Contrast-enhanced mammography-guided biopsy: technical feasibility and first outcomes

This was a retrospective review of a cohort prospectively enrolled in a pre-marketing clinical validation and feasibility study performed at Hospital del Mar, a university hospital in Barcelona, Spain (ClinicalTrials.​gov ID: NCT05250674). We included all consecutive adult women presenting with a clinical indication for CEM-guided breast biopsy and considered eligible for this procedure as per standard of care between October 2019 and September 2021. CEM-guided biopsy was clinically indicated for suspicious enhancing lesions, not reliably identified by other morphological breast imaging techniques, including ‘second-look’ US or FFDM. Exclusion criteria were contraindications to iodinated contrast media. The study was approved by the hospital ethics committee (protocol number 2019/8890) and all patients provided written informed consent. A pre-marketing version of the Pristina Serena Bright™ GE Healthcare system had been under clinical validation in this study since 2019. The system received CE mark and Food and Drug Administration (510k) pre-marketing notification [20]. All the biopsy procedures were carried out using a 7- or 10-gauge (G) vacuum-assisted biopsy (VAB) needle.

The procedure workflow is depicted in Fig. 1.

The first step of any mammographic-guided procedure should be a prior complete review of the case to define the optimal approach, based on the location of the lesion in the breast and the patient’s characteristics, such as breast thickness and any potential physical limitations. As the evaluated equipment is an upright biopsy add-on unit, the patient can be seated or lying in a lateral decubitus position, with the help of a decubitus breast imaging table. The same dedicated needle support can be used for both the vertical and horizontal approach. The optimal approach should be planned on the diagnostic CEM that leads to the indication for CEM-guided biopsy. When the suspicious enhancing finding was originally seen on MRI, a pre-procedural CEM was performed to allow correct planning.

CEM-guided biopsy uses the principle of conventional stereotactic guidance, with the additional preliminary step of intravenous iodinated contrast media administration. As with diagnostic CEM, a 2-min interval between contrast administration and compression is necessary to avoid interference with blood inflow to the breast, allowing adequate lesion enhancement. The focal compression applied during the biopsy procedure seems to minimise lesion washout and contrast uptake is visible for up to 10 min, which is sufficient for proper targeting.

At the end of the required 2-min interval, the patient is positioned, the area where the target lesion is expected to be found is centred on the compression paddle, and a dual-energy image is acquired at 0°. Once the inclusion of the enhancement in the scout image has been confirmed, a pair of dual-energy angled views is acquired at ± 15°. In each of the angled images, the target is identified and marked, so that the device can calculate the 3 location coordinates (X, Y and Z), allowing the biopsy needle holder to be automatically positioned.

After antisepsis and local anaesthesia, the needle is inserted into the breast, until reaching the limit point defined by the support itself, when the tip of the needle would face the target. A pre-fire stereo pair of dual-energy views is then performed to confirm that the target is in front of the needle tip; otherwise, minimal coordinate adjustments may improve precision. Next, once the fire-forward is activated, the sampling is carried out with the available vacuum system device. After sampling, the biopsy bed will be contaminated by blood with contrast media, so the pre-fire image will be the last to benefit from the dual technique.

It is highly important to place a radiological marker on the biopsy bed in order to evaluate procedural effectiveness and enable further localisation if the results reveal malignancy. It is recommended to check marker release with a DBT scout image, since this allows its precise location with a single acquisition, and therefore with a lower average glandular dose (AGD) than stereotactic acquisitions [21]. At the end of the procedure, the correct location of the radiological marker is confirmed, with the usual 2 mammographic views (mediolateral and craniocaudal), which are compared with the pre-procedural low-energy images.

Clinical data of the included patients and suspicious lesions were collected, including patient age, breast density, background parenchymal enhancement (BPE) and the size and type of enhancement of the target lesion. Breast density was visually assessed on the low-energy image of the diagnostic CEM according to breast composition categories of the American College of Radiology (ACR) BI-RADS® mammography lexicon [22] and collected from the radiology report. BPE was visually graded in the diagnostic recombined image into 4 categories (minimal, mild, moderate and marked), according to the BI-RADS® MRI lexicon [23], applied for CEM [24]. Similarly, the reported size of the suspicious lesion was collected and the type of enhancement was classified in focus or foci (< 5 mm), mass enhancement (> 5 mm) and non-mass enhancement [23, 24].

Procedural characteristics, i.e. patient position (seated/lying), the kind of approach (horizontal/vertical) and needle gauge, were collected for each procedure.

For successful procedures, we collected the histopathological results on the biopsy specimen, as well as the kind of treatment undergone by patients. For patients who underwent surgery, we collected the pathological results of the surgical specimen.

To evaluate the feasibility of CEM-guided biopsy, various endpoints were taken into account.

The success rate was calculated. Procedural success was defined as the visualisation of the suspicious enhancing lesion during procedural CEM, allowing adequate targeting and sampling, together with post-procedure imaging showing biopsy changes at the site of the intended lesion and with the clip in place or nearby in sequential mammography. Other cases were classified as cancelled, when the enhancing lesion visible at diagnostic CEM was not visible on procedural CEM, leading to non-performance of biopsy and short-term interval follow-up. Biopsies were classified as unsuccessful when the enhancing lesion was less visible on all procedural CEM views, limiting proper target selection, but the biopsy could still be performed by using another reference and guidance.

Procedural time was registered, considering the time span from contrast administration to clip placement scout view, immediately prior to breast decompression. Moreover, the number of scout acquisitions needed to represent the enhancing target in the field of view was collected. The complication rate was calculated, including immediately visible hematomas, vasovagal reactions, allergic reactions and late complications such as hematomas or infection.

For patients who underwent surgery or percutaneous VAE, the upgrade rate was calculated as the percentage of cases originally graded as B3 on the biopsy specimen, according to the NHS Breast Screening Programme pathology classification [25, 26], and then found to be malignant on the surgical or excision specimen, as well as cases diagnosed as ductal carcinoma in situ (DCIS) on biopsy and upstaged to invasive ductal carcinoma (IDC) on the surgical specimen.

From October 2019 to September 2021, 64 patients were referred for CEM-guided breast biopsy. Of these, 2 underwent 2 biopsies for 2 different suspicious lesions (performed on the same day, in one patient with only one contrast administration), making a total of 66 procedures (Table 1). The median age of the included patients was 59 years, and breast density was predominantly classified as B (35.9%) and C (59.4%), whilst BPE was mostly minimal (48.4%) or mild (32.8%). Suspicious enhancing lesions were mass enhancement in 43.9% of the procedures, focus/foci in 18.2% and non-mass enhancement in 37.9% (Figs. 2 and 3). The median size of the suspicious lesion was 8.5 mm. Median time from diagnostic CEM to CEM-guided biopsy was 12 (6–21) days. For the 10 patients with an MRI-detected suspicious lesion, the CEM-guided biopsy was performed 15 (10–23) days after the MRI.

Of the 66 suspicious lesions with an indication for CEM-guided biopsy, the procedure was successful in 63 (95.4%) lesions (Table 2). For the remaining 3 lesions, the enhancement classified as suspicious at diagnostic CEM was not clearly visible at procedural CEM. In 2 of these 3 patients, the biopsy was cancelled. In the remaining patient, who had a mild BPE, although the enhancement was not clearly visible in all views, a small distortion was seen at the time of the procedure with a complementary DBT scout, and biopsy was performed under DBT guidance, resulting in IDC (Fig. 4). Of the 2 patients with cancelled biopsies, one had a further follow-up CEM after 6 months showing no suspicious findings, whilst the second had no available follow-up CEM, although there was no evident suspicious enhancement at the CEM retest performed on the same day as the planned procedure (Supplementary Fig. 1). The original suspicious findings in these 2 patients were a 55-mm non-mass enhancement and a 6-mm mass enhancement, and BPE was moderate in the first and mild in the second.

As for the procedural characteristics of successful CEM-guided biopsies, the patients were most frequently seated (n = 52, 82.5%) and the preferred approach was horizontal (n = 48, 76.2%). The median total time per procedure was 15 min. The number of acquisitions needed before the targeting was between 1 and 4 (median 2 acquisitions). In most procedures (n = 4, 69.8%), no complications were registered. Self-limiting early hematomas were recorded in 14 (22.2%) patients and vasovagal reactions in 2 (3.2%). Late complications consisted of only moderate hematomas (3/63, 4.8%). No severe complications or allergic reactions were registered.

Of the 63 lesions that underwent successful CEM-guided biopsy, 2 (3.2%) were B1 on biopsy specimen, 26 (41.3%) were B2 (predominantly fibroadenomas, fat necrosis and adenosis/sclerosis), 10 (15.9%) were B3 (predominantly atypical ductal hyperplasia) and 25 (39.7%) were B5, including IDC (n = 10), DCIS (n = 7), a combination of IDC and DCIS (n = 1) and invasive lobular carcinoma (n = 7) (Fig. 5). The median size of the suspicious enhancement according to histopathological diagnosis is reported in Table 3. After radiological-pathological correlation, only 2 patients with discordant results were referred to short-term follow-up; one of these showed no suspicious enhancement after 6 months and the other showed remaining enhancement at 6 months, resulting in a re-biopsy close to the clip, still with B2 histopathological results. Of the 35 lesions that were treated, whether surgically or through percutaneous VAE, 6 (17.1%) were upgraded, including 4/10 B3 lesions which were upgraded to malignancy, and 2/7 DCIS which were upstaged to IDC (Table 3).