Ductal Carcinoma In Situ (DCIS) Diagnosed by MRI-Guided Biopsy among BRCA1/BRCA2 Mutation Carriers

Background While BRCA1/BRCA2 pathogenic sequence variants (PSVs) clearly confer an increased risk for invasive breast cancer, the extent to which these mutant alleles increase DCIS risk is less clear. Objective To assess the rate of detection over a 5-year period, and MRI imaging features of pure noncalcified DCIS in a cohort of Israeli BRCA1/BRCA2 PSV carriers attending a high-risk clinic from 2015 to 2020. Materials and Methods All female BRCA1/BRCA2 PSV-carriers followed at the Meirav High-risk clinic from 2015 to 2020 were eligible if they underwent semiannual breast imaging (MRI/mammography) and MRI-guided biopsy-proven pure DCIS. Clinical data, pathology information, and imaging characteristics were retrieved from the computerized archiving system. Results 18/121 (15.2%) participating BRCA1 PSV carriers and 8/81 (10.1%) BRCA2 PSV-carriers who underwent MRI-guided biopsy were diagnosed with DCIS. The median age of BRCA1 carriers and BRCA2 carriers was 49.8 years and 60.6 years, respectively (p = 0.55). Negative estrogen-receptor tumors were diagnosed in 13/18 (72%) BRCA1 and 2/8 (25%) BRCA2 PSV carriers (p < 0.05). Thirteen (13/18–72%) BRCA1 carriers had intermediate to high-grade or high-grade DCIS compared with 4/8 (50%) of BRCA2 carriers (p = 0.03). Over the 5-year study period, 29/1100 (2.6%) BRCA1/BRCA2 PSV carriers were diagnosed with DCIS seen on MRI only. Conclusion MRI-detected noncalcified DCIS is more frequent in BRCA1 PSV carriers compared with BRCA2 carriers, unlike the BRCA2 predominance in mammography-detected calcified DCIS. BRCA1-related DCIS is diagnosed earlier, more likely to be estrogen receptor-negative and of higher grade compared with BRCA2-related DCIS. Future prospective studies should validate these results and assess the actual impact they might have on clinical management of BRCA PSV carriers.


Introduction
Women harboring germline BRCA1 or BRCA2 pathogenic sequence variants (PSVs) are at a substantially increased risk for developing breast (and ovarian) cancer (BC). While actual breast cancer risk differs by study design and case selection, the most comprehensive updated risk estimates are 72% and 69% lifetime risk for BRCA1 and BRCA2 PSV carriers, respectively [1].
While risk is clearly increased for invasive BC, the extent to which these mutant alleles are associated with an increased risk for DCIS is less clear. Originally, a lower incidence of pure DCIS or DCIS accompanying invasive BC in BRCA1/BRCA2 PSV carriers was found compared with non-BRCA carriers/sporadic cases of BC [2][3][4]. In fact, DCIS was reportedly rare in BRCA1 carriers, or even not considered as part of the cancer spectrum risk conferred by BRCA1 PSV [2,5]. In the Breast Cancer Linkage Consortium (BCLC), carriers of BRCA1 mutations showed less DCIS around the invasive lesion compared with controls (sporadic BC cases) (41 vs 56%) [3]. More recent studies reported higher prevalence of DCIS incidentally diagnosed during risk reducing mastectomy in BRCA1 and BRCA2 PSV carriers-Kauff and coworkers compared BRCA carriers with age-and race-matched controls without a known cancer predisposition. ey found 3 DCIS cases in 24 reducing mastectomy (13%) in BRCA carriers vs. none of the 48 controls-p � 0.034 [6]. Moreover, Bayraktar et al. [7] showed that 10% (12/118) and 17% (20/118) of their studied DCIS cases harbored BRCA1 and BRCA2 PSV, respectively. Hwang reported a cohort of 129 PSV carriers and 269 noncarriers. 48 (37%) PSV carriers had DCIS (with or without invasive cancer) compared with 92 noncarriers (34%). DCIS was as equally as prevalent in patients that are PSV carriers, to those that are noncarriers, but occurred at an earlier age in PSV carriers. [8]. e risk of developing invasive breast cancer after a diagnosis of DCIS ranges from 14% to 60% at 10 years [9]. When DCIS is diagnosed at an early age (premenopausal women), the prognosis is worse [9,10].
MRI is the most sensitive breast imaging tool, especially in young BRCA1/BRCA2 PSV carriers, with a sensitivity between 75.2 and 100%, generally over 80%, and specificity between 83% and 98.4% [15]. MRI imaging of DCIS may reveal a host of imaging abnormalities: foci, non-mass enhancement, and masses.
MRI has reportedly outperformed either mammography or ultrasound in assessing the extent of disease in DCIS cases: MRI sensitivity for accurate assessment of DCIS extent was 89% compared with only 55% and 47% for mammography and ultrasound, respectively [16].
To the best of our knowledge, a comparison of the incidence of noncalcified DCIS between BRCA1 and BRCA2 mutation carriers, detected by MRI, has never been reported.
e aim of the current study was to assess the rate of detection over a 5-year period and the imaging features of MRI-detected pure noncalcified DCIS among Israeli BRCA1/ BRCA2 PSV carriers who attend a high-risk clinic in a single medical center between 2015 and 2020 and to compare the incidence of noncalcified DCIS by mutated gene.

Materials and Methods
All female BRCA1 and/or BRCA2 PSV carriers followed at the Meirav high-risk clinic, Sheba Medical Center, Tel HaShomer, Israel from 2015 to 2020 were eligible if they underwent regular breast imaging protocol offered (see below) and underwent an MRI-guided biopsy for a radiologically suspicious lesion. For each study participant, clinical data such as ethnicity, specific BRCA gene PSV, type and size of the imaging findings, grade of DCIS, and receptor status were retrieved from the computerized archiving system: Picture Archiving and Communication System (PACS), Radiology Information System (RIS), and the central database of the medical center (Camelon). For each MRI, a BI-RADS score for MRI result that required a biopsy (4 or 5) was assigned, and pathology reports were individually retrieved and reviewed.
Imaging protocol surveillance of BRCA PSV carriers in our institution during the study period included annual MRI starting at 25 years of age and adding annual mammogram starting at age 30 years, alternating with MRI so that breast imaging is performed every six months. Between the ages of 30 and 40 years, a one-view only mammography for each breast (MLO view) is performed to minimize ionizing radiation exposure. Between ages of 40 and 50 years, performing one-or two-view mammography depends on the radiologist's preference. Between 25 and 29 years of age, an ultrasound alternating with MRI is also performed. Any suspicious imaging finding on MRI, such as a non-mass enhancement (NME) or a focus, has led to a second-look mammography to look for microcalcifications. If a mass or a large NME was detected, a second-look ultrasound was also performed. MRI-guided vacuum biopsy was performed when there was no mammography or ultrasound correlate with the MRI finding.
Breast MRI scans were performed on a 1.5-T system (Signa Excite HDx; GE Healthcare) with a dedicated double breast coil equipped with 8 channels. A standard dynamic contrastenhanced MRI protocol was via axial vibrant multiphase 3D DCE T1-weighted sequence with the following parameters: echo time (TE)/ repetition time (TR) of 2.6/5.4 ms; flip angle 15°; bandwidth 83.3 kHz; matrix, 512 × 364; field of view (FOV) 340 mm; slice thickness of 2 mm. Images were acquired prior to and four times after an automated injection of contrast agent bolus (0.1 ml/kg at 2 ml/sec Dotarem (gadoterate meglumine, Guerbet)) followed by a 20-ml saline flush. Postcontrast images were acquired, with the first acquisition centered at 1 : 25 minutes after injection and the delayed acquisition centered at 7 : 35 minutes after injection, as was previously described [17].
All MRI-guided biopsies were performed using a Sentinelle Vanguard ™ Breast MRI coil in GE Signa 1.5 T scanner (described above), with a vacuum-assisted biopsy needle 2 e Breast Journal (Hologic Suros ATEC ® Sapphire breast biopsy system for stereotactic biopsy). All MRI-guided biopsies performed during the study period in our institution in the context of the high-risk clinic were retrieved and reviewed by two independent physicians. ose with a pathology result of pure DCIS (meaning only ductal carcinoma in situ, without the evidence of invasive component) were included in the study. Patients with calcifications in a mammography before the MRI biopsy were excluded from the study.
All MRI and mammography interpretation and MRIguided biopsies were carried out by a fellowship-trained radiologist.

Statistical Analysis.
Comparison of the rate of DCIS between groups, comparison of negative estrogen receptor, and number of intermediate to high-grade and high-grade DCIS, as well as the median size of the lesions, were performed using Fisher's exact test. All statistical analyses were calculated using SPSS V27. P-values were considered significant at p ≤ 0.05.
is study was ethically approved by the Sheba Institutional Review Board with an exemption from obtaining written informed consent from the study participants.

Results
Overall, during the study period (2015 to 2020), 1100 BRCA PSV carriers were eligible for participation: 629 BRCA1 PSV carriers and 447 BRCA2 PSV carriers, and in the remaining cases (n = 24) the mutated gene could not be clearly ascertained, or they carried a PSV in both genes. During the study period, 227 MRI-guided biopsies were performed in 209 BRCA PSV carriers: 121 BRCA1 carriers, 81 BRCA2 carriers, 2 were double BRCA1 and BRCA2 PSV carriers, and in 5 the mutated gene could not be ascertained.
During the study period, 29/209 biopsied different cases were diagnosed as DCIS, all not calcified on mammography. In two of these cases, patients harbored both BRCA1 and BRCA2 PSVs and were excluded from further analysis, and an additional case (with a nonassignable mutated gene) was lost to follow-up. us, subsequent analysis was focused on the remaining 26 DCIS cases. Of 121 BRCA1 PSV carriers who underwent biopsy, 18 were diagnosed with DCIS (15.2%), and 8/81 BRCA2 PSV carriers were also diagnosed with DCIS (10.1%) (p > 0.05) (Figure 1).
Relevant clinical and tumor specific features (e.g., median age at diagnosis, estrogen receptor status, and grade of DCIS) by mutated gene are shown in Table 1.

Discussion
To the best of our knowledge, the current study is the first to report the rates of MRI-only diagnosed occult DCIS in BRCA1/BRCA2 PSV carriers-2.6% (29/1100) over a 5-year surveillance in a single medical center in Israel, and the rate of biopsy-diagnosed "pure" DCIS was 13.9% (29/209).
Previous studies on pure DCIS or DCIS adjacent to IDC diagnosed by any breast imaging technique (primarily mammography) reported that these diagnoses were more prevalent in BRCA2 PSV carriers than in BRCA1 PSV carriers-Arun et al. analyzed 73 BRCA1/2 PSV carriers-70% of BRCA2 carriers had preinvasive lesion adjacent to the IDC, in comparison to 52% of BRCA1 [18].
Rijnsburger et al. reported the MRI sensitivity was superior to that of mammography for invasive cancer (77.4% v 35.5%;P < 0.00005) but not for DCIS [19]. In contrast, for DCIS cancers only, the sensitivity of mammography (69.2%) was much higher than that of MRI sensitivity (38.5%) in their study. e present study, that is based on MRI imaging only, shows that noncalcified DCIS seen on MRI is not rare, implying another important role of MRI imaging on PSV mutation carriers, in contrast to Rijnsburger's study.
Warner et al. [20] reported that all 6 out of 236 BRCA carriers diagnosed with pure DCIS were BRCA2 PSV     Similarly, the study by Liu and co-workers [21] reported that among women diagnosed with DCIS and genotyped for BRCA PSVs, the incidence of DCIS in BRCA2 PSVs was higher than that of BRCA1 PSV mutation (27.5%-8/29 versus 3%-1/ 33) (p � 0.009) [21]. ese data are inconsistent with the data presented herein, where most DCIS cases were diagnosed in BRCA1 PSV carriers. e inconsistent results may stem from the lack of calcifications of the tumors reported herein and targeting MRI-only detected tumors compared with previous studies that have analyzed all DCIS-diagnosed cases, including those that were detected by mammography and those that were IDC adjacent. Indeed, BRCA2-associated DCIS are more likely to present with calcifications that are detected by mammography [14]. In the present study, calcified tumors were excluded from the analysis. us, the current study establishes the fact that the absence of calcifications on mammography does not indicate that BRCA1 PSV carriers do not develop DCIS, only that its radiological features may be different from those in BRCA2 carriers. Yang et al. reported that pure DCIS was diagnosed in 15/ 71 BRCA1 patients (21.1%) undergoing surgery for BC and 10/43 BRCA2 patients (23.3%), in addition to DCIS associated with IDC diagnosed in both carrier groups (45/ 71-63.4% and 26/43-60.5%, respectively) [22]. us, most BRCA-associated invasive BC had adjacent DCIS, arguing for a DCIS-associated premalignant pathway. Originally, DCIS-associated preinvasive pathway was presumed to be absent in BRCA-associated BC, in contrast to sporadic BC [3]. However, later studies showed pure DCIS and IDC associated with DCIS in BRCA-associated BC, supporting the notion that DCIS is a precursor to invasive BC in BRCA PSV carriers, like sporadic cases [22]. Furthermore, it has also been initially reported that DCIS is less often found with IDC in BRCA PSV carriers compared with sporadic cases [2][3][4]. Yet, subsequent studies showed that DCIS can also be found in risk-reducing mastectomy in BRCA PSV carriers [6,[23][24][25]. Currently, and consistent with the study reported herein, it is well established that BRCA PSV carriers may be diagnosed with DCIS-only cancers.
Our study shows that BRCA1 PSV carriers also develop DCIS that can be detected by MRI and not only by mammography. e present MRI-only based study, that demonstrated that noncalcified DCIS is not rare in BRCA PSV carriers, further supports the importance of periodic MRI imaging as an effective screening modality in BRCA PSV carriers. ese MRI effectiveness results in detecting DCIS in BRCA PSV carriers are in line with a previous study by Warner [26], who reported that 10/445 BRCA PSV carriers were diagnosed with DCIS by MRI (4/10-BRCA1 PSV carriers and 6/10-BRCA2 PSV carriers).
Importantly, phenotypic and clinical differences were noted by the mutated gene: DCIS was diagnosed at a younger age (not statistically significant), with a higher rate of intermediate-to-high-grade DCIS and ER negativity in BRCA1 PSV carriers compared with BRCA2 PSV carriers. ese features and differences in phenotypes between female carriers of the two mutated genes have also been reported for invasive BC [19], further supporting the notion that DCIS can be regarded as a precursor for invasive BC.
In the current study, MRI-detected DCIS was diagnosed at a younger age and higher grade in BRCA1 carriers than in BRCA2 carriers. Krammer  showed that breast cancers in BRCA1 mutation carriers are associated with more aggressive tumor characteristics than that of BRCA2, similar to our study, and are less well seen on mammography [14]. In patients that undergo annual MRI, the objective of mammography is to detect calcifications that could represent DCIS that was not seen on MRI. Since mammography rarely identified cancers not visible on MRI in their study, and since our study shows MRI-detected DCIS seen more commonly in BRCA1 than BRCA2 patients, the omission of mammography in BRCA1 mutation carriers screened with MRI that was suggested in their study might be extrapolated to our study, although this comparison is limited since we did not include DCIS diagnosed with mammography. BRCA1 carriers are sensitive for radiation; therefore, we suggest that mammography could be omitted, or it could be performed once every two years, with a one view only. e current study has obvious limitations and caveats. It is a retrospective, single-center study, with a small number of patients diagnosed with DCIS and a short follow-up. Yet, the fact that the spectrum of PSV is limited in Israeli (mostly Jewish) BRCA PSVs and the fact that the surveillance scheme was carried out by the same team using the same equipment help in reducing any biases but does not eliminate them in total.
In conclusion, noncalcified MRI-only diagnosed DCIS is an important entity that is diagnosed by adhering to an annual MRI screening of female BRCA PSV carriers, more frequently in BRCA1 than in BRCA2, and in these BRCA1 PSV carriers, the tumor is more likely to be estrogen receptor negative and displays high-grade phenotype. If these data are validated in an extended study, these realities that may clinically translate to a more aggressive treatment approaches to these tumors should be discussed with female BRCA PSV carriers during their follow-up sessions.

Data Availability
Access to data is restricted due to patient privacy, as per the regulations of the ethics committee that approved this study.

Conflicts of Interest
e authors declare that there are no conflicts of interest.