False-negative findings of malignant breast lesions on preoperative magnetic resonance mammography

The Breast (2001) 10, 131–139 # 2001 Harcourt Publishers Ltd doi:10.1054/brst.2000.0232, available online at http://www.idealibrary.com on

False-negative findings of malignant breast lesions on preoperative magnetic resonance mammography S. Wurdinger, S. Kamprath, D. Eschrich, A. Schneider and W. A. Kaiser Institute of Diagnostic and Interventional Radiology, Department of Gynaecology, Friedrich Schiller University, Jena, Germany S U M M A R Y . The aim of this study was to evaluate causes and percentages of false negative diagnoses of malignant breast lesions on preoperative dynamic magnetic resonance mammography (MRM). MRM was performed in 223 patients with 234 histopathologically proven malignant breast lesions (193 invasive carcinoma, 41 CIS) which were analyzed prospectively by routine analysis prior to surgery and re-analyzed by specialists, retrospectively. False negative findings were re-evaluated with respect to contrast enhancement, size and shape of lesions, reading errors, and technical problems. Preoperative analysis missed 27 of 234 malignant breast lesions (sensitivity 88.5%) including 15 of 193 invasive cancers (sensitivity 92%) and 12 of 41 CIS (sensitivity 71%). Five of 193 invasive cancers (four invasive lobular, one invasive tubular carcinoma) and five of 41 CIS lesions were missed due to delayed or no contrast enhancement. The remaining 17 false negative diagnoses were due to reading errors (n=8), previous core biopsies (n=3), metal induced artefacts (n=3), localization outside the field of view (n=1), incorrect injection (n=1) or movement artefacts (n=1). Using dynamic MR mammography, there were 4.3% slow contrast enhancing malignant breast lesions and a maximum sensitivity of 95.7% for detection of all malignant breast lesions (97.4% for invasive breast cancer, 87.8% for carcinoma in situ) can be achieved in a preselected preoperative population. # 2001 Harcourt Publishers Ltd

yielding sensitivities above 94% for all malignant lesions.3,7,16–18 Careful analysis of early contrast-enhancement patterns increase specificity on dynamic MR mammography.4,6 However, differentiation malignant from benign lesions by dynamic MR mammography remains a contentions issue19–22 and it remains unclear if there is a high percentage of atypically slow enhancing malignant breast lesions. The present retrospective study was performed to evaluate the sensitivity of MR mammography in a large group of preoperative patients under routine conditions and to determine the causes of failing to diagnose invasive breast cancer and carcinoma in situ as well as to determine the possible limits of the sensitivity of the technique.

INTRODUCTION MR mammography (MRM) has been in clinical use since 1985.1,2 Several steps have been performed in order to increase the accuracy of this diagnostic method including the development of a double breast coil,3 injection of contrast medium,2 and the application of temporal resolution for contrast enhancement on dynamic MR imagning.4–8 Before MRM can be used for routine work-up of breast disease, technique and interpretation of data should be standardized.9,10 Several authors have reported an excellent sensitivity for detection and identification of malignant breast lesions: sensitivity rates ranging between 82%11–13 and 100%8,14 for invasive cancer and from 65%14 to 94%15 for carcinoma in situ, with the majority of studies

MATERIALS AND METHODS Address correspondence to: Dr Susanne Wurdinger, Institute of Diagnostic and Interventional Radiology, Friedrich Schiller University, Bachstr, 18, D-07740 Jena, Germany. Tel.: +49 (03641) 934005; Fax: +49 (03641) 934070; E-mail: susanne.wurdinger@med. uni-jena.de [email protected]

From December 1994 to October 1998, a total of 1388 patients were examined using dynamic MRM including 424 consecutive patients (16–86 years, mean age 54 131

132

The Breast

years) of the Department of Gynaecology of our University Hospital. These 424 patients were scheduled for surgery of breast lesions. They were referred with palpable and/or mammographically and/or sonographically suspicious breast lesions to the Department of Gynaecology. All patients underwent routine MR mammography prior to excisional biopsy. In patients with lesions highly suggestive of malignancy established by clinical, mammographic, and/or sonographic criteria, MR mammography was performed to determine multifocality, multicentricity including assessment tumours in the contralateral breast, pectoral attachment and tumour extention. In cases with findings suspicious of multifocality on MR mammography, core biopsy or open biopsy prior to mastectomy was performed for histopathological confirmation. Patients with probably benign or probably malignant lesions underwent MR mammography to attempt to differentiate between benign and malignant lesions. Patients with a high probability of having benign lesions underwent open biopsy afterwards on request by the patient, for cosmetic reasons or due to the recommendation of the referring gynaecologist or radiologist. Patients gave their informed consent about MR imaging and agreed to the use their personal data for research purposes. MR images were obtained on a 1.5 T imager (Gyroscan ACS II, Philips, Hamburg, Germany) using a dedicated double breast coil. All patients were examined according to an identical protocol. Prior to imaging an intravenous injection system with saline was introduced into a cubital vein or, if not possible, into a dorsal vein of the hand. Patients were examined in the prone position with their breasts placed into the double breast coil without deformation. This procedure makes it possible to localize small lesions with reference to a coordinate system when the nipple is located in the centre. Patients were instructed to hold as still as possible during the whole imaging session which took about 30 min. The examination protocol is listed in Table 1. After an initial scout, a T1-weighted fast-field-echo sequence was

obtained in coronal or axial orientation. Precontrast scans were inspected with respect to inhom ogeneous fat signal or susceptibility artefacts. If such inhom ogenities were detected, the precontrast scan was repeated after a second tuning. For the dynamic study 2D multi-slice images were obtained using a fast-field-echo T1weighted sequence images with the following parameters: TR 97, TE 5.0; flip angle 808; slice thickness 4.0 mm, gap 0.4 mm; field of view 3156350 mm2; matrix 2306256 (90% rectangular FOV) corresponding to an in-plane scan resolution of 1.37 mm61.37 mm, axial orientation, 24 slices covering both breasts. After acquisition of a native scan 0.1 mmol/per kg body weight Gd-DTPA (Magnevist, Schering, Berlin, Germany) was administered intravenously as a rapid bolus within 10 seconds followed by 20 ml saline flush (30 ml saline flush if a dorsal hand vein was used). After bolus and saline administration dynamic scanning was continued with the same sequence parameters and under identical tuning conditions at 1-min intervals for a total of 8 min. After the dynamic scan the first coronal or axial T1-weighted scan was repeated approximately 10 min after contrast administration, followed by axial T2-weighted turbo-spin-echo images in identical slice positions. Postprocessing of the dynamic study included calculation of time-signal intensity curves and subtraction of precontrast images from the postcontrast dynamic images. In cases of large motion artefacts, enhancing lesion were only detected by comparison of nonsubtracted pre- and postcontrast images. For the calculation of time-signal intensity curves and time-signal intensity data built in scanner software were used. Regions of interest (ROI) were placed in parts of the lesions showing the strongest signal increase within the first 2 min after bolus injection. Between 2 and 20 different time-signal intensity curves were analyzed for each lesion. The main evaluation criterion for a malignant lesion was a focal signal enhancement of 90% or more (relative to the precontrast signal intensity) within the first 2 min after bolus injection followed by a subsequent signal

Table 1 Examination protocol of dynamic MRM Orientation TR in msec TE in msec Flip angle in 8 Slice thickness in mm Gap in mm Field of view in mm Number of slices T1 T1 T1 T1 T2

SE Transversal FFE Coronal FFE* Transversal FFE Coronal TSE Transversal

121 96 97 96 4000

13 5 5 5 300

90 80 80 80 90

5 4 4 4 4

3 0.4 0.4 0.4 0.4

450 350 350 350 350

5 24 24 24 24

*Dynamic study, postcontrast images after i.v. application of 0.1 mmol/per kg body weight Gd-DTPA within 10 s followed by 20/30 ml saline flush.

False-negative findings of malignant breast lesions on preoperative MR mammography plateau level or a decrease of signal intensity (wash-out phenomenon). Wash-out phenomenon was considered as highly suspicious for malignancy. Additional secondary criteria for malignancy were spiculated margins of the lesion, early centripetal, complete or inhomogeneous contrast enhancement and a low signal intensity on T2weighted images post-contrast. The extent, number, and morphology of suspicious lesions were recorded and described. The position of the lesion was determined by considering the breast in the prone position as a half spherical mass with the nipple on top and marking the distance from the lesion to the nipple in millimetres. Lesion seen only on MR and lesions which were not reliably localized when compared to the findings of mammography or sonography were marked under MRguidance using a MR-compatible double-breast device23 and titanium wires (MR eyeTM needle, Cook, Mo¨nchengladbach, Germany). MR mammograms were evaluated prospectively by different resident physicians and were checked by an experienced general radiologist. The MR interpretation was performed with the knowledge of the diagnosis made by the referring gynaecologist. In general, Mammographic and sonographic images of lesions were not available. During a second review by residents, the diagnoses of MRM were compared with the histopathological finding. MRM-findings which confirmed with histopathological results were re-evaluated by an experienced radiologist knowing the histopathological results.

RESULTS Histopathologic results A total of 488 lesions in 424 patients were diagnosed and examined histologically after biopsy and/or mastectomy. Histological evaluation revealed a total of 234 malignant lesions in 223 patients. There were seven patients with multicentric malignant lesions in one breast (distance between lesions more than 3 cm) and four patients with malignant lesions in both breasts. Fifty-eight patients had multifocal breast cancer localized in one quadrant of the breast counting as one lesion, respectively. The histologic results are summarized in Table 2. One hundred and ninety-three invasive cancers with or without associated carcinoma in situ and 41 pure carcinoma in situ lesion, were found. The histologically verified tumour size, based on the largest diameter, was classified corresponding to postoperative TNM-stages. Among the 193 invasive lesions 184 were primary cancers, six were recurrences, two were

133

Table 2 Distribution of histopathological tumour types of all malignant breast lesions Histopathologic diagnosis Invasive carcinoma (n=193) Ductal carcinoma Lobular carcinoma Mixed ductal/lobular carcinoma Tubular carcinoma Medullary carcinoma Mucinous carcinoma other (metaplastic carcinoma, metastases, angiosarcoma) Carcinoma in situ (n=41) Ductal carcinoma in situ Lobular carcinoma in situ

No. of lesions 114 37 18 13 3 2 6

40 1

metastases and one was an angiosarcoma. The majority of invasive breast cancers were staged as pT1c (1.1– 2.0 cm) (44%). Six per cent of tumours were detected in stage pT1a (up to 0.5 cm), 18% in stage pT1b (up to 1.0 cm), 25% in stage pT2 (2.1–5 cm), 3% in stage pT3 and 4% in stage pT4. Individual tumour size ranged from 3 mm to 15 cm.

Results on MRM Twenty-seven malignant lesions (15 invasive cancers and 12 carcinoma in situ (CIS)) were not described in the preoperative evaluation of MRM, which results in an overall sensitivity of 88.5% (92.3% for invasive cancers and 70.7% for CIS). Complementary, histopathologic diagnoses of benign lesions were compared to diagnoses of MRM:174 of 254 benign lesions were described as benign in MR mammography. Accordingly, overall specificity was 68.5% and the accuracy was 78.7% in this study.

Lesions with delayed contrast enhancement on MRM After postoperative re-evaluation of the MRM, a total of 10 malignant lesions were not detected because of a delayed (Fig. 1) or no contrast enhancement (sensitivity 95.7%) including five of 193 invasive cancers (sensitivity 97.4%) and five of 41 CIS (sensitivity 87.8%) (Table 3). The five invasive cancers which were not detected after re-evaluation were invasive lobular (n=4) and invasive tubular (n=1) carcinoma. The invasive lobular cancers were graded from G2 (n=3) (moderately differentiated) to G3 (n=1) (poorly differentiated). One of the patients with invasive lobular cancer had undergone adjuvant high dose chemotherapy 6 months prior to MR imaging for a cancer of the contralateral

134

The Breast

Fig. 1 MR images of a 74-year-old woman with a suspicious density on X-ray mammogram and suspicious lesion on sonography in the inner upper quadrant of the right breast; negative in MRM due to slow contrast enhancement. (A) Dynamic MR images prior (upper left) to contrast administration and subtracted images 1 min (upper right), 2 min (lower right) and 7 min (lower left) after contrast administration showed a slow contrast enhancement of an 15 mm invasive lobular carcinoma. (B) Time signal-intensity curve of the lesion demonstrates a delayed signal enhancement and signal increase less than 90% after contrast administration.

Table 3 Causes of false-negative results of 234 histologically verfied malignant breast lesions in dynamic MR mammography Percentage of false-negative diagnoses Causes of false-negative diagnoses Delayed or no contrast enhancement of lesions Reading errors Bleeding after biopsy Overlap by metal-induced extinction artefacts Other technical problems

breast. Lesion sizes were 4 mm (n=1), 15 mm (n=2), 30 mm (n=1) and 35 mm (n=1). Twelve of the 41 CIS lesions were not described in preoperative evaluation. On re-evaluation five CIS showed no or delayed contrast enhancement. Grading of these CIS was G1 in two cases (well differentiated), G2 in one case, and G3 in one case (poorly differentiated); one CIS was not classified. Excisional biopsy of these lesions was indicated due to a suspicious X-ray mammogram in four of five invasive cancers, due to a palpable mass in one invasive cancer, due to suspicious microcalcifications in X-ray mammography in four of five carcinoma in situ and due to palpable mass in one carcinoma in situ (Table 4).

Invasive carcinoma (n=193) 2.6% 2.6% 1.0% 0.5% 1.0%

(n=5) (n=5) (n=2) (n=1) (n=2)

Carcinoma in situ (n=41) 12.2% (n=5) 7.3% (n=3) 2.4% (n=1) 4.8% (n=2) 2.4% (n=1)

Other reasons for false negative MRM As shown in Table 3, other reasons for false negative diagnoses of invasive cancers included reading errors (n=5), haemorrhagic artifacts caused by previous core biopsy (n=2), overlap by metal-induced artifact (n=1), localization outside the field of view (n=1), and moving artifact (n=1). Figure 2 shows a 3 mm ductal carcinoma which was misinterpreted as an inflamed cyst. Three CIS cases were missed due to reading errors: one patient was taking hormone replacement therapy and two lesions were classified as proliferative hyperplasia. Two CIS were not diagnosed due to an overlap by metal-induced artifacts after former biopsy. One

Table 4 Malignant tumours not described by dynamic MR mammography with histology, pT-stages, and different examination modalities Histopathologic diagnosis

pT-stage Indication of excisional biopsy Palpation X-ray mammography Sonography

Invasive tubular carcinoma ILC ILC ILC 3.5 cm recurrence of ILC DCIS DCIS DCIS DCIS DCIS

PT1a PT1c PT1c PT2(m) n.c. pTis pTis pTis pTis pTis

X-ray mammography X-ray mammography X-ray mammography X-ray mammography Palpation X-ray mammography X-ray mammography X-ray mammography X-ray mammography Palpation

0 0 0 0 1 0 0 0 0 1

1 1 1 1 0 1 1 1 1 0

1 1 1 1 1 0 0 0 0 1

IDC IDC+DCIS IDC+LCIS inv. medullar carcinoma+LCIS inv. carcinoma n.c.+LCIS DCIS DCIS LCIS

pT1b pT1a(m) pT1b pT1c pT1a pTis pTis pTis

Palpation X-ray mammography Palpation X-ray mammography X-ray mammography X-ray mammography Palpation X-ray mammography

1 0 0 0 0 1 1 0

0 1 0 1 1 1 0 1

0 0 0 1 0 1 0 0

2.0 cm recurrence of IDC IDC DCIS

n.c. pT3 pTis

Palpation Palpation Palpation

1 1 1

0 1 0

1 1 0

IDC+DCIS DCIS DCIS

pT1a pTis pTis

X-ray mammography X-ray mammography X-ray mammography

0 0 0

1 1 1

0 1 0

ILC/IDC pT1b X-ray mammography 1.5 cm and 1.0 cm recurrence IDC n.c. Palpation DCIS pTis (m) Change of implants

0 1 0 9

1 0 0 18

0 0 0 11

Delayed contrast or no contrast enhancement of lesion

Reading errors

Bleeding artefact after biopsy

Overlap by metal-induced artefacts

Other technical problems

Total

Table footnote ILC=invasive lobular carcinoma; IDC=invasive ductal carcinoma; DCIS=ductal carcinoma in situ; LCIS=lobular carcinoma in situ; n.c.=not classified; 1=detected; 0=not detected.

False-negative findings of malignant breast lesions on preoperative MR mammography

Causes of false-negative diagnosis

135

136

The Breast

Fig. 2 Images of a 62-year-old woman with a 3 mm invasive ductal carcinoma located in upper the outer quadrant of the left breast, detected by X-ray mammography, which was classified as an inflamed cyst on MRM. (A) Dynamic MR images prior (upper left) to contrast application and 1 min (upper right), 2 min (lower right) and 7 min (lower left) after contrast application. (B) T2-weighted image shows high signal intensity, atypical for malignant lesions.

patient showed a generally delayed contrast enhancement caused by an incorrect bolus administration of contrast medium. One patient had undergone core biopsy prior to MRM and a bleeding artifact was visible in MR imaging. Table 4 shows that excisional biopsy was indicated by suspicious X-ray mammogram in five of the 10 invasive cancers and by a suspicious palpable mass in four cases. One invasive lobular carcinoma was classified as a fibroadenoma on MRM, X-ray mammography, and sonography. Surgery was indicated by suspicious microcalcifications on X-ray mammography in four of seven patients with carcinoma in situ and due to a palpable mass in two of these seven patients. One DCIS was detected by chance while changing silicone implants.

DISCUSSION MR mammography (MRM) is increasingly integrated in the preoperative diagnostic workup of women diagnosed with abnormal findings of the breast. Some of the most important advantages of MRM are to determine multifocality, small foci and localization of breast cancer.5,11,16,24 In previous studies sensitivities of more than 98% for the detection of breast cancer have been reported for dynamic MR mammography.14,16,18,25 In our study a

sensitivity of 88.5% for all malignant lesions (92.3% invasive cancer, 70.7% carcinoma in situ) was found in the preoperative evaluation. The choice of a selected study population with a high prevalence of disease could have influenced the sensitivity. However, preselection of patients is recommended for MRM to achieve an adequate specificity and a high cost effectiveness.10 Thus, we evaluated a preselected preoperative patients undergoing routine diagnostic tests. Although this study is not a prospective trial, the patient population studied is representative of the kinds of consecutive large population, examination using of identical protocol for MRM and surgery performed with the knowledge of clinical, mammographic, sonographic and MR mammographic features. The sensitivity of dynamic MRM is based on the observation that the predominant majority of invasive cancers show rapid contrast enhancement after contrast injection.3,6,16 The underlying biologic basis of dynamic MRM is the increased number of vessels with abnormal basement membranes in invasive breast cancer.26 Several studies have shown increased microvessel density in breast tumours leading to an increased initial rate of contrast medium enhancement.13,22,27 In this study five invasive cancers (four lobular, one tubular carcinoma) were not detected due to slow contrast enhancement. Bone et al.17 reported eight false negatives (five invasive lobular cancers, one invasive tubular, one mucinous carcinoma and one angiosarcoma) in a group

False-negative findings of malignant breast lesions on preoperative MR mammography of 155 malignant lesions. Compared with our technique the authors used a lower temporal resolution in their dynamic study with an acquisition time of 6 min 23 seconds for postcontrast images in favour of high spatial resolution. They examined one patient with angiosarcoma. Angiosarcomas contain a high number of vessels and this may have led to a decreased signal intensity due to early washout. Boetes et al.28 identified one invasive lobular carcinoma and one mucinous carcinoma among 129 invasive breast cancers with atypically low rate of contrast uptake using a TurboFLASH technique. Gilles et al.29 found two slowly enhancing invasive lobular carcinomas in a series of 64 malignant breast lesions. In a series of 23 invasive lobular carcinomas, Sittek et al.30 found four false negatives using a dynamic 3D-FLASH sequence. Despite differences in technique, several authors have described atypically slow enhancing invasive lobular, tubular or mucinous carcinomas. This might be explained by the different tissue architecture of some lobular, tubular and mucinious carcinomas compared to invasive ductal carcinoma with small cells diffusely invading stroma, entrapping normal structures (invasive lobular carcinoma), reactive appearing fibroblastic stroma (invasive tubular carcinoma) or large amounts of extracellular mucus in contact with the stroma (invasive mucinous carcinoma).31 Furthermore five of 41 carcinoma in situ (CIS) of our series showed no or late contrast enhancement. Results of previous studies have confirmed a lower detectability rate for CIS using dynamic MR mammography. Fischer et al.32 found that six of 35 foci of DCIS had no contrast enhancement using a MR technique similar to ours. Among 36 women with DCIS, Gilles et al.15 demonstrated two cases without early contrast enhancement, Bone et al.17 detected three slowly enhancing CIS in a series of 17 cases. Although some malignant lesions are not detectable on MRI, in our study they were recognized by X-ray mammography or palpation. Four of five slowly enhancing CIS in our study had suspicious microcalcifications and four of five invasive carcinomas showed suspicious densities. The remaining two lesions were detected by palpation. A combination of diagnostic methods is recommended to attain the highest possible sensitivity during preoperative workup.33 In our study, 17 malignant lesions were detected on review and the sensitivity increased following re-evaluation from 88.5% (92.3% invasive cancer, 70.7% carcinoma in situ) to 95.7% (97.4% invasive cancer, 87.8% carcinoma in situ). Eight of 234 malignant tumours (3.4%) were not recognized due to reading errors in the preoperative evaluation under routine

137

conditions. Patel and Withman34 reviewed 496 X-ray mammograms of breast cancer patients and found that one-third of former false-negatives were visible in retrospect. This is in accordance with our findings (eight/27 false negatives). On the other hand, two carcinomas of 4 mm and 3 mm, respectively, were detected in retrospect. Possibly, these lesions would have been easier to detect if the slice thickness of our dynamic scans was less than 4 mm. For detection of multifocality and extent MR imaging should be performed with bilateral volume coverage. Fast dynamic imaging is supposed to have advantages with regard to both sensitivity and specificity. Compared to our study, MR dynamic images obtained at intervals of 2.3 seconds for 2 min resulted in a similar sensitivity of 95% and a higher specificity of 86%.35 According to recommendations of the Working Group for Breast MRI volume dynamic scanning requires a minimum temporal resolution of 90 seconds.36 Therefore decreasing temporal resolution form 60 seconds to 90 seconds while increasing spatial resolution with a section thickness of 3 mm or less may be a potential improvement to the scanning protocol used in this study. Furthermore, two cancers were diagnosed incorrectly as fibroadenomas. Fibroadenomas are described as round or lobulated lesions with central, increasing contrast enhancement and occasionally internal septa.37 High signal intensity of fibroadenomas in T2-wieghted images may help to distinguish fibroadenomas from carcinomas.38 In our study, these two lesions had a high signal intensity on T2-weighted images, but the shape of the time-signal intensity curve with rapid contrast enhancement followed by wash-out was suspicious of malignancy.37,39 There were three false negative findings due to signal changes following biopsies. Signal enhancement of lesions may be caused by haemorrhage and or oedematous changes in surrounding tissue. Therefore, MR mammography should be performed prior to biopsy. In our study the error rate increased from 2.6% up to 3.4% if carcinoma in situ (CIS) was included. This was due to the difficulty in interpreting some cases of CIS: carcinoma in situ may appear as a reticular segmental or a small spotting area of enhancement on MRM.15,32 In our study, multiple enhancing spots in one patient takes hormone replacement therapy (HRT) covered the segmental suspicious enhancement of one case lobular carcinoma in situ. By quantitative evaluation of signal changes in breast parenchyma Reichenbach et al. (40) found an increase of parenchymal tissue/fat ratio during application of an oestrogen/progestagen combination but not during oestrogen-based hormone replacement

138

The Breast

therapy. Consequently, in cases of changed appearance of parenchyma on HRT malignancy cannot be excluded. In these patients follow-up is recommended after discontinuing hormone treatment.41 Three malignant lesions were classified as false negatives because of extinction artefacts. Metal particles may cause signal extinction on MRI. Most likely, these metal-induced artefacts were caused by metallic rub off particles remaining from previous surgery. These microparticles were not visible on X-ray mammography. Haemosiderin from previous bleeding may also cause extinction areas. However, to our knowledge a systematic study of extinction artefacts following breast surgery does not exist. The problem is, that small malignant lesions cannot be excluded in such areas on MR images. In conclusion, 207 of all 234 malignant lesions were described in routine preoperative dynamic MR mammography. There were 4.3% of slow contrast enhancing malignant breast lesions in a preselected preoperative population. The remaining 7.2% false negative findings were caused by reading errors and by technical aspects like extinction artefacts (hemorrhage after core biopsy, metallic particles). The influence of second reading on sensitivity of MRM needs further evaluation; however, performance and evaluation of MRM in specialized departments could keep the rate of reading errors and technical problems on a low level. If a combination of diagnostic methods including MR mammography is used the greatest possible sensitivity and exact localization of breast cancer will be achieved for almost all patients, prior to surgery. References 1. Kaiser W A, Zeitler E. Kernspintomographie der Mamma — Erste klinische Ergebnisse. Ro¨ntgenpraxis 1985; 38(7): 256–262. 2. Heywang S H, Hahn D, Schmid H et al. MR imaging of the breast using gadolinium-DTPA. J Comput Assist Tomogr 1986; 10: 199–204. 3. Kaiser W A, Kess H. Prototyp-Doppelspule fu¨r die Mamma-MRMessung. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1989; 151: 103–105. 4. Kaiser W A, Zeitler E. MR imaging of the breast: fast imaging sequences with and without Gd-DTPA. Radiology 1989; 170: 681–686. 5. Kaiser W A. MR-Diagnostik der Mamma – Erfahrungen nach 253 Untersuchungen. Dtsch Med Wochenschr 1989; 114(36): 1351– 1357. 6. Stack J P, Redmond O M, Codd M B, Dervan P A, Ennis J T. Breast disease: tissue characterization with gadopentate dimeglumine enhancement profiles. Radiology 1990; 174: 491–494. 7. Fischer U, von-Heyden D, Vosshenrich R, Vieweg I, Grabbe E. [Signal characteristics of malignant and benign lesions in dynamic 2D-MRT of the breast] Signalverhalten maligner und benigner Lasionen in der dynamischen 2D-MRT der Mamma. RofoFortschr-Geb-Rontgenstr-Neuen-Bildgeb-Verfahr 1993; 158(4): 287–292. 8. Boetes C, Mus R D, Holland R, et al. Breast tumours: comparative accuracy of MR imaging relative to mammography

9. 10. 11. 12.

13. 14.

15. 16.

17.

18. 19. 20. 21.

22.

23. 24.

25. 26. 27.

28. 29.

and US for demonstrating extent. Radiology 1995; 197(3): 743–747. Piccoli C W. The specificity of contrast-enhanced breast MR imaging. Magn Reson Imaging Clin N Am 1994; 2(4): 557–571. Heywang-Koebrunner SH. Breast MRI proves worth but lacks of standardization. Diagnostic Imaging Europe 1999; 5: 33–48. Oellinger H, Heins S, Sander B et al. Gd-DTPA enhanced MR breast imaging: the most sensitive method for multicentric carcinomas of the female breast. Eur Radiol 1993; 3(3): 223. Fobben ES, Rubin CZ, Kalisher L, Dembner AG, Seltzer MH, Santoro EJ. Breast MR imaging with commercially available techniques: radiologicpathologic correlation. Radiology 1995; 196(1): 143–152. Hulka CA, Smith BL, Sgroi DC, et al. Benign and malignant breast lesions: differentiation with echo-planar MR imaging. Radiology 1995; 197(1): 33–38. Stomper PC, Herman S, Klippenstein DL, et al. Suspect breast lesions: findings at dynamic gadolinium-enhanced MR imaging correlated with mammographic and pathologic features. Radiology 1995; 197(2): 387–395. Gilles R, Zafrani B, Guinebretiere JM, et al. Ductal carcinoma in situ: MR imaging-histopathologic correlation. Radiology 1995; 196(2): 415–419. Harms SE, Flaming DP, Hesley KL, et al. MR imaging of the breast with rotating delivery of excitation off resonance: clinical experience with pathologic correlation. Radiology 1993; 187: 493. Bone B, Aspelin P, Bronge L, Isberg B, Perbeck L, Veress B. Sensitivity and specificity of MR mammography with histopathological correlation in 250 breasts. Acta Radiol 1996; 37: 208–213. Heywang-Koebrunner SH, Viehweg P, Heinig A, Ku¨chler C. Contrast-enhanced MRI of the breast: Accuracy, value, controversies, solutions. Eur J Radiol 1997; 24: 94–108. Heywang-Koebrunner SH. Contrast-enhanced magnetic resonance imaging of the breast. Invest Radiol 1994; 29: 94–104. Orel SG, Schnall MD, LiVolsi VA, Troupin RH. Suspicious breast lesions: MR imaging with radiologic-pathologic correlation. Radiology 1994; 190: 485–493. Schnall M, Orel S, Muenz L. Analysis of time-intensity curves for enhancing lesions (abstr). In: Book of abstracts: Society of Magnetic Resonance in Medicine. Berkeley: Society of Magnetic Resonance in Medicine, 1993; 120. Buadu LD, Murakami J, Murayama S, Hashiguchi N et al. Breast lesions: correlation of contrast medium enhancement pattern on MR images with histopathologic findings and tumour angiogenesis. Radiology 1996; 200: 639–649. Wurdinger S, Noras H, Straube K, Michaelsen S, Kaiser WA. New MRI-breast coil for imaging and doublesided intervention: preliminary results. Eur Radiol 1997; 7(suppl): 243. Esserman L, Hylton N, Yassa L, Barclay J, Frankel S, Sickles E. Utility of magnetic resonance imaging in the management of breast cancer: evidence for improved preoperative staging. J Clin Oncol 1999; 17: 110–119. Heywang-Ko¨brunner SH. Brustkrebsdiagnostik mit MR: U¨berblick nach 1250 Patientenuntersuchungen. Electromedica 1993; 61(2): 243–252. Weidner N, Semple JP, Welch WR, Folkman J. Tumour angiogenesis and metastasis: correlation in invasive breast carcinoma. N Engl J Med 1991; 342: 1–8. Van Dijke CF, Brasch RC, Roberts TPL. Mammary carcinoma model: correlation of macromolecular contrast-enhanced MR imaging characterizations of tumour microvasculature and histologic capillary density. Radiology 1996; 198: 813–818. Boetes C, Strijk SP, Holland R, Barentsz JO, Van der Sluis RF, Ruijs JHJ. False-negative MR imaging of malignant breat tumours. Eur Radiol 1997; 7: 1231–1234. Gilles R, Guinebretiere J, Lucidarme O et al. Nonpalpable breast tumours: diagnosis with contrast-enhanced subtraction dynamic MR imaging. Radiology 1994; 191: 625.

False-negative findings of malignant breast lesions on preoperative MR mammography 30. Sittek H, Perlet C, Untch M, Kessler M, Reiser MF. Dynamic MR mammography in invasive lobular breast cancer. Radiology 1998; 209(suppl): 467. 31. Tavassoli FA. Infiltrating lobular carcinomas, common and familiar special types. In: Tavassoli FA, ed. Pathology of the breast. Norwalk: Appleton & Lange; 293–347. 32. Fischer U, Westerhof JP, Brinck U, Korabiowska M, Schauer A, Grabbe E. Das duktale In-situ-Karzinom in der dynamischen MRMammographie bei 1,5 T. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1996; 164(4): 290–294. 33. Mueller-Schimpfle M, Stoll P, Stern W, Kurz S, Dammann F, Claussen CD. Do mammography, sonography, and MR mammography have a diagnostic benefit compared with mammography and sonography? AJR 1997; 168: 1323–1329. 34. Patel MR, Withman GJ. Negative mammograms in symptomatic patients with breast cancer. Acad Radiol 1998; 5(1): 26–33. 35. Boetes C, Barentsz JO, Mus RD et al. MR characterization of suspicious breast lesions with a gadolinium-enhanced TurboFLASH substraction technique. Radiology 1994; 193(3): 777–781.

139

36. Plewes DB, Leach MO, Barentsz JO. Protocol Requirements. Summary of the International Working Group on Breast MR Imaging. Radiology 1999; 213(suppl): 53. 37. Kaiser WA. False-positive results in dynamic MR mammography (MRM) — frequency, causes and methods to avoid it. Magn Reson Imaging Clin N Am 1994; 2(4): 539–555. 38. Kuhl CK, Klaschik S, Mielcarek P, Gieseke J, Wardelmann E, Schild HH. Do T2-weighted pulse sequences help with the differential diagnosis of enhancing lesions in dynamic breast MRI? J Magn Reson Imaging 1999; 9(2): 187–196. 39. Kuhl CK, Mielcareck P, Klaschik S et al. Dynamic breast MR imaging: are signal intensity time course data useful for differential diagnosis of enhancing lesion? Radiology 1999; 211(1): 101–110. 40. Reichenbach JR, Przetak C, Klinger G, Kaiser WA. Assessment of breast tissue changes on hormonal replacement therapy using MRI: a pilot study. J Comput Assist Tomogr 1999; 23(3): 407–413. 41. Viehweg P, Paprosch I, Strassinopoulou M, Heywang-Ko¨brunner S. Contrast-enhanced magnetic resonance imaging of the breast: interpretation guidelines. Top Magn Reson Ima 1998; 9(1): 17–42.