of the fibroadenomas was 22 mm (range mm). All images were acquired at 1.5 T (Picker Edge EPI 11), using a commercially available dedicated receiveonly double breast coil (Picker). The patients were examined in the prone position, and the breasts were gently cushioned to reduce patient motion. Dynamic, contrast-enhanced TI -weighted images were obtained using a 3D radio frequency spoiled gradient echo sequence (RF-FAST) with TR/TE/ flip angle 9 ms/3.8 ms/30", 1 excitation, FOV 200 mm, acquisition matrix 128x256, and with a temporal resolution of 50 s. This 3D sequence covered the whole breast, and was performed with 40 partitions, corresponding to an effective slice thickness of 3 mm. The 3D sequence was repeated continuously 9 times, and at the start of the third set of images, a bolus injection of 0.1 mmol/kg b.w. gadodiamide (Omniscan, Nycomed Amersham) was administered through a needle in the antecubital vein, followed by a 20 ml flushing bolus of isotonic saline solution. The total injection time was 10 s. A region-of-interest (ROI) was positioned in the part of the tumor that enhanced most by visual inspection, and time-versus-signal intensity curves were obtained. As an approximation to the relative enhancement velocity, the signal intensity increase within the first 90 s after the end of the gadodiamide/saline injection was calculated as a percentage of the precontrast signal intensity. Immediately following the dynamic contrast enhanced T1 -weighted sequence, a single slice in the most enhancing part of the tumor was chosen for dynamic T2*-weighted first pass perfusion imaging. In this study, a gradient echo sequence with TR/TE/flip angle 54 ms/35 ms/lo", 1 excitation, FOV 200 mm, acquisition matrix 92x256, section thickness 5 mm, and with a temporal resolution of 4.8 s was used. The sequence was repeated 60 times. After the first 10 repetitions, a rapid (<3 s) bolus injection of 0.1 mmol/kg b.w. gadodiamide was administered. A ROI was positioned in the part of the tumor that enhanced most on the T1- weighted images, and time-versus-signal intensity curves were obtained. The maximal signal intensity loss within the first 30 s after bolus injection was calculated as a percentage of the baseline signal intensity values from the ten precontrast images. The time between the contrast agent injection for the T1-weighted sequence and the contrast agent injection for the T2*-weighted sequence was approximately 15 min. The relative enhancement velocity in the dynamic T1-weighted sequence and the maximal signal intensity loss in the T2*-weighted first pass perfusion sequence were compared statistically between the group of carcinomas and the group of fibroadenomas. Because the variance for both parameters was different in the two groups, the nonparametric Mann-Whitney test was used for statistical comparison between the two groups of patients. Pearson's correlation coefficient was calculated for carcinomas and fibroadenomas separately to correlate signal intensity changes in both imaging sequences to tumor size. 46

4 T2* PERFUSION IMAGING IN BREAST LESIONS Fig. 1. Invasive ductal carcinoma in a 72-year-old woman. a) Signal intensity-versus-time curve of the shown ROI in dynamic contrast-enhanced TI-weighted imaging. The signal increase during the first 90 s after contrast injection was 214% compared to the precontrast value. b) Signal intensity-versus-time curve of the shown ROI in T2*-weighted first pass perfusion imaging. The signal intensity loss during the first 30 s after contrast injection was 41% compared to the baseline value. Fig. 2. Fibroadenoma (+) in a 38-year-old woman. a) Signal intensity-versus-time curve of the shown ROI in dynamic contrast-enhanced T1-weighted imaging. The signal increase during the first 90 s after contrast injection was 457% compared to the precontrast value. b) Signal intensity-versus-time curve of the shown ROI in T2*-weighted first pass perfusion imaging. The signal intensity loss during the first 30 s after contrast injection value was close to the noise level, and was estimated to be 6% compared to the baseline value. Results Examples of T1- and T2*-weighted images with time-signal intensity curves of a carcinoma (Fig. 1) and a fibroadenoma (Fig. 2) are shown. In the T1- weighted dynamic contrast-enhanced 3D sequence, the mean signal intensity increase in carcinomas (n=9) measured 90 s after the end of contrast injection was 213% (95% confidence interval (CI) for 47

5 K. A. KVISTAD ET AL. mean 174%; 251%) compared to the precontrast signal intensity values (Fig. 3a). The mean signal intensity increase in the fibroadenomas (n= 10) was 190% (95% CI for mean 114%; 265%) (Fig. 3a). There was no statistically significant difference in the signal intensity increase in carcinomas and fibroadenomas (p=0.121). In the T2*-weighted first pass perfusion sequence, the mean signal intensity loss in the carcinomas during the first 30 s after contrast injection was 21.3% (95% CI for mean 13.5%; 29.2%) compared to the baseline value (Fig. 3b). The mean signal intensity loss in fibroadenomas was 6.1% (95% CI for mean 3.2%; 8.2%). The difference be- Signal intensity increase (Yo) (mean + 95% C I) 8"", I I I a Carcinomas Fibroadenomas Signal intensity loss (%) (mean + 95% CI) 104 O f I I 1 b Carcinomas Fibroadenomas Fig. 3. a) Mean values with 95% CI of the signal intensity increase as a percentage of the precontrast value in the dynamic T1-weighted sequence. There is considerable overlap between the signal intensity increase after contrast injection in carcinomas (n=9) and fibroadenomas (n=10). b) Mean values with 95% CI of the maximal signal intensity loss as a percentage of the baseline value in the T2*-weighted first pass perfusion sequence. A highly significant difference between carcinomas and fibroadenomas is seen (p=0.0004). tween the signal intensity reduction in carcinomas and in fibroadenomas was statistically significant (p = ). There was no correlation between tumor size and signal intensity increase in the T1-weighted sequence (r=0.072, p=0.86 for carcinomas, and r= -0.31, p=0.42 for fibroadenomas), nor between tumor size and signal loss in the T2*-weighted sequence (r=0.61, p=0.084 for carcinomas, and r= 0.26, p=0.49 in fibroadenomas). The typical time duration from the bolus contrast injection to a signal intensity change could be observed in the T2*-weighted sequence was 20 s. In all of the carcinomas except one, the signal intensity loss was very rapid, typically reaching its maximum negative value in the next image in the dynamic sequence of 4.8 s time resolution. After the maximum signal loss was reached, a slow increase of the signal intensity towards baseline was observed, but the baseline value was not reached during the duration of the image acquisition (4 min 30 s after contrast injection) in any of the carcinomas. Discussion A considerable overlap in the signal intensity increase in carcinomas and fibroadenomas after contrast injection was found in the dynamic contrastenhanced T1-weighted sequence (Fig. 3a). This is in accordance with previous studies (7, 12). Since most invasive carcinomas enhance contrast rapidly, the velocity of the signal intensity increase in the tumor after contrast agent injection is the most important diagnostic element in the evaluation of dynamic contrast-enhanced MR imaging of the breast. The velocity of the tumor contrast enhancement has been evaluated in different ways. Some authors use an 80-90% signal intensity increase compared to the precontrast values during the first minute after contrast injection as a threshold for malignancy (13, 18, 21). In other studies, a more qualitative estimation of the contrast enhancement in a lesion during the first 1.5 min after contrast injection has been used for diagnosis of malignancy (1 1). However, all the fibroadenomas in the present study would have been considered suspicious of malignancy judged on the signal intensity increase in the T1-weighted sequence alone, regardless of the method used for evaluation of contrast enhancement velocities. The addition of morphologic criteria like the presence of smooth, lobulated contours and internal septa, can to some degree increase the differentiation between carcinomas and contrast-enhancing fibroadenomas (6, 12). 48

6 T2* PERFUSION IMAGING IN BREAST LESIONS The T2*-weighted sequence used in this study was a single slice sequence with a high temporal resolution and a low image resolution. The positioning of the single slice was dependent on the detection of a contrast-enhancing lesion in the TIweighted sequence, which covered the entire breast. In the T2*-weighted sequence, the carcinomas demonstrated a significantly more pronounced signal intensity reduction after contrast injection than the fibroadenomas. There was only a small overlap between carcinomas and contrast-enhancing fibroadenomas in this selected material (Figs. 3b and 4). These results are in accordance with the findings of KUHL et al. (20), but in the present material the signal intensity loss found in the carcinomas with the T2*-weighted sequence was less pronounced. This may be explained by the lower dose of contrast agent used in our study (0.1 mmol/kg b.w. compared to 0.2 mmol/kg b.w.). Expansion of a carcinoma to a size more than 1-2 mm depends on the creation of new vessels by angiogenic factors produced by the cancer cells, a process called tumor angiogenesis (8, 9). Dynamic T2*-weighted imaging is considered to be well suited for in vivo evaluation of tumor angiogenesis since the degree of signal loss depends on the fractional volume of the intravascular space within the tumor, and on the concentration of an injected contrast agent within the blood (25). Solid malignant tumors are usually highly vascularized. Not only the number, but also the size of the vessels have been found to increase with increased tumor angiogenesis (28). This may be a possible explanation for the increased capillary perfusion in carcinomas compared to fibroadenomas indicated by the difference in signal loss in the T2*-weighted sequence. In the T2*-weighted sequence, the carcinoma with a signal intensity loss of only 9% in the T2*- weighted sequence demonstrated a high amount of sclerotic connective tissue on histological examination. In most of the fibroadenomas, the signal intensity loss measured after contrast injection was so small that it was difficult to discern from image noise, while in one patient, who was receiving hormone substitution therapy, there was a more pronounced signal loss of 11%. Also in TI-weighted imaging a correlation between the velocity of contrast enhancement and microvessel density has been found (5, 9, 17). The much lower time resolution in the T1-weighted sequence used in the present study (50 s) compared to the single slice T2*-weighted sequence (4.8 s) may have contributed to the difference in the ability to distinguish between malignant and benign lesions. Rapid single slice TI -weighted imaging Signal intensity increase ("/.) o Fibroadenomas 0 Carcinomas I 01 I, 7, I,,, Signal intensity loss ("lo) Fig. 4. Scatterplot of the correlation between the signal intensity increase in dynamic T1-weighted contrast-enhanced imaging and signal intensity loss in T2*-weighted first pass perfusion imaging for carcinomas and fibroadenomas. with a time resolution of 2.3 s has been reported to differentiate between malignant and benign lesions, based on the time of contrast appearance in the tumor, with a specificity of 86% (2). This specificity is better than in most studies where multislice T1 -weighted image sequences, which cover the whole of one or both breasts, are used. Coverage of one or both breasts with thin slices is not possible with a time resolution lower than s without the use of echo-planar imaging (EPI), and EPI of the breast is problematic because of the high fat content of the breast. The great disadvantage of a single slice Tlweighted sequence is the difficulty in finding the exact location of the tumor for slice positioning. Often the tumor is impossible to distinguish from surrounding glandular tissue on the precontrast images, and T1 -weighted contrast-enhanced images that cover the entire breast have to be obtained. Wash-out of contrast agent from the tumor is usually slow, and the patient will have to be scheduled for a second examination with the rapid single slice technique at a later time. In our experience, it is difficult to exactly locate the tumor on the second examination because an identical positioning of the breast between two examinations is hard to achieve. However, in the T2*-weighted first pass perfusion technique used in the present study, signal intensity losses could be identified even though contrast medium was present in the tumor from the TI-weighted sequence performed 15 min prior to the perfusion sequence. This offers the practical opportunity to first perform a dynamic contrastenhanced TI-weighted sequence that covers the entire breast for the identification and localisation of 49

7 K. A. KVISTAD ET AL. contrast-enhancing lesions, immediately followed by a rapid T2*-weighted single slice first pass perfusion sequence located in the contrast-enhancing tumor for better differentiation between benign and malignant lesions. In dynamic T2*-weighted brain studies an intact blood-brain barrier causes the commonly used small molecular MR contrast agents to behave as true intravascular agents and the signal intensityversus-time curve will return to values close to the baseline after the first passage of the contrast agent bolus through the capillaries (1, 25). Under these circumstances, the signal intensity-versus-time curve has been used to quantify the regional blood volume and the mean transit time (I, 25). In other tissues like the breast, the contrast agent will leak through the capillary endothelium during the first passage and make quantification of tumor blood volume impossible. True intravascular contrast agents ( blood pool agents ) are under development, but not yet in clinical use (24). These novel agents may increase the ability to quantify flow and perfusion in organs other than the brain. Conclusion: Invasive breast carcinomas and contrast-enhancing fibroadenomas demonstrate a great overlap in their signal intensity increase after contrast injection in dynamic TI-weighted MR imaging. Our results indicate that these tumors can be better differentiated with a single slice T2*- weighted first pass perfusion sequence with a high time resolution. This offers the practical opportunity to first perform a dynamic TI-weighted sequence that covers the entire breast for identification of a possible contrast-enhancing lesion, immediately followed by a single slice T2*-weighted sequence positioned in the lesion for better differentiation between benign and malignant tumors. ACKNOWLEDGMENT This study was supported by a grant from the Norwegian Cancer Society, Oslo, Norway. REFERENCES 1. ARONEN H. J., GAZIT I. E., LOUIS D. N. et al.: Cerebral blood volume maps of gliomas. Comparison with tumor grade and histologic findings. Radiology 191 (1994), BOETES C., BARENTSZ J. O., Mus R. D. et al.: MR characterization of suspicious breast lesions with a gadolinium-enhanced turbo-flash subtraction technique. Radiology 193 (1994), Bo~k B., ASPELIN I?, BRONGE L., ISBERG B., PERBECK L. & ~RESS B.: Sensitivity and specificity of MR mammography with histopathological correlation in 250 breasts. Acta Radio]. 37 (1996), BONE B., PENTEK Z., PERBECK L. & VERESS B.: Diagnostic accuracy of mammography and contrast-enhanced MR imaging in 238 histologically verified breast lesions. Acta Radio]. 38 (1997), BUADU L. A,, MURAKAMI J., MURAYAMA S. et al.: Breast lesions. Correlation of contrast medium enhancement patterns on MR images with histopathologic findings and tumor angionesis. Radiology 200 (1996), FISCHER U. V., HEYDEN D. V., VOSSHENRICH R. et al.: Signal/ time relation of benign and malignant,jesions in dynamic 2D-MR imaging of the breast. ROFO 158 (1993), FLICKINGER E W., ALLISON J. D., SHERRY R. M. et al.: Differentiation of benign from malignant breast masses by time-intensity evaluation of contrast enhanced MRI. Magn. Reson. Imaging 11 (1993), FOLKMAN J. & SHING Y.: Angiogenesis. J. Biol. Chem. 267 (1992), FROUGE C., GUINEBRETIERE J. M., CONTESSO G. et al.: Correlation between contrast enhancement in dynamic magnetic resonance imaging of breast and tumor angiogenesis. Invest. Radio]. 29 (1994), GILLES R., GARNIER C., MEINGAN F! et al.: Fibroadenomas of the breast. Histopathological/dynamic contrast-enhanced MR correlation. Eur. Radio]. 5 (1995), GILLES R., MEUNIER M., LUCIDARME 0. et al.: Clustered breast calcifications. Evaluation by dynamic contrast-enhanced subtraction MRI. J. Comput. Assist. Tomogr. 20 (1996), GREENSTEIN OREL S., SCHNALL M. D., LIVOLSI V. A. & T~OUPIN R. H.: Suspicious breast lesions. MR imaging with radiologic-pathologic correlation. Radiology 190 (1994), GRIBBESTAD I. S., NILSEN G., FJ0SNE H. et al.: Contrastenhanced magnetic resonance imaging of the breast. Acta Oncol. 31 (1992), GRIEBEL J., MAYR N. A., DE VRIES A. et al.: Assessment of tumor microcirculation - a new role of dynamic contrast MR imaging. J. Magn. Reson. Imaging 7 (1997), HARMS S. E., FLAMIG D. P, HESLEY K. L. et al.: MR imaging of the breast with rotating delivery of excitation off resonance. Clinical experience with pathologic correlation. Radiology 187 (1993), HEYWANG S. H., WOLF A., PRUSS E. et al.: MR imaging of the breast with Gd-DTPA. Use and limitations. Radiology 171 (1989), HULKA C. A., SMITH B. L., SGROI D. C. et al.: Benign and malignant breast lesions. Differentiation with echo-planar MR imaging. Radiology 197 (1995), KAISER W. A.: MR imaging promises earlier breast cancer diagnosis. Diagn. Imaging Int. 8 (1992), KESSLER M.: Debate grows over breast imagings future. Diagn. Imaging Eur. 13 (1997), KUHL C. K., BIELING H., GIESEKE J. et al.: Breast neoplasms. T2* susceptibility-contrast, first pass perfusion MR imaging. Radiology 202 (1997), KUHL C. K., BIELING H. B., GIESEKE J. et al.: Healthy premenopausal breast parenchyma in dynamic contrast enhanced MR imaging of the breast. Normal contrast medium enhancement and cyclical-phase dependency. Radiology 203 (1997), MULLER-SCHIMPFLE M., STOLL P, STERN W. et al.: Do mammography, sonography, and MR mammography have a diagnostic benefit compared with mammography and sonography. AJR. 168 (1997), REMPP K. A., BRIX G., WENZ E et al.: Quantification of regional cerebral blood flow and volume with dynamic susceptibility contrast-enhanced MR imaging. Radiology 193 (1994), RINCK I? A.: Some future perspectives of contrast agent development. Clin. MRI 5 (1995),

8 T2* PERFUSION IMAGING IN BREAST LESIONS 25. SORENSEN A. G., TIEVSKY A. L., OSTERGAARD L. et al.: Contrast agents in functional MR imaging. J. Magn. Reson. Imaging 7 (1997), STOMPER? C., HERMAN S., KLIPPENSTEIN D. L. et al.: Suspect breast lesions. Findings at dynamic gadolinium-enhanced MR imaging correlated with mammographic and pathologic features. Radiology 197 (1995), Su S., SHEN J., TIAN G. et al.: A reevaluation of T1- and T2*-weighted imaging methods for myocardial perfusion. (Abstract, p. 684.) Proc. SOC. Magn. Reson. 1995, Berkeley, CA VAN DIJKE C. E, BRASCH R. C., ROBERTS T. P. L. et al.: Mammary carcinoma model - correlation of macromolecular contrast-enhanced MR imaging characterizations of tumor microvasculature and histologic capillary density. Radiology 198 (1996),