Clinical implications of germline mutations in breast cancer: TP53

The frequency of germline pathogenic variants in TP53 in the general population is unknown and has been estimated using penetrance figures. The estimates vary between 1 in 5000 [7] and 1 in 20,000 [8]. Gonzalez and colleagues estimated the frequency of TP53 germline mutations in the general population using the frequency of specific cancers (breast cancer age ≤ 30 years and adrenocortical carcinoma) in the general population and the frequency of that cancer being due to TP53 germline mutations. This gave an estimated frequency of 1 in 17,000 to 1 in 23,000 people. A recent study of germline variation in cancer-susceptibility genes in a healthy cohort found 15 TP53 missense variants and did not find any nonsense or frameshift variant in 681 individuals [9]; one missense variant was likely pathogenic and the others were variants of unknown clinical significance.

The prevalence of TP53 mutations among women with early onset breast cancer has been studied in various populations [10, 11]. McCuaig et al. estimated that 5–8% of women with breast cancer diagnosed under age 30 years and no pathogenic variant in BRCA1 or BRCA2 will have a pathogenic variant in TP53, and a smaller proportion of women with breast cancer diagnosed aged 30–39 years [12]. The likelihood of having a TP53 mutation is increased if there is a family history of LFS-related cancers, or a personal history of an additional LFS-related cancer.

In a series of patients who had a germline TP53 mutation ascertained due to having a young onset cancer, it was estimated that 7–20% of the mutations were de novo [13]. This is in contrast to hereditary breast and ovarian cancer syndrome due to BRCA1 or BRCA2 gene mutations, where de novo mutations are exceedingly rare. This observation supports testing very young onset breast cancer patients for TP53, even in the absence of family history.

The prevalence of TP53 mutations among women who have had panel testing is low at under 1% in four recent studies. Buys et al. found 61 women with TP53 mutations among 35,409 women with breast cancer who had testing using a panel of 25 cancer genes (0.17%) [14]. Moran et al. detected one TP53 mutation among 190 breast cancer patients with a strong family history and previous negative BRCA1/BRCA2 testing using a protein truncation test (0.53%) [15]. Kapoor et al. found one TP53 mutation among 377 women who were offered gene testing by breast surgeons using multigene panels (5–43 genes, average 14.7) (0.27%) [16]. Susswein et al. reported results for over 10,000 consecutive cases referred for evaluation of germline cancer genes. They reported nine pathogenic and one likely pathogenic TP53 mutation among 3315 women with breast cancer who had not had previous BRCA1/BRCA2 testing (0.30%), and three pathogenic and one likely pathogenic TP53 mutation among 1894 women with breast cancer who had previous BRCA1/BRCA2 testing (0.21%) [17].

There have been several studies looking at genotype–phenotype correlations in families with LFS and LFS-like syndrome. Birch and colleagues’ study of 34 families (20 LFS and 14 LFL) showed that individuals with missense mutations in the DNA-binding region had higher overall rates of cancer with significantly higher rates of breast cancer and central nervous system tumours compared to individuals with missense mutations in other parts of the gene or protein-truncating mutations [29]. A study by Olivier et al. analysed the IARC database (including 1068 individuals from 265 families) and found that missense mutations outside the DNA-binding region are more commonly associated with adrenocortical carcinoma compared to missense mutations in the DNA-binding domain. They also noted that individuals with missense mutations in the DNA-binding domain were more likely to have early onset breast cancer compared to those with missense mutations outside the DNA-binding domain (32 years vs. 42 years) and that mutations leading to a TP53 null phenotype are associated with earlier onset brain tumours [23].

The option for risk-reducing bilateral mastectomy or breast screening should be considered in women without cancer with a mutation in the TP53 gene. In the UK, annual MRI breast screening is recommended from age 20 to 49 years and should be considered between 50 and 69 years. Mammography is not recommended [30]. In the USA, National Comprehensive Cancer Network guidelines recommend annual breast MRI 20–29 years and annual MRI and mammography from 30 to 75 years [31]. In Australia, national guidelines recommend that bilateral mastectomy should be offered, otherwise annual breast MRI is recommended from 20 to 50 years [32]. Based on the finding that breast cancer risk increases significantly after the second decade [22], bilateral mastectomy should be considered from age 20 (in line with NCCN guidelines [31]).The annual breast cancer risk peaks at around age 40–45 and then decreases [22]. Bilateral mastectomy is less likely to benefit women over 60 years.

Mastectomy rather than lumpectomy is recommended to reduce the risks of a second primary breast cancer and to avoid radiotherapy where possible. Bilateral mastectomy should also be considered due to the risk of a contralateral breast cancer. Contralateral breast cancer risk will depend on the patient’s age, but there are no clear figures from the literature.

There are concerns about increased risk of radiation-induced second primary tumours. Heymann et al. [33] describe a series of 8 patients with germline TP53 mutations who were treated for breast cancer between 1997 and 2007 from 47 documented Li–Fraumeni families; three underwent conservative breast surgery with post-operative radiotherapy, three had mastectomy and radiotherapy and two had mastectomy with no radiotherapy. Among the six who received radiotherapy, there were three ipsilateral breast recurrences, three contralateral breast cancers, two radio-induced cancers (one breast histiocytoma fibrosarcoma and one chest wall angiosarcoma) and three new primaries (including one papillary thyroid carcinoma which developed inside the radiation field after 2 years). One contralateral breast cancer occurred in the two patients who did not have radiotherapy, with a median follow-up of 6 years. Despite the small sample size, this study does suggest that radiotherapy should be avoided or used with extreme caution in TP53 mutation carriers after very careful consideration of the risks and benefits.

A study of the breast cancer histopathology in TP53 carriers showed that most invasive ductal carcinomas and ductal carcinomas in situ (DCIS) in LFS are hormone receptor positive and/or HER-2 positive [34]. In this study, there were 32 invasive ductal carcinomas in 30 women with confirmed germline TP53 mutations. 84% of the tumours stained positive for oestrogen receptors, 72% for progesterone receptors and 63% showed HER-2 amplification and/or overexpression. Melhem-Bertrandt et al. [35] have also reported that women who are TP53 mutation carriers are more likely to have HER2 amplification and/or overexpression (present in 67% of cases and 25% of controls in this study).

These histopathological findings have management implications as treatment is more likely to include Trastuzumab (Herceptin) and hormone therapy. There are no published data on treatment response to chemotherapeutic agents so standard chemotherapy regiments should be used.

There is no international consensus about the best surveillance for TP53 mutation carriers. In the UK, breast MRI screening is the only recommendation. In the USA, annual complete physical examination including neurologic and skin examination is recommended, and 2–5 yearly colonoscopy from age 25 [31]. In the Netherlands, annual breast surveillance from age 20 to 25 and an optional annual physical examination is recommended [36]. The Australian guidelines recommend annual clinical review, and to consider 2–5 yearly colonoscopy from 25 years if there is a family history of colorectal cancer, or 2–5 yearly endoscopy if there is a family history of gastric cancer [32].

An intensive and comprehensive surveillance programme has been proposed by Villani et al. for children and adults who carry a TP53 germline mutation [37]. This uses multiple modalities (physical examination, blood tests, ultrasound, mammography, MRI and colonoscopy). The results of a non-randomised trial showed good long-term compliance with the protocol. It reported a significantly improved five-year overall survival in the surveillance group. However, there was a very high rate of symptomatic tumours in the group who initially declined surveillance. A larger-scale randomised controlled trial is required to evaluate the protocol further.

Neonatal testing for the c.1010G > A, p.(Arg337His) mutation and subsequent surveillance for adrenocortical tumours in mutation carriers has been evaluated in a large, non-randomised clinical trial in Southern Brazil [38]. Adrenocortical tumours diagnosed in the surveillance group were smaller and had an improved clinical outcome. Although the results are not directly transferrable to other populations of TP53 carriers with different mutations, it does suggest that surveillance for adrenocortical tumours in children should be considered.

The UK SIGNIFY study [39] investigated the role of one-off whole body non-contrast MRI screening in asymptomatic TP53 mutation carriers. Four malignancies were diagnosed among 44 TP53 mutation carriers and none in matched controls. Two malignancies were not picked up on scan—one patient developed leukaemia and one became symptomatic from a mediastinal liposarcoma which was thought to be a pericardial cyst on scan. The difference between cancer detection in carriers and controls was not statistically significant, but the trend of the results suggests that whole body MRI is a useful investigation.

There are several ongoing trials using whole body MRI screening (reviewed in McBride 2014 [40] and Ballinger 2015 [41]). The LIFSCREEN trial [42] at the Institut Gustave Roussy (Villejuif, France) is a randomised trial with annual whole body MRI screening for three years in the intervention arm and standard care in the control arm. It is recruiting individuals with TP53 mutations (age 5–71 years). The Surveillance of Multi-Organ Cancer Prone Syndrome (SMOC) trial [43] in Australia is enrolling individuals with a germline mutation in a cancer risk gene (including TP53) and those at 50% risk aged 18–70 years for a period of 3 years. It is a non-randomised trial using annual physical examination, full blood count, whole body MRI, breast MRI, ultrasound and mammography (in females) with additional procedures of gastroscopy and colonoscopy based on family history. The Dana Faber Cancer Institute whole body MRI study involves annual physical examination and whole body MRI and is currently recruiting children only.

Based on the observation that metformin has been associated with reduced cancer risk in several epidemiological studies [44], a pilot study for chemoprevention using metformin has been commenced which assesses the safety and tolerability of the drug over 14 weeks, measuring IGF-1, insulin and IGFBP3 levels in blood at baseline, and weeks 0 and 8 [45].

Some women may require additional psychological support. There are several patient support groups including the Li–Fraumeni Syndrome Association <http://​www.​lfsassociation.​org> based in the USA, and the George Pantziarka TP53 Trust <http://​www.​tp53.​co.​uk> based in the UK.