(Radiology. 2001;220:31-39.)
© RSNA, 2001
MR Imaging–guided Large-Core (14-Gauge) Needle Biopsy of Small Lesions Visible at Breast MR Imaging Alone1
Christiane K. Kuhl, MD,
Nuschin Morakkabati, MD,
Claudia C. Leutner, MD,
Alexandra Schmiedel, MD,
Eva Wardelmann, MD and
Hans H. Schild, MD
1 From the Department of Radiology (C.K.K., N.M., C.C.L., A.S., H.H.S.) and Institute of Pathology (E.W.), University of Bonn, Sigmund-Freud-Strasse 25, D-53105 Bonn, Germany. Received May 19, 2000; revision requested July 12; final revision received January 5, 2001; accepted February 6. Address correspondence to C.K.K. (e-mail: kuhl@uni-bonn.de).
 |
ABSTRACT
|
|---|
PURPOSE: To report our experience with magnetic resonance (MR) imaging–guided large-core breast biopsy of lesions visible at breast MR imaging only.
MATERIALS AND METHODS: Stereotactic large-core (14-gauge) needle biopsy of 78 lesions visible at MR imaging only was performed with MR imaging guidance in 59 patients. Results were validated with excisional biopsy or mastectomy in 42 lesions and with radiologic-pathologic correlation and/or follow-up MR imaging for at least 2 years in another 17 lesions. The accuracy of MR imaging–guided core biopsy was determined for those 59 lesions with established validation. The effect on patient treatment was evaluated by comparing the prebiopsy treatment plan with the ultimate treatment.
RESULTS: Histologic diagnosis from core biopsy was possible in 77 (99%) of 78 lesions. In the 59 lesions with established validation, the diagnostic accuracy of MR imaging–guided core biopsy was 98% (58 of 59). Successful MR imaging–guided core biopsy findings changed treatment in 70% (54 of 77) of lesions. Difficulties were due to the unsatisfactory performance of earlier types of MR imaging–compatible biopsy guns and decreasing target visibility during intervention.
CONCLUSION: MR imaging–guided large-core stereotactic breast biopsy is sufficiently accurate for obtaining histologic proof of lesions visible only at MR imaging. It can change patient treatment by reducing unnecessary surgical biopsy and can enable one-step surgery for breast cancers.
Supplemental material: radiology.rsnajnls.org/cgi/content/full/220/1/31/DC1.
Index terms: Breast, biopsy, 00.1267, 00.1269 • Breast, diseases, 00.719, 00.722, 00.725, 00.89 • Breast, MR, 00.121411, 00.121412, 00.121415, 00.12143, 00.12149 • Breast neoplasms, 00.311, 00.322, 00.324, 00.327 • Magnetic resonance (MR), guidance, 00.12149
|
INTRODUCTION
|
|---|
During the past years, breast magnetic resonance (MR) imaging has emerged as a valuable adjunct to conventional imaging modalities such as mammography and high-frequency breast ultrasonography (US). There is broad agreement that breast MR imaging offers high sensitivity for invasive malignant lesions (1–5). At the time this article was written, breast MR imaging had been used mainly for clarification of conventionally equivocal lesions, particularly in follow-up after breast-conserving therapy. More recently, breast MR imaging has increasingly been used to enhance detection of primary breast cancer (6–8). It is used to identify early breast cancer in high-risk patients or in patients before surgery who have a solitary suspicious lesion at conventional imaging, in whom multicentric primary lesions should be ruled out before breast-conserving therapy is initiated. According to results in a recent series, breast MR imaging can depict conventionally invisible but therapeutically relevant additional breast cancer foci in up to 10% of cases (6).
The main use of breast MR imaging in these situations is, of course, to identify breast cancer foci that are invisible at mammography and breast US. However, if such an "MR imaging–only" lesion (ie, a suspicious lesion without correlation at conventional mammography or breast US) is detected, histologic proof is not easy to obtain. MR imaging–guided stereotactic hook wire placement has been developed and rapidly introduced into clinical practice (9–13). However, the disadvantage of such a technique is that it requires subsequent open surgery; hence, it is associated with substantial costs and possible morbidity and requires substantial logistic effort to coordinate the timing of wire placement and surgery.
Core-needle biopsy with mammographic or US guidance has been established as a safe and cost-effective tool to obtain histologic clarification of nonpalpable lesions, thus avoiding excisional biopsy (14–23). With the advent of MR imaging–compatible core-needle biopsy guns, authors of early reports (24–27) have confirmed that MR imaging guidance of direct tissue sampling is technically feasible.
The purpose of our study was to report our experience with MR imaging–guided stereotactic large-core breast biopsy of lesions visualized only at breast MR imaging with regard to procedural accuracy, specific procedure-related difficulties, and procedural effect on clinical patient treatment. The issue of cost-effectiveness was not addressed.
|
MATERIALS AND METHODS
|
|---|
Our prospective study was approved by our institutional review board; biopsies were performed after obtaining patient informed consent.
Patients and Inclusion Criteria
Our study included 59 patients (mean age, 45.6 years; age range, 22–70 years). Mammography with two views plus additional views when appropriate (Mammomat 3000; Siemens Medical Systems, Erlangen, Germany) and high-frequency (7.5- or 12-MHz) whole-breast US (Elegra; Siemens Medical Systems) were performed in all patients. Patients were included only if, at MR imaging, a lesion was shown that had no correlate at mammography or high-frequency breast US. The decision to recommend biopsy with MR imaging guidance was made only if, after retrospective analysis of the mammograms and after repeat directed and whole-breast US, the MR imaging–visible lesion could not (or not without doubt) be identified. In addition, two patients were enrolled (Table E1, radiology.rsnajnls. org/cgi/content/full/220/1/31/DC1) who had been scheduled to undergo excisional biopsy of a solitary mammographically visible lesion ("index lesion") and who underwent preoperative breast MR imaging. Breast MR imaging revealed additional lesions suspected of representing further multicentric breast cancer foci. MR imaging–guided biopsy of the "MR imaging–only" lesions and mammographically visible index lesion were performed in the same session. Table 1 shows the indications for diagnostic breast MR imaging.
In the 59 patients, 78 lesions were identified at diagnostic breast MR imaging (Table E1, radiology.rsnajnls.org/cgi/content/full/220/1/31/DC1): In 16 patients (19 lesions), diagnostic breast MR imaging had revealed lesions rated as probably benign but in which, owing to a specific personal or family history or the patients’ request, histologic proof was desired. Another 20 patients (23 lesions) underwent MR imaging–guided core biopsy owing to lesions rated as equivocal in which histologic proof was needed for clarification. In another 23 patients (36 lesions), MR imaging–guided biopsy was initiated to confirm the presumed malignant nature of a lesion to allow a one-step surgical approach.
Equipment
Stereotactic large-core needle biopsy was performed with a 1.5-T short-bore closed magnet (ACS-NT with POWERTRAK 6000 gradient system or Gyroscan INTERA; Philips Medical Systems, Best, the Netherlands). The stereotactic device used for biopsy has been described elsewhere (12); it consists of a support that keeps the patient in a semiprone position and two compression plates featuring an MR imaging–visible fiducial system. The compression plates are perforated, with 2-mm-wide holes every 2.5 mm, thus accommodating needles of up to 13 gauge. For image generation, a regular flexible circular surface coil was placed around the breast.
For breast biopsy, fully MR imaging–compatible 14-gauge core-needle systems from different vendors (Daum Medical Systems, Schwerin, Germany; Guerbet, Würzburg, Germany; Somatex, Berlin, Germany) were used. All systems were single-use semiautomatic biopsy guns with a long throw and a 20-mm biopsy notch. One of the biopsy systems (Guerbet) allowed cores to be obtained repeatedly with a backloading core needle; the other systems allowed coaxial tissue sampling with a 13-gauge coaxial introducer needle.
Core Biopsy Procedure
The patients were positioned on the stereotactic unit until a comfortable position was reached. The breast was gently pulled through the opening of the circular coil and placed between the plates, with compression in a mediolateral direction. To avoid nonenhancing targets (28), only gentle compression was applied—just enough to immobilize and stabilize the breast. Core biopsy was performed with the patients under shallow sedation with 3–5 mg of midazolam (Dormicum; Hoffman-LaRoche, Grenzade-Wyhlen, Germany) or 5–10 mg of diazepam (Valium; Hoffman-LaRoche) intravenously administered just prior to the contrast material–enhanced series and with a dosage titrated to make the patients feel comfortable and relaxed yet awake and responsive.
Contrast-enhanced (intravenously injected 0.1 mmol per kilogram of body weight gadopentetate dimeglumine, Magnevist; Schering, Berlin, Germany) dynamic breast MR imaging was performed to reidentify the target lesion; parameters were kept equivalent to those used during diagnostic breast MR imaging (repetition time msec/echo time msec, 280/4.6; flip angle, 90°) (7,12). The field of view was adjusted to the single-breast imaging setting and set to 220 mm with an 80% rectangular field of view, yielding a 4-minute acquisition time. In addition, T2-weighted turbo spin-echo (SE) imaging (3,000/120) was performed with the same geometric parameters (acquisition time, 90 seconds). When the target lesion was identified on the postcontrast subtraction images, every attempt was made to identify the same lesion on the corresponding T2-weighted turbo SE images. Stereotactic coordinates of the lesion with respect to the external fiducial system were defined as described in reference 12, and an appropriate needle trajectory was calculated.
After thorough administration of superficial and deep local anesthesia (10–15 mL of lidocaine [Xylocaine; Astra Zeneca, Wedel, Germany] or bupivacaine [Carbostesin; Astra Zeneca]), the coaxial needle was placed through a small skin nick. Correct placement of the biopsy needle was verified with the T2-weighted turbo SE sequence. Depending on the yield of the individual needle passes, five to 11 cores were obtained, always after turning the biopsy notch in a clockwise direction. If the position of the needle with respect to the target lesion was inconclusive, a second contrast-enhanced series (T1-weighted turbo SE imaging before and after contrast material injection [350/10]; acquisition time, 90 seconds) was obtained with the core needle in place; alternatively, contrast-enhanced T1-weighted turbo SE images with spectral-selective fat suppression were used (acquisition time, 2 minutes 15 seconds).
Validation of Core Biopsy Diagnoses
Information on diagnosis validation and follow-up in all patients is given in Table E1 (radiology.rsnajnls.org/cgi/content/full/220/1/31/DC1). For all lesions, a thorough radiologic-pathologic correlation was obtained in close cooperation with an experienced breast pathologist (E.W.) who interpreted the biopsy specimens. However, the core biopsy diagnoses were considered validated only if either (a) excisional biopsy or mastectomy was performed on the lesion; (b) a follow-up period of 2 years, including breast MR imaging follow-up, was available; or (c) core biopsy resulted in a distinct and unambiguous histologic diagnosis that undoubtedly correlated with the MR imaging findings. In patients who received a diagnosis of breast cancer after MR imaging–guided core biopsy and were scheduled to undergo breast conservation therapy, the site of the lesions was localized with a hook wire with MR imaging guidance prior to oncologic surgery.
In one patient with one lesion rated as probably malignant, the core biopsy diagnosis could not be established owing to technical failure of the biopsy needle. Subsequent excisional biopsy after MR imaging–guided hook wire placement revealed invasive ductal cancer.
In the 58 patients (77 lesions) who underwent technically successful MR imaging–guided core biopsy, core biopsy results were validated as follows:
1. In 30 patients (42 lesions), excisional biopsy or mastectomy was performed. They included 22 patients with 28 breast cancers in whom excisional biopsy and/or mastectomy were performed for definitive oncologic surgery. In four of these 22 patients, biopsy was performed in four additional lesions, with benign results, but the lesions were resected because mastectomy or wide excision was performed for cancer elsewhere in the same breast.
In five patients with seven lesions with benign core biopsy diagnoses, secondary MR imaging–guided hook wire placement and excisional biopsy of the same lesions were performed later. This was done because at radiologic-pathologic correlation, the lesions’ core diagnoses were thought to not entirely explain the MR imaging findings (ie, radiologic-pathologic mismatch was suspected), such that needle misplacement could not be excluded.
In one patient who had a probably benign lesion, core biopsy revealed fibroadenoma plus lobular carcinoma in situ within the fibroadenoma, and the patient opted for resection of the fibroadenoma.
For reasons unrelated to breast MR imaging findings, one patient with a familial history of breast cancer underwent mastectomy 25 months after MR imaging–guided core biopsy.
In one high-risk patient who had a familial history of breast cancer, excisional biopsy was performed 20 months after MR imaging–guided biopsy of an equivocal lesion. This was done because at US there was a suspicious finding at the MR imaging–guided biopsy site. The patient opted for excisional biopsy of the US lesion.
2. In seven patients with seven lesions with benign findings at core biopsy, validation was obtained with a follow-up period of at least 2 years; follow-up included clinical, mammographic, breast US, and MR imaging studies.
3. In 10 patients with 10 lesions, validation was obtained with an unambiguous radiologic-pathologic correlation. This group included nine patients with nine lesions suspected of representing fibroadenomas at diagnostic MR imaging in whom the histologic core diagnosis also was fibroadenoma. One patient had bilateral well-circumscribed enhancing lesions, and core biopsy was performed on the right and excisional biopsy on the left; both revealed lymph node hyperplasia.
Thus, validation for the core diagnoses is offered for 47 patients with 59 lesions. These 59 lesions were used to calculate the sensitivity, specificity, and positive and negative predictive values of MR imaging–guided core biopsy. In 18 lesions, validation was considered insufficient. These lesions were not included in the calculation of the diagnostic accuracy of MR imaging–guided biopsy but were used to assess mean lesion size, procedure-related difficulties, and complications.
Data Analysis
For all patients who had breast cancer, the pT stage was recorded. The diagnostic yield of the MR imaging–guided core biopsies was evaluated by recording how often a pathologic diagnosis was established on the basis of the core material. The target lesion size, procedure duration (time in the magnet), and complications were evaluated by the radiologas discussed with the referring physician. The findings on the diagnostic MR images were discussed with the referring physician by one of the interpreting radiologists (C.K.K., N.M., or C.C.L.), a recommendation for further treatment was given, and the decision was registered in a spreadsheet. Thereafter, core biopsy was discussed as another possibility. After core biopsy, the actual patient treatment was recordeist who performed the biopsy (C.K.K., N.M., A.S., or C.C.L.) and were entered in a spreadsheet (N.M., A.S.). In each case, patient treatment wd in the same spreadsheet by the radiologist who had performed the biopsy (C.K.K., A.S., or N.M.). The intended treatment and the actual treatment were compared. The sensitivity, specificity, positive and negative predictive values, and diagnostic accuracy of MR imaging–guided core biopsy were calculated for the 59 lesions with established validation.
|
RESULTS
|
|---|
Size of Lesions in which Biopsy was Performed in Study Cohort
The average diameter of all target lesions was 14.6 mm (range, 6–30 mm). Five of the 78 lesions corresponded to diffusely enhancing areas suspected of representing ductal carcinoma in situ (n = 3) or diffusely infiltrating cancer (n = 2). If these five lesions were disregarded, the average size of the remaining focal lesions was 11.3 mm (size range, 6–18 mm).
Core Biopsy Findings
All findings are summarized in Table E1 (radiology.rsnajnls.org/cgi/content/full/220/1/31/DC1). Of the 58 patients who underwent technically successful core biopsy, 36 had core biopsy diagnoses of benign lesions (Fig 1), and 22 had breast cancer. Six of these 22 patients also had one or more benign lesions. When considered according to the number of lesions, core biopsy enabled identification of 50 benign lesions (including one borderline lesion, a radial scar) and 27 breast cancers. Bilateral breast cancer, with multicentric foci on the left side, was present in one patient; in another two patients, bilateral multicentric breast cancer was suspected, but core biopsy revealed breast cancer on only one side. Four patients received a diagnosis of breast cancer with multicentric growth; in four other patients, multicentric breast cancer was suspected, and core biopsy was performed in different quadrants of the same breast but revealed cancer in only one quadrant.
View larger version (98K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (104K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (135K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1c. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (116K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1d. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (100K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1e. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (88K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1f. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (93K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1g. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1h. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
View larger version (100K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1i. MR imaging-guided 14-gauge core biopsy of the right breast in a 48-year-old patient; an 8-mm equivocal lesion had been detected at diagnostic breast MR imaging (diagnostic image not shown). Histologic examination of the core biopsy specimen revealed myxoid fibroadenoma. Follow-up was 18 months. (a) Transverse precontrast T1-weighted gradient-echo MR image (280/4.6; flip angle, 90°). Arrowheads = fiducial system of stereotactic unit. (b) Transverse T1-weighted MR image (280/4.6) acquired after injection of 12 mL gadopentetate dimeglumine with the same parameters as in a shows the small well-circumscribed lesion (arrowheads) as having strong enhancement. (c) Transverse postcontrast subtraction image (b - a) shows the lesion as having high contrast. (d) Corresponding transverse T2-weighted turbo SE image (3,000/120) shows the lesion (arrowheads) with high signal intensity against the surrounding low-signal-intensity parenchyma. (e) Transverse postcontrast T1-weighted turbo SE image (350/10) obtained after attempted introduction of a 14-gauge core biopsy needle system shows the subcutaneous tissue (arrowheads) as displaced by the needle and shows that the target lesion itself is not visible (vanishing target). (f) Transverse T2-weighted turbo SE image (3,000/120) corresponding to e shows that the target lesion is still visible as a roundish hyperintense mass (arrowheads), as compared with the lesion in d. This suggests that the target position did not change despite the tissue shift at the needle insertion site. (g) Transverse T1-weighted turbo SE image (350/10) obtained after needle reinsertion and biopsy notch advancement through the calculated position of the target lesion (arrowheads) shows that the lesion itself is not visible owing to the vanishing target phenomenon. (h) Transverse subtraction T1-weighted turbo SE image with the same needle position and acquisition parameters as in g, obtained after a second injection of gadopentetate dimeglumine, shows the enhancing target lesion (arrowheads). On this image, the biopsy needle notch passing through the lesion is visible as faint signal void (arrows). This documents the correct needle position within the target lesion. (i) Transverse T2-weighted turbo SE image (3,000/120) obtained with the same needle position as in g with anteroposterior phase-encoding direction shows the reduced diameter of the needle-induced signal void (arrows), as compared with that in g. However, blurred image contours, probably owing to respiratory motion, also are seen. The target lesion is visible as a high-signal-intensity lesion (arrowheads).
|
|
On the basis of the final oncologic surgery, breast cancer stage pTis was present in three patients; stage pT1, in 15; stage pT2, in four (including the two patients with a mammographically visible index lesion); and pT4, (inflammatory type) in one. Twenty-three cancer stages were seen owing to the patient who had bilateral disease. The patient with stage pT4 disease was clinically suspected of having mastitis or inflammatory breast cancer. She had undergone repeated US- and clinically guided core biopsies with results that had remained negative. MR imaging–guided core biopsy revealed diffusely infiltrating poorly differentiated ductal carcinoma.
In one patient, MR imaging–guided core biopsy of the mammographically visible index lesion revealed a radial scar. After excisional biopsy, a radial scar was found at the biopsy site, but the lesion was surrounded by a pT2 ductal invasive breast cancer.
In 42 patients, 50 benign lesions were found; included are the 36 patients with 42 lesions rated as probably benign or equivocal and the six patients who had eight benign lesions and underwent biopsy to rule out multicentric or contralateral breast cancer. Table 2 shows the histologic diagnoses in the benign lesions.
In all five patients who had benign core biopsy diagnoses in seven lesions and underwent secondary excisional biopsy owing to suspected radiologic-pathologic mismatch, excisional biopsy results confirmed the core biopsy diagnosis. In one patient who had a core biopsy diagnosis of fibroadenoma plus incidental lobular carcinoma in situ, excisional biopsy was performed, and findings confirmed lobular carcinoma in situ within a fibroadenoma. One patient developed a US-visible lesion at the MR imaging–guided biopsy site 20 months after MR imaging–guided biopsy and underwent excisional biopsy, although MR imaging follow-up had revealed no changes. Findings at excisional biopsy corresponded entirely to those obtained at core biopsy 20 months earlier.
Diagnostic Yield and Accuracy of Core Biopsy Specimens
In 77 (99%) of the 78 lesions, core biopsy yielded material rated as sufficient to establish a firm histologic diagnosis. If we disregard the core biopsies of lesions without appropriate validation, we may calculate the diagnostic accuracy of core biopsy in 47 patients with 59 lesions. If we consider the patient who received a core biopsy diagnosis of radial scar and an excisional biopsy diagnosis of radial scar plus surrounding invasive breast cancer as a false-negative core biopsy, the following results emerge: True-positive diagnoses were obtained in 27 lesions; true-negative diagnoses, in 31 lesions; false-negative diagnosis, in one lesion; and false-positive diagnoses, in no lesion. Thus, the following diagnostic indices for MR imaging–guided core biopsy can be calculated: Sensitivity was 96% (27 of 28 lesions); specificity, 100% (31 of 31 lesions); positive predictive value, 100% (27 of 27 lesions); negative predictive value, 97% (31 of 32 lesions); and diagnostic accuracy, 98% (58 of 59 lesions).
Effect on Patient Treatment
In the 16 patients with 19 probably benign lesions, the lesions would have been managed with follow-up if no core biopsy facility had been available. Accordingly, the negative core biopsy findings could have obviated further follow-up MR imaging. Still, follow-up MR imaging studies were performed or are scheduled as part of the follow-up protocol of this trial. Thus, core biopsy results did not affect patient treatment. However, in one of the 16 patients, core biopsy revealed lobular carcinoma in situ within a fibroadenoma, which changed treatment from follow-up to excision. One patient who was suspected of having a radiologic-pathologic mismatch opted for excision instead of follow-up.
In the 20 patients with 23 equivocal lesions, these patients would have been treated with excisional biopsy if no core biopsy facility had been available. In 16 of the 20 patients, with 17 equivocal lesions, the benign histologic core diagnoses obviated surgical biopsy. In the remaining four patients with six lesions, secondary excisional biopsy was performed to definitively rule out a false-negative core biopsy result.
In the 22 patients with 35 probably malignant lesions, MR imaging–guided core biopsy revealed breast cancer in 27 lesions and benign changes in eight. Because the histologic diagnosis of breast cancer was available, treatment was changed in all patients. A one-step surgical approach was made possible in 21 patients; in the patient who had inflammatory breast cancer, induction chemotherapy was initiated. In four patients, biopsy was performed on multicentric foci, and treatment was changed from breast conservation to mastectomy with immediate reconstructive surgery. In two patients who had unilateral core biopsy–proved breast cancer, contralateral breast cancer had been suspected at preoperative staging breast MR imaging. However, core biopsy revealed only benign changes, such that the patients opted for follow-up instead of contralateral surgery.
In summary, MR imaging–guided core biopsy changed patient treatment in that it resulted in excision instead of follow-up in two patients and obviated surgical biopsy in 16 patients with 17 equivocal lesions and excisional biopsy in one patient who had inflammatory breast cancer. It also changed the surgical approach in 22 patients who had 27 breast cancers and eight benign lesions by allowing one-step surgery, change in the surgical approach from breast conservation to mastectomy, and/or obviation of contralateral surgery.
Accordingly, MR imaging–guided core biopsy resulted in a change in treatment or management plans in 40 (69%) of 58 patients and 54 (70%) of 77 lesions.
Specific Difficulties and Complications
The procedure was tolerated well by all 59 patients. No major complication occurred; one minor complication arose in one patient: A significant hematoma developed at and around the biopsy site, with diffuse skin discoloration; however, the hematoma subsided without specific therapy. Difficulties encountered during the MR imaging–guided procedures were caused by the specific features of the MR imaging–compatible biopsy devices, as well as by the "vanishing target" (12) phenomenon.
With the exception of the semiautomatic 14-gauge coaxial system, the performance of the MR imaging–compatible biopsy guns was poor. This was particularly true for one coaxial system, in which the backloading core needle tended to get stuck in the outer cannula, such that the biopsy specimen was sheared off the biopsy notch. In addition, because a majority of needles were relatively blunt, they delivered empty biopsy notches in a substantial number of needle passes. Therefore, several (up to 12) needle passes were necessary to obtain sufficient material. Because the blunt needles tended to push the parenchyma rather than punch it (Fig 2), severe displacement of the parenchyma occurred in six of 59 patients, thus invalidating the previously calculated stereotactic coordinates. In these cases, the procedure had to be discontinued until the lesion’s location was reidentified with another contrast-enhanced series. Since a newer semiautomatic biopsy gun (Daum Medical Systems) became available, difficulties related to the insufficient cutting capabilities have been substantially reduced.
View larger version (94K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. MR imaging-guided 14-gauge core biopsy of the left breast in a 49-year-old high-risk patient with a familial history suggestive of breast cancer. A 6-mm lesion with suggestive enhancement had been detected at screening breast MR imaging (screening image not shown). MR imaging-guided core biopsy revealed a 6-mm focus of adenosis and epitheliosis without atypia. Because, owing to the vanishing target, the lesion was invisible during intervention, and because radiologic-pathologic mismatch could not be excluded, secondary excisional biopsy was performed after MR imaging-guided hook wire placement. Excisional biopsy findings confirmed the core biopsy diagnosis of focal adenosis without atypia. Follow-up was 8 months. (a) Transverse precontrast T1-weighted gradient-echo image (280/4.6; flip angle, 90°) from the dynamic series obtained prior to biopsy. As in Figure 1a, the breast is immobilized in the mediolateral direction by the two compression plates. The white dots (arrowheads) medial and lateral to the breast are part of the stereotactic unit’s fiducial system. (b) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a shows a 6-mm enhancing target lesion (arrowheads). (c) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a and b, obtained after administration of local anesthetic and placement of the needle phantom (arrowheads) to simulate the calculated needle trajectory, shows the lesion as only faintly visible with reversed contrast (vanishing target) because of progressive contrast enhancement in the adjacent parenchyma combined with a rapid washout of contrast material in the lesion. (d) Transverse T2-weighted turbo SE image (3,000/120) obtained after introduction of the 14-gauge core biopsy needle shows that the target at the calculated stereotactic coordinates (arrowheads) is not visible. No tissue shift is induced with the needle. (e) Transverse T2-weighted turbo SE image (3,000/120) obtained after firing the biopsy gun shows that the needle has passed exactly through the calculated position of the target (arrowheads) and that the target itself is not visible.
|
|
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. MR imaging-guided 14-gauge core biopsy of the left breast in a 49-year-old high-risk patient with a familial history suggestive of breast cancer. A 6-mm lesion with suggestive enhancement had been detected at screening breast MR imaging (screening image not shown). MR imaging-guided core biopsy revealed a 6-mm focus of adenosis and epitheliosis without atypia. Because, owing to the vanishing target, the lesion was invisible during intervention, and because radiologic-pathologic mismatch could not be excluded, secondary excisional biopsy was performed after MR imaging-guided hook wire placement. Excisional biopsy findings confirmed the core biopsy diagnosis of focal adenosis without atypia. Follow-up was 8 months. (a) Transverse precontrast T1-weighted gradient-echo image (280/4.6; flip angle, 90°) from the dynamic series obtained prior to biopsy. As in Figure 1a, the breast is immobilized in the mediolateral direction by the two compression plates. The white dots (arrowheads) medial and lateral to the breast are part of the stereotactic unit’s fiducial system. (b) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a shows a 6-mm enhancing target lesion (arrowheads). (c) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a and b, obtained after administration of local anesthetic and placement of the needle phantom (arrowheads) to simulate the calculated needle trajectory, shows the lesion as only faintly visible with reversed contrast (vanishing target) because of progressive contrast enhancement in the adjacent parenchyma combined with a rapid washout of contrast material in the lesion. (d) Transverse T2-weighted turbo SE image (3,000/120) obtained after introduction of the 14-gauge core biopsy needle shows that the target at the calculated stereotactic coordinates (arrowheads) is not visible. No tissue shift is induced with the needle. (e) Transverse T2-weighted turbo SE image (3,000/120) obtained after firing the biopsy gun shows that the needle has passed exactly through the calculated position of the target (arrowheads) and that the target itself is not visible.
|
|
View larger version (113K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2c. MR imaging-guided 14-gauge core biopsy of the left breast in a 49-year-old high-risk patient with a familial history suggestive of breast cancer. A 6-mm lesion with suggestive enhancement had been detected at screening breast MR imaging (screening image not shown). MR imaging-guided core biopsy revealed a 6-mm focus of adenosis and epitheliosis without atypia. Because, owing to the vanishing target, the lesion was invisible during intervention, and because radiologic-pathologic mismatch could not be excluded, secondary excisional biopsy was performed after MR imaging-guided hook wire placement. Excisional biopsy findings confirmed the core biopsy diagnosis of focal adenosis without atypia. Follow-up was 8 months. (a) Transverse precontrast T1-weighted gradient-echo image (280/4.6; flip angle, 90°) from the dynamic series obtained prior to biopsy. As in Figure 1a, the breast is immobilized in the mediolateral direction by the two compression plates. The white dots (arrowheads) medial and lateral to the breast are part of the stereotactic unit’s fiducial system. (b) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a shows a 6-mm enhancing target lesion (arrowheads). (c) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a and b, obtained after administration of local anesthetic and placement of the needle phantom (arrowheads) to simulate the calculated needle trajectory, shows the lesion as only faintly visible with reversed contrast (vanishing target) because of progressive contrast enhancement in the adjacent parenchyma combined with a rapid washout of contrast material in the lesion. (d) Transverse T2-weighted turbo SE image (3,000/120) obtained after introduction of the 14-gauge core biopsy needle shows that the target at the calculated stereotactic coordinates (arrowheads) is not visible. No tissue shift is induced with the needle. (e) Transverse T2-weighted turbo SE image (3,000/120) obtained after firing the biopsy gun shows that the needle has passed exactly through the calculated position of the target (arrowheads) and that the target itself is not visible.
|
|
View larger version (90K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2d. MR imaging-guided 14-gauge core biopsy of the left breast in a 49-year-old high-risk patient with a familial history suggestive of breast cancer. A 6-mm lesion with suggestive enhancement had been detected at screening breast MR imaging (screening image not shown). MR imaging-guided core biopsy revealed a 6-mm focus of adenosis and epitheliosis without atypia. Because, owing to the vanishing target, the lesion was invisible during intervention, and because radiologic-pathologic mismatch could not be excluded, secondary excisional biopsy was performed after MR imaging-guided hook wire placement. Excisional biopsy findings confirmed the core biopsy diagnosis of focal adenosis without atypia. Follow-up was 8 months. (a) Transverse precontrast T1-weighted gradient-echo image (280/4.6; flip angle, 90°) from the dynamic series obtained prior to biopsy. As in Figure 1a, the breast is immobilized in the mediolateral direction by the two compression plates. The white dots (arrowheads) medial and lateral to the breast are part of the stereotactic unit’s fiducial system. (b) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a shows a 6-mm enhancing target lesion (arrowheads). (c) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a and b, obtained after administration of local anesthetic and placement of the needle phantom (arrowheads) to simulate the calculated needle trajectory, shows the lesion as only faintly visible with reversed contrast (vanishing target) because of progressive contrast enhancement in the adjacent parenchyma combined with a rapid washout of contrast material in the lesion. (d) Transverse T2-weighted turbo SE image (3,000/120) obtained after introduction of the 14-gauge core biopsy needle shows that the target at the calculated stereotactic coordinates (arrowheads) is not visible. No tissue shift is induced with the needle. (e) Transverse T2-weighted turbo SE image (3,000/120) obtained after firing the biopsy gun shows that the needle has passed exactly through the calculated position of the target (arrowheads) and that the target itself is not visible.
|
|
View larger version (86K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2e. MR imaging-guided 14-gauge core biopsy of the left breast in a 49-year-old high-risk patient with a familial history suggestive of breast cancer. A 6-mm lesion with suggestive enhancement had been detected at screening breast MR imaging (screening image not shown). MR imaging-guided core biopsy revealed a 6-mm focus of adenosis and epitheliosis without atypia. Because, owing to the vanishing target, the lesion was invisible during intervention, and because radiologic-pathologic mismatch could not be excluded, secondary excisional biopsy was performed after MR imaging-guided hook wire placement. Excisional biopsy findings confirmed the core biopsy diagnosis of focal adenosis without atypia. Follow-up was 8 months. (a) Transverse precontrast T1-weighted gradient-echo image (280/4.6; flip angle, 90°) from the dynamic series obtained prior to biopsy. As in Figure 1a, the breast is immobilized in the mediolateral direction by the two compression plates. The white dots (arrowheads) medial and lateral to the breast are part of the stereotactic unit’s fiducial system. (b) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a shows a 6-mm enhancing target lesion (arrowheads). (c) Transverse postcontrast T1-weighted gradient-echo image with acquisition parameters equivalent to those in a and b, obtained after administration of local anesthetic and placement of the needle phantom (arrowheads) to simulate the calculated needle trajectory, shows the lesion as only faintly visible with reversed contrast (vanishing target) because of progressive contrast enhancement in the adjacent parenchyma combined with a rapid washout of contrast material in the lesion. (d) Transverse T2-weighted turbo SE image (3,000/120) obtained after introduction of the 14-gauge core biopsy needle shows that the target at the calculated stereotactic coordinates (arrowheads) is not visible. No tissue shift is induced with the needle. (e) Transverse T2-weighted turbo SE image (3,000/120) obtained after firing the biopsy gun shows that the needle has passed exactly through the calculated position of the target (arrowheads) and that the target itself is not visible.
|
|
In 49 of the 78 lesions, the vanishing target phenomenon (Figs 1, 2) necessitated an additional contrast-enhanced dynamic series to verify the position of the core needle with respect to the target lesion. In the remaining cases, T2-weighted turbo SE imaging was used to document correct needle placement within the lesion.
Owing to the broad 4-mm signal void caused by the core biopsy system, small (6–10-mm) lesions were obscured once the biopsy needle was placed into the lesion (Fig 2e). If the phase-encoding direction was chosen perpendicular to the needle’s long axis, the signal void was only about 2 mm in diameter; however, in this setting, images are blurred because of translated respiratory and cardiac motion artifacts, and, more important, the needle position may be displayed up to 2 mm away from its actual position in vivo. Therefore, in the cases with small target lesions (<10 mm), the needle position was controlled with the central core needle extended (ie, with the biopsy notch placed within the lesion and with parallel phase-encoding direction). With the extended biopsy notch, the signal void of the biopsy needle was much smaller, such that a direct verification of correct needle placement was feasible even in small lesions (Fig 1g, 1h).
The average time in the magnet for the complete procedure was 60 minutes (range, 45–100 minutes [the latter for two biopsies]). This time frame included repeated verification of accurate needle position after several passes.
|
DISCUSSION
|
|---|
Imaging-guided localization and fine-needle or core biopsy of nonpalpable lesions are established techniques in mammography and breast US. They allow definite and tissue-sparing clarification of suspicious lesions that are clinically occult (14–23). Since it is the declared goal to identify breast cancers at even earlier stages, the need to diagnose clinically occult lesions is ever increasing. The necessity of histologic verification of lesions visible only at breast MR imaging has grown in parallel with the increasing availability of breast MR imaging and the increasing demand for preoperative MR imaging (6,7).
Yet, as in conventional imaging, once a nonpalpable suspicious lesion has been identified at breast MR imaging alone, histologic proof of the lesion must be obtained before it can be allowed to effect any therapeutic decisions. It is clearly the responsibility of the radiologist not only to detect and classify subclinical lesions, but also to ensure secure but tissue-sparing histologic verification of these lesions.
Until recently, most experience consisted of MR imaging–guided needle localization and excisional biopsy of MR imaging–only suspicious lesions. This technique is simple, rapid, and safe; it is meanwhile used as a routine clinical tool on a broad scale (9–13). However, because surgical biopsy must be performed for the actual tissue sampling, this technique does not seem to be appropriate for the many MR imaging cases of equivocal or probably benign lesions, in which, for one reason or another, histologic proof of the lesion is thought necessary. In addition, MR imaging–guided lesion localization is no solution in patients who have probably malignant MR imaging–only lesions, in whom preoperative tissue sampling is desired to allow a one-step surgical approach.
MR imaging–guided vacuum-assisted biopsy has been described recently as allowing direct incisional biopsy of lesions with MR imaging guidance (29); however, this approach requires extensive dedicated technical equipment that may not achieve widespread use in the foreseeable future. With the development of MR imaging–compatible biopsy guns and dedicated stereotactic breast biopsy devices, MR imaging–guided stereotactic large-core biopsy has become feasible. Still, reports on this technique are few. Fischer et al (10) discussed MR imaging–guided core biopsy in five patients; a majority of the remainder of the literature on MR imaging–guided breast biopsy concerns lesion localization. This study served to investigate the accuracy and clinical usefulness of MR imaging–guided large-core (14-gauge) breast biopsy and assess its effect in terms of clinical patient treatment.
It is widely accepted that large-core needle biopsy with mammographic or US guidance reaches diagnostic accuracy equivalent to that of imaging-guided localization and excisional biopsy procedures (14–23). The influence of core-needle size on the reliability of core-needle histologic findings has been extensively investigated (30,31); investigators in these studies have documented the superiority of large-core (11- and 14-gauge) over small-core biopsy. Hence, it is generally recommended that, to maximize diagnostic safety, stereotactic breast biopsy be performed with large-core needles (33,34). Because MR imaging–guided core biopsy is a cost-intensive procedure, it seems even more important to provide an optimum diagnostic yield.
Our results indicate that MR imaging–guided large-core biopsy allows histologic verification of subclinical MR imaging–only lesions with sufficient accuracy and reliability. It seems noteworthy that a definitive histologic diagnosis was achieved in 77 (99%) of the 78 lesions, which suggests that the cores retrieved with the 14-gauge needles provide quantitatively and qualitatively satisfactory tissue specimens. The success rate of MR imaging–guided core biopsy matches well with that of previous reports on mammography- or US-guided large-core biopsy (14,33).
It is encouraging that among the 47 patients with validated diagnoses in 59 lesions, core results matched the patients’ final diagnoses in all but one case. In the patient with discrepant results, MR imaging–guided core biopsy had revealed a radial scar, whereas excisional biopsy revealed a radial scar plus surrounding invasive breast cancer. Invasive breast cancer and radial scars have been reported to exhibit equivalent MR imaging features; therefore, we feel that this false-negative core biopsy finding was a matter of sampling error rather than of needle placement inaccuracy. Because radial scar is considered a high-risk lesion that usually is treated with excisional biopsy, this false-negative core biopsy finding had no adverse effect on patient treatment.
Because in 10 patients, validation was based merely on radiologic-pathologic correlation, there is a slight possibility that a false-negative core result was not identified. However, in a patient suspected of having fibroadenoma at MR imaging and with fibroadenoma tissue seen at core histologic examination, this is only possible if the needle missed the actual target and instead sampled a fibroadenoma that was not visible at breast MR imaging. Because this is improbable, we think it is justifiable to include these lesions as sufficiently validated.
MR imaging–guided core biopsy results affected patient treatment in 40 (69%) of 58 patients and 54 (70%) of 77 lesions: It resulted in surgery in two patients but was used to avoid surgical biopsy in 16 patients, allowed immediate induction chemotherapy in one patient, and made one-step oncologic surgery feasible in another 21 patients. In addition, in 22 patients with breast cancer, core biopsy obviated surgical biopsy in two patients suspected of having contralateral breast cancer and in four suspected of having multicentric breast cancer.
It is important to note that this study focused on the feasibility and diagnostic accuracy of MR imaging–guided core biopsy. Therefore, and for patient safety reasons, follow-up MR imaging was or will be performed in all patients with benign core biopsy diagnoses; thus, no change in treatment has been achieved in these patients. Once the diagnostic reliability of MR imaging–guided core biopsy is established, benign core biopsy results in lesions rated as probably benign can probably be used to avoid follow-up MR imaging. Accordingly, we believe that our present data should underestimate the possible effects of MR imaging–guided core biopsy on patient treatment.
In five patients with seven lesions, secondary excisional biopsy was performed because we were in doubt as to whether the target had been missed. This was in part caused by the fact that our confidence in the accuracy of MR imaging–guided core biopsy was affected by the various difficulties encountered during this procedure: First, with one exception of a recently designed biopsy gun (Baum Medical Systems, Schwerin, Germany), MR imaging–compatible core biopsy needles were relatively blunt. Therefore, the parenchyma was displaced rather than penetrated (Fig 1e, 1f). This led to a gross change of the configuration of the entire breast, including the position of the target lesion. In such a situation, the stereotactic coordinates had to be redefined.
However, redefinition of coordinates was difficult or even impossible owing to fading lesion visibility over time (vanishing target; Figs 1e, 2c) (12). The vanishing target is not problematic once the stereotactic coordinates of a lesion have been determined and as long as these coordinates are valid. It causes major difficulties, however, if target reidentification becomes necessary during the intervention owing to, for example, inadequate performance of MR imaging–compatible biopsy devices. A solution is to try to reidentify the target lesion on the corresponding T2-weighted turbo SE images, on which image contrast remains constant. Because target visibility is maintained for only a few minutes, repetitive control imaging is unreliable for documenting the position of the needle with respect to the target. Accordingly, freehand real-time approaches (13) may be inadequate for MR imaging–guided core biopsy, and stereotactic techniques are preferable. For the same reason, we think that MR imaging–guided breast biopsy may not profit from an open magnet configuration. Open magnets allow direct needle manipulation with MR fluoroscopic control; however, this is only advantageous as long as the lesion can be visualized as a target. If, as in breast MR imaging, lesion visibility is preserved for only a short time, then a stereotactic approach with closed-bore magnets seems more suitable.
Another difficulty, and clearly a drawback of the large-core biopsy approach, was the signal void caused by the large-core biopsy systems. Although the signal void was relatively small considering the actual diameter of the coaxial biopsy systems, it was still broad enough to completely obscure small enhancing lesions (Fig 2e). Therefore, needle position should be verified with the core-needle notch extending into the lesion, because the signal void caused by the notch is only half the size of the entire biopsy system’s susceptibility artifact.
It is encouraging that, in spite of all the mentioned difficulties, the average lesion size was 11.3 mm, with a minimum lesion size of 6 mm; this suggests that MR imaging–guided large-core biopsy is sufficiently accurate to depict lesions at a prognostically relevant stage.
Concerning the objectives of our study, the following statements can be made: 14-gauge large-core breast biopsy is feasible with MR imaging guidance in terms of technical and clinical success. The technique seems adequately accurate for use as a clinical tool in patients who have MR imaging–only visible lesions, provided that thorough radiologic-pathologic correlation is established and appropriate action is taken if mismatch is suspected.
The most important difficulty specifically associated with MR imaging guidance of large-core breast biopsy has been the unsatisfactory performance of the fully MR imaging–compatible biopsy devices. Combining this with the fading visibility of the target lesions during the intervention meant a substantial threat of missing the target. Although, eventually, adequate histologic specimens of the target lesion were obtained in 77 of the 78 lesions, these difficulties led to substantial prolongation of the entire procedure and gave rise to serious doubt of the credibility of the biopsy results. Currently, much effort is being spent to improve the performance of MR imaging–compatible biopsy devices so that these difficulties can be anticipated to subside in the near future.
Already in this preliminary study with emphasis on feasibility and patient safety, MR imaging–guided core biopsy enabled treatment change in 40 (69%) of 58 patients by obviating surgical biopsy and allowing a one-step surgical procedure in patients with breast cancers. These results suggest that MR imaging–guided core biopsy can be integrated into clinical patient treatment. This seems even more important if MR imaging is used as a screening tool in high-risk patients (7). In this context, MR imaging–guided core biopsy may become an indispensable tool to help distinguish fibroadenomas from medullary breast cancer.
With MR imaging, a technique is at hand that allows the diagnosis of invasive breast cancer with almost unlimited sensitivity. However, this extraordinary diagnostic tool can only be used to the advantage of our patients if the many sma
TOP
ABSTRACT
INTRODUCTION
