Outcome of High-Risk Lesions at MRI-Guided 9-Gauge Vacuum- Assisted Breast Biopsy
Abstract
OBJECTIVE. The purposes of this study were to determine the frequency of underestimation of high-risk lesions at MRI-guided 9-gauge vacuum-assisted breast biopsy and to determine the imaging and demographic characteristics predictive of lesion upgrade after surgery.
MATERIALS AND METHODS. We retrospectively reviewed consecutively detected lesions that were found only at MRI and biopsied under MRI guidance from May 2007 to April 2012. Imaging indications, imaging features, and histologic findings were reviewed. The Fisher exact test was used to assess the association between characteristics and lesion upgrade. Patients lost to follow-up or who underwent mastectomy were excluded, making the final study cohort 140 women with 151 high-risk lesions, 147 of which were excised.
RESULTS. A database search yielded the records of 1145 lesions in 1003 women. Biopsy yielded 252 (22.0%) malignant tumors, 184 (16.1%) high-risk lesions, and 709 (61.9%) benign lesions. Thirty of the 147 (20.4%) excised high-risk lesions were upgraded to malignancy. The upgrade rate was highest for atypical ductal hyperplasia, lobular carcinoma in situ, and radial scar. No imaging features were predictive of upgrade. However, there was a significantly higher risk that a high-risk lesion would be upgraded to malignancy if the current MRI-detected high-risk lesion was in the same breast as a malignant tumor previously identified in the remote history, a recently diagnosed malignant tumor, or a high-risk lesion previously identified in the remote history (p = 0.0001). The upgrade rate was significantly higher for women with a personal cancer history than for other indications combined (p = 0.0114).
CONCLUSION. The rate of underestimation of malignancy in our series was 20%. No specific imaging features were seen in upgraded cases. Surgical excision is recommended for high-risk lesions found at MRI biopsy and may be particularly warranted for women with a personal history of breast cancer.
Breast MRI is increasingly used to screen women at high risk of breast cancer and for a variety of diagnostic purposes, including evaluation of women with newly diagnosed breast cancer, imaging problem solving, and postoperative evaluation [1–4]. As more women undergo MRI, the number of lesions detected only on MR images and the number of resultant MRI-guided biopsies are also increasing [5]. High-risk lesions—including atypical ductal hyperplasia (ADH), atypical lobular hyperplasia (ALH), lobular carcinoma in situ (LCIS), papillary lesions, radial scar, and flat epithelial atypia—and mucocelelike lesions and phyllodes tumors are not infrequently found after MRI-guided biopsy.
Multiple studies of MRI vacuum-assisted biopsy (VAB) results have shown high-risk lesion incidence and upgrade rates as a subset group of all biopsy results [5–12]. In addition, few studies have been conducted to directly evaluate high-risk lesions initially detected at MRI [13–15]. These studies were retrospective and often yielded relatively small numbers of each type of lesion. Most work is united in showing no distinct morphologic or kinetic characteristics predictive of high-risk pathologic features or upgrade of a high-risk lesion to malignancy. Studies have also shown that ADH is the high-risk lesion most commonly found at initial biopsy and may have the highest upgrade rate [13].
Management of certain high-risk lesions of the breast remains controversial, but surgical excision has usually been recommended for high-risk lesions diagnosed at percutaneous stereotactic or sonographic breast biopsy because of the potential for a lesion upgrade to malignancy. As has been well discussed in the literature, multiple factors are thought to contribute to such underestimation, including differences in pathologic interpretation, sampling error, and the type of biopsy device. Appropriate management of high-risk lesions detected and biopsied with MRI is not always clear; recommendations are based on the in itself controversial ultrasound and mammographic literature [16, 17].
The primary purpose of this retrospective study was to evaluate the imaging characteristics and features of high-risk lesions on MR images and to provide further data on the frequency and upgrade of high-risk lesions detected at MRI and biopsied under MRI guidance. Another purpose was to examine whether any imaging or demographic factor may be predictive of upgrade to malignancy after excision.
Materials and Methods
Case Selection
Institutional review board approval was obtained for this HIPAA-compliant study. The requirement for informed consent for this study was waived by the institutional review board at our cancer center. We performed a retrospective review of our database of all consecutive MRI examinations performed from May 2007 through April 2012 and identified examinations that led to MRI-guided biopsy. All lesions were initially detected at MRI. We excluded women who were lost to follow-up, who underwent less than 2 years of imaging follow-up, and who did not undergo surgical excision after biopsy. Women with more than 2 years of lesion stability on images were assumed to have benign findings for the purpose of analysis. We also excluded all women who underwent mastectomy without imaging-guided localization of the lesion identified at MRI.
MRI Acquisition Parameters
All MRI examinations were performed with a 3-T system (Tim Trio, Siemens Healthcare) and a seven-element breast coil with the patient prone. An unenhanced sagittal T2-weighted sequence (TR/TE, 8990/107) preceded a sagittal fat-suppressed T1-weighted 3D volume-interpolated breath-hold examination (VIBE) sequence (TR/TE, 4.01/1.52; resolution, 1.4 × 0.9 × 1.5 mm). The latter sequence was performed before and four times after injection of gadopentetate dimeglumine (0.1 mM/kg body weight, Magnevist, Bayer Healthcare) at 2 mL/s followed by saline flush with a power injector (Spectris Solaris, Medrad). Image acquisition was begun immediately after administration of the contrast material and saline bolus. The first contrast-enhanced dynamic sequence was performed approximately 100 seconds after contrast injection and followed by three more consecutive sequences. The total duration of the dynamic study was approximately 6–7 minutes.
MRI Interpretation
The findings of MRI examinations and MRI-guided biopsies were interpreted by 1 of 14 radiologists who were either fellowship-trained breast imagers or had more than 10 years’ experience in breast imaging. MR images were evaluated and interpreted in accordance with the American College of Radiology BI-RADS MR lexicon [18]. The morphologic features were assessed by evaluation of the lesion type. All lesions were classified as a focus, mass, or nonmass enhancement. The kinetic evaluation was performed by characterization of time–signal intensity curves. The delayed phase was subdivided into three categories: persistent (type 1), plateau (type 2), or washout (type 3). These time–signal intensity curves were assessed with computer-assisted detection software (DynaCad, Invivo). In addition, maximum-intensity-projection images were routinely generated for postprocessing.
The level of suspicion was reported according to BI-RADS category as follows: BI-RADS 1, negative; BI-RADS 2, benign finding; BI-RADS 3, probably benign finding; BI-RADS 4, suspicious finding; and BI-RADS 5, finding highly suggestive of malignancy. Mammographic density and background parenchymal enhancement were recorded. Indications for breast MRI, morphologic and kinetic features, size and location of the concerning abnormality, and histologic findings were reviewed and noted for each examination.
MRI-Guided Biopsy Technique
MRI-guided biopsies were pursued in cases in which the abnormality detected at MRI was mammographically and sonographically occult or lesions were considered likely to be sonographically occult because of small size or lesion type. All biopsies were performed with the computer-aided detection software (DynaCAD) to target the lesion during the biopsy. Patients were prone for biopsy, which was performed with the 3-T system, the seven-element breast coil, and a grid-localizing system (Biopsy Breast Device, Invivo). Biopsies were performed with an MRI-compatible coaxial 9-gauge VAB system and either a 20-mm (standard) or 12-mm (petite) core (ATEC console, handpieces, and disposable introducer sets, Suros Surgical Systems–Hologic).
Localizing sequences were followed by a sagittal T1-weighted fat-suppressed 3D VIBE sequence (section thickness, 2 mm; matrix, 384 × 290) before and after IV administration of gadopentetate dimeglumine (0.1 mM/kg body weight Magnevist), which was injected at 2 mL/s and followed by saline flush with the power injector.
Unenhanced and contrast-enhanced images were evaluated to identify fiducial markers, biopsy grid, and parenchymal landmarks indicating the expected location of the target. Repeat T1-weighted fat-suppressed images in the sagittal and axial planes were obtained to verify proper placement of the biopsy introducer. If positioning was satisfactory, 10–12 core samples were obtained at the discretion of the radiologist. After tissue sampling, additional sagittal and axial T1-weighted fat-suppressed images were obtained to check tissue sampling and postbiopsy changes. Once the procedure was complete, an MRI-compatible titanium marker (Trimark, Suros Surgical Systems–Hologic) was deployed at the biopsy site, and MRI was repeated. For all patients a postbiopsy mammogram confirmed the position of the clip relative to the biopsy site.
Data Collection
Demographic data, including age, menopausal status, and family history, were obtained from a questionnaire each woman completed before undergoing MRI. Previous breast surgery, radiation therapy, chemotherapy, and use of antiestrogen agents were recorded. MR images were interpreted in accordance with the BI-RADS MRI lexicon [18]. Mammographic density and background parenchymal enhancement were recorded. Indications for breast MRI, morphologic and kinetic features, size and location of the concerning abnormality, and histologic findings were reviewed. Lesion size on MR images was measured in the anteroposterior, craniocaudal, and transverse dimensions, and greatest single dimension was used for data collection. Biopsy pathologic results and excision pathologic results were reviewed and entered into a Microsoft Excel for Mac 2011 spreadsheet.
Statistical Analysis
The Fisher exact test was used to assess the association of each characteristic with respect to whether a lesion was upgraded to cancer at surgery. Exact 95% CIs based on the binomial distribution were obtained for the percentage of women with a lesion upgraded to cancer among women with specific attributes. Logistic regression was used to identify sets of factors representing independent predictors that a lesion would be upgraded to cancer. All tests were conducted with SAS software (version 9.3, SAS Institute) at the two-sided 5% significance level.
Radiologic-Pathologic Correlation
The pathologic findings of all core biopsies were reviewed interdepartmentally and by one of eight breast pathologists (6–15 years’ experience) at our institution. All histopathologic results were correlated with MR images (and with mammograms and ultrasound images, if available) by the breast imaging radiologist who performed the procedure. Surgical excisions were performed with mammographically guided needle localization of the metallic clip deployed at the conclusion of the MRI-guided biopsy.
Of 151 lesions, 147 (97%) were further evaluated with needle localization and surgical excision. For lesions that contained multiple high-risk pathologic entities, a single pathologist identified the highest-risk lesion from the pathology report for statistical analysis. For example, lesions that included ADH were categorized as ADH and excluded from other high-risk categories. Histopathologic findings of MRI-guided biopsy were compared with the final surgical findings to determine the upgrade rate. During our data review, we specifically searched the final pathology report to identify whether the initial biopsy site was included in the final excision specimen. In addition, the breast pathologists at our institution routinely include the presence or absence of the biopsy site and biopsy track in their surgical pathology report.
Results
Using our institutional database, we identified 6123 consecutive breast MRI examinations performed from May 2007 to April 2012. From these examinations, we identified all patients who had a BI-RADS 4 or 5 final category assessment and underwent MRI-guided biopsy. In this group of 1003 women, we retrospectively identified 1145 breast lesions consecutively detected at MRI and then biopsied under MRI guidance.
Of 1145 lesions biopsied, 252 (22.0%) were malignant, 184 (16.1%) were high risk, and 709 (61.9%) were benign. From the 184 high-risk lesions, we excluded 33 lesions. Sixteen women were either lost to follow-up or did not undergo surgery and had less than 2 years of follow-up. An additional 17 high-risk lesions in 17 women were not localized before surgery because the patient went on to mastectomy and was therefore excluded.
The final study cohort consisted of 140 women with 151 high-risk lesions (13% of all biopsies during the period of the study). Eleven of these women had two high-risk lesions at MRI biopsy. In accordance with our current policy, surgical excision was recommended for all high-risk lesions. In this group, four lesions were not surgically excised because of individual preference but were followed up with imaging for at least 3 years. The mean follow-up time for these patients was 44 months (range, 36–54 months).
The median age of the 140 women was 50 years (range, 26–84 years). Seventy-six (54.3%) were premenopausal and 64 (45.7%) were postmenopausal. Among the 140 women, six (4.3%) had American College of Radiology type 2 mammographic density; 122 (87.1%), type 3; and 12 (8.6%), type 4. Background parenchymal enhancement was type 1 in 12 (8.6%) women, type 2 in 100 (71.4%), type 3 in 27 (19.3%), and type 4 in one (0.7%). The indications for MRI were classified into screening and diagnostic settings. Screening included women without symptoms but with a personal history of breast cancer, women with a strong family history, and women with a history of a high-risk lesion. Diagnostic indications included women with known breast cancer undergoing evaluation for extent of disease and women undergoing problem solving (Table 1). Upgrade rates were similar, approximately 20%, in the two categories. Examples of histologically upgraded lesions are illustrated in Figures 1 and 2.
| MRI Indication | Fraction Upgraded | 95% CI (%) |
|---|---|---|
| Diagnostic setting | ||
| Newly diagnosed cancera | 14/66 (21.2) | 12.1–32.4 |
| Problem solving | 1/9 (11.1) | 0.1–44.4 |
| Screening setting | ||
| Personal history of cancera | 10/25 (40) | 22.2–61.3 |
| Family history | 3/28 (10.7) | 3.0–27.5 |
| History of high-risk lesion | 2/23 (8.7) | 1.6–26.8 |
Note—Values in parentheses are percentages.
a
Percentage of women with a lesion upgraded to cancer was significantly higher for women with a remote or recent personal history of cancer than for the other indications combined (p = 0.0114).
Fig. 1A —54-year-old-woman with strong family history of breast cancer and normal mammographic findings.
A, Sagittal T1-weighted fat-suppressed 3D gradient-echo subtraction MR image shows linear clumped nonmass enhancement (arrow) with plateau enhancement kinetics.
Fig. 1B —54-year-old-woman with strong family history of breast cancer and normal mammographic findings.
B, Axial T1-weighted fat-suppressed 3D gradient-echo MR image obtained during biopsy shows obturator in place (arrow).
Fig. 1C —54-year-old-woman with strong family history of breast cancer and normal mammographic findings.
C, Sagittal T1-weighted fat-suppressed 3D gradient-echo postbiopsy MR image shows lesion has been adequately sampled and obturator is in place (arrow). MRI-guided vacuum-assisted biopsy revealed radial scar. Grade 1 ductal carcinoma in situ was found at definitive surgery.
Fig. 2A —49-year-old-woman with ipsilateral breast cancer.
A, Sagittal T1-weighted fat-suppressed 3D gradient-echo subtraction MR image shows 1.3-cm irregular mass (short arrow) with persistent enhancement kinetics in distinct quadrant from known cancer (long arrow).
Fig. 2B —49-year-old-woman with ipsilateral breast cancer.
B, Sagittal T1-weighted volume-interpolated breath-hold examination postbiopsy MR image shows lesion has been adequately sampled and obturator is in place (short arrow). Known cancer (long arrow) is also evident. MRI-guided vacuum-assisted biopsy revealed atypical lobular hyperplasia. Histopathologic diagnosis at surgery was grade 3 ductal carcinoma in situ.
Frequency and Upgrade Rate of High-Risk Lesions
Most of the 151 high-risk lesions identified (23.2% [35/151]) were ADH, followed by LCIS (19.9% [30/151]) and papillary lesions (19.9% [30/151]). If ALH and LCIS were considered on a lobular neoplasia continuum, then lobular neoplasia was the most common high-risk lesion (29.8% [45/151]). For the 151 high-risk lesions found at biopsy, final pathologic analysis revealed 30 (19.9%) malignant tumors, 78 (51.6%) high-risk lesions, and 43 (28.5%) benign lesions. Of the benign lesions, 39 of the 43 had final surgical pathologic results, and the other four lesions, including two cases of ADH, one case of ALH, and one papillary lesion, were considered benign after stable imaging follow-up findings. These four lesions had at least 3 years’ imaging stability (Table 2). Of 30 malignant tumors, 19 (63%) were ductal carcinoma in situ (DCIS) and 11 (37%) were upgraded to invasive ductal carcinoma (IDC) or invasive lobular carcinoma (ILC) (Table 3). ADH had the highest upgrade rate (34.3% [12/35]; 95% CI, 19.1–52.2%) followed by LCIS (26.7% [8/30]; 95% CI, 13.1–44.0%) and radial scar (24% [6/25]; 95% CI, 11.0–43.1%) (Tables 2 and 3). No cases of flat epithelial atypia were upgraded to malignancy.
| Lesion Type | Lesion Frequency | Upgrade Rate | 95% CI (%) | Ductal Carcinoma in Situ | Invasive Ductal Carcinoma | Invasive Lobular Carcinoma | High Risk at Surgery | Benign at Surgery |
|---|---|---|---|---|---|---|---|---|
| Atypical ductal hyperplasia | 35/1145 (3.06) | 12/35 (34.3) | 19.1–52.2 | 9 | 3 | 0 | 11 | 10 (2 lesions follow-up > 2 y) |
| Lobular carcinoma in situ | 30/1145 (2.62) | 8/30 (26.7) | 13.1–44 | 4 | 3 | 1 | 19 | 3 |
| Atypical lobular hyperplasia | 15/1145 (1.31) | 2/15 (13.3) | 2.4–38.5 | 0 | 1 | 1 | 6 | 6 (1 lesion follow-up > 2 y) |
| Papillary lesiona | 30/1145 (2.62) | 2/30 (6.7) | 1.2–20.5 | 2 | 0 | 0 | 21 | 6 (1 lesion follow-up > 2 y) |
| Radial scarb | 25/1145 (2.18) | 6/25 (24) | 11.0–43.1 | 4 | 2 | 0 | 13 | 6 |
| Flat epithelial atypia | 16/1145 (1.4) | 0/16 (0) | 0–20.6 | 0 | 0 | 0 | 8 | 8 |
| Total | 151 | 30 | 19 | 9 | 2 | 78 | 43 |
Note—Except for 95% CI, values are numbers of lesions with percentages in parentheses.
a
Six of 30 papillary lesions had atypia; none of these were upgraded to malignancy after surgery.
b
Five of 25 radial scar cases had atypia; of these three of five were upgraded to malignancy after surgery. Twenty of 25 radial scar cases were without atypia; of these, three of 20 were upgraded to malignancy after surgery.
| Patient No. | Age (y) | Clinical Indication | Lesion Type | Lesion Kinetic Type | Lesion Size at MRI (cm) | MRI Core Needle Biopsy Histologic Diagnosis | Surgical Histologic Diagnosis |
|---|---|---|---|---|---|---|---|
| 1 | 49 | History of breast cancer | Mass | 2 | 0.7 | LCIS | IDC |
| 2 | 54 | Family history | NME | 2 | 1.7 | Radial scar | DCIS |
| 3 | 68 | Known breast cancer | Mass | 2 | 0.8 | ADH | DCIS |
| 4 | 48 | Known breast cancer | Mass | 2 | 0.7 | LCIS | IDC |
| 5 | 46 | Known breast cancer | NME | 1 | 3 | ADH | DCIS |
| 6 | 67 | Known breast cancer | Mass | 1 | 0.8 | ADH | DCIS |
| 7 | 49 | Known breast cancer | Mass | 1 | 1.3 | ALH | IDC |
| 8 | 40 | Known breast cancer | Mass | 3 | 1.5 | LCIS | DCIS |
| 9 | 49 | History of breast cancer | Mass | 3 | 2 | Radial scar | DCIS |
| 10 | 62 | Known breast cancer | NME | 1 | 1.4 | Radial scar | IDC |
| 11 | 42 | Known breast cancer | Mass | 3 | 2 | ADH | IDC |
| 12 | 45 | Known breast cancer | NME | 2 | 0.6 | LCIS | DCIS |
| 13 | 68 | Known breast cancer | Mass | 3 | 0.7 | Radial scar | IDC |
| 14 | 62 | History of breast cancer | NME | 2 | 0.6 | ADH | IDC |
| 15 | 62 | History of breast cancer | Mass | 2 | 0.5 | ADH | IDC |
| 16 | 63 | Problem solving | Mass | 3 | 3.8 | Radial scar | DCIS |
| 17 | 46 | Family history | NME | 1 | 0.8 | LCIS | ILC |
| 18 | 49 | Known breast cancer | Mass | 2 | 0.7 | LCIS | IDC |
| 19 | 58 | Known breast cancer | NMLE | 2 | 0.7 | ALH | ILC |
| 20 | 84 | Known breast cancer | Mass | 2 | 0.7 | Papilloma | DCIS |
| 21 | 72 | History of breast cancer | NME | 2 | 1.3 | LCIS | DCIS |
| 22 | 73 | Family history | NME | 2 | 0.7 | Radial scar | DCIS |
| 23 | 56 | Known breast cancer | NME | 1 | 2 | Papilloma | DCIS |
| 24 | 46 | History of high-risk lesion | Mass | 2 | 0.8 | LCIS | DCIS |
| 25 | 50 | History of breast cancer | Mass | 2 | 0.7 | ADH | DCIS |
| 26 | 58 | History of high-risk lesion | NME | 1 | 2.6 | ADH | DCIS |
| 27 | 61 | Known breast cancer | NME | 1 | 6.2 | ADH | DCIS |
| 28 | 66 | History of breast cancer | NME | 1 | 0.9 | ADH | DCIS |
| 29 | 44 | History of breast cancer | NME | 2 | 1.1 | ADH | DCIS |
| 30 | 39 | History of breast cancer | NME | 1 | 2.5 | ADH | DCIS |
Note—LCIS = lobular carcinoma in situ, IDC = invasive ductal carcinoma, NME = nonmass enhancement, DCIS = ductal carcinoma in situ, ADH = atypical ductal hyperplasia, ALH = atypical lobular hyperplasia, ILC = invasive lobular carcinoma, nonmasslike enhancement.
Lesion Imaging Characteristics Predictive of Upgrade
Lesion types were mass (62/151 [41.0%]), nonmass enhancement (78/151 [51.6%]), and focus (11/151 [7.3%]). The median size of upgraded high-risk lesions as measured on MR images was 1.8 cm (range, 0.7–3.8 cm) compared with 1.1 cm (range, 0.4–5.2 cm) for nonupgraded lesions. The median sizes in the two groups were not significantly different. Among the 30 upgraded high-risk lesions, 13 were masses, 14 were nonmass enhancement, and three were enhancing foci (Table 3). The kinetic features varied (10 lesions had type 1 kinetic curves, 15 had type 2 kinetic curves, and five had type 3 kinetic curves) (Table 3). No kinetic curve type had significant predictive value for upgrade (Table 4). The Fisher exact test p values were used to assess the association between each characteristic and whether lesions were upgraded to cancer at surgery (Table 4). Neither lesion morphologic features nor kinetics were predictive of upgrade. More specifically, neither lesion size nor lesion type based on MRI appearance was predictive of upgrade. Within mass lesions and foci, neither margins nor enhancement pattern (homogeneous versus heterogeneous or rim enhancement) were found to be predictive of upgrade. For nonmass lesions, neither internal enhancement pattern nor distribution was a predictive criterion.
| Lesion Characteristic | No. of Biopsied Lesions | No. of Benign or High-Risk Final Histopathologic or Normal Imaging Follow-Up Findings | No. of Malignant Lesions | p |
|---|---|---|---|---|
| Lesion size on MR image (cm) | 0.6877 | |||
| ≤ 1 | 75 | 59 (78.7) | 16 (21.3) | |
| > 1 | 76 | 62 (81.6) | 14 (18.4) | |
| Lesion type | 0.6293 | |||
| Focus | 11 | 10 (90.9) | 1 (9.1) | |
| Mass | 62 | 48 (77.4) | 14 (22.6) | |
| Nonmass | 78 | 63 (80.8) | 15 (19.2) | |
| Mass lesion margins | 0.5551 | |||
| Smooth | 35 | 26 (74.3) | 9 (25.7) | |
| Irregular or spiculated | 27 | 22 (81.5) | 5 (18.5) | |
| Mass lesion enhancement | 0.3804 | |||
| Homogeneous | 29 | 24 (82.8) | 5 (17.2) | |
| Heterogeneous or rim | 33 | 24 (72.7) | 9 (27.3) | |
| Nonmass distribution | 1.0000 | |||
| Linear, ductal | 59 | 48 (81.4) | 11 (18.6) | |
| Segmental | 19 | 15 (79.0) | 4 (21.0) | |
| Nonmass internal enhancement pattern | 1.0000 | |||
| Homogeneous, stippled | 38 | 31 (81.6) | 7 (18.4) | |
| Heterogeneous, clumped | 40 | 32 (80.0) | 8 (20.0) | |
| Focus lesion margins | 1.0000 | |||
| Smooth | 2 | 2 (100) | 0 (0) | |
| Irregular or spiculated | 9 | 8 (88.9) | 1 (11.1) | |
| Focus lesion enhancement | 1.0000 | |||
| Homogeneous | 9 | 8 (88.9) | 1 (11.1) | |
| Heterogeneous or rim | 2 | 2 (100) | 0 (0) | |
| Time–signal intensity curve | 0.7198 | |||
| Persistent | 43 | 33 (76.7) | 10 (23.3) | |
| Plateau | 85 | 70 (82.4) | 15 (17.6) | |
| Washout | 23 | 18 (78.3) | 5 (21.7) |
Note—Values in parentheses are percentages.
Patient Characteristics Predictive of Upgrade
There was no significant difference between patients with and those without upgraded lesions with regard to menopausal status (p = 0.3040), breast density on mammogram (p = 0.5588), and background parenchymal enhancement on MRI (p = 0.3501). However, there was a significantly higher risk that a high-risk lesion would be upgraded to malignancy if the current MRI-detected high-risk lesion was in the same breast as remotely previously identified cancer, recently diagnosed cancer, or a remotely previously identified high-risk lesion (p = 0.0001). Seventy-percent (21/30) of lesions upgraded to cancer were in women whose original lesion was in the ipsilateral breast as a remote or current pathologic finding.
Indication and Upgrade
According to results of a post hoc Fisher exact test, the percentage of cases with an upgraded lesion was significantly higher among women with a personal history of cancer (remote and recent) than for the other indications combined (p = 0.011). Of the 30 upgraded lesions, 15 (50%) were upgraded in women with newly diagnosed disease (Table 3). Of these 15 malignant tumors, one (6.7%) was in the contralateral breast, eight (53.3%) were in a separate quadrant, and six (40.0%) were in the same quadrant but more than 4 cm from the known malignant tumor. These six lesions occurred at a site where documentation of additional cancer might alter surgical planning. Wire localization of the high-risk lesion specifically was performed in all 15 cases. Of the 15 malignant tumors, eight (53.3%) were DCIS, six (40%) were IDC, and one (6.7%) was ILC.
Of the 30 upgraded lesions, 30% (9/30) were in women with a remote history of cancer (Table 3): IDC in five (55.6%) and DCIS in four (44.4%) women. Four of the nine women (44.4%) were taking hormonal therapy, and three (33.3%) women underwent preoperative breast MRI. The mean time from the original diagnosis of breast cancer to diagnosis of the high-risk lesion was 31 months (range, 10–59 months). Three malignant tumors (33.3%) were in the contralateral breast; one (11.1%) was in a separate quadrant; and five (55.6%) were in the same quadrant and consistent with local recurrence. Six (66.7%) malignant tumors were DCIS, and three (33.3%) were IDC.
Discussion
High-risk lesions detected exclusively with MRI have increased in number as the use of breast MRI continues to rise. The imaging features of these high-risk lesions overlap with those of malignant tumors, and there is a chance that the lesion may be upgraded to malignancy at surgery owing to undersampling and sampling errors. Given these limitations, surgical excision is recommended. However, there is also concern about reducing the rate of false-positive findings at breast MRI and a desire to identify demographic and imaging features that may be associated with upgrade to malignancy. To determine appropriate management of high-risk lesions, risk of upgrade to malignancy at MRI must be understood.
In addition to the small number of studies focusing specifically on characteristics of high-risk lesions detected with MRI, a growing number of studies of the general outcome of MRI-guided breast biopsy have included data on the rates and underestimation of high-risk lesions [19]. Some of these studies, however, focused only on certain high-risk pathologic features, making it unclear whether the other high-risk pathologic features were not diagnosed in the populations studied or were classified as benign. Perlet et al. [10], for example, listed data for ADH and papilloma but did not specifically mention other high-risk lesions. In most studies papillomas are classified as high risk. Han et al. [20], however, included only atypical papillomas in their high-risk category, and Orel et al. [5] classified papillomas as benign. Meta-analysis of these studies to reach a consensus recommendation is thus limited by these differences and uncertainty about which lesions were included in the high-risk category.
Despite the aforementioned limitations, the reported incidence of MRI-detected high-risk lesions as a group ranges from 3% to 21% [19]. Our findings were within this range: 16.1% (184/1145) of all consecutive MRI-guided biopsies yielded high-risk lesions.
Studies in the literature have shown a broad range of underestimation rates (13% to 57%) for all MRI-detected high-risk lesions combined [19]. Our overall high-risk lesion upgrade rate for cases of surgical excision was 20%, within the reported range [6, 10, 13] but higher than that of stereotactic biopsy. The difference may be due to the overall higher risk of breast cancer in the population undergoing breast MRI and the heterogeneous and noncontiguous nature of high-risk lesions [13]. It is also possible that the population of high-risk lesions detected with MRI—lesions that are contrast enhancing—may be a distinct group of lesions with a distinct pathophysiologic makeup compared with high-risk lesions found with conventional imaging.
When individual high-risk lesion statistics are evaluated, our results are in line with other work identifying ADH as the lesion with the highest rate of upgrade. Liberman et al. [13] examined the outcome of ADH diagnosed at MRI-guided 9-gauge VAB and encountered 15 (6.3%) cases of ADH among 237 MRI-guided biopsies. Surgical excision was performed in 13 cases, and five (38%) were upgraded to DCIS. In several studies [5–7, 10, 11, 13–15, 20–25], the reported rate of underestimation of ADH among lesions detected with MRI biopsy performed with a 14-gauge or larger biopsy device ranged from 11% to 67%. In all but two of those studies, relatively low numbers of cases of ADH were found. The exceptions were the studies by Perlet and colleagues [10], who found 17 cases of ADH and an upgrade rate similar to that in our study (5/17 [29%], all to DCIS), and Strigel and colleagues [15], who identified 51 cases of ADH in total with an upgrade of 11 cases (six to DCIS, four to IDC, and one to ILC.) This frequency of ADH underestimation at MRI-guided VAB is slightly higher than that reported for stereotactic biopsy, which has an average upgrade rate of 20% for 11-gauge VAB accumulated from published studies and 20–30% for 9-gauge VAB [26]. Another cause of this underestimation rate is likely the high-risk population. In addition, our data support previous findings of upgrade to both IDC and ILC [7, 12, 14, 15, 27].
LCIS also had a high upgrade rate (26.7% [8/30]; 95% CI, 13.1–44%). If ALH and LCIS are considered together on a continuum as lobular neoplasia, the upgrade rate was still high at 22% (10/45). At present, relatively few absolute reported cases of either ALH or LCIS with surgical follow-up are cited in the MRI biopsy literature. Crystal and colleagues [14] noted only four cases of lobular neoplasia upgraded among eight cases (three cases to DCIS and one to invasive carcinoma). Mahoney [7] cited one of one case of LCIS upgraded to IDC, and Strigel et al. [15] detailed one of two cases of ALH upgraded to DCIS and neither of two cases of LCIS upgraded. Rauch et al. [12] reported one of six cases of LCIS and none of 12 cases of ALH upgraded to ILC at surgery. In a more recent report of the upgrade rate of lobular neoplasia that incorporated lesions biopsied under MRI, ultrasound, and stereotactic guidance and included pleomorphic LCIS, the upgrade rate was 16% (11/67 patients) with upgrade to both invasive and in situ disease [28].
Twenty percent (5/25) of radial scar lesions in our cohort exhibited atypia. Of this group, three of five (60%) were upgraded at surgical excision: two cases to DCIS and one case to IDC. If we exclude the cases of atypia, 3 of 20 (15%) cases of radial scar were upgraded. Both rates are higher than those reported in the stereotactic biopsy literature, in which an upgrade rate of 3.9–4.0% has been found in lesions without associated atypia. The upgrade rate of 28% (8/29) is seen in radial scar lesions with associated atypical hyperplasia [29]. To our knowledge, published reports of high-risk lesions consecutively detected with MRI and of results of MRI VAB have detailed few cases of radial scar with no upgrades reported [19]. Our results might have been skewed by the small sample size and the overall low incidence of radial scar [29, 30].
Using MRI as a problem-solving tool, Linda et al. [31] analyzed 54 radial scar lesions detected with ultrasound and mammography and found that 2 of 54 (3.7%) of these lesions were upgraded, one to ILC and one to DCIS. Those authors looked at MRI characteristics of high-risk lesions and found a high negative predictive value for both radial scar (97.6%) and papilloma without atypia (97.4%). Two of 54 cases of radial scar were upgraded at surgery (3.7%). One case of radial scar was correctly identified as suspicious at MRI (BIRADS category 4) and was found to be ILC at surgery. The other case of radial scar was not enhancing and was incorrectly assigned to BIRADS category 1. This lesion was upgraded to low-grade DCIS. Linda et al. noted that lack of enhancement on MR images may suggest lack of invasive disease in the context of high-risk lesions found with other modalities, but that more studies are needed to confirm this.
Conversely, we had low upgrade rates for papillary lesions. Two of 30 papillary lesions with and without atypia were upgraded (6.7%; 95% CI, 1.2–20.5%). In total 6 of 30 (20%) of the papillary lesions in our study exhibited atypia, and none of these were upgraded at surgical excision. If the lesions with atypia are excluded, 8% of papillary lesions were upgraded, both cases to DCIS. Our papillary lesion upgrade rate is similar to that in a 2012 study by Brennan et al. [32] that focused on the incidence and upgrade rates of papillary lesions both with and without atypia at MRI-guided 9-gauge VAB. They found papillomas in 75 of 1487 biopsies, 25 with atypia and 50 without. At surgery, 9% (2/23) of lesions with atypia and 5% (2/44) without atypia were upgraded to malignancy. All cases were DCIS.
We had no upgraded cases of flat epithelial atypia (0/16; 95% CI, 0–20.6%). This result is consistent with the findings in the literature. Additional studies are necessary to determine whether this type of lesion may be safely followed rather than excised. At present we recommend excision given the small sample size of cases in the literature.
Our study showed no imaging features to be predictive of upgrade. MRI enhancement characteristics may be useful as a problem-solving tool: a means of predicting upgrade to invasive disease among high-risk lesions detected initially with mammography or sonography [27, 31]. At present, however, there are no definite predictive features among enhancing high-risk lesions detected with MRI. This has been shown in multiple other studies of high-risk lesions on MRI [19]. A quantitative analysis of contrast uptake may be able to identify a subgroup that is more likely to be upgraded to malignancy.
Among women with a personal history of breast cancer, tumor recurrence rates after breast conservation therapy have historically been estimated at 1–2% per year [33]. With the use of hormonal therapy and recent improvements in chemotherapy, the recurrence rates at 10 years are now less than 10%. Despite the low recurrence rates, our data are consistent with the growing body of literature that supports annual screening MRI for women with a personal history of breast cancer. For example, DeMartini and colleagues (DeMartini WB, et al., presented at the 2010 meeting of the Radiological Society of North America) found that the cancer yield among women with a personal history of breast cancer (3.1%) in screening MRI examinations was double that of women with a genetic or family history (1.5%). Brennan et al. [34] identified 18 cases of cancer among 17 women with a personal history of cancer alone (12% of women in the study) through screening MRI. Sardanelli and colleagues [35] looked at women at high risk with personal histories of breast cancer who were also BRCA mutation carriers. The cancer incidence among women with a history of cancer was 4.3%, borderline significantly higher than that among women without such a history (2.5%).
Limitations
Our study was extensive compared with other analyses of high-risk lesions, but the numbers of cases still was limited. Although most lesions were surgically excised, another limitation was that four cases were followed up only; however, all of the lesions exhibited at least 36-month stability at follow-up MRI. We also did not investigate the upgrade rate of lesions in cases in which second-look ultrasound depicted a correlate. Finally, the one feature we found to be predictive of upgrade was a history of ipsilateral pathologic findings. This may be in keeping with the results of multiple studies that have shown high risk of ipsilateral cancer among women with known recently identified or remote malignancy [12, 20, 24]. Although our results might have suggested that the findings represented extension of known disease rather than distinct multifocal or multicentric lesions, we performed radiologic-pathologic correlation and identified MRI-guided biopsy sites in final excision specimens whenever possible to minimize this possibility. In addition, of the 15 high-risk lesions upgraded to malignancy in women with newly diagnosed cancer, all were at least 4 cm distant from the known index lesion as measured on MR images, in a distinct quadrant, or in the contra-lateral breast.
Management of high-risk lesions has been based primarily on literature findings regarding stereotactically or sonographically guided biopsies. Underestimation of high-risk lesions diagnosed with these traditional percutaneous biopsy methods has established the recommendation for surgical excision in most cases, although the data are in themselves controversial [16]. Investigators working on high-risk lesions detected on MR images have recommended excision of ADH and lobular neoplasia, and our results confirm the potential for high upgrade rates for ADH and LCIS-ALH (lobular neoplasia). The risk of upgrade of papillomas without atypia was 8%, in keeping with the literature findings on this topic. We also identified radial scar as a lesion with a 15% risk of underestimation, even excluding radial scar with atypia.
Conclusion
We obtained a diagnosis of a high-risk lesion in 16% of all lesions on which MRI-guided VAB was performed. Upgrade to malignancy occurred in 20% of those high-risk lesions with surgical follow-up. We identified no specific imaging feature predictive of upgrade to malignancy, but we found a significantly higher risk of upgrade among women with a personal history of breast cancer and women with an imaging indication (remote or recent history of cancer or previous high-risk lesion) in the ipsilateral breast as the high-risk finding.
Although further study with larger numbers is warranted to more accurately guide treatment of these lesions, we recommend excision of all high-risk lesions found at MRI. In particular, we found high underestimation rates for ADH and LCIS-ALH at MRI-guided biopsy and for radial scar (both with and without atypia). Our study showed a relatively low upgrade rate for papillary lesions but sufficiently high upgrade rates to warrant the recommendation for excision. Finally, although we found that no cases of flat epithelial atypia were upgraded, the small sample size of the subset of women with this lesion ensures a recommendation of excision. We also suggest that particular attention be paid to high-risk lesions detected in women with a remote or recent history of cancer.
References
1.
Kuhl CK, Schmutzler RK, Leutner CC, et al. Breast MR imaging screening in 192 women proved or suspected to be carriers of a breast cancer susceptibility gene: preliminary results. Radiology 2000; 215:267–279
2.
Lee SG, Orel SG, Woo IJ, et al. MR imaging screening of the contralateral breast in patients with newly diagnosed breast cancer: preliminary results. Radiology 2003; 226:773–778
3.
Schnall MD, Blume J, Bluemke DA, et al. Diagnostic architectural and dynamic features at breast MR imaging: multicenter study. Radiology 2006; 238:42–53
4.
Lehman CD, DeMartini W, Anderson BO, Edge SB. Indications for breast MRI in the patient with newly diagnosed breast cancer. J Natl Compr Canc Netw 2009; 7:193–201
5.
Orel SG, Rosen M, Mies C, Schnall MD. MR imaging-guided 9-gauge vacuum-assisted core-needle breast biopsy: initial experience. Radiology 2006; 238:54–61
6.
Liberman L, Bracero N, Morris E, Thornton C, Dershaw DD. MRI-guided 9-gauge vacuum-assisted breast biopsy: initial clinical experience. AJR 2005; 185:183–193
7.
Mahoney MC. Initial clinical experience with a new MRI vacuum-assisted breast biopsy device. J Magn Reson Imaging 2008; 28:900–905
8.
Malhaire C, El Khoury C, Thibault F, et al. Vacuum-assisted biopsies under MR guidance: results of 72 procedures. Eur Radiol 2010; 20:1554–1562
9.
Noroozian M, Gombos EC, Chikarmane S, et al. Factors that impact the duration of MRI-guided core needle biopsy. AJR 2010; 194:[web]W150–W157
10.
Perlet C, Heywang-Kobrunner SH, Heinig A, et al. Magnetic resonance-guided, vacuum-assisted breast biopsy: results from a European multicenter study of 538 lesions. Cancer 2006; 106:982–990
11.
Perretta T, Pistolese CA, Bolacchi F, Cossu E, Fiaschetti V, Simonetti G. MR imaging-guided 10-gauge vacuum-assisted breast biopsy: histological characterisation. Radiol Med (Torino) 2008; 113:830–840
12.
Rauch GM, Dogan BE, Smith TB, Liu P, Yang WT. Outcome analysis of 9-gauge MRI-guided vacuum-assisted core needle breast biopsies. AJR 2012; 198:292–299
13.
Liberman L, Holland AE, Marjan D, et al. Underestimation of atypical ductal hyperplasia at MRI-guided 9-gauge vacuum-assisted breast biopsy. AJR 2007; 188:684–690
14.
Crystal P, Sadaf A, Bukhanov K, McCready D, O'Malley F, Helbich TH. High-risk lesions diagnosed at MRI-guided vacuum-assisted breast biopsy: can underestimation be predicted? Eur Radiol 2011; 21:582–589
15.
Strigel RM, Eby PR, DeMartini WB, et al. Frequency, upgrade rates, and characteristics of high-risk lesions initially identified with breast MRI. AJR 2010; 195:792–798
16.
Berg WA. Image-guided breast biopsy and management of high-risk lesions. Radiol Clin North Am 2004; 42:935–946
17.
Georgian-Smith D, Lawton TJ. Controversies on the management of high-risk lesions at core biopsy from a radiology/pathology perspective. Radiol Clin North Am 2010; 48:999–1012
18.
D'Orsi CJ, Mendelson, EB, Ikeda DM, et al. Breast Imaging Reporting and Data System: ACR BI-RADS—breast imaging atlas. Reston, VA: American College of Radiology, 2003
19.
Heller SL, Moy L. Imaging features and management of high-risk lesions on contrast-enhanced dynamic breast MRI. AJR 2012; 198:249–255
20.
Han BK, Schnall MD, Orel SG, Rosen M. Outcome of MRI-guided breast biopsy. AJR 2008; 191:1798–1804
21.
Chen X, Lehman CD, Dee KE. MRI-guided breast biopsy: clinical experience with 14-gauge stainless steel core biopsy needle. AJR 2004; 182:1075–1080
22.
Ghate SV, Rosen EL, Soo MS, Baker JA. MRI-guided vacuum-assisted breast biopsy with a handheld portable biopsy system. AJR 2006; 186:1733–1736
23.
Lehman CD, Deperi ER, Peacock S, McDonough MD, Demartini WB, Shook J. Clinical experience with MRI-guided vacuum-assisted breast biopsy. AJR 2005; 184:1782–1787
24.
Liberman L, Morris EA, Dershaw DD, Thornton CM, Van Zee KJ, Tan LK. Fast MRI-guided vacuum-assisted breast biopsy: initial experience. AJR 2003; 181:1283–1293
25.
Tozaki M, Yamashiro N, Sakamoto M, Sakamoto N, Mizuuchi N, Fukuma E. Magnetic resonance-guided vacuum-assisted breast biopsy: results in 100 Japanese women. Jpn J Radiol 2010; 28:527–533
26.
Eby PR, Ochsner JE, DeMartini WB, Allison KH, Peacock S, Lehman CD. Frequency and upgrade rates of atypical ductal hyperplasia diagnosed at stereotactic vacuum-assisted breast biopsy: 9-versus 11-gauge. AJR 2009; 192:229–234
27.
Londero V, Zuiani C, Linda A, Girometti R, Bazzocchi M, Sardanelli F. High-risk breast lesions at imaging-guided needle biopsy: usefulness of MRI for treatment decision. AJR 2012; 199:[web] W240–W250
28.
Niell B, Specht M, Gerade B, Rafferty E. Is excisional biopsy required after a breast core biopsy yields lobular neoplasia? AJR 2012; 199:929–935
29.
Brenner RJ, Jackman RJ, Parker SH, et al. Percutaneous core needle biopsy of radial scars of the breast: when is excision necessary? AJR 2002; 179:1179–1184
30.
Douglas-Jones AG, Denson JL, Cox AC, Harries IB, Stevens G. Radial scar lesions of the breast diagnosed by needle core biopsy: analysis of cases containing occult malignancy. J Clin Pathol 2007; 60:295–298
31.
Linda A, Zuiani C, Furlan A, et al. Nonsurgical management of high-risk lesions diagnosed at core needle biopsy: can malignancy be ruled out safely with breast MRI? AJR 2012; 198:272–280
32.
Brennan SB, Corben A, Liberman L, et al. Papilloma diagnosed at MRI-guided vacuum-assisted breast biopsy: is surgical excision still warranted? AJR 2012; 199:[web]W512–W519
33.
Fisher B, Redmond C, Poisson R, et al. Eight-year results of a randomized clinical trial comparing total mastectomy and lumpectomy with or without irradiation in the treatment of breast cancer. N Engl J Med 1989; 320:822–828
34.
Brennan S, Liberman L, Dershaw DD, Morris E. Breast MRI screening of women with a personal history of breast cancer. AJR 2010; 195:510–516
35.
Sardanelli F, Podo F, Santoro F, et al. Multicenter surveillance of women at high genetic breast cancer risk using mammography, ultrasonography, and contrast-enhanced magnetic resonance imaging (the High Breast Cancer Risk Italian 1 Study): final results. Invest Radiol 2011; 46:94–105
Information & Authors
Information
Published In
Copyright
© American Roentgen Ray Society.
History
Submitted: January 10, 2013
Accepted: April 11, 2013
First published: December 26, 2013
Keywords
Authors
Metrics & Citations
Metrics
Citations
Export Citations
To download the citation to this article, select your reference manager software.