High rate of occult cancer found in prophylactic mastectomy specimens despite thorough presurgical assessment with MRI and ultrasound: findings from the Hereditary Breast and Ovarian Cancer Registration 2016 in Japan

Purpose Prophylactic surgery is a preemptive strategy for hereditary breast and ovarian cancer (HBOC). Prophylactic mastectomy (PM) reduces breast cancer risk by > 90%. The aim of our study is to analyze the information of the Japanese pedigrees and to utilize the results for clinical practice. Methods We statistically analyzed records of HBOC registrees who had undergone BRCA1/2 genetic testing at seven medical institutions up until 2016. In the cases of PM, we examined breasts with the use of mammography (MMG), ultrasound (US), and magnetic resonance imaging (MRI) before surgery. After PM, the specimens were divided about 1 cm serially and examined in their entirety. Results Of 1527 registrees who underwent BRCA testing, 1125 (73.7%) were negative for BRCA1/2 mutation, 297 (19.5%) were positive for BRCA1/2 mutation (BRCA1/2MUT+), and 105 (6.9%) had uncertain results. To decide whether to undergo total mastectomy vs. breast-conserving surgery (BCS), 370 registrees underwent presurgical genetic testing. During the follow-up period, four new-onset breast cancers were found among the 55 non-affected BRCA carriers. Among the 73 BRCA1/2MUT+ carriers who underwent BCS, 3 were found to have ipsilateral breast cancer. Of 189 BRCA1/2MUT+ carriers with unilateral breast cancer, 8 were found to have contralateral breast cancer. Of 53 PM specimens, 6 (11.3%) were found to have occult breast cancer despite using MMG, US, and MRI. Conclusions Our report showed a relatively higher incidence rate of occult cancer at 11.3% in PM specimens despite thorough pre-operative radiological evaluations, which included a breast MRI. Considering the occult cancer rates and the various pathological methods of our study and published studies, we propose the necessity of a histopathological protocol.


Introduction
The breast cancer rate tends to peak at a younger range in Japan than in Western countries [1][2][3]. About half of breast cancer diagnoses in Japan are for patients in their 30s-50s [2]. As this pattern probably reflects their genetic background, investigations of hereditary breast and ovarian cancer (HBOC) are important for Japanese women. The Japanese nationwide HBOC registration system aims to clarify clinical and genetic features of Japanese HBOC and to improve its medical treatment.
The Japanese HBOC Consortium (JHC) was established in December 2012. We established a registration committee for JHC in October 2013 and promoted it as a nationwide registration project. The registered subjects were all Japanese individuals who underwent BRCA1/2 genetic testing (including individuals in which no mutation was detected) [4]. Here, we report results of the HBOC Registration from its establishment until 2016. The objective of the current study is to analyze the information of the Japanese 1 3 pedigrees, who underwent BRCA1/2 genetic testing, and to make use of the results in clinical practice.
All subjects, who received genetic counseling and underwent genetic testing of their own free will in clinical practice, were those who had been provided explanations of the HBOC risk in accordance with Genetic/Familial High-Risk Assessment: Breast and Ovarian in NCCN Guidelines [16]. Most of genetic testing with sequencing and large rearrangement analysis was performed at Myriad Genetic Laboratories or FALCO Biosystems. Detected variants were interpreted by the criteria of Myriad Genetic Laboratories. We entered information for BRCA1/2 genetic testing and clinicopathological findings of breast cancer, ovarian cancer, and other cancers in the original electronic template. All data except sex were anonymously registered in each institution. Dates of birth only included year and month [4].
In the cases of PM, we examined breasts with the use of mammography (MMG), ultrasound (US), and magnetic resonance imaging (MRI) before surgery. After PM, the specimens were processed by a pathologist. Although a surgicopathological protocol for occult cancer in the PM specimens does not exist, the specimens were divided about 1 cm serially and examined in their entirety.
Among the 51 patients who underwent PM (Table 6), 6 had specimens in which occult breast cancer was found, including 1 with a BRCA1 mutation and 5 with BRCA2 mutations. All six patients had undergone extensive imaging prior to PM, using MMG, US, and breast MRI (Tables 7, 8). In our study, the rate of occult cancer among total removed breasts by PM was 6/53 = 11.3%.

Discussions
We report herein one of the highest incidence rates in the literature: 11.3% of occult cancer in PM specimens, despite thorough presurgical assessment with MRI, US, and MMG, compared with previously reported rates of 0.5-9.9% (Table 9). We reviewed several factors thought to influence occult cancer occurrence, including (a) rates of bilateral prophylactic mastectomy (BPM), (b) pre-PM examination methods, (c) BRCA1/2 MUT+ rates among subjects, and (d) pathological methods.

Rates of BPM
Rates of BPM among subjects in the first three studies of Table 9 are higher (at or near 100%) than in the other studies. The retrospective study of Hartmann et al. [5] included all women with family histories of breast cancer who underwent BPM in USA between 1960 and 1993. They found only 0.5% of occult cancer after BPM, though the rate of BRCA mutations among their subjects was not available. Meijers-Heijboer et al. [6] conducted a prospective study of 139 women with pathogenic BRCA1 or BRCA2 mutations who were enrolled in a breast-cancer surveillance program, Netherlands [12] in UK between 1985 and 2010, considered whether CPM improves overall survival, and found 5.7% occult cancer in 105 women with BRCA1/2 mutations and unilateral breast cancer who underwent CPM. By comparing the 100% (or near-100%) BPM cohorts with 100% CPM cohorts, we see that high rates of BPM might be associated with lower rates of occult cancer. Therefore, even among BRCA carriers, detection rate of occult cancer may have been influenced by the status whether affected or non-affected.

Pre-PM examination methods
Black et al. [15] of USA reviewed occult malignancy in 192 PMs in 173 patients treated from 1999 to 2005, to compare pre-operative MRI with sentinel lymph node biopsy (SLNB), and found that MRI (performed in 59 patients) missed three of four total occult cancers. In the study of McLaughlin et al. [11] (USA) of 529 patients who underwent 613 PMs between 1999 and 2006, both pre-operative MRI and SLNB were performed selectively at the discretion of the surgeon; they reported the sensitivity of MRI for detecting occult cancers to be 78%. In a 2015 study, Riedl et al. [17] insisted that the use of MRI to screen women at increased risk for breast cancer improved   [19] reported with respect to BRCA1/2 MUT carriers, that the sensitivity of US was 77%, and that of MRI was 100%. Ohuchi et al. [20] from Japan reported that the sensitivity of MMG + US for asymptomatic women aged 40-49 years with no history of any cancer in the previous 5 years was 91.1%.
With regard to MRI + MMG + US (yearly MRI, MMG, and biannual US), van Zelst et al. [21] reported the sensitivity to be 76.3% for surveillance of BRCA1/2 Mut+ women and their first-degree untested relatives. Riedl et al. [17] reported the sensitivity to be 95.0%, among BRCA1/2 Mut+ carriers and women with a familial risk > 20% (US was offered to BRCA mutation carriers). Kuhl et al. [22] reported the sensitivity to be 100%, in a high-risk population (370 of 687 patients underwent US). Our study

BRCA1/2 MUT+ rates
BRCA mutation rates and occult cancer rates do not seem to be related in the studies cited in Table 9. For example, in the study of Burger et al. [8] on women who underwent PM (n = 83 in 71 patients) and SLNB (n = 1522 in 1498 patients) between 2005 and 2010 in UK, the rate of BRCA mutation in the 71 patients was 8.5% and the occult cancer rate was 4.8%, which is similar to 5% reported by van Sprundel et al.
[10] among a 100% BRCA1/2 MUT+ population. Kauff et al. [14] compared prevalence of histopathologic lesions in PM (performed between 1987 and 2001 in USA) specimens from women with BRCA mutations and in age and race-matched cadaver mastectomy specimens and found that high-risk epithelial proliferative lesions (including DCIS) are more common in the unaffected breasts of women with known BRCA mutations than in women of the comparison group. However, they said that determining whether these lesions are more common in women with BRCA mutations than in those without will require direct comparison to women without mutations or with low risk for carrying mutations.

Pathological methods
Pathological examination methods vary among the papers cited in Table 9. Some studies evaluated 2-4 sections per quadrant of the breast and a section of the nipple, and another evaluated them by 5 mm slices and radiological examination; the methods of the others are not known. Boughey et al. [9] examined specimens (from PMs, between 2000 and 2005, USA) of at least 2 sections per each quadrant and nipple; the specimens were also macroscopically sliced and any areas found abnormal by palpation were evaluated further at the pathologists' discretion. They noted a 5% occult cancer rate, including 2 IDCs. In the study of Hoogerbrugge et al. [13], the specimens (from PM between 1989 and 2001, Netherlands) were cooled and sliced in serial sections with approximately 5 mm intervals. Radiographs were made from the tissue slices. Suspicious lesions and randomly selected areas from each quadrant and the nipple were sampled, with a mean number of 19 samples per specimen. With this method, they detected   [23]. Similarly, standardized guidelines for examining PM specimens may be required, as we might easily have missed the aforementioned-5-mm occult cancer. In addition to this, occult cancer should be defined in the protocol, for example, as to whether LCIS can be included in the occult cancer. This investigation is limited by the fact that it is a registration study from seven institutions and may not fairly reflect the entire population of Japanese BRCA mutation carriers.

Conclusions
Our report showed a relatively higher incidence rate of occult cancer at 11.3% in PM specimens despite thorough pre-operative radiological evaluations, which included a