Women with nonpalpable suspicious breast lesions (BI-RADS category 3, 4, or 5) detected on mammography or breast ultrasound, who are referred for histological analysis of the lesion, are eligible for the study. Exclusion criteria are age below 18 or above 75 years, breast surgery or radiation therapy of the breast less than 9 months prior to inclusion, pregnancy or lactation, claustrophobia, severe obesity (> 130 kg), general contraindications for MRI (i.e. cardiac pacemaker, metal implants or history of severe allergic reaction after administration of contrast agent), inability to maintain in prone position for one hour, medically unstable patients and severe coagulopathies or use of anti-coagulants that cannot be discontinued. Written informed consent will be obtained from all patients. The study has been approved by the ethical boards of the participating hospitals.
The MONET – study is a randomized controlled trial. Patients will be recruited in 3 hospitals: two large community teaching hospitals and one University Medical Center. In all participating centers, Breast MRI is not part of routine clinical practice in the work-up of patients with suspicious nonpalpable breast lesions. We aim to include a total of 500 patients with nonpalpable, mammographically suspicious breast lesions who are referred for biopsy. After giving informed consent, patients will be randomized in the control group or the MRI group. Patients in the control group (n = 250) will undergo routine medical care. This includes mammography, ultrasound, and the lesion is sampled by means of large core needle biopsy. In case of malignancy (DCIS or invasive carcinoma), the lesion is surgically removed. The MRI group (n = 250) will undergo routine clinical care and in addition, contrast enhanced Breast MRI in the University Medical Center prior to the large core needle biopsy. The MRI scan will be performed within one week after mammography and before LCNB which will be performed within one week after MRI to avoid a diagnostic delay.
Randomization is stratified by hospital. In each hospital, randomization within strata is blocked with a fixed block size. Randomization is performed by an independent trial center. If a patient meets the inclusion criteria and has provided informed consent, the physician contacts the trial center by phone. The trial center will perform the randomization of the patients.
Patient recruitment will take place between 2006 and 2008. Between January and November 2006, each month approximately 15 patients were included in the MONET – study.
Breast MRI will be performed on a 3-T clinical MR system (Achieva, Philips, Best, The Netherlands). Patients will be placed in prone position. A dedicated phased-array bilateral breast coil (MRI devices, Würzburg, Germany) will be utilized for simultaneous imaging of both breasts. The scan protocol will include a transverse high-resolution T1-weighted fast gradient echo fat-suppressed series (TE/TR 1.7/4.5 msec; inversion delay SPAIR 130 msec; flip angle 10°; FOV 340 × 340 mm2, acquired voxel size 0.66 × 0.66 × 1.6 mm3, reconstructed voxel size 0.66 × 0.66 × 0.80 mm3) and a transverse T2-weighted fat suppressed spin echo series (TE/TR 120/9022 msec; inversion delay SPAIR 125 msec; flip angle 90°; FOV 340 × 340 mm2, acquired voxel size 1.01 × 1.31 × 2.0 mm3, reconstructed voxel size 0.66 × 0.66 × 2.00 mm3). Both series will be used to study the morphology of the lesion. A diffusion-weighted fat-suppressed series (TE/TR 61/5000 msec; inversion delay SPAIR 70 msec; flip angle 90 °; FOV 320 × 320 mm2; acquired voxel size 2.22 × 2.52 × 4.00 mm3, reconstructed voxel size 1.33 × 1.33 × 4.00 mm3; b-values 0, 150, 499 and 1500 seconds/mm2) will be acquired to assess the cellularity of the lesion. Finally, dynamic contrast-enhanced fat-suppressed T1-weighted gradient echo images (TE/TR 1.3/3.4 msec; flip angle 10°; FOV 320 × 320 mm2, acquired voxel size 0.91 × 0.91 × 2.00 mm3, reconstructed voxel size 0.83 × 0.83 × 1.00 mm3; dynamic scan duration 60 sec) will be acquired before and immediately after administration 0.1 mmol/kg Gadolinium-DTPA (Magnevist, Schering, Germany) to study the contrast enhancement of the lesions and herewith the perfusion of the lesion. All patients will receive this scan-package with a total scan duration of less than 30 minutes.
Interpretation of Breast MRI
MR images of the breast will be interpreted on soft copy using a Picture Archiving and Communications System (PACS, Philips, Best, The Netherlands) that allows manual window level setting. MR images will be interpreted by two radiologists independently and in case of discrepancy, consensus will be sought. The radiologists will have access to the mammograms and ultrasound examination, but will be blinded for the results of the LCNB, following clinical practice. The MR images will be interpreted following the guidelines of the BI-RADS-MRI classification system proposed by the American College of Radiology . Classification of the lesions will be based on lesion morphology, enhancement pattern and enhancement kinetics (persistent, plateau, or washout) [19, 20]. Level of suspicion will be reported on a scale of 0 – 6: 0 additional imaging required; 1 normal; 2 benign; 3 probably benign, 6-month follow-up MRI recommended; 4 suspicious for malignancy; 5 highly suggestive of malignancy; 6 known malignancy.
LCNB will be performed under ultrasound guidance or stereotactically. A minimum of four 14 Gauge biopsy specimens per lesion is acquired. The histological diagnosis of the tissue sampled by means of the LCNB provides the definitive diagnosis. Based on histology, patient management will be planned: 'benign lesions' require mammographic follow-up, 'normal breast tissue' requires a repeat LCNB or open breast biopsy and in case of a 'high risk lesion' (i.e. lobular carcinoma in situ, atypical ductal or lobular hyperplasia), 'in-situ carcinoma' or 'invasive carcinoma' require image-guided needle localization followed by surgical excision with appropriate axillary staging (i.e. sentinel node biopsy or axillary lymph node dissection). Patient management will be the same for the MRI and the control group.
If in the intervention group an additional, mammographically occult lesion is detected on the MR images that requires tissue sampling, second-look ultrasound will be performed to determine whether the lesion is ultrasonographically evident and suitable for ultrasound-guided tissue sampling. If the lesion cannot be visualised by means of second-look ultrasound, MRI-guided large core needle biopsy will be performed.
Depending on the size of the tumor, the size of the breast and patient preferences, the lesion will be removed following the institutions' guidelines for breast cancer management. MR images will be shown to and discussed with the surgeon before the operation. The number and type of surgical procedures (lumpectomy, re-excision, mastectomy, sentinel node biopsy or axillary lymph node dissection) will be reported until 1 year after the LCNB.
To evaluate whether performing preoperative Breast MRI will improve breast cancer management, we will collect data on all biopsies (including biopsies performed for lesions that were detected on MRI only) and all surgical procedures during 1 year of follow-up after the LCNB. The number of biopsies and surgical procedures in patients in the MRI group will be compared to the number of procedures in the control group.
Previous studies have shown that the positive predictive value of nonpalpable, BI-RADS 5 lesions detected on mammography is 96% . If these lesions are classified as BI-RADS-MRI 5 as well, we expect the positive predictive value to increase further. In the 250 women who will undergo Breast MRI (intervention group), the positive and negative predictive values of Breast MRI in combination with mammography and ultrasound will be determined. If the predictive values are high (in certain subgroups, e.g patients with microcalcifications and dense breast lesions on mammography), LCNB could be omitted in the future in some women. Since all patients in the MONET – study will undergo LCNB, following clinical practice, the number of patients in whom LCNB could possibly be omitted in the future will be estimated based on the calculated predictive values of the MRI.