Fax: (216) 445-1068
Preoperative breast magnetic resonance imaging in early breast cancer†
Implications for partial breast irradiation
Article first published online: 17 FEB 2009
Copyright © 2009 American Cancer Society
Volume 115, Issue 8, pages 1621–1630, 15 April 2009
How to Cite
Tendulkar, R. D., Chellman-Jeffers, M., Rybicki, L. A., Rim, A., Kotwal, A., Macklis, R. and Obi, B. B. (2009), Preoperative breast magnetic resonance imaging in early breast cancer. Cancer, 115: 1621–1630. doi: 10.1002/cncr.24172
Presented as an oral presentation at the 49th Annual American Society for Therapeutic Radiology and Oncology (ASTRO) Meeting, October 28-November 1, 2007, Los Angeles, California.
- Issue published online: 6 APR 2009
- Article first published online: 17 FEB 2009
- Manuscript Accepted: 8 OCT 2008
- Manuscript Revised: 29 SEP 2008
- Manuscript Received: 25 AUG 2008
- breast cancer;
- magnetic resonance imaging;
- partial breast irradiation
Accelerated partial breast irradiation (APBI) of patients with early breast cancer is being investigated on a multi-institutional protocol National Surgical Adjuvant Breast and Bowel Project (NSABP) B-39/RTOG 0413. Breast magnetic resonance imaging (MRI) is more sensitive than mammography (MG) and may aid in selection of patients appropriate for PBI.
Patients with newly diagnosed breast cancer or ductal carcinoma in situ (DCIS) routinely undergo contrast-enhanced, bilateral breast MRI at the Cleveland Clinic. We retrospectively reviewed the medical records of all early-stage breast cancer patients who had a breast MRI, MG, and surgical pathology data at our institution between June of 2005 and December of 2006. Any suspicious lesions identified on MRI were further evaluated by targeted ultrasound ± biopsy.
A total of 260 patients met eligibility criteria for NSABP B-39/RTOG 0413 by MG, physical exam, and surgical pathology. The median age was 57 years. DCIS was present in 63 patients, and invasive breast cancer was found in 197 patients. MRI identified suspicious lesions in 35 ipsilateral breasts (13%) and in 16 contralateral breasts (6%). Mammographically occult, synchronous ipsilateral foci were found by MRI in 11 patients (4.2%), and in the contralateral breast in 4 patients (1.5%). By univariate analysis, lobular histology (infiltrating lobular carcinoma [ILC]), pathologic T2, and American Joint Committee on Cancer stage II were significantly associated with additional ipsilateral disease. Of patients with ILC histology, 18% had ipsilateral secondary cancers or DCIS, compared with 3% in the remainder of histologic subtypes (P = .004). No patient older than 70 years had synchronous cancers or DCIS detected by MRI.
Breast MRI identified synchronous mammographically occult foci in 5.8% of early breast cancer patients who would otherwise be candidates for APBI. Cancer 2009. © 2009 American Cancer Society.
Whole breast radiation therapy has been an accepted standard of care after breast conserving surgery for early stage breast cancer and ductal carcinoma in situ (DCIS). Accelerated partial breast irradiation (APBI) is currently being investigated in patients with early breast cancer or DCIS on a randomized multi-institutional protocol, the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-39/Radiation Therapy Oncology Group (RTOG) 0413. Patients eligible for this protocol include women aged 18 years or older, with unicentric invasive carcinoma or DCIS ≤ 3 cm in size, lymph node stage N0-N1, and negative surgical margins. The potential goals of APBI are to improve treatment-related adverse effects, cosmesis, quality of life, and convenience of care, while providing local tumor control comparable to treatment with whole breast radiotherapy. Early results of select APBI series have demonstrated favorable outcomes to date, however, long-term data are not yet available.
Breast magnetic resonance imaging (MRI) has emerged as a sensitive tool for evaluation of local breast cancer extent compared with conventional imaging methods of mammography and ultrasonography, and the role of MRI in patient management continues to evolve.1-5
At our institution, patients with newly diagnosed breast cancer undergo routine dynamic, contrast-enhanced MRI of the bilateral breasts as part of the standard preoperative workup. We hypothesized that breast MRI may aid in the selection of patients appropriate for APBI by identifying or ruling out additional mammographically occult, ipsilateral invasive cancers or DCIS. We initiated this retrospective study with the specific aim of defining the incidence of occult synchronous invasive or intraductal breast tumors detected by MRI among patients eligible for protocol NSABP B-39/RTOG 0413.
MATERIALS AND METHODS
With approval from our institutional review board, we retrospectively reviewed the medical records of all consecutive patients with a new diagnosis of invasive breast cancer or DCIS who had a breast MRI conducted at the Cleveland Clinic between June 2005 and December 2006.
Patients recently diagnosed with invasive breast cancer or DCIS by an abnormal screening mammogram or a palpable mass were included. All mammograms were conducted or reviewed by dedicated breast radiologists at our institution. Diagnostic mammography included standard craniocaudal and medial lateral oblique views, with spot compression and magnification views as clinically indicated. Both film mammography and digital mammography were used, depending on the site where the diagnostic study was performed. Supplemental ultrasound was conducted at the time of diagnostic mammogram when clinically indicated.
During the study period, all patients with a new diagnosis of invasive breast cancer or DCIS seen in the Cleveland Clinic Breast Center underwent dynamic, contrast-enhanced bilateral breast MRI in the prone position. Each MRI study was conducted with a 1.0 Tesla scanner (Harmony, Siemens Medical Systems; Erlangen, Germany), a dedicated breast-surface coil, and 1 series of images obtained before and 5 series of images obtained after the administration of contrast material, with 3-dimensional, T1-weighted, gradient-echo sequences. Initial and delayed images were obtained within 4 and 8 minutes after the injection of contrast material. Spatial resolution criteria included voxels smaller than 0.9 mm in the frequency-encoding direction, smaller than 1.8 mm in the phase-encoding direction, and 3 mm or smaller in the slice direction, providing full coverage of the breast. All procedures used 20 mL of bolus-injected gadopentetate dimeglumine (Magnevist; Berlex; Wayne, NJ) at 2 mL per second followed by a 20-mL saline flush. Initial coronal short T1 inversion recovery (STIR) images at 6 mm and axial STIR images at 4 mm were followed by the dynamic series. Dedicated breast radiologists interpreted all breast MRI procedures with the assistance of computer-aided detection on an independent workstation. CADstream software (Confirma, Inc.; Bellevue, Wash) was used before summer of 2006, and DynaCAD software (MRI Devices Corporation; Waukesha, Wis) was used thereafter. Any suspicious lesions identified on MRI were further evaluated by a targeted ultrasound, with a subsequent directed biopsy if the lesion was still considered suspicious by the radiologist. MRI-directed stereotactic biopsy was available if the lesion was not sonographically visible.
Patients included in this analysis were required to have a macroscopically unifocal tumor based on physical exam, mammography, and ultrasonography, with age ≥ 18 years, pathologic index tumor size ≤ 3 cm, no more than 3 involved lymph nodes, no distant metastases, and negative surgical margins. Patients were excluded from the study for any of the following reasons: 1) no mammogram available for comparison; 2) no surgical pathology data available; 3) multifocal, multicentric, bilateral, or diffuse disease based on physical exam or mammographic findings; 4) index tumor size > 3 cm; 5) involvement of skin or chest wall; 6) positive final surgical margin; 7) use of neoadjuvant chemotherapy; 8) >3 involved lymph nodes; 9) inadequate axillary staging; 10) prior ipsilateral breast cancer or DCIS; or 11) Paget disease of the breast. These criteria were selected based on the eligibility and exclusion criteria for the ongoing NSABP B-39/RTOG 0413 clinical trial.
Pathologic staging was defined per American Joint Committee on Cancer (AJCC) 2002 criteria. Pathologic tumor size was defined as the largest single histologic dimension noted on the pathology report. Histology was coded as either DCIS, infiltrating ductal carcinoma (IDC), infiltrating lobular carcinoma (ILC), or mixed/other carcinoma, which were comprised of infiltrating mammary carcinomas with both ductal and lobular features, mucinous carcinomas, and tubular carcinomas. Index lesions were defined as the mammographically visible, histologically confirmed primary cancer or DCIS lesion. A mammographically occult, MRI-detected lesion was defined as a synchronous lesion if it was both radiographically and histologically discontiguous from the index lesion. The radiographic size of the synchronous lesion and its distance from the index tumor (ie, the number of centimeters of intervening normal breast parenchyma between the nearest edges of the lesions) were calculated using the MRI. Menopausal status was obtained by review of the medical history.
Univariate analysis was conducted using the chi-square test to identify risk factors for biopsy-proven synchronous disease. The factors that were considered included age, menopausal status, clinical T stage, clinical N stage, primary tumor histology, pathological T stage, pathological N stage, overall pathological AJCC stage, estrogen receptor (ER) status, progesterone receptor (PR) status, Her2/neu expression, Bloom-Richardson grade, lymphovascular invasion (LVI), and extensive intraductal component (EIC). A meaningful multivariate analysis could not be conducted due the small number of events and a large number of potential risk factors, and thus low statistical power.
A total of 532 patients with a new diagnosis of nonmetastatic and noninflammatory breast cancer underwent bilateral breast MRI during the 1.5-year study period. Of these, 260 patients with early stage breast cancer and DCIS met the aforementioned eligibility criteria for analysis based on diagnostic mammogram/ultrasound, physical exam, and surgical pathology (exclusive of MRI findings).
Patient characteristics are summarized in Table 1. The median age was 57 years (range, 33-86 years). Postmenopausal women comprised 68% of the cohort. The primary tumor was palpable in 77 cases (30%). The median time from initial biopsy-proven diagnosis of the index breast cancer or DCIS to conduction of the MRI study was 12 days. Partial mastectomy was conducted in 77%, whereas 23% elected mastectomy either by patient preference or when clinically indicated by MRI demonstrating additional suspicious disease.
|Median [range]||57 [33-86 y]||—|
|Breast with index cancer|
|Fine needle aspiration||3||1|
|Surgical resection type|
The index tumor characteristics are summarized in Table 2. The majority of cases were IDC (60%), followed by DCIS (24%), mixed/other histologies (9%), and ILC (7%). The majority of cases were positive for estrogen (82%) and progesterone (60%) receptors by immunoperoxidase staining, and HER2/neu was overexpressed in 14% by fluorescence in situ hybridization.
|Histology of index lesion|
|Pathologic T stage|
|Pathologic N stage|
|Pathologic AJCC stage|
|Invasive cancer characteristics|
|Her2/neu receptor status|
Breast MRI findings are shown in Figures 1 and 2. MRI identified an additional suspicious abnormality in 35 (13%) ipsilateral breasts and 16 (6%) contralateral breasts. Close observation of the ipsilateral breast was recommended in cases of a targeted ultrasound or percutaneous biopsy demonstrating benign findings. In 12 cases, the patient directly proceeded to surgical evaluation of the suspicious lesion (wide excision or mastectomy), which demonstrated biopsy-proven synchronous foci in 3 patients. Overall, a histologic diagnosis of synchronous ipsilateral invasive breast cancer or DCIS was identified in 11 patients (4.2%). In the contralateral breast, a percutaneous biopsy demonstrated 4 cases (1.5%) of biopsy-proven synchronous breast cancer or DCIS. Overall, 15 patients (5.8%) were found by MRI to have macroscopic, mammographically occult synchronous tumors. Among all the suspicious abnormalities observed on MRI, the overall yield of detecting biopsy-proven synchronous foci was 29%.
The characteristics of the biopsy-proven, mammographically occult, synchronous lesions are summarized in Table 3. These ranged in size from 0.5 to 3.3 cm by MRI, with no invasive cancers greater than 2 cm in size detected. The synchronous lesions occurred at a median distance of 3.0 cm of intervening tissue between lesions on MRI (range, 1.0 to 5.4 cm). In addition, there was 1 case of high-grade DCIS in multiple quadrants throughout the breast, which was not detected mammographically. The incidence of finding ipsilateral or contralateral disease by various risk factors are summarized in Table 4. By univariate analysis, index tumor factors significantly associated with finding occult ipsilateral disease included ILC histology (P = .043), pathologic T2 stage (P = .038), and pathologic AJCC stage II (P = .002). Pathologic lymph node positivity was of borderline statistical significance (P = .06), whereas age, menopausal status, clinical T stage, ER status, PR status, Her2/neu expression, Bloom-Richardson grade, presence of LVI, and presence of EIC were not found to be significantly different. In finding contralateral synchronous tumors, only pathologic T2 stage of the index lesion was statistically significant (P = .048).
|Patient||Age, y||Index Tumor Size||Index Tumor Histology||Synchronous Tumor Histology||Synchronous Tumor Distance From Index||Synchronous Tumor Size on MRI|
|1||42||3.0 cm||IDC||IDC||2.2 cm||1.4 cm|
|2||58||1.2 cm||ILC||ILC||2.7 cm||1.0 cm|
|4||58||2.2 cm||ILC||ILC||3.5 cm||1.3 cm|
|5||49||1.8 cm||ILC||ILC||1.0 cm||0.7 cm|
|6||54||1.2 cm||IDC||IDC||3.7 cm||1.1 cm|
|7||63||2.4 cm||IDC||IDC||3.0 cm||0.7 cm|
|8||66||1.8 cm||IDC||IDC||3.0 & 1.8 cm||1.3 & 1.5 cm|
|9||51||2.5 cm||IDC||IDC||5.4 cm||1.5 cm|
|10||65||0.7 cm||IMC||DCIS||5.3 cm||3.3 cm|
|11||54||2.5 cm||DCIS||DCIS||1.0 cm||1.3 cm|
|12||62||1.5 cm||IDC||IDC||CL||0.5 cm|
|13||52||1.4 cm||IDC||IDC||CL||1.5 cm|
|14||62||2.5 cm||IDC||IDC||CL||0.9 cm|
|15||55||3.0 cm||IDC||DCIS||CL||2.1 cm|
|Ipsilateral Biopsy-proven Tumor||%||Contralateral Biopsy-proven Tumor||%|
|Clinical T stage||P=.64||P=.21|
|Pathological T stage||P=.038||P=.048|
|Pathological N stage||P=.06||P=.79|
|Pathological AJCC stage||P=.002||P=.29|
A synchronous ILC was discovered in the ipsilateral breast in 3 of 17 (18%) cases with a primary ILC, significantly higher than that found in the aggregate of non-ILC histologies (3%, P = .004). Ipsilateral synchronous tumor foci were found in 2 of 63 (3%) cases of DCIS, 5 of 157 (3%) cases of IDC, and 1 of 23 (4%) cases of mixed/other invasive histologies. Among the 4 patients diagnosed with a synchronous contralateral breast cancer, all occurred in patients with an index lesion consisting of IDC. None of the 47 patients aged 70 years or older was diagnosed by MRI with a mammographically occult synchronous ipsilateral or contralateral tumor.
The role of breast MRI in the management of breast cancer patients continues to evolve. Breast MRI may be used to screen women with a high lifetime risk of breast cancer, evaluate suspicious lesions seen with conventional imaging, identify an unknown primary presenting with axillary metastases, monitor response to neoadjuvant chemotherapy, and determine the extent of high-grade DCIS and ILC that are less well-observed by mammography.1-5 In addition, MRI has greater sensitivity than mammography in detecting multifocal or multicentric breast cancers, especially in patients with dense breasts.5-8
However, the routine use of MRI for determining local therapy of biopsy-proven early stage breast cancer remains controversial. The rationale for pretreatment breast MRI is that accurate delineation of tumor extent may allow the surgeon to achieve a negative-margin resection with a single operation. This has been the standard practice at our institution with rare exceptions, and thus our series represents a relatively unselected population of breast cancer patients. Published series suggest that pretreatment MRI can identify breast tumor foci in addition to the index lesion in 15% to 37% of ipsilateral breasts and in 3% to 5% of contralateral breasts,5, 6, 9-14 which ultimately alters the clinical management in 10% to 31% of cases.9, 10, 15-21
Our analysis suggests that approximately 4% of our patients eligible for APBI according to the ongoing NSABP B-39/RTOG 0413 protocol have macroscopic ipsilateral synchronous foci of invasive cancer or DCIS that are only detected by MRI when used in addition to conventional imaging methods. The synchronous tumors tended to be small, with no invasive tumor greater than 2 cm in size, and at a median distance of 3 cm away from the index lesion. In concordance with other series,22, 23 most of these occult lesions were in relative proximity to the index lesion. Whether these occult lesions would have been identified or removed intraoperatively, or identified pathologically by finding a positive margin, would depend on the extent of resection. Regardless, these additional lesions could at least theoretically serve as a nidus for ipsilateral breast tumor recurrence (IBTR). Studies have demonstrated that patients with macroscopically multicentric cancers treated with breast conservation have a higher rate of IBTR than those with unifocal disease.24, 25
Special consideration must be given to patients with ILC, in whom there was a higher rate of mammographically occult, synchronous ipsilateral lesions compared with the other histologies. It has been documented that this histologic subtype is much more accurately observed on MRI than with mammography.5, 6, 12, 26, 27 It may be argued that given the frequent pattern of multicentricity and bilaterality of ILC in general, these patients may not be ideal candidates for APBI. However, patients with ILC are not specifically excluded from the NSABP B-39/RTOG 0413 protocol. Our data suggest that, if APBI is being considered for an individual patient with ILC, a pretreatment breast MRI may prove valuable in excluding additional disease.
Although pathologic T2 stage and AJCC stage II were also associated with synchronous cancers, these factors would not aid in preoperative decision making as the final pathologic stage would still be unknown. Pretreatment clinical tumor staging was not beneficial in our population.
EIC is an established risk factor for residual disease in the breast and for the development of local recurrence after breast conservation therapy.28, 29 There were 19 patients in our cohort with EIC, none of whom were found to have an occult ipsilateral synchronous lesion.
Women younger than 50 years of age were not at higher risk of occult synchronous lesions compared with those between 50 and 70 years. Notably, none of the 47 women aged 70 or older had additional lesions in either breast on MRI. Despite the relatively low number of older women in our cohort, this observation may suggest that clinical evaluation and traditional imaging modalities of mammogram and ultrasound may be sufficient in older women. Further recommendations for the utility of MRI in this population based on age criteria would likely require a prospective evaluation.
Our results represent an unselected cohort of early breast cancer patients at a single large institution, specifically examining the role of preoperative MRI in patients eligible for the NSABP B-39/RTOG 0413 trial. The 5.8% rate of finding additional ipsilateral or contralateral breast lesions is somewhat lower than the 10% to 16% quoted by others using similar eligibility criteria.30-32 This finding may be explained by several potential reasons. First, our institution routinely conducts breast MRI in virtually all patients with newly diagnosed breast cancer, as opposed to reserving it for patients in whom additional disease may already be suspected. Second, by only including the relatively favorable subset of early breast cancer and DCIS patients being studied on NSABP B-39/RTOG 0413, multicentric or multifocal disease in this cohort may be less likely compared with the overall population of breast cancer patients. Third, given the retrospective nature of this study, a detailed pathologic assessment of the entire tissue specimen was not routinely conducted. Finally, some may argue that our use of a 1.0 Tesla magnet may lack the sensitivity of a higher magnet strength MRI unit and may yield more false-negative results. On these grounds, we believe our detection rate of mammographically occult lesions by MRI (4.2% in the ipsilateral breast, 1.5% in the contralateral breast) represents the baseline minimum incidence of synchronous breast cancers or DCIS that would be expected in this cohort of patients.
In conclusion, pretreatment breast MRI should be considered for select patients being treated with breast conservation therapy, especially in cases of ILC histology. MRI may identify a small but potentially meaningful number of patients with occult breast cancer or DCIS at some distance away from the index lesion. Defining the exact population of patients who would benefit most from pretreatment breast MRI may require a prospective evaluation.
Conflict of Interest Disclosures
The authors made no disclosures.
- 9Clinical utility of breast magnetic resonance imaging in patients presenting with primary breast cancer. Proc (Bayl Univ Med Cent). 2007; 20: 227-230., , .
- 32Pre-operative breast magnetic resonance imaging (MRI) influences patient selection for partial breast irradiation [abstract]. J Clin Oncol. 2006; 24: 604., , , , , .