contents TIB Abstract Breast MR is a sensitive but nonspecific imaging investigation to detect breast can- cer. MR imaging strengths lie in the accu- rate staging of the primary tumour, detect- ing recurrent cancer following lumpectomy and radiation therapy, problem solving in cases where there are equivocal mammo- graphic findings, and screening for breast cancer in younger women with familial breast cancer. Interpretation of MR images requires a meticulous imaging technique including the use of contrast enhancement and fat suppression MR sequences using a good breast coil. Introduction The role of MR imaging in the diagno- sis of breast cancer is not clearly defined, however this modality is becoming impor- tant in breast imaging to solve problematic cases where the mammogram is inconclu- sive.1 MR imaging is highly sensitive to the detection of focal breast masses, approach- ing 100%, but not very specific for cancer, varying from 37% to 70% in most series.1 This is the reason why breast MR imaging is relegated to a second-line imaging inves- tigation. To interpret MR breast studies accurately it is important to understand the MR appearance of normal breast tissue, the enhancement pattern following gadolini- um contrast injection and the specific MR techniques used to obtain these images. Technique Phased array surface breast coils are essential to improve the signal-to-noise ratio (Fig. 1). We use spin echo (SE) T1 and T2 STIR sequences in the transverse planes. These are repeated following gadolinium enhancement using fat suppression in the transverse and sagittal planes. Fat suppres- sion is essential as both normal breast glan- dular tissue and breast cancer enhance fol- lowing contrast injection and this enhance- ment is easily obscured by the high intensi- ty of normal fat on T1-weighted images (Fig. 2a-d). Normal breast glandular tissue enhancement is minimal during days 7 - 20 of the menstrual cycle.2 This is the best period of time to image the breast in pre- menopausal women.2 In perimenopausal women focal enhancement of involuting breast parenchyma is a normal appear- ance.3 Contrast enhance- ment Gadolinium DTPA is injected intra- venously as a bolus of 20 ml at 3 ml/sec (0.15 mmol/kg) and signal intensity is mea- sured over 5 minutes. However breast can- cer enhances within the first 120 seconds of a contrast injection while normal glandular tissue enhances later than 120 seconds (Figs 2c, d). A mean curve function using regions of interest (ROI) over the first 5 minutes post contrast injection is then generated automatically. Although the shape of the curve, which measures contrast enhance- ment as a change of signal intensity over time, is useful in improving specificity of a focal lesion, the curve cannot be used to localise lesions for biopsy. The curve can be broken down into 2 components: the initial rise in contrast enhancement and the delayed phase. The initial rise can be slow, medium or fast. The delayed phase can be persistent, plateau or washout in character. Breast cancer has a rapid initial rise in con- trast enhancement due to tumour neovas- cularity (Figs 2c,d, 3a-e), however in one- third of lobular cancers and in patients with ductal carcinoma in situ (DCIS) there is a slow rise in enhancement. In the delayed phase there is persistent or plateau curve with breast cancer while normal glandular tissue shows a washout curve. 4 SA JOURNAL OF RADIOLOGY • October 2005 REVIEW ARTICLE Magnetic resonance imaging of invasive breast cancer P Corr FFRad (D) SA S Panday FCRad (D) P Seolall Nat Dip Rad (D) H Booth Nat Dip Rad (D) Department of Radiology Nelson Mandela School of Medicine and Inkosi Albert Luthuli Hospital Durban Fig.1. Breast coil used at Inkosi Albert Luthuli Hospital. Both breasts can be imaged simulta- neously in the prone position. Fig 2a,b. Patient with a breast cancer involving the medial aspect of the breast (arrow) is more conspicuous on the fat-saturated T1 STIR post- contrast sequence. pg4-6 TIB 10/8/05 11:10 AM Page 4 REVIEW ARTICLE 5 SA JOURNAL OF RADIOLOGY • October 2005 Morphological signs of breast masses Focal mass As in film screen mammography, the presence of a mass is confirmed by its mass effect and architectural distortion. Most malignant breast masses have a low intensi- ty on T1 and T2-weighted scans. Simple breast cysts, fat necrosis and intra-mamma- ry lymph nodes have a high intensity on T1-weighted scans. Myxoid fibroadeno- mas, fat necrosis and lymph nodes have a high intensity on T2-weighted scans. It is important to appreciate that focal contrast enhancement may not be due to a focal mass but it could rather represent normal glandular tissue in a peri- menopausal patient, fibrocystic disease of the breast or localised DCIS. This is called 'non mass' enhancement. Shape and margins of the mass Masses may be round, oval, lobular, Figs 2c,d. T1 STIR post gadolinium-enhanced image of the same patient as in Fig.2a demonstrates focal enhancement of the can- cer (region of interest 1) in Fig. 2d, and the cor- responding curve for this lesion (continuous line curve, arrow) shows the rapid initial wash in rise and the delayed washout typical of can- cer. Figs 3a,b. Mammogram of the right breast of a 30-year-old woman who had a right lumpectomy for breast cancer demonstrates a dense parenchymal pattern with a suspicious mass in the outer lateral quadrant (arrow) on the magnified view, Figs 3c,d. T1 and T1 fat-saturated contrast-enhanced transverse images of the right breast demonstrate the focal mass as a low-density spiculated lesion on T1 (arrow 3c) which enhances markedly after con- trast injection (arrow 3d) which is suspicious for a new cancer. Fig. 3e. Dynamic contrast curve for region 1 over the suspicious lesion (arrow) demonstrates the typical curve for a cancer with a rapid initial rise and slow washout of contrast with time. pg4-6 TIB 10/8/05 11:10 AM Page 5 irregular, smooth or spiculated as detected mammographically. However the most predictive sign of cancer on breast MR imaging is spiculation (positive predictive value of 80 - 91%) while the presence of a 'halo' of surrounding breast parenchyma, rim and central enhancement and ductal distribution of enhancement have a lower positive predictive value varying from 40% to 86%.4 Mass architecture and con- trast enhancement patterns Contrast enhancement within the mass can be focal, diffuse or segmental in nature. Segmental or branching enhancement rep- resents ductal pathology and is commonly detected in DCIS. Focal clumped enhance- ment is also found in DCIS. Heterogeneous focal enhancement is seen in cancers and fibroadenomas. Rim or edge enhancement is found in cancer. Masses with internal septations are found in fibroadenomas. Interpretation of MR images It is always good practice to assess a mass on its morphological appearances and use the contrast curves as secondary evi- dence. MR images must be read in con- junction with mammograms and ultra- sound examinations.1 Indications for breast MR imaging Preoperative staging Determination of the size of the cancer and the presence or absence of multifocal and multicentric cancers is critical in deter- mining the type of surgical procedure or treatment offered. MR is superior to both mammography and ultrasound in deter- mining the true extent of a cancer and is the most accurate imaging investigation when compared with the histological tumour extent following resection.5 Postoperative assessment of residual cancer The detection of residual cancer fol- lowing lumpectomy is important if breast conservation is to be considered. Assess- ment of histological tumour margins and detection of residual malignant microcalci- fications on postoperative mammograms are often inaccurate. MR imaging is more accurate with a sensitivity varying from 89% to 94%.6 Specificity improves if the MR scan is performed at least 28 days fol- lowing lumpectomy and at this time is 70%.7 MR has been found to be useful in differentiating scar from recurrent tumour in those patients who have had lumpec- tomies and/or radiotherapy treatment. However it is important to remember that surgical scars can enhance up to 6 months postoperatively and if radiotherapy is given then up to 18 months post treatment. Lobular breast cancer MR is especially useful in detecting lob- ular cancer, which occurs in 10% of women with breast cancer. This cancer infiltrates along ducts without a desmoplastic response making it difficult to detect by mammography. Lobular cancer is often bilateral and multicentric making this can- cer more easily detected by MR imaging.1,8 Cancer in mammographi- cally dense breasts MR imaging detects more extensive tumour as well as multifocal and multicen- tric cancer in patients with newly diag- nosed cancer than mammography. This is especially true in those patients with mam- mographically dense breasts.9 This has been demonstrated to change surgical manage- ment in up to 51% of patients.10 Screening in familial breast cancer Breast MR imaging has been demon- strated to be more sensitive than screening mammography in the detection of familial cancer which may be multifocal in patients who are BRCA1 or 2 positive or who have a strong family history of breast cancer.11 MR detects between 1% and 4% more cancers than mammography in these patients.1 Conclusions Breast MR imaging is a new advance in the diagnosis of breast cancer and in screening for cancer in high-risk women. Attention to MR technique is essential for the correct interpretation of findings. As with all forms of breast imaging, there is a steep learning curve for the radiologist to take to become proficient in interpretation. References 1. Lee C. Problem solving MR imaging of the breast. Radiol Clin North Am 2004; 42: 919-934. 2. Muller-Schimpfle M, Ohmenhauser K, Stoll P, Dietz K, Clauseen CD. Menstrual cycle and age: influence on parenchymal contrast enhancement in MR imaging of the breast. Radiology 1997; 203: 145-149. 3. Kuhl C, Bieling HB, Gieseke J, Kreft BP, Sommer T, Lutterbey G. Healthy premenopausal breast parenchyma in dynamic contrast enhanced MR imaging of the breast: normal contrast medium enhancement and cyclical phase dependency. Radiology 1997; 203: 137-144. 4. American College of Radiology. Breast Imaging Reporting and Data System (BI-RADS). Reston, USA: ACR, 2003. 5. Boetes C, Mus RD, Holland R, Barentsz J, Strijk S, Wobbs T. Breast tumors: comparative accuracy of MR to mammography and ultrasound in deter- mining extent. Radiology 1995; 197: 743-747. 6. Frei KA, Kinkel K, Bonel HM, Lu Y, Esserman LJ, Hylton NM. MR imaging of the breast in patients with positive margins after lumpectomy: influence of the time interval between lumpectomy and MR imaging. Am J Roentgenol 2000; 175: 1577-1584. 7. Lee JM, Orel SG, Czerniecki BJ, Solin LJ, Schnall MD MRI before re-excision surgery in patients with breast cancer. Am J Roentgenol 2004; 182: 473- 480. 8. Weinstein WP, Orel SG, Heller R, Reynolds C, Czerniecki B, Solin L. MR imaging of the breast in patients with invasive lobular carcinoma. Am J Roentgenol 2001; 176: 399-406. 9. Hlawatsch A, Tiefke A, Schmidt M, Thelen M. Preoperative assessment of breast cancer: sonogra- phy vs MR imaging. Am J Roentgenol 2002; 179: 1493-1501. 10. Bedrosian I, Mick R, Orel SG, Schnall M, Reynolds C. Changes in the surgical management of patients with breast carcinoma based on preoperative MR imaging. Cancer 2003; 98: 468-473. 11. Kuhl CK, Schmutzler RK, Leutner CC, Kempe A, Wardelman E, Hocke A. Breast MR imaging screening in 192 women proved or suspected to be carriers of a breast cancer susceptibility gene; pre- liminary results. Radiology 2000; 215: 267-269. REVIEW ARTICLE 6 SA JOURNAL OF RADIOLOGY • October 2005 pg4-6 9/28/05 2:31 PM Page 6