Introduction

Breast cancer is the most common cancer among women in the Netherlands, accounting for 28% of all newly diagnosed cancers in women in 2016 [1]. Although most cases of breast cancer are sporadic, an estimated 5–10% are attributable to inherited germline mutations in single genes, particularly the BRCA1 or BRCA2 genes [2, 3]. Since early onset of breast cancer is an indicator of genetic susceptibility [4], diagnostic BRCA1/BRCA2 DNA testing is offered to all women diagnosed with breast cancer before the age of 40 (in the Netherlands) or 46 (in the U.S.A.) [5, 6]. In addition, routine parallel genetic testing for CHEK2*1100delC mutations has been performed in the Netherlands since September 2014 [7]. However, there are some other cancer predisposition genes that are associated with early-onset breast cancer, including TP53, which has been linked to Li Fraumeni syndrome (LFS) [6].

LFS is a rare autosomal dominant inherited cancer syndrome [8]. This syndrome is characterized by a high risk for a wide variety of early-onset neoplasms [9]. The National Cancer Institute showed a cumulative cancer incidence of 50% by age 31 years among female TP53 germline mutation carriers and an incidence of nearly 100% by age 70 years [10]. Sarcoma, breast cancer, brain tumours and adrenocortical carcinoma are the most frequently observed tumours in patients with germline TP53 mutations [9, 11].

Over the years, several different sets of criteria have been established to identify individuals and families at risk for a germline TP53 mutation [11,12,13,14,15,16]. The strictest criteria are the clinical criteria for classical LFS and Li–Fraumeni-like syndrome (LFL) [12,13,14]. However, TP53 mutations were also identified in families who had not fulfilled these clinical criteria due to another tumour spectrum, age at diagnosis or sporadic occurrence of cancer [9, 11]. It is estimated that at least 14% of germline TP53 mutation carriers have a de novo mutation [17]. Therefore, lack of a positive family history does not exclude the possibility of a TP53 mutation. To cover the different clinical presentations of LFS, and to facilitate the identification of germline TP53 mutations, Chompret and colleagues have sequentially updated the testing criteria [11, 16]. Recent multigene panel studies suggest that LFS may have a broader phenotypic spectrum than previously reported because TP53 germline mutations have been identified in individuals who do not fulfill established clinical criteria recommended for LFS testing [18,19,20]. TP53 mutation carriers detected by multigene panel testing (MGPT) are notably older at breast cancer diagnosis [18].

Breast cancer is the most frequently observed cancer in women with TP53 germline mutations [21]. Several studies have described TP53 germline mutations in women with early-onset breast cancer (age of diagnosis ≤ 30 years). Reported mutation detection rates varied between 0% and 8.5%, but these numbers are largely based on small studies [9, 11, 22,23,24,25,26]. In 2015, the Chompret TP53 testing criteria were revised to include early-onset breast cancer [11]. In addition to an early age of onset, amplification of HER2 has been reported as more common in TP53-mutation-associated breast cancer [27,28,29].

According to the third (2005) and fourth (2010) edition of the Dutch national consensus-based guideline for genetics professionals (STOET/VKGN), TP53 germline mutation testing could be considered in women with breast cancer diagnosed before the age of 30 [30, 31]. The latest guideline (2017) recommends TP53 genetic testing in all breast cancer patients diagnosed before the age of 31, following the 2015 version of the Chompret criteria [32]. We conducted a nationwide retrospective laboratory records review in order to evaluate the prevalence of TP53 germline mutations among early-onset breast cancer patients in the Netherlands. We also examined genetics professionals’ experiences and attitudes regarding the timing and content of genetic counselling for LFS. The results of this survey have been published [33]. Our overall aim is to gain insight into the genetic counselling and testing of young breast cancer patients for LFS in order to make clinical recommendations regarding the most appropriate counselling strategy for these women.

Methods

Subjects

We retrospectively reviewed the laboratory records of all women diagnosed with invasive breast cancer or carcinoma in situ before the age of 30 who had been tested for TP53 germline mutations in the Netherlands between January 2005 and December 2016. Patients were referred for TP53 genetic testing on the basis of eligibility criteria that were part of the national guidelines of that time [30, 31]. In the Netherlands, TP53 mutation analysis is performed in six DNA-diagnostics laboratories. Early-onset breast cancer patients were identified by checking the TP53 DNA testing application forms of all six laboratories for personal clinical history of cancer and age at diagnosis. These data were provided by genetics professionals at the time of testing. If information about clinical history of cancer was not available on the application forms, medical records from the genetics departments were evaluated.

Additional data extracted from the TP53 DNA testing application forms or medical records from the genetics departments included the requested timing of DNA test results (i.e. treatment-focused DNA testing, which is used when decisions about primary breast cancer treatment could be impacted by genetic test results [34], or regular DNA testing), BRCA status (i.e. whether BRCA 1/BRCA2 testing was performed, including the results) and family history. Information about the family history was coded into three categories: (1) ‘family history suggestive of LFS’: first-degree relative (FDR) and/or second-degree relative (SDR) with a sarcoma, brain tumour or adrenocortical carcinoma (core LFS cancers); (2) ‘family history non-suggestive of LFS’: no FDR and/or SDR with a sarcoma, brain tumour or adrenocortical carcinoma (based on the family pedigree provided by the genetic professional or the notification ‘family history is non-contributory’); (3) ‘no information’: no information available about the family history of cancer. For TP53 mutation carriers, additional information was obtained regarding the histological type of breast cancer and the oestrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2) status. The Medical Research Ethics Committee of the University Medical Center Utrecht concluded that the Medical Research Human Subject Acts (WMO) does not apply.

Mutation analysis

Sanger sequence analysis of all coding exons (2–11) and exon/intron boundaries was performed for TP53 (NM_000546.5) to identify small nucleotide changes and for the identification of larger rearrangement. Multiplex ligation-dependent probe amplification (MLPA) was undertaken using the SALSA MLPA kit P056 (MRC-Holland, Amsterdam, the Netherlands). Sequence variants were classified into five categories based on the degree of likelihood of pathogenicity: (1) ‘benign’, (2) ‘likely benign’, (3) ‘uncertain significance’, (4) ‘likely pathogenic’, (5) ‘pathogenic’ [35]. Classes 3–5 were reported to the genetic professional involved.

Results

In total, 370 women diagnosed with breast cancer before the age of 30 years were tested for TP53 germline mutations in the Netherlands between 2005 and 2016, of whom 32% (119) had a treatment-focused genetic test. Since 2011, there has been a notably steep increase in the number of women tested (Fig. 1). Characteristics of the patients tested are described in Table 1. Mean age at breast cancer diagnosis was 26 years (range 15–29 years). Information about BRCA status was available for 356 patients (96%), of whom 15 were found to carry a BRCA1 pathogenic mutation and four a BRCA2 pathogenic mutation. Information about family history was available for 284 patients (77%), of whom the majority had a family history non-suggestive of LFS (255/284; 90%). Five patients had a history of LFS-spectrum cancers in addition to young onset breast cancer.

Fig. 1
figure 1

Number of young breast cancer  patientsi tested for germline TP53 mutations in the Netherlands 2005–2016

Full size image
Table 1 Clinical characteristics of young breast cancer patients tested for germline TP53 mutations in the Netherlands 2005–2016
Full size table

Mutation prevalence

Of the 370 early-onset breast cancer patients analysed for TP53 germline mutations, test results were available for 364 women. In six women, DNA analysis could not be performed due to insufficient quality of DNA. In eight women (2.2%), a (likely) pathogenic TP53 germline sequence variant was identified.

Of these variants, six were definitely pathogenic (c.503A > G, c.524G > A, c.637C > T, c.723delC, c.916C > T, and 376 − ?_(993 + ?)del; del exon5-exon9) and two were likely pathogenic (c.587G > A and c.749C > T). In addition, one woman was tested for a specific TP53 germline mutation that was already known to exist in her family. The majority of carriers of a (likely) pathogenic TP53 germline sequence variant (6/8) had been referred for treatment-focused genetic testing. The youngest TP53 mutation carrier was diagnosed with breast cancer at 19 years of age. Histology reports and information about family history of LFS-related cancer were available for seven carriers of a (likely) pathogenic TP53 germline mutation. Five breast cancers were identified as HER2 positive and two as triple positive (i.e. ER, PR and HER2 positive). Three carriers of a (likely) pathogenic TP53 germline sequence variant had a family history suggestive of LFS, and three carriers had developed another LFS-related tumour (2 rhabdomyosarcoma, 1 osteosarcoma) prior to their breast cancer diagnosis.

Discussion

This study was designed to evaluate the diagnostic yield of TP53 genetic testing in early-onset breast cancer patients. Our findings, the largest such study to date, show a steep increase in the number of TP53 DNA testing applications since 2011. We assume that this increase is the result of changes in national guidelines and the results of international studies [6, 21, 24, 32]. A family history suggestive of LFS was no longer required for TP53 genetic testing. Consequently, TP53 DNA genetic testing is now offered to women with early-onset breast cancer more often than it was previously. However, another possible explanation could be the increase of referrals for treatment-focused genetic counselling and testing.

Our results also show that the TP53 mutation prevalence among early-onset breast cancer patients was lower than most studies previously reported (Table 2) [9, 11, 22,23,24,25]. We found (likely) pathogenic TP53 germline mutations in 2.2% (8/370) of women diagnosed with breast cancer before the age of 30, unselected for family history. To evaluate the clinical utility of ‘early-onset breast cancer’ as a criterion for identifying TP53 carriers, it is necessary to investigate the likelihood of a germline TP53 mutation in young breast cancer patients who did not fulfil other established criteria for TP53 testing. McCuaig et al. reported the results of TP53 testing in a very small sample; they found a TP53 mutation was present in one of 13 (7.7%) cases of breast cancer diagnosed before age 30 who did not meet established criteria for TP53 testing [24]. More recently, Bougeard et al. reported on a larger series of 333 early-onset breast cancer patients (≤ 30 years) [11]. Patients had been selected for TP53 genetic testing on the basis of clinical criteria that corresponded, in most cases, with the Chompret criteria (version 2009). Of these women, 14% were found to have a TP53 mutation. Among the patients who did not fulfil the 2009 version of the Chompret criteria, the mutation detection rate was 6%. However, the exact number of women who did not fulfil these criteria was not reported. In our cohort, six of eight carriers of a (likely) pathogenic TP53 mutation did meet other established criteria for TP53 testing, in addition to ‘early-onset breast cancer’. The TP53 mutation prevalence was 0.9% (2/233) among early onset breast cancer BRCA1/BRCA2 mutation negative patients without additional features indicative of LFS (family history suggestive of LFS and/or a personal history of multiple LFS-related tumours).

Table 2 Results of studies exploring TP53 germline mutation prevalence among early-onset breast cancer patients
Full size table

Our study highlights the importance of critically reviewing established testing criteria. Women with breast cancer diagnosed before the age of 30 without additional features indicative of LFS can be reassured that they have a very low chance of being a TP53-mutation carrier. This raises the question of whether age of onset actually is an appropriate criterion for TP53 genetic testing.

Several studies have reported the frequency of TP53 mutations in different cohorts of breast cancer patients [16, 20, 26, 36,37,38,39]. For women unselected by family history, with a personal history of breast cancer up to age 36 years of age and who are BRCA1/BRCA2 mutation negative, the proportion of TP53 germline mutation carriers ranged between 2.3% and 4.2% [16, 36,37,38]. Sample size varied from 30 to 128. Two studies explored TP53 genetic testing in patients with HER2-amplified breast cancer, unselected for family history suggestive of LFS [26, 39]. Eccles and colleagues reported a TP53 mutation prevalence of 3% (9/304) among women diagnosed with HER2+ breast cancer (ER/PR status was not uniform) under the age of 41, and 8.5% (5/71) for women diagnosed before 31 years [26]. In a cohort of women with HER2+ breast cancer diagnosed ≤ 50 years without a known family history, consistent with Chompret criteria, TP53 mutations were identified in 1/195 (0.5%) [39]. In a review of 37 multigene panel studies (including mainly patients who were referred due to personal or family history of breast cancer), the overall TP53 mutation prevalence ranged from 0 to 4.4% [20]. For BRCA1/BRCA2 mutation negative patients of European ancestry undergoing breast cancer multigene panel testing (MGPT), the weighted frequency of TP53 mutation carriers was 0.3%. Although patients were diverse in age of onset and family history, it is noteworthy that approximately half of the TP53 mutation carriers detected by MGPT did not meet NCCN criteria for LFS testing (e.g. classic LFS criteria or Chompret 2015 criteria) [6]. The findings of these studies suggest that it may be beneficial to consider an expanding age of onset or HER2-positive tumour status for LFS testing criteria. However, larger studies are needed to provide more precise estimates of TP53 germline mutations in different cohorts of breast cancer patients.

An evaluation of the benefits and concerns raised by offering TP53 genetic testing is also warranted. The most reported justifications for the utility of TP53 genetic testing in patients with a low a priori probability of carrying a pathogenic TP53 variant, concerns potential clinical implications [20, 24, 40, 41]. First, there are important consequences for primary breast cancer treatment. If a woman is proven to be a TP53 mutation carrier, she may opt for bilateral mastectomy to reduce the risk of a second primary breast cancer. In addition, radiotherapy should be avoided if possible because radiation therapy may increase future cancer risks [42]. Patients without a genetic predisposition can be reassured that their risk of a secondary breast cancer is not substantially increased. There are also important potential implications for family planning and the use of prenatal diagnostics (e.g. pre-implantation genetic testing (PGD)). Lammens and colleagues evaluated the attitudes towards PGD among high risk family members from LFS and Von-Hippel Lindau families [43]. Approximately half of those contemplating a future pregnancy (23/48) would consider the use of PGD. Of note, over a period of 9 years (2009–2017), 12 couples were referred for counselling on PGD for LFS in the Netherlands, of whom 50% (6/12) started with PGD-treatment [44].

In contrast, there are other issues that complicate genetic testing for TP53 in clinical practice. Although TP53 mutation carriers may participate in a comprehensive screening protocol [10, 45,46,47], determining the optimal risk management is complex because of the wide range of cancer predisposition involved [6, 48].This applies especially to TP53 mutation carriers identified in non-classic LFS-families due to lack of information about actual phenotypic spectrum and cancer risks in these families [18]. The benefits of intensive annual screening programs must be weighed against the possibility of false positive findings that require additional diagnostic procedures [45, 49]. In addition, a few studies have reported that a substantial proportion of the individuals in families with TP53 mutations exhibit psychological distress [50,51,52]. Another relevant aspect, which was highlighted by Azzolini and colleagues, concerns patients who willingly decline TP53 DNA-analysis following genetic counselling [53]. Almost one-fourth (23.4%; 11/47) of patients decided not to undergo TP53 testing. Also, in newly diagnosed breast cancer patients who needed to define treatment options, 13% (2/23) consented to BRCA1/BRCA2 genetic testing but declined TP53 analysis.

In one genetic department in the Netherlands, we have kept records about declined testing. Here we find that 40% (16/39) of early-onset breast cancer patients who were offered TP53 genetic testing declined to be tested, and of these a considerable proportion (28%; 7/25) occurred in a setting of treatment-focused genetic counselling. These findings emphasize the necessity of thorough genetic counselling and confirm the importance of the recommendations we have made previously about the ten information items about LFS that should be discussed during pre-test counselling [33].

There are some limitations to be reported. First, this is not a population-based series of cases, but rather a nationwide clinical testing cohort. Only young breast cancer patients who were referred for genetic counselling and testing and who had chosen TP53 mutation testing were included. Between 2005 and 2016, 965 women were diagnosed with breast cancer before the age of 30 in the Netherlands [1]. Our findings show that 370 early-onset breast cancer patients were tested for TP53 germline mutations during this 12-year period. This selection might have increased the likelihood of a suggestive family history or personal history of multiple LFS-related tumours. In that case, the actual TP53 mutation prevalence among early onset breast-cancer patients may be even lower than we report. Furthermore, the degree to which information about family history was mentioned in the DNA-testing application forms differed. These data were provided by genetics professionals at the time of testing and are based on the patient’s own knowledge of their family history. Particularly in the case of treatment-focused testing, family cancer history will not be verified by pathology reports. Therefore, we may have missed information regarding family history. Another issue concerns mosaic TP53 mutations. High-coverage NGS provides new opportunities to identify germline or somatic mosaicism that was undetectable or nearly undetectable with Sanger sequencing [17]. Since we have performed Sanger sequence analysis, we may have missed mosaic TP53 mutations. However, as mosaicism in TP53 is probably very rare [54], additional NGS-results are unlikely to have a large impact on the overall TP53 mutation prevalence among early-onset breast cancer patients.

In conclusion, for early-onset breast cancer patients without additional features indicative of LFS (i.e. family history suggestive of LFS or a personal history of multiple LFS-related tumours) the probability of finding a germline TP53 mutation is exceedingly low. This is useful information for genetic counselling. However, since TP53 mutation prevalence was comparable or even higher in some studies selecting patients with breast cancer at an older age or HER2-positive breast cancers, the question is raised whether very early age of onset is an appropriate LFS testing criterion. Further research is warranted to determine the most appropriate TP53 testing strategy for women diagnosed with breast cancer. In the interim, a careful approach to genetic counselling for LFS is needed as we have noted that a substantial number of breast cancer patients who were offered TP53 DNA-analysis following genetic counselling declined to be tested.