Magnetic resonance imaging (MRI) has been used to supplement screening mammography and clinical breast examination (CBE) in women who are at high risk of developing breast cancer. In this study, the authors investigated the efficacy of alternating screening mammography and breast MRI every 6 months in women who had a genetically high risk of developing breast cancer.
A retrospective chart review was performed on all women who were seen in a high-risk breast cancer clinic from 1997 to 2009. Patients with breast cancer gene (BRCA) mutations who underwent alternating screening mammography and breast MRI every 6 months were included in the study. Mammography, ultrasonography, MRI, and biopsy results were reviewed.
Of 73 patients who met the study criteria, 37 had BRCA1 mutations, and 36 had BRCA2 mutations. Twenty-one patients (29%) completed 1 cycle of mammography and MRI surveillance, 23 patients (31%) completed 2 cycles, 18 patients (25%) completed 3 cycles, and patients 11 (15%) completed ≥4 cycles. The median follow-up was 2 years (range, 1-6 years). Thirteen cancers were detected among 11 women (15%). The mean tumor size was 14 mm (range, 1-30 mm), and 2 patients had bilateral cancers. Twelve of 13 cancers were detected on an MRI but not on the screening mammography study that was obtained 6 months earlier. One cancer (a 1-mm focus of ductal carcinoma in situ) was an incidental finding in a prophylactic mastectomy specimen. One patient had ipsilateral axillary lymphadenopathy identified on ultrasonography but had no evidence of lymph node involvement after neoadjuvant chemotherapy and surgery.
Women who carry a breast cancer gene (BRCA) mutation have a lifetime risk as high as 60% to 90%1-4 of developing breast cancer. These women also are at increased risk of developing ovarian cancer or a second breast cancer after the first breast cancer, and they tend to develop cancer at a younger age than women who do not have a gene mutation.1 In addition, the tumor volume doubling time is shorter in patients with breast cancer who have a BRCA mutation than in breast cancer patients without a gene mutation.2 Options available for reducing the risk of invasive breast or ovarian cancer in women with a BRCA mutation include intense imaging surveillance, chemoprevention with tamoxifen, bilateral prophylactic mastectomy, and prophylactic oophorectomy.3 Mammography has been the primary imaging modality used for breast cancer screening over the past 30 years. The use of screening mammography in the general population has reduced the mortality rate associated with breast cancer by 40% to 45%.4 However, annual screening mammography has low sensitivity for the detection of breast cancer in women who are at high risk of developing breast cancer. The dense breast tissue in younger women may be 1 reason for mammography's low sensitivity in this group, with the result that breast cancer often is detected at a later stage.4-11
Contrast-enhanced magnetic resonance imaging (MRI) has greater sensitivity than mammography for detecting breast cancer (71%-100% vs 16%-40%, respectively)12-17 and, thus, has been introduced to supplement mammography and clinical breast examination (CBE) for the screening of women who are at high risk of developing breast cancer from genetic mutations. Although the specificity of MRI is lower than that of mammography, resulting in a higher false-positive rate with MRI, recent advances in breast MRI techniques may improve the specificity. Indeed, 1 study demonstrated that the specificity of MRI was not significantly different from that of mammography.12
In previously published studies12-17 on screening high-risk women, mammography and MRI studies were obtained annually, with MRI performed within 90 days of mammography. Because previous research indicated that the rate of interval cancers may be greater in BRCA mutation carriers, alternating mammography and MRI every 6 months has been proposed to facilitate early detection of breast cancer. Therefore, we investigated the efficacy of alternating screening mammography and breast MRI every 6 months in women who had a genetically high risk of developing breast cancer.
MATERIALS AND METHODS
This was a retrospective, single-institution study of patients who attended the high-risk breast cancer clinic at The University of Texas MD Anderson Cancer Center between January 1997 and March 2009. These patients were entered in a prospective research database. The institutional review board approved this Heath Insurance Portability and Accountability Act-compliant study, and a waiver of informed consent was obtained. To be included in this study, patients were required to be at least 18 years of age, to have undergone alternating screening mammography and breast MRI every 6 months at our institution, and either to be confirmed carriers of a BRCA1 or BRCA2 mutation or to be the first-degree relative of a BRCA1 or BRCA2 carrier. Women who had a history of unilateral breast cancer, who had a calculated lifetime risk of breast cancer >20%, or who did not undergo a screening MRI were excluded. Women who did not choose screening mammography or MRI but, instead, chose chemoprevention or bilateral prophylactic mastectomies also were excluded. Other exclusion criteria are listed in Table 1. All available radiologic studies (mammography, ultrasonography [US], and MRI) were reviewed by 2 radiologists (H.T.L. and G.J.W). The overall mammographic breast density was determined according to the American College of Radiology (ACR) Breast Imaging Reporting and Data System (BI-RADS) density categories18: Type 1, the breast is entirely fatty; Type 2, the breast has scattered, fibroglandular densities; Type 3, the breast is heterogeneously dense; or Type 4, the breast is extremely dense. Mammographic and sonographic abnormalities were recorded using the BI-RADS lexicon.18 The MRI findings and the corresponding kinetic curve assessments were recorded using the terminology described in the ACR BI-RADS MRI lexicon.18 The clinical and pathologic records of all patients who chose alternating screening mammography and MRI at 6-month intervals were reviewed by the radiologists (H.T.L. and G.J.W), whereas the records for the remaining patients were reviewed by the medical oncologist and research team (B.K.A., D.P.A., and A.G-.B.).
Table 1. Patients Who Were Excluded From the Study
Reason for Exclusion
No. of Patients (%)
Magnetic resonance imaging could not be performed because of the patient's body habitus and/or weight.
In the majority of patients, the screening mammography was obtained by using a Lorad M3 unit (Hologic, Inc., Bedford, Mass) or a General Electric DMR unit (General Electric Medical Systems, Milwaukee, Wis). The screening mammography examination consisted of at least 2 views, including craniocaudal and mediolateral views. Additional mammographic views were obtained as clinically indicated.
Real-time gray-scale and color-Doppler US studies were obtained to evaluate abnormal mammographic or MRI findings, as recommended in the mammography report or the MRI report, using a Siemens Elegra or Antares unit (Siemens Medical Solutions, Issaquah, Wash) with a 13.5-MHz linear-array transducer. The lesions were scanned in the transverse and sagittal planes.
Screening MRI studies were obtained with the patient prone on a commercially available, 1.5-Tesla system (Signa EXCITE HD; GE Healthcare, Waukesha, Wis) using either a dedicated 7-channel breast coil (InVivo Corp., Orlando, Fla) or an 8-channel, high-density breast array coil (GE Healthcare). Patients were imaged in the bilateral mode with the following sequences: 1) localizer pulse sequence; 2) sagittal, fat-suppressed, T2-weighted, fast-spin-echo pulse sequence (repetition time/echo time [TR/TE] = 4000/85 msec); 3) sagittal, dynamic contrast-enhanced sequence; and 4) delayed, axial, contrast-enhanced, T1-weighted, 3-dimensional (3D) fast spoiled gradient-echo pulse sequence. The dynamic sequences consisted of 1 sequence before and 4 sequences after the bolus injection of gadopentetate dimeglumine (0.1 mmol/L per kilogram of body weight; Magnevist, Bayer Healthcare Pharmaceuticals, Wayne, NJ), which was administered using an MRI-compatible injector (Spectris; Medrad Inc., Pittsburgh, Penn), followed by a 20-mL saline flush. The T1-weighted, 3D, fast spoiled gradient-echo pulse sequence (Volume Imaging for Breast Assessment [VIBRANT]; GE Healthcare) uses the following scan parameters: flip angle = 15°, receiver bandwidth = ±50 ± 62.5 kHz, TR/TE = 6/2 msec, acquisition matrix = 256 × 256, field of view = 16 cm to 22 cm, and slice thickness = 0.9 mm to 3.0 mm. Parallel imaging with an acceleration factor of 2 was used to reduce the total scan time, which was kept to less than 2 minutes per temporal phase in all studies. Fat suppression for the VIBRANT sequence was achieved using a spectral inversion pulse. Time-signal intensity curves were used to characterize the enhancement pattern of lesions.18 MRI findings were described using the BI-RADS breast MRI lexicon.18
Statistical analysis for this retrospective chart review was mainly descriptive. We calculated means and standard deviations for continuous variables and frequency for categorical variables. To determine the sensitivity and specificity for mammography or MRI, we defined true-negative (TN) results as those from participants who had negative mammography and MRI studies and demonstrated no evidence of malignancy on a follow-up clinical and imaging examinations 12 months later. We defined false-positive (FP) results as those from participants who had suspicious findings on screening mammography or MRI examinations that resulted in further imaging or a biopsy with a final benign assessment or patients who had a follow-up clinical and imaging examination either 6 months or 1 year later with benign results. Sensitivity was computed using the formula (true positive [TP])/(TP + false negative [FN]). Specificity was computed as TN/(FP + TN). The cancer-detection rate or overall cancer yield rate was defined as TP results/number of screening participants. One cycle of surveillance was defined as 1 mammogram and 1 MRI examination (roughly 6 months apart from each other) with no evidence of cancer at 1-year follow-up. Two cycles of screening were defined as alternating mammography and MRI at 6-month intervals for 2 years with 1-year follow-up after the last MRI examination.
The records of 321 women with BRCA1 or BRCA2 mutations who attended the high-risk breast cancer clinic from January 1, 1997, to March 31, 2009, were reviewed. Of these women, 73 who received alternating screening mammography and breast MRI every 6 months and met all other inclusion criteria were identified and considered as the study group in this report. The records of these 73 women were analyzed. The remaining 248 women were excluded from the study because of the factors listed in Table 1.
Of the 73 women, 64 women (88%) were white, 5 women (7%) were African American, 3 women (4%) were Hispanic, and 1 woman (1%) was Asian. Thirty-seven women (51%) had BRCA1 mutations, and 36 women (49%) had BRCA2 mutations. The median age at the time of the initial consultation was 44 years (range, 23-75 years) (Table 2).
Table 2. Characteristics of Women Who Were Included in the Study (n=73)
No. of Patients (%)
BRCA1 indicates breast cancer 1 gene.
Age; Median [range], y
No. of surveillance cycles completed
Race or ethnic group
All 73 women had a clinical breast examination every 6 months. During the study period, 21 women (29%) completed 1 cycle of screening, 23 women (31%) completed 2 cycles, 18 women (25%) completed 3 cycles, 9 women (12%) completed 4 cycles, and 2 women (3%) completed 5 or 6 cycles. The median number of screening cycles was 2 (range, 1-6 cycles).
On the basis of the first screening mammography study examination performed during the study period, breast density was categorized as extremely dense in 7 patients (10%), heterogeneously dense in 41 patients (56%), and scattered fibroglandular tissue in 25 patients (34%). Of the 73 women who received alternating screening mammogram and breast MRI every 6 months, 70% had undergone bilateral prophylactic oophorectomy at the time of the chart review process. The remaining 26% received screening serum cancer antigen 125 (CA 125) and pelvic ultrasound, 3% had no available gynecologic history, and 1% developed metastatic breast cancer with no gynecologic history.
Thirteen cancers (7 invasive ductal carcinomas, 3 invasive lobular carcinomas, and 3 ductal carcinomas in situ [DCIS]) were detected in 11 patients; 2 of those women had bilateral cancers (Table 3). Twelve of 13 cancers were identified on MRI but had not been identified on a mammography study that was obtained 6 months earlier. Five of the 13 cancers (38%) were identified on both the screening MRI examination and the subsequent targeted US examination. One cancer was identified only on pathology after prophylactic left mastectomy in a patient who underwent a right mastectomy for the treatment of a screening MRI-detected invasive lobular carcinoma. In this patient, the left breast cancer was a 1-mm focus of DCIS. No cancers were detected only by screening mammography. None of the 13 cancers were palpable at the 6-month physical examination. Five of the 11 patients with screen-detected cancers had their cancer detected after 1 cycle of screening, 2 patients had cancer detected after 2 cycles, 3 patients had cancer detected after 3 cycles, and 1 patient had cancer detected after 4 cycles.
Table 3. Characteristics of the Patients Who Had Cancers Detected (n=11)
Mammogram 6 Months Before MRI
BRCA1 indicates breast cancer 1 gene; BRCA2, breast cancer 2 gene; MRI, magnetic resonance imaging; US, ultrasound; scattered, scattered fibroglandular tissue; IDC, invasive ductal carcinoma; DCIS, ductal carcinoma in situ; Hetero, heterogeneously dense; ILC, invasive lobular carcinoma.
This patient underwent prophylactic left mastectomy and had an incidental finding of a 1-mm focus of DCIS.
This patient chose undergo prophylactic left mastectomy instead of image-guided biopsy of the 5-mm lesion that was visible on MRI.
The mean size of the tumors on MRI was 14 mm (range, 1-30 mm). The most common morphologic MRI characteristics, which were described in 6 of 13 tumors (46%), was nonmass-like enhancement in a linear or focal distribution (Fig. 1) with clumped internal enhancement, according to the ACR BI-RADS MRI lexicon.18 The other tumors presented as irregular, enhancing breast masses with spiculated margins (Fig. 2) and heterogeneous, internal enhancement or rim enhancement (4 patients; 31%), a lobular enhancing mass with irregular margins and heterogeneous internal enhancement (1 patient; 8%), and linear or ductal enhancement (1 patient; 8%). One tumor, a 1-mm focus of DCIS (8%), was not detected on MRI or mammography from 6 months earlier, but it was detected on final pathology after the patient underwent prophylactic mastectomy.
The right breast was involved in 5 women (46%), the left breast was involved in 4 women (36%), and both breasts were involved in 2 women (18%). An enlarged ipsilateral axillary lymph node was identified at the time the breast cancer was diagnosed in 1 patient who had a biopsy-confirmed, metastatic adenopathy; she had invasive ductal carcinoma that was negative for estrogen receptor, progesterone receptor, and human epidermal growth receptor 1 (HER2) (triple-negative breast cancer).
Of the 11 women who had breast cancers diagnosed, the mean age at the time of diagnosis was 51 years (range, 43-64 years). Seven of 13 cancers (54%) were triple-negative. Three cancers (23%) were positive for estrogen receptor, 2 cancers (15%) were positive for progesterone receptor, and 2 (15%) were positive for HER2-neu. Both contralateral carcinomas that were diagnosed at the time of prophylactic mastectomies were DCIS that measured <1 cm in greatest dimension, and no tumor marker results were available for those 2 lesions.
Nine of the 11 women ho were diagnosed with breast cancer chose to undergo bilateral mastectomies. Two cancers were detected in the contralateral prophylactic mastectomy specimens and measured 1 mm and 5 mm (both were DCIS). The 5-mm DCIS was detected on MRI but not on a mammography study that was obtained 6 months earlier. This patient declined image-guided biopsy and chose to undergo bilateral mastectomies without a preoperative diagnosis of a contralateral breast cancer. Nine of the 11 patients underwent sentinel lymph node biopsy at the time of mastectomy and had 1 to 4 sentinel lymph nodes removed. All of the sentinel lymph nodes were negative for malignancy. One patient had a 3-cm, nonpalpable, malignant-appearing lymph node on staging US, but no positive lymph nodes were identified at surgery, probably because the patient received preoperative neoadjuvant chemotherapy. This was the only patient who received neoadjuvant chemotherapy. No patient had other associated findings, such as chest wall involvement, nipple or skin retraction, contralateral axillary lymphadenopathy, or distant metastasis.
Twenty of 73 women (27%) women underwent biopsy, which identified 11 breast cancers in 10 women. One woman underwent MRI-guided biopsy of a suspicious MRI finding that had a preoperative diagnosis of focal atypical hyperplasia. This patient declined additional biopsy and proceeded directly to bilateral prophylactic mastectomies without a preoperative diagnosis of invasive breast cancer. The final pathology at mastectomy revealed a 12-mm, invasive ductal carcinoma with associated DCIS in the same breast as the MRI finding. The overall biopsy yield for MRI was 50% (10 of 20 women). Five of the 10 biopsies that yielded a breast cancer diagnosis were performed under MRI guidance, and the other 5 biopsies were performed under US guidance.
Eight of 73 women had FP findings on MRI, requiring an additional imaging workup with targeted US examination in 4 women, targeted US examination followed by MRI-guided biopsy in 2 women, and only an MRI-guided biopsy without targeted US examination in 2 women. An FP finding on mammography was observed in 15% of women (n = 11) and required further evaluation with US in 11%(n = 8) and stereotactic biopsy in 4% (n = 3). The mean follow-up for the women with FP findings was 1.7 years (range, 1-3 years) from the time of the examination that produced a suspicious imaging finding.
The sensitivity of MRI in this study was 92% (95% confidence interval [CI], 0.76%-1.00%), and the specificity was 87% (95% CI, 0.79%-0.95%). For the current study, the sensitivity of mammography could not be determined, and the specificity was 82% (95% CI, 0.72%-0.92%). No interval cancers were detected on mammography studies that were obtained between 2 breast MRI examinations.
Seventy-three women with confirmed genetic mutations (BRCA1 or BRCA2) underwent imaging screening with alternating mammography and MRI every 6 months, which resulted in an overall cancer yield of 15%. Our cancer yield rate is higher than the reported overall cancer yield of 3% to 6% in multi-institutional trials.15, 19-22 This higher cancer yield may be because of our patient population, which included only BRCA mutation carriers, who have a greater risk of developing breast cancer than women without genetic risk factors. Five of the 11 cancers were detected on the first cycle of screening MRI and probably were a component of prevalent cancers.
In published trials,15, 19-22 the screening mammography and MRI studies were obtained at a single time point. The effect of the alternating annual mammography and MRI screening regimen versus a nonalternating regimen on the overall cancer yield needs to be assessed by a randomized controlled trial comparing the 2 screening regimens. At our facility, the practice of alternating annual mammography and MRI every 6 months with a physical examination every 6 months is an institutional practice. There was no control group of women who underwent annual mammography and MRI at the same time point during the study period. When mammography and MRI studies were obtained at the same time point, the interval cancer rate was <3%.12-13, 15 In view of the already low interval cancer rate, the alternating screening regimen is unlikely to significantly lower the interval cancer rate. However, there may be psychological benefits for patients who undergo a screening test every 6 months (mammography alternating with MRI) rather than every 12 months. The shorter screening interval may enable women to feel more comfortable and less anxious regarding the screening process.
In the patients who were diagnosed with breast cancer during the screening period, the mammography study that was obtained 6 months before diagnosis was either benign or negative. Unfortunately, mammography was not repeated in the majority of these women after MRI detected a suspicious finding. Only 4 of the 11 women with cancer had mammography studies obtained after a screening MRI detected cancer; 2 of those 4 women demonstrated a mammographic abnormality that corresponded to the screening MRI-detected cancer (microcalcifications). Therefore, the difference in sensitivity between mammography and MRI could not be calculated. However, reports comparing the sensitivity of mammography with that of MRI have been published by other authors, and the combination of mammography with MRI offered higher sensitivity than either imaging modality alone.12 To our knowledge, we are the first to report findings of alternating screening mammography and MRI at 6-month intervals in combination with physical examination. Despite adding MRI to mammography screening, 1 woman had lymph node-positive disease identified at initial diagnosis in our study. Lymph node-positive disease at presentation was observed in 1 of 11 patients (9%) in our study, which is less than the reported 14% to 16% of patients in 2 other series.12, 15 The cumulative risk of contralateral breast cancer in patients who have breast cancer with BRCA mutations reportedly is 16% to 29.5%.23, 24 In our study, contralateral breast cancer was observed in 2 of 13 patients (15%) who had cancer.
Because this was a retrospective study, we did not have a control population. Also, because mammography studies were not obtained after cancer was detected on MRI in each patient, the sensitivity and specificity of mammography could not be determined. In addition, we did not assess the psychological effects of breast cancer screening every 6 months. A future multi-institutional trial not only could evaluate the psychological effects of screening every 6 months but also could evaluate breast density as 1 of the objectives in the detection of breast cancer on MRI, because high breast density has been suggested as a risk factor for breast cancer.25, 26
The use of US as a screening examination was not evaluated in the current study. However, previously published data indicate that US has lower sensitivity (range, 33%-44%) than MRI (range, 77%-91%) for the detection of breast cancer and that the addition of US to screening provides no additional benefit over screening with mammography and MRI.12, 13 However, we frequently use US to further evaluate suspicious breast lesions identified on MRI in our practice despite a report that US detected only approximately half (46%) of the lesions observed by MRI.23 Our findings were similar, with second-look or targeted US detecting only 36% of the cancers that were detected on MRI. The follow-up in our study was 2 years.
The sensitivity and specificity for MRI observed in our study were within the reported range for sensitivity of breast MRI.12-15 The rate of FP findings on MRI (11%) was lower than the rate of FP findings on mammography (15%) in our study. We speculate that this was because higher spatial resolution images, advances in MRI technology (such as improvement in breast coils), and a more standardized imaging protocol and high-field magnetic system in the current study than were available in earlier published studies.
Imaging screening typically has been performed on an annual basis in patients who have a genetic predisposition for breast and ovarian cancers. However, in published studies, lymph node-positive disease and interval cancers still were identified in women with BRCA1 mutations who had annual mammography and MRI screenings and underwent prophylactic oophorectomy.23, 27 In 2009, the US Preventive Services Task Force published a recommendation of biennial screening mammography in women ages 50 to 74 years24; however, that recommendation does not apply to high-risk patients like those in our study, and it does not apply to screening MRI.
In conclusion, the results from this retrospective descriptive study suggest that alternating MRI and mammography for breast cancer screening may be an option for women who are at increased risk of breast cancer. Larger multi-institutional trials comparing annual and semiannual screening intervals are needed to determine the optimal protocols for women who are at genetically high risk of developing breast cancer.
CONFLICT OF INTEREST DISCLOSURES
Dr. Litton has worked as a PI for Novartis and Bi-Par Sciences, but has not received honoraria for consultant services. Dr. Hortobagyi has worked as a consultant for Novartis, Merck, and Sanofi Aventis.