Deciphering dual clinical entities associated with TP53 pathogenic variants: Insights from 53,085 HBOC panel analyses in French laboratories

Abstract

TP53 is included in most cancer predisposition multigene panels, especially those exploring Hereditary Breast and Ovarian Cancer (HBOC) predisposition. The purpose of this study was to define the contribution of TP53 pathogenic variants (PV) to the HBOC phenotype by collecting genotypes and phenotypes of 398 patients harboring a TP53 variant identified by 53,085 HBOC panel sequencing in 15 French laboratories. Heterozygous TP53 variants were identified in 0.44% of HBOC panels, evenly distributed between PV and VUS. Breast cancers associated with TP53 were predominantly triple positive, particularly Her2+ breast cancer, in situ cancer, or phyllodes tumors (p < 0.0001 for both). Interestingly, TP53 PV were identified across all ages in breast cancer patients, with enrichment before 36y. We demonstrated that null variants were linked with the HBOC phenotype, and missense variants, especially with a dominant negative effect, with the LFS phenotype (p = 0.0096). Patients with breast cancer harboring null variants displayed an earlier age of onset compared to missense (p = 0.0030). Surprisingly, we identified, in late‐onset cancer patients, TP53 hotspot PV usually identified in classic LFS, which underlines variable penetrance. Thus, this study suggests the existence of two phenotypic entities associated with TP53 PV: clinical LFS and TP53‐related breast cancer. The type of TP53 variant, as well as modifying factors reflected in family history, may influence these phenotypes, and both should be considered to define the clinical follow‐up of patients and relatives.

What's new?

TP53 is included in most cancer predisposition multigene panels, but analyzing the contribution of TP53 variants to cancer predisposition beyond the Li–Fraumeni syndrome remains challenging. This study, based on 53,085 hereditary breast and ovarian cancer panel analyses, explores the genotype–phenotype correlation of TP53 variants and argues for the existence of two clinical entities: the classic Li–Fraumeni syndrome associated with missense variants, and TP53‐related breast cancer associated with null variants. The results also suggest the phenotypes are influenced by the type of TP53 variant and modifying factors, calling for these two parameters to be considered in clinical management decisions.

Abbreviations

CHIP

clonal hematopoiesis of indeterminate potential

ctDNA

circulating tumor DNA

DNE

dominant‐negative effect

EOBC

early‐onset breast cancer

HBOC

hereditary breast and ovarian cancer

LFS

Li–Fraumeni syndrome

LOF

loss‐of‐function

LP

likely pathogenic

P

pathogenic

PV

pathogenic variant

VAF

variant allele frequency

VUS

variant of unknown significance

1. INTRODUCTION

Germline alterations of TP53 are associated with Li–Fraumeni syndrome (LFS), a severe cancer predisposition with a wide range of tumors and early onset following Chompret's nosological criteria (Table S1). ¹ , ² , ³ LFS was initially associated with rare tumors, such as rhabdomyosarcoma or adrenocortical carcinomas, mainly in children, and more frequent tumors such as early onset breast cancers (EOBC) before age 31 or sarcomas in young adults. ⁴ Cancer incidence by age 70 approaches 100% for both genders. ⁵ , ⁶ Genotype–phenotype analyses, performed in LFS families, showed that missense variants with dominant negative effects (DNE) are linked to severe phenotypes and pediatric tumors, while loss‐of‐function (LOF) variants are associated with later onset. ⁷ This disparity in severity is underpinned by the biological effect of the variant. ⁸ , ⁹

As a key player in oncogenetics, TP53 is frequently included in multigene panels worldwide, particularly for assessing hereditary breast and ovarian cancer (HBOC) predisposition. Analyzing TP53 beyond the classic LFS clinical context presents challenges in variant classification and expands the range of phenotypes associated with TP53 alterations, highlighting variability in TP53 variant expressivity and penetrance. ¹⁰ , ¹¹ , ¹² Consequently, clinical management uncertainties arise for patients not exhibiting classic LFS, especially considering the burden of surveillance recommendations for the index case and related carriers, including children. ³ , ¹³ , ¹⁴ , ¹⁵

Through the French UNICANCER Groupe Génétique et Cancer (GGC) network, data from 53,085 HBOC panels across 15 laboratories were collected. The great number of analyzed panels, the consistency in indications, the uniformity of analyzed genes, and the assessment of personal and family cancer history using the Manchester score ensure the reliability of the results presented hereafter.

Thus, this dataset offers comprehensive insights into the contribution of TP53 in breast cancer patients and provides new perspectives on genotype–phenotype correlations, aiding personalized clinical management for carriers.

2. PATIENTS AND METHODS

2.1. Inclusion criteria

All patients included in this study met HBOC criteria for germline testing (Table S2) and underwent French HBOC consensus panel analysis including 13 genes, that is, BRCA1, BRCA2, PALB2, TP53, CDH1, PTEN, RAD51C, RAD51D, MLH1, MSH2, MSH6, PMS2, and EPCAM. ¹⁶

2.2. Data collection

A comprehensive set of 53,085 germline HBOC panels conducted between 2016 and 2021 was collected from 15 French genetic laboratories (Figure S1). The participating laboratories carried out high‐throughput sequencing using libraries obtained by capture, including at least the coding regions of the 13 genes included in the French HBOC consensus panel. Genotype data included TP53 variants, allelic ratios, classifications by submitting laboratories, and other relevant variants identified. As circulating tumor DNA (ctDNA) and clonal hematopoiesis of indeterminate potential (CHIP) could mimic a de novo TP53 variant, only variants with a variant allele frequency (VAF) above 40% were taken into account. In addition, NGS analyses of genes frequently mutated in CHIP were performed at the same time, and whenever possible, for older patients, tumor analyses were performed to confirm the germline status of the variant. Phenotypic details included birth year, gender, cancer types, onset age, and family cancer history. Additional breast cancer characteristics (e.g., histopathology, pathological tumor stage, molecular subtypes) were recorded when available (Table S3). Manchester scoring system stratified phenotypic severity. ¹⁷ , ¹⁸

2.3. Local databases

In this study, two different local databases were used. The first one, corresponding to our local TP53 database, was used for correlation studies and corresponds to a national genotype–phenotype database maintained by the Rouen laboratory as the reference laboratory for TP53 analysis (Table S4). On one hand, this local TP53 database was used in addition to the international TP53 database (The TP53 Database [R20, July 2019]: https://tp53.cancer.gov) to define if the variant was previously identified or not. On the other hand, to explore a potential genotype–phenotype correlation, we use this local TP53 database to enrich our collection of HBOC patients with patients carrying the same null or missense variants identified within this French HBOC patients cohort. The second one, our local breast cancer database, corresponds to data from 8018 consecutive breast cancer patients attending oncogenetic counseling in Rouen University Hospital prior to any genetic analysis. Our local breast cancer database was used to compare breast cancer characteristics with TP53 carriers of this study. Most of these patients were treated outside our center; in these cases, we have only partial access to clinical data. Thus, in this local breast cancer database, histopathological subtypes were available for 2327 patients, tumor grade for 1214 patients, and hormonal and Her2 status for 1915 patients.

2.4. Variant interpretation

Variants were described according to the NM_000546.5 and submitted to ClinVar and FROG (French OncoGenetics) databases. Classifications followed ACMG/AMP guidelines for TP53 germline variants (version 1.4.0 https://cspec.genome.network/cspec/ui/svi/doc/GN009?version=1.4.0). ¹⁹ For some variants, the Rouen oncogenetic laboratory has performed functional assays monitoring mutant p53 transcriptional activity from patients' blood samples or patients' lymphoblastoid cell lines. ⁸ , ²⁰ Unpublished data, obtained by the Rouen oncogenetic laboratory, were reported in Table S5 as « lab expertise ». Variants were categorized into: (i) null variants including nonsense and frameshift variants and large deletions; (ii) missense variants with DNE ⁸ , ²¹ , ²² , ²³ , ²⁴ ; (iii) other missense variants without DNE; (iv) splice variants affecting consensus splice sites −1, −2 or +1, +2 or with a demonstrated effect on splice by RNA analysis; (v) intronic variants without splicing analysis; (vi) silent variants; (vii) in‐frame deletion or duplication variants; and (viii) UTR variants.

2.5. Statistical analysis

To evaluate genotype–phenotype correlation, the Pearson's Chi‐squared test with a Yates continuity correction was performed. We used Kaplan–Meier survival analyses, independent two‐sided t‐tests, Chi‐squared, and Fisher exact test. All analyses were done using R software version 4.3.2, MedCalc® Statistical Software version 22.009, and biostaTGV website (https://biostatgv.sentiweb.fr/).

3. RESULTS

Fifteen French laboratories have performed HBOC panel analyzes in 53,085 patients meeting HBOC criteria, leading to the identification of a TP53 variant in 398 patients (Figure 1). Among them, 52 patients with a variant allelic frequency (VAF) below 40% were excluded from following analyses to minimize the risk of mosaicism, ctDNA, or clonal hematopoiesis of indeterminate potential (CHIP) issues. ²⁵ Prior to any exclusion, we checked for short insertion, deletion, or duplication that could mimic mosaicism due to calling issues. Two patients with CHIP were also excluded from following analyses. Reclassification according to ACMG‐TP53 guidelines excluded 110 patients carrying a benign or likely benign variant, resulting in 110 variants of uncertain significance (VUS) carriers and 124 carriers of pathogenic and likely pathogenic variants, grouped as pathogenic variants (PV). Then, we used our local TP53 database and the international TP53 database (The TP53 Database [R20, July 2019]: https://tp53.cancer.gov) (i) to search for variants previously identified in a patient meeting Chompret criteria (n = 96), and (ii) to separate patients carrying variants previously identified only in patients with breast cancer (hereinafter designated as HBOC‐associated variants), that is, without any other LFS tumor (n = 40) from variants previously identified in patients with LFS phenotype (hereinafter designated as LFS‐associated variants) (n = 84). Notably, 77 of these 84 LFS‐associated variants have been detected in children and seven in adults only.

FIGURE 1.

Study flow diagram.

3.1. Description of TP53 variants identified in HBOC panels

The mutation frequency for TP53 variant was 0.44% in HBOC panel (234/53,085), that is, 0.23% of PV and 0.21% of VUS. The vast majority (89%) of PV was located in the DNA binding domain (DBD) of the protein, whereas VUS were more evenly distributed with 59% located in DBD and 41% in other functional domains (Figure S2). Among the 124 PV carriers, we counted 96 patients with missense (65 DNE/31 others), 18 with null variants, eight with splice variants, and two with in‐frame deletion/duplication. Among the 110 VUS carriers, we described 42 patients with other missense variants, 31 with silent, 29 with intronic, six with UTR variants, and two with in‐frame deletion/duplication. Overall, 78 different TP53 PV and 81 VUS have been identified in these 124 and 110 carriers, respectively (Table S5). We identified 10 recurrent variants, reported at least in four patients: seven PV, that is, six missense (c.329G>A, c.473G>A, c.541C>T, c.733G>A, c.799C>T and c.847C>T) and one null variant (c.586C>T), and three VUS, that is, two missense (c.460G>A, c.467G>A) and one silent (c.618G>A). These 10 variants were present in 25% of patients (60/234) of this series. Thus, among patients tested on French HBOC panels, approximately one in 200 harbored a TP53 heterozygous variant, half of which were PV, mostly DNE missense variants located in DBD, and half of which were VUS, mostly other missense variants evenly distributed on p53 protein.

3.2. Phenotypes associated with TP53 PV or VUS

In this series, 226 women carried a TP53 variant (120 PV, 106 VUS) and eight men (4 PV, 4 VUS). Phenotypes of patients with PV and VUS are depicted in Figure 2. PV carriers developed breast cancer at a mean age of 40 [range: 22;82], significantly earlier than VUS carriers, 46 years [27;82] (t = 3.345, mean age difference 95% CI = [2.4814; 9.6106], p = 0.00098). With regards to ovarian cancer, mostly high‐grade serous carcinoma, no significant difference was found with a mean age of 68 years [51;81] for PV carriers and 55.8 years [23;74] for VUS carriers (t = −1.783, mean age difference 95% CI = [−2.0306; 26.4306], p = 0.0891). We reported 17 patients with two different tumors with at least one in the HBOC spectrum (11 with PV and 6 with VUS).

FIGURE 2.

Phenotypes of patients harboring a TP53 variant. Phenotypes of the 124 patients with a TP53 PV (A) and of the 110 patients with a TP53 VUS (B).

Manchester score was employed to integrate personal and family HBOC history and stratify phenotypes. No significant difference was found between patients with PV (mean score = 14.5 [2;37]) and patients with VUS (13.4 [1;47]) (t = −1.128, mean age difference 95% CI = [−0.8590; 3.1582], p = 0.2606).

In this series, 58 patients met the Chompret criteria (13 VUS and 45 PV carriers), mainly EOBC. In other words, 63% of patients (78/123) carrying a TP53 PV did not meet the Chompret criteria.

3.3. Phenotypes associated with TP53 variant types

Null variants were significantly more frequent among patients with HBOC‐associated variants, while missense variants were significantly more frequent in patients with LFS‐associated variants (χ² = 4.482, p = 0.0343), especially DNE missense variants (χ² = 6.711, p = 0.0096) (Figure 3 [A]). Null variants were associated with early onset compared to missense variants (χ² = 8.3633, p = 0.0153), particularly in the context of breast cancer (χ² = 11.6169, p = 0.0030**) (Figure 3B,C). No significant age difference was found between DNE and other missense variants (t = 0.271, mean age difference 95% CI = [−5.3792; 7.0778]; p = 0.7871).

FIGURE 3.

LOF TP53 variants are associated with HBOC‐associated variants and with early onset breast cancer. (A) Percentage of the different types of TP53 variants according to their proteic impact in patients with a LFS associated variant or a HBOC associated variant. We observe a significantly higher proportion of LOF variants in the HBOC associated variants group of patients compared to missense variants and particularly to DNE missense variants which seem associated to the LFS associated variants group of patients. (B) and (C) Age of first cancer (B) or first breast cancer (C) development in HBOC patients carrying a LOF, DNE missense or other missense TP53 PV. Patients with LOF variants (green) develop cancer sooner than those carrying a DNE missense variant (blue) or other missense variant (orange).

To deepen our understanding of these associations and avoid HBOC selection bias, we queried our local TP53 database for the TP53 PV identified in this study: 229 patients with missense and 24 patients with null variant were added to the 95 missense and the 18 null variants of this series (Table S4). Density plots depicted that null variants exhibit a bimodal distribution of age at first cancer development, with a first peak during childhood and a major peak in early adulthood between 20 and 40 years (Figure 4). In contrast, missense variants are more evenly distributed across ages, with a maximum risk between 35 and 45 years.

FIGURE 4.

Cancer distribution according to the age in patients carrying a null or missense TP53 PV. Density plots representing the cancer risk depending on age for null (in pink) and missense (in blue) TP53 variants. This graph has been obtained from collected data in this series and data from the Rouen laboratory local TP53 database. We reported the age of the first cancer for 324 patients carrying one of the 95 missense variants identified in this study and for 42 patients with one of the 18 null variants documented in this work.

Comparison of Manchester scores for patients carrying an HBOC‐associated variant versus an LFS‐associated variant showed no difference; mean scores were 13.5 [4;35] and 15.1 [ ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ ], respectively (t = 1.107, 95% CI[−1.2611;4.4598], p = 0.2705). Similarly, comparison of the localization of the variants within the different functional domains of the p53 protein showed no difference; all were primarily located in the DBD.

Altogether, we document an enrichment of TP53 null variants in HBOC‐associated variants, and these null variants were associated with an earlier age of breast cancer onset compared to missenses.

3.4. Further breast cancer considerations related to TP53 pathogenic variants

As expected, breast cancer was the predominant type of cancer in this series (n = 111 with PV). Median age for breast cancer development was 40 years old [22;82], and 39.7% of breast cancer occurred before or at age 30, 16.2% between ages 31 and 35, 18% between ages 36 and 40, 8.1% between ages 41 and 45, 9% between ages 46 and 50 and 18.9% after or at age 51. These observations demonstrate that even individuals with late‐onset breast cancer may harbor a TP53 PV (Figure S3), subject to verification of the absence of CHIP. Nevertheless, comparison with data from the local breast cancer database showed an enrichment of TP53 PV in patients with breast cancer before the age of 31 (OR = 11.9004, 95%CI = [7.5349;18.4676], p < 0.0001***), which corresponds to the recommended cut‐off age for TP53 analysis in sporadic cases ³ (Table 1). We also demonstrated an enrichment of TP53 PV in patients with breast cancer between 31 and 35 years old (OR = 2.5673, 95%CI = [1.4472;4.3212], p = 0.0011*). However, at older ages, we did not find any significant enrichment (OR = 1.7092, 95%CI = [0.9876;2.8302], p = 0.0486 for patients with breast between ages 36 to 40, OR = 0.5446, 95%CI = [0.2414;1.0792], p = 0.0953 for patients with breast cancer between ages 41 to 45, OR = 0.4701, 95%CI = [0.2182;0.9036], p = 0.0219 for patients with breast cancer between ages 46 to 50 and OR = 0.2784, 95%CI = [0.1641;0.4526], p < 0.0001 for patients with breast cancer after age of 50). Concerning other breast cancer specifications, in most cases only partial data were available for the patients of our series; however, we obtained, at least, one information for 93.6% of patients. Our data showed that, compared to the 8018 patients from the local breast cancer database from Rouen oncogenetic counselling, patients with TP53 PV are more likely to develop (i) in situ carcinoma than invasive carcinoma (OR = 4.2504, 95%CI = [2.0657;8.1872], p < 0.0001***) especially ductal in situ carcinoma before age of 41 years (OR = 4.2505, 95%CI = [1.4288;12.7181], p < 0.0056**); (ii) phyllodes tumors (OR = 10.8299, 95%CI = [1.0768;58.5601], p = 0.0219*); (iii) triple positive breast cancers (OR = 3.5034, 95%CI = [1.7765;6.5593], p = 0.0002***) and particularly before the age of 41 years (OR = 2.236, 95%CI = [0.9662;4.8421], p = 0.0387*); (iv) Her2 positive breast cancers (OR = 4.398, 95%CI = [2.529;7.6203], p < 0.0001***) also particularly before the age of 41 years (OR = 2.9724, 95%CI = [1.5346;5.7016], p = 0.0009***). Thus, patients developing before the age of 41 years old, breast phyllodes tumor or in situ carcinoma, Her2 positive or triple positive breast cancer, as well as patients with any types of breast cancer before age of 36, are indicative of a TP53 alteration.

TABLE 1.

Characteristics of breast cancer for patients with TP53 variants and patients from our local breast cancer database from Oncogenetic consultation.

		% of patients with TP53 variants (number of patients)							% of patients from oncogenetic consultation (number of patients)
		Total	<31y	[31–35]	[36–40]	[41–45]	[46–50]	≥51y	Total	<31y	[31–35]	[36–40]	[41–45]	[46–50]	≥51y
Breast cancer histopathological subtypes TP53 patients = 63 Oncogenetic B. ca. patients = 2327	IDC	79.3% (50)	27% (17)	11.1% (7)	16.4% (11)	7.5% (5)	4.5% (3)	11.1% (7)	85.1% (1980)	3.4% (79)	7.48% (174)	12.2% (283)	13.1% (305)	15.8% (368)	33.1% (771)
	ISDC	17.5% (11)	3% (2)	3% (2)	3% (2)	3% (2)	3% (2)	1.5% (1)	5.5% (127)	0.2% (4)	0.3% (7)	0.5% (12)	0.7% (17)	1.2% (29)	2.5% (58)
	ILC	6.3% (4)	0	0	1.5% (1)	1.5% (1)	1.5% (1)	1.5% (1)	12.4% (288)	0.1% (3)	0.3% (7)	1% (23)	2% (47)	2.2% (52)	6.7% (156)
	ISLC	3.2% (2)	0	0	0	0	1.5% (1)	1.5% (1)	0.3% (7)	0	0.04% (1)	0.04% (1)	0.04% (1)	0.09% (2)	0.09% (2)
	Phyllode tumor	3.2% (2)	1.5% (1)	0	1.5% (1)	0	0	0	0.3% (7)	0.04% (1)	0	0.09% (2)	0.09% (2)	0.04% (1)	0.04% (1)
	Medullar carcinoma	0% (0)	0	0	0	0	0	0	0.9% (20)	0.1% (3)	0.04% (1)	0.2% (4)	0.2% (5)	0.04% (1)	0.3% (6)
	Sarcoma	0% (0)	0	0	0	0	0	0	0.1% (3)	0	0	0	0	0.04% (1)	0.09% (2)
Grade TP53 patients = 41 Oncogenetic B. ca. patients = 1214	I	0% (0)	0	0	0	0	0	0	13.0% (158)	0.1% (1)	0.5% (6)	1.2% (14)	2.8% (30)	2.2% (27)	6.6% (80)
	II	48.8% (20)	14.6% (6)	9.8% (4)	12.2% (5)	2.4% (1)	4.9% (2)	4.9% (2)	45.2% (549)	1% (13)	2.8% (34)	6.6% (80)	6.8% (83)	8.2% (99)	19.8% (240)
	III	51.2% (21)	22% (9)	7.3% (3)	9.8% (4)	4.9% (2)	2.4% (1)	4.9% (2)	41.8% (507)	1.8% (22)	4% (49)	6.8% (83)	7.1% (86)	7.3% (89)	14.7% (178)
Immunochemistry TP53 patients = 67 for HR status. 61 for Her2 status and 61 for both Oncogenetic B. ca. patients = 2369 for HR status. 1939 for Her2 status and 1915 for both	HR positive	41.8% (28)	9% (6)	6% (4)	9% (6)	3% (2)	3% (2)	7.5% (5)	58.4% (1383)	1.8% (42)	3.9% (92)	7.9% (188)	9.6% (228)	11.1% (264)	24% (569)
	HR negative	50.7% (34)	20.1% (14)	7.5% (5)	7.5% (5)	6% (4)	3% (2)	6% (4)	32.2% (763)	13% (31)	3.3% (77)	4.9% (117)	5.2% (123)	5.5% (130)	12% (285)
	Her2 positive	47.5% (29)	16.4% (10)	6.6% (4)	11.5% (7)	3.3% (2)	1.6% (1)	8.2% (5)	17.1% (331)	1.1% (22)	2.4% (47)	2.9% (56)	2.8% (55)	2.7% (53)	5.1% (98)
	Her2 negative	52.5% (32)	16.4% (10)	6.6% (4)	11.5% (7)	6.6% (4)	4.9% (3)	6.6% (4)	82.9% (1608)	2.6% (50)	5.7% (110)	11% (213)	13.6% (263)	15.3% (296)	34.9% (676)
	Triple positive	24.6% (15)	4.9% (3)	4.9% (3)	8.2% (5)	0	0	6.6% (4)	8.5% (163)	0.6% (11)	1.3% (24)	1.8% (32)	1.3% (25)	1.6% (30)	2.1% (41)
	Triple negative	26.2% (16)	8.2% (5)	4.9% (3)	3.3% (2)	3.3% (2)	1.6% (1)	4.9% (3)	29.4% (564)	0.1% (19)	2.7% (52)	4.2% (81)	5% (95)	5.4% (104)	11.1% (213)
Total number of patients with breast cancer		111	29.7% (33)	16.2% (18)	18% (20)	8.1% (9)	9% (10)	18.9% (21)	8018	3.4% (275)	7% (562)	12.6% (1012)	13.9% (1118)	17.4% (1395)	45.6% (3656)

Note: Numbers of patients with available data are indicated in italics in the first column.

Abbreviations: IDC. Intraductal carcinoma; ILC, intralobular carcinoma; ISDC, in situ ductal carcinoma; ISLC, in situ lobular carcinoma.

4. DISCUSSION

TP53 gene is involved in breast cancer predisposition, but the delicate classification of its variations and the wide phenotypic variability raise many questions. Below, we discuss difficulties encountered in classifying TP53 variations, genotype–phenotype correlation, types of breast cancer suggestive of a TP53 alteration, and clinical management of patients arising from the identification of this alteration.

Inclusion of TP53 in French HBOC panel has led to identification of a TP53 variant in 0.44% of analyzed patients, 0.23% with PV and 0.21% with VUS. This finding aligns with previous studies reporting a mutation frequency from 0.1% to 1.6% in multigene panel analysis, varying according to ethnic origin and inclusion criteria. ¹⁰ , ¹² , ²⁶ , ²⁷ However, previous studies reported twice as many VUS as PV, ¹⁰ , ²⁶ underlining the importance of efficient variant curation, functional assays, and systematic comprehensive splicing analyses, especially as 54% of VUS identified in this series corresponds to intronic and silent. ²⁸ While TP53 is a tumor suppressor gene, it is characterized by a high rate of missense variants (above 70% according IARC database). Numerous functional assays have been developed and all missense variants of this series have benefited from these assays in yeast ²⁹ and human cell lines, ²⁴ both exploring transcriptional activity of p53 protein, mainly supported by the DBD. As VUS were located throughout p53, new functional assays dedicated to other p53 activities may help in their classification.

TP53 is a genuine gene of hereditary predisposition to breast cancer, and according to this work, not only in patients with EOBC, indeed, in our study, 63% of patients harboring a TP53 PV are not fulfilling Chompret criteria. Nevertheless, for older patients not meeting Chompret criteria, the presence of the TP53 PV at the germline level has to be confirmed by analyzing another tissue, especially to avoid the risk of clonal hematopoiesis. Since the identification of a TP53 PV may have an impact on clinical decisions (e.g., surgery, radiotherapy), the question arises as to whether TP53 status should be known prior to any treatment and in which patients. Our results showed an enrichment in TP53 PV in patients with breast cancer before the age of 36, and in patients with Her2‐positive or triple‐positive breast cancer, in situ carcinoma or phyllodes tumor before the age of 41, as previously suggested by Sandoval and collaborators. ³⁰ Therefore, these patients should benefit from TP53 analysis prior to any therapeutic decision. This highlights the variable penetrance of TP53 variants, the associated challenges for proper clinical monitoring, and the need for large, in‐depth genotype–phenotype studies.

PV type contributes to phenotypic variability. Here, we demonstrated an association between null variants and EOBC. Previous studies have suggested this association, but results were not significant. ¹¹ , ¹² Whereas missense variants, particularly with DNE, were mostly associated with classic LFS. ⁷ Notably, in HBOC patients, we observed no difference in terms of age of onset between patients with missense variants with or without DNE. The age‐dependent distribution of tumor development for null and missense variants suggests that two different entities relying on PV type could coexist (Figure 4). The first one, linked to missense PV, would correspond to classic LFS with a wide tumor spectrum in children and adults, including EOBC. The second one, linked to null PV, would correspond to breast cancer in early adulthood. These results propose that mammary tumorigenesis could rely on a different mechanism than other LFS tumors.

However, the type of variant is far from explaining the whole picture, as illustrated by the c.1010G>A variant with a founder effect in Brazil ³¹ and our own findings on prototypic missense recurrent variants. Hence, variants such as c.329G>A and c.847C>T, already described in children with cancer and exhibiting intermediate effects in functional assays, were found in patients of this series presenting a great diversity of phenotypes from EOBC to late peritoneal carcinosis. Unexpectedly, bona fide TP53 PV with abrogated functions through functional testing, including DNE, were found in our series associated with attenuated LFS as described by Kratz and collaborators. ³² Thus, the c.524G>A, c.799C>T, c.818G>A, and c.844C>T variants have been identified in five patients (highlighted in Table S3 with a *) with breast cancer between 41 and 62 years or even asymptomatic. This phenotypic variability associated with variants with intermediate or drastic effects highly suggests the involvement of modifying factors, which could be both genetic (polygenic risk factors, SNP, modifiers genes, second allele), epigenetic (methylation), physiological (breast density) and/or environmental (genotoxic exposure, lifestyle, …). ³³ , ³⁴ , ³⁵ Interestingly, in Montellier et al., breast cancer‐associated variants were mostly variants with intermediate activities in Kato et al.'s functional assay. ³⁶ Unlike our study based on HBOC patients from routine consultations, Montellier et al.'s study was based mainly on LFS families included in the IARC database. This testifies to the importance of considering the phenotypic gateway leading to the TP53 variant identification and the need to take family and individual factors into account when considering tumor risk. ³⁷ Progress in understanding these modifying factors seems crucial to explain the phenotypic variability associated with TP53 variants and predict the cancer risk of each carrier. Thus, TP53 variants, whether null or DNE missense, could act as permissive variants and, associated with exacerbating or protecting modifying factors, could influence the location and kinetics of tumor development, explaining the variable penetrance observed with TP53 variants between families or within one family.

Given the considerable burden of medical follow‐up of patients with TP53 PV, it is desirable to establish appropriate genotype–phenotype associations, to tailor it as closely as possible to the patient's needs. Our data, although limited by patient follow‐up information, showed a drastic decrease in the number of first tumor in patients over 50 years old carrying a TP53 null variant (Figure 4). This provides new arguments for carrying out a genuine risk study taking into account the type of TP53 PV in order to stratify the clinical management and follow‐up of patients carrying the variant. Contribution of modifying factors also pleads for integration of family history of cancer and other personal features in the clinical management of the patient and its relatives, either to alleviate or reinforce the medical follow‐up, especially in patients harboring a null TP53 PV. It requires verifying that the absence of family history is not linked to a de novo TP53 variant. ³⁸ This de novo status could not be verified in our cohort, but, in any case, has no impact on our results.

Inclusion of TP53 in HBOC panels has proven valuable in identifying PV, offering medical support to a significant proportion of patients not fulfilling traditional nosological criteria, but also raising questions about actual associated risk. The findings also highlight the need for careful variant classification, considering the challenges posed by VUS. The study underscores TP53 as a genuine gene of hereditary breast cancer predisposition regardless of age of onset, but with an enrichment in patients before the age of 36, and in patients with Her2‐positive or triple‐positive breast cancer, in situ carcinoma, or phyllodes tumor before the age of 41, suggesting that these patients should benefit from TP53 analysis prior to any therapeutic decision. Finally, our results suggest the existence of two phenotypic entities: classic LFS with a broad tumor spectrum, including EOBC, in adults and children, and TP53‐related breast cancer in early adulthood. These two entities rely on TP53 variant type and the variable penetrance we observed with recurrent variants or hotspot mutations highlights the effect of modifying factors that we could approach nowadays thanks to family history of cancers and other personal factors. The classification of the variant and its molecular effect must therefore be distinguished from each patient's risk of developing cancer, and both aspects must be considered in order to establish the most appropriate follow‐up recommendations for patients with a TP53 PV.

AUTHOR CONTRIBUTIONS

Edwige Kasper: Conceptualization; investigation; writing – original draft; methodology; validation; visualization; writing – review and editing; formal analysis; data curation; supervision. Flavie Boulouard: Validation; writing – review and editing; data curation. Noémie Basset: Validation; writing – review and editing; data curation. Lisa Golmard: Validation; writing – review and editing; data curation. Hela Sassi: Data curation; validation; writing – review and editing. Emilie Bouvignies: Validation; writing – review and editing; data curation. Maud Branchaud: Validation; writing – review and editing; data curation. Camille Charbonnier: Methodology; validation; writing – review and editing. Nathalie Parodi: Validation; writing – review and editing; data curation. Marion Rolain: Validation; writing – review and editing; data curation. Juliette Albuisson: Validation; writing – review and editing. Ayman al Saati: Validation; writing – review and editing. Patrick Benusiglio: Validation; writing – review and editing. Pascaline Berthet: Validation; writing – review and editing. Marie Bidart: Validation; data curation; writing – review and editing. Céline Bonnet: Validation; writing – review and editing. Ahmed Bouras: Validation; writing – review and editing. Nadia Boutry‐Kryza: Data curation; validation; writing – review and editing. Fanny Brayotel: Data curation; validation; writing – review and editing. Virginie Bubien: Validation; writing – review and editing. Adrien Buisson: Validation; writing – review and editing. Laurent Castéra: Validation; writing – review and editing. Olivier Caron: Validation; writing – review and editing. Chrystelle Colas: Validation; writing – review and editing. Florence Coulet: Validation; writing – review and editing. Capucine Delnatte: Validation; writing – review and editing. Valentin Derangère: Validation; writing – review and editing. Alice Fievet: Validation; writing – review and editing. Céline Garrec: Data curation; validation; writing – review and editing. Marion Gauthier‐Villars: Validation; writing – review and editing. Mathilde Gay‐Bellile: Data curation; validation; writing – review and editing. Vincent Goussot: Validation; writing – review and editing. Jessica le Gall: Validation; writing – review and editing. Mathis Lepage: Validation; writing – review and editing. Anna Lokchine: Data curation; validation. Alexandre Perrier: Validation; writing – review and editing. Etienne Rouleau: Validation; writing – review and editing. Nicolas Sevenet: Data curation; validation; writing – review and editing. Dominique Stoppa‐Lyonnet: Validation; writing – review and editing. Jean‐Marie Ravel: Validation; writing – review and editing. Pierre Vande Perre: Validation; writing – review and editing. Dominique Vaur: Validation; writing – review and editing. Paul Vilquin: Data curation; validation; writing – review and editing. Gaëlle Bougeard: Validation; writing – review and editing. Stéphanie Baert‐Desurmont: Supervision; validation; writing – review and editing. Jean‐Christophe Thery: Supervision; validation; writing – review and editing. Claude Houdayer: Supervision; validation; writing – review and editing; methodology; writing – original draft.

CONFLICT OF INTEREST STATEMENT

The authors have no conflict of interest to declare.

ETHICS STATEMENT

The study was approved by the ethics committee of Rouen University Hospital (E2023‐68) and informed consent was taken from all individual participants.

Supporting information

DATA S1. Supporting information.

Kasper E, Boulouard F, Basset N, et al. Deciphering dual clinical entities associated with TP53 pathogenic variants: Insights from 53,085 HBOC panel analyses in French laboratories. Int J Cancer. 2025;157(5):897‐907. doi: 10.1002/ijc.35475

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.