This rule code can only be used to compare variants asserted as pathogenic or likely pathogenic following the ClinGen TP53 VCEP’s specifications. Must confirm there is no difference using RNA data or SpliceAI (SpliceAI < 0.2). Caveat: If both PS1 and PM5 are met, apply the strongest weight possible for each rule code not to exceed a combined strength of strong (4 points in total).

De novo points should be tallied using the table for tallying proband points based on whether maternity and paternity have been confirmed and the type of cancer(s) seen in the proband. This includes probands that are confirmed constitutional mosaics (low TP53 VAF on blood or buccal testing with the mutation detected in non-lymphocyte tissue and/or segregating in children) which may be counted as a confirmed de novo case. For probands with multiple cancers, use the most specific/highest weight cancer to determine point application for that proband. Points for all probands should be tallied to determine the strength of PS2 code application, consistent with SVI guidance. To avoid redundancy and increase consistency, the TP53 VCEP has opted to drop PM6 and use PS2 exclusively for de novo evidence.

A Table for LFS Cancers for PS2 (and PP1) code application is included below

Kato et al., 2003 (PMID: 12826609) systematic data performed best on our test set of reference variants. These data thus remain the primary functional assay underlying the classification. Giacomelli et al., 2018 (PMID: 30224644) assays are also systematic and are available for all p53 missense variants. When using cut-offs derived from original publication data (optimal cut-offs separating silent and common cancer variants), they show good concordance with other assays. Giacomelli LOF class can thus be used to support and complement Kato data. If Kato data is supported by Kawaguchi et al., 2005 (PMID: 16007150) in the tetramerization domain and tetramerization is affected, this can be used to apply PS3_Supporting. Kotler et al., 2018 (PMID: 29979965) data are available for a large number of variants with different effects, but only for those within the DNA binding domain. They may be used as an additional non-systematic missense LOF assay or for small deletions. Caveat: Do not apply PS3 at any weight for “missense” variants using assays done at the protein level (such as Kato et al. or Giacomelli et al.) if PP3 is applied based on SpliceAI. If there is any laboratory evidence, including RNA-seq data, of splicing aberration for the genetic variant being assessed, for which PVS1_Variable Weight (RNA) might be considered instead.

Data Supporting Functional Classes:

Kato et al. 2003 (PMID: 12826609) Transactivation Class: Classification based on the median transactivation activity using eight promoters in yeast. Values can be found in the NCI TP53 Database.

Non-functional: ≤ 20% activity

Partially-functional: > 20% and ≤ 75% activity

Functional : > 75% activity (variants showing supertransactivation are treated as Functional)

Giacomelli et al., 2018 (PMID: 30224644): Classification based on results from growth suppression assays in A549 human cells.

LOF: Etoposide Z-score ≤ -0.21

No LOF: Etoposide Z-score > -0.21

Kawaguchi et al., 2005 (PMID: 16007150): Classification based on the ability to form an oligomer in yeast.

Abnormal: Monomer/dimer

Normal: Tetramer

Other assays: In vitro growth assays in H1299 human cells from Kotler et al., 2018 (PMID: 29979965) with RFS score ≥ -1.0 for LOF and RFS score < -1.0 for noLOF. Or colony formation assays, growth suppression assays, apoptosis assays, tetramer assays, or knock-in mouse models.

Non-systematic assays are harder to calibrate, but if they meet Brnich et al., 2019 (PMID: 31892348) recommendations for the application of functional evidence and they are in agreement with Kato et al., 2003 , they should be taken into account. A large proportion of these assays are documented in the NCI TP53 database and thus can easily be found by curators. Second assays that may be considered include colony formation assays, apoptosis assays, tetramer assays, knock-in mouse models, and growth suppression assays.

This rule should be used and weighted appropriately for variants with functional evidence of loss of function. Follow SVI guidance regarding control numbers for functional studies. Downgrade to PS3_Moderate if PVS1_Strong is applied. Do not apply PS3 at any strength if PVS1 is applied at full strength.

See Functional Flowchart for more information and guidance on application of functional rule codes

There are two widely used clinical criteria for assessing the likelihood of Li Fraumeni syndrome - Classical and Chompret criteria - with the Chompret criteria being less restrictive. Individuals who meet the Revised Chompret criteria have an estimated ~30% risk of harboring a pathogenic TP53 variant (Bougeard et al., 2015; PMID: 26014290). Members of the TP53 VCEP calculated likelihood ratios (LRs) for patients meeting Classic LFS or Revised Chompret criteria (excluding confirmed constitutional mosaics and carriers of pathogenic variants in other cancer predisposition genes) using multigene panel testing from Ambry Genetics laboratory. Our data demonstrated that individuals meeting Revised Chompret criteria had a LR of > 2.08 to ≤ 4.3 and individuals meeting Classic LFS criteria had a LR of > 4.3 to ≤ 18.7. Therefore, we recommend that probands with TP53 germline variants meeting Revised Chompret should be given 0.5 point and probands meeting Classic LFS criteria should be given 1 point.  Do not apply this code for probands with de novo TP53 variants, in which case PS2_Variable Weight should be applied instead.

Early-onset breast cancer is the most common malignancy in women with LFS. Breast tumors from TP53 carriers are more likely to be HER2+ than those of non-carriers. Fortuno et al., 2020 (PMID: 32485079) investigated if breast tumor HER2 status has utility as a predictor of TP53 germline variant pathogenicity considering age at diagnosis. Their results showed that the identification of HER2+ breast tumors diagnosed before age 40 equated to Supporting level towards pathogenicity and therefore can be incorporated into TP53 criteria. Unrelated probands who are diagnosed with a HER2+ breast cancer below the age of 40 should be conservatively given 0.5 point. 

Phenotype points in unrelated probands should be tallied using the simplified table for tallying PS4 proband points.

Caveats: Points attributed to HER2 status may only be applied in unrelated individuals who underwent multigene panel testing with no other pathogenic/likely pathogenic variants in cancer predisposition genes; individuals who underwent targeted TP53 single gene testing may not count towards applied points.Variant must meet PM2_Supporting in order for PS4 to be applied at any strength.

See simplified table for tallying probing points for PS4

There are several known major hotspots for the TP53 gene. This code can be used for variants within the following codons using canonical transcript NM_000546.4: 175, 245, 248, 249, 273, 282

This code can also be used for germline missense variants seen in cancerhotspots.org (v2) with ≥ 10 somatic occurrences for the same amino acid change. This follows the recommendation from the ClinGen Germline/Somatic Variant Curation Subcommittee (PMID: 30311369).

This code can be applied for a missense change at an amino acid residue where one or more pathogenic/likely pathogenic variants have been identified. The other variant must be interpreted as pathogenic or likely pathogenic following the ClinGen TP53 VCEP’s specifications. The previously established pathogenic/likely pathogenic variant must reach a classification of pathogenicity without PM5. 

Grantham should be used to compare the variants. The variant being evaluated must be equal or worse (value is greater than) than the known pathogenic variant (i.e. the variant residue should be equally chemically different or more chemically different than the known pathogenic residue in comparison to the wild type residue). Splicing should be ruled out with either RNA data or SpliceAI (SpliceAI < 0.2).

Caveats: If both PS1 and PM5 are met, apply the strongest weight possible for each rule code not to exceed a combined strength of strong (4 points in total).

Meioses should be counted for individuals who both carry the variant and have a relevant cancer (see LFS cancers table). Meioses can be counted through unaffected obligate carriers. Caution should be used in counting meioses across many families where breast cancer is the only cancer present as this is a common cancer type. It is preferable that breast cancer predisposition syndromes have been ruled out with genetic testing, but this is not required to apply meioses.

In cases where multiple individuals in a family have a relevant cancer and only tumor testing demonstrating the variant (no germline data), meioses may be applied if the variant has been demonstrated in the germline in at least one individual in the family. (Caveat: Positive tumor testing must exist in multiple family members; meioses should not be applied if there is only positive tumor testing in a single individual. If the variant allele fraction in the tumor is not consistent with the variant being heterozygous it should not count towards meioses. Use caution if the family does not meet Classic LFS criteria). 

Do not apply PP1 if variant meets BA1/BS1 criteria.

See Table of LFS cancers for PP1 (and PS2) code application.

According to the published study by Fortuno et al., 2018 (PMID: 29775997) comparing the performance of different bioinformatics tools for TP53, the tools selected are aGVGD (not available for single amino acid in-frame deletions) and BayesDel. To investigate potential effects on splicing for intronic, silent, and apparent missense variants, the SpliceAI tool was selected based on recommendations from the ClinGen SVI Splicing Subgroup. All variants should be assessed to consider if there are splicing effects predicted. PP3 should not be used in combination with PVS1.

Missense variants (See Flowchart for application of PP3 and BP4 rule codes for missense variants)

• PP3_Moderate: aGVGD Class C65 and BayesDel score ≥ 0.16
• PP3: aGVGD class C25-C55 and BayesDel score ≥ 0.16

Single amino acid in-frame deletions (See single aa BayesDel spreadsheet)

• PP3: BayesDel score ≥ 0.16

Exonic (including silent variants and apparent “missense” variants or “single amino acid in-frame deletions” for which there is a predicted splice effect) or Intronic Splice Variants (excluding ± 1,2 positions):

• PP3: SpliceAI ≥ 0.2

The frequency of likely somatic variants in blood among patients undergoing multigene panel testing is high for variants in TP53 (PMID: 29189820). TP53 variants observed at a low variant allele fraction (VAF) may be due to true constitutional mosaicism (which can be confirmed by observing the variant in other non-lymphocyte tissues; in the tumor at higher VAF; and/or segregating in other family members); technical assay issues; a clone driven by underlying malignancy or previous treatment with chemotherapy; or clonal hematopoiesis of indeterminate potential (CHIP). 

Positive selection has been proposed to be a mechanism driving clonal hematopoiesis (CH). Fortuno et al., 2022 (PMID: 34906512) demonstrated that the observation of TP53 variants at low VAF is a significant predictor of variant pathogenicity. Likelihood ratios toward pathogenicity associated with a VAF 5-25% corresponded to the ACMG-AMP strength level of moderate, and supporting with VAF 25-35%. Code-weighting for this rule was derived from datasets that are equivalent to the information available to diagnostic laboratories with the aim that this would be accurate for interpretation for low VAF variants in a real world testing situation. Uncertainty about the variant truly being the result of CHIP is built into the code strengths assigned, which therefore excludes confirmed constitutional mosaicism. 

Caveats: This evidence code assumes a somatic origin of the TP53 variant. PP4 and points towards any phenotype-based rule codes (e.g., PS4, PS2, PP1) cannot be applied in the same individual in combination. PP4 and PS2 cannot be applied together in the same individual. Do not apply this code if variant meets BA1 or BS1. Variant must have been detected on MGPT in order for this code to be applied.