Clinicopathological and imaging features of ductal carcinoma in situ in BRCA1/2 mutation carriers

In this study, we investigated the clinicopathological and imaging findings of BRCA-associated DCIS. Understanding the clinicopathological and imaging characteristics of BRCA-associated DCIS is helpful in informing surveillance in women with BRCA mutations.

BRCA-associated DCIS was more commonly detected in women with mutations in BRCA2 than in those with mutations in BRCA1. Previous reports have reported similar results [5,9,17]. In a single-center BRCA-associated DCIS survey conducted by Liu et al. [9], there were nine DCIS cases among BRCA1/2 mutation carriers. Of these, only one BRCA1-associated DCIS was observed, and it was characterized as NG3 and TN. Warner et al. reported that one of eight pure DCIS cases diagnosed in patients with BRCA1/2 mutations was BRCA1-associated [20].BRCA1-associated cancers have aggressive characteristics (TN type and high grade) and easily become invasive, making the occurrence of BRCA1-associated DCIS rare. In our study, 8 of the 30 DCIS cases were BRCA1-associated, and this prevalence was higher than that reported in previous studies. This may be because five of the eight cases were occult cancers. Six (75%) of the eight BRCA1-associated DCISs were ER positive, and only two were TN DCIS. Among DCIS cases, unlike in IBC cases, the TN type was not common. This may also be due to the rarity of high-grade DCIS cases.

The most common imaging findings of the cancers were calcification on MG, non-mass enhancement on MRI, and hypoechoic area on US. These imaging features were similar to those of sporadic DCIS. Some studies have found that BRCA1-associated IBC shows benign imaging features more often than sporadic IBC. In contrast, the imaging features of BRCA2-associated IBCs are typical malignant findings [4,17]. This has also been observed in Japanese women [10]. The imaging features of IBCs with BRCA mutations reflect the pathological findings. That is, many BRCA1 cancers are high-grade, such as TN, whereas BRCA2 cancers are often hormone receptor positive. In this study, BRCA-associated DCISs were often hormone receptor positive and had low NG. These pathological findings influenced the BRCA-associated DCIS imaging findings. We found no differences in DCIS imaging findings between the BRCA1 and BRCA2 cases.

The detection rates of MG, MRI, and US in our study were 64.3%, 72.0%, and 64.0%, respectively. Previous studies have reported that the detection rate of IBC in high-risk women and sporadic DCIS by MRI was higher than that of MG and US [8,13,14,17,21]. Reports of the detection rate of BRCA-associated DCIS vary [4,13,19,20]. Warner et al. [20] reported that the detection rates of MG, MRI, and US were 10%, 100%, and 0%, respectively, whereas Ha et al. [4] reported detection rates of 50.0%, 85.7%, and 78.6%, respectively. MRI is useful for BC surveillance of women with BRCA1/2 mutations because of its high sensitivity. Annual MRI surveillance for high-risk women is recommended by some guidelines, such as the American Cancer Society Guideline for Breast Screening with MRI as an Adjunct to Mammography (2007) (ACS 2007) [15], National Comprehensive Cancer Network Guidelines Version 2. 2022 for Genetic/Familial High-Risk Assessment (http://www.nccn.org), and Guidelines for Diagnosis and Treatment of Hereditary Breast and Ovarian Cancer 2021 by the Japanese Organization of Hereditary Breast and Ovarian Cancer. The addition of MG and US to MRI screening may have limited utility for high-risk women because of the lower sensitivities of MG and US than that of MRI. However, in this study, the detection rates of MG, MRI, and US were comparable for BRCA-associated DCIS. One of the reasons for discrepancies in detection rates between modalities is the difference in patient backgrounds. This is because patient characteristics such as age, breast density on MG, and background parenchymal enhancement on MRI can lead to different rates of lesion identification [11,16,18]. In addition, as a reason for the high detection rate of US, there is data that the US was useful for early detection of BC in Japanese women [12]. Another possible explanation for the observed high sensitivity of MG and US may be related to diagnostic suspicion bias in retrospective studies. It is necessary to further study the optimal examination modalities for each situation.

In our study, there were differences in the detection rates between genotypes. The detection rate of BRCA1-associated DCIS on MRI was significantly lower than that of BRCA2. A similar trend was observed in MG (28.6% and 76.2%, respectively) and US (28.6% and 72.2%, respectively); however, no statistically significant difference was observed. All five occult cancers that underwent the three examinations (MG, MRI, and US) were BRCA1-associated. Yamauchi et al. showed a relatively higher incidence rate of occult cancer (11.3%) in risk-reducing mastectomy specimens from Japanese women [23]. Of the five occult DCISs detected in this study, four were associated with BRCA2. Our results differ from these findings. In our study, occult cancers were <6 mm in size, and the size of DCISs detected by any modality was ≥6 mm. Size is considered to be related to lesion detection. We suggest that it is difficult to detect DCIS in women with BRCA1 mutations because of the small size of these lesions. This may be because BRCA1-associated cancers easily progress to invasive carcinoma owing to their aggressive nature.

Currently, individualization of surveillance methods according to BRCA genotype is being discussed because BRCA1- and BRCA2-associated IBCs have different characteristics [7,19]. In BRCA-associated DCIS, there were not many differences between BRCA genotypes in pathological and imaging findings; however, the detection rates of BRCA1-associated DCIS were lower than those of BRCA2. In BRCA1 mutation carriers, it was difficult to detect cancer in the DCIS state before it became invasive. The results of this study may change the weighting of options for BRCA mutation carriers when choosing between imaging surveillance and risk-reducing mastectomy, depending on the mutation type: BRCA1-positive patients are less likely to have DCIS detected by routine imaging surveillance and, therefore, may be more likely to elect for risk-reducing mastectomy; for BRCA2, routine imaging surveillance may detect BC at an earlier stage and, therefore, may prefer surveillance over risk-reducing mastectomy. The survival benefits of early DCIS detection in BRCA1/2 mutation carriers are unclear, and further investigation is required to determine the prognostic impact of surveillance and risk-reducing mastectomy for each genotype. However, for women who are more likely to have high-grade BC, diagnosing early DCIS may be more beneficial because it avoids additional therapy. The results of this study may provide a more tailored option for BRCA mutation carriers when considering the individual surveillance methods.

This study has some limitations. This retrospective study was conducted at a single institution. Not all patients underwent MG, MRI, or US, which could potentially lead to a selection bias. However, we believe that the effect on the results is minimal because the missing data for each method represented only a small percentage of the lesions. Because the examinations were performed over a long period, the results may be somewhat different due to the use of less technologically advanced imaging in earlier cases. All diagnostic imaging equipment have undergone technological innovations over the past 17 years, from 2003 to 2020. For the equipment used in this study, the MG pixel size was changed from 100 to 70 μm, and digital conversion was also changed from indirect conversion to direct conversion. DBT has been introduced since then. DBT has been performed at our hospital since 2012, and there is a possibility of detection bias depending on whether DBT was performed. However, there was no difference in the detection of malignant findings among the 11 cases in which DBT was performed in the current study. This is due to the fact that most of the MG findings were calcified lesions. This may be because DBT is superior in detecting masses and architectural distortion but is comparable with MG in detecting calcifications [3]. MRI has progressed from 1.5T to 3.0T and from 4 to 8 and 32 channels. For US systems, the frequency of the probe has been improved from 10 to 14 MHz, and the spatial resolution has been improved as a result of the detection capability of superficial US. It is possible that changes in these imaging techniques may be related to lesion detection. However, the quality control of the images from the early years was also well maintained. In addition, only 3 of the 30 cases were older than 2009, suggesting that the image quality from the early years was similar or not significantly different for the purposes of our study.

5. Conclusion

Our study showed the clinicopathological and imaging features of BRCA-associated DCIS at our institution. BRCA-associated DCIS was more strongly associated with BRCA2-positive, NG1, and hormone receptor-positive cases. The imaging features were similar to those observed in sporadic DCIS. The detection rates of DCIS in MG, MRI, and US were comparable. Our results are helpful in informing surveillance in women with BRCA mutations.