A substantial proportion of apparently heterozygous TP53 pathogenic variants detected with a next‐generation sequencing hereditary pan‐cancer panel are acquired somatically

Abstract Previous analysis of next‐generation sequencing (NGS) hereditary pan‐cancer panel testing demonstrated that approximately 40% of TP53 pathogenic and likely pathogenic variants (PVs) detected have NGS allele frequencies between 10% and 30%, indicating that they likely are acquired somatically. These are seen more frequently in older adults, suggesting that most result from normal aging‐related clonal hematopoiesis. For this analysis, apparent heterozygous germline TP53 PV carriers (NGS allele frequency 30–70%) were offered follow‐up testing to confirm variant origin. Ninety‐eight probands had samples submitted for follow‐up family member testing, fibroblast testing, or both. The apparent heterozygous germline TP53 PV was not detected in 32.6% (15/46) of submitted fibroblast samples, indicating that it was acquired somatically, either through clonal hematopoiesis or via constitutional mosaicism. Notably, no individuals with confirmed germline or likely germline TP53 PVs met classic Li–Fraumeni syndrome (LFS) criteria, only 41% met Chompret LFS criteria, and 59% met neither criteria, based upon provider‐reported personal and family cancer history. Comprehensive reporting of TP53 PVs detected using NGS, combined with follow‐up analysis to confirm variant origin, is advised for clinical testing laboratories. These findings underscore the investment required to provide individuals and family members with clinically accurate genetic test results pertaining to their LFS risk.

pathogenic variants (PVs) in TP53 are the most common cause of LFS.
Children who carry a germline PV in TP53 have an approximately 42% risk of developing a TP53-related cancer before age 16 (Le Bihan, Moutou, Brugières, Feunteun, & Bonaïti-Pellié, 1995), and adult men have an approximately 73% lifetime risk of cancer (Chompret et al., 2000). Women who have germline TP53 PVs have nearly a 100% chance of developing breast cancer during their lifetimes (Chompret et al., 2000). For these reasons, recent expert panel recommendations emphasize comprehensive, lifelong screening across numerous cancer types, beginning as soon as an individual has been confirmed to carry a TP53 PV or meets the "classic clinical definition" of LFS with no TP53 PV. This follows a modified version of the "Toronto protocol" (Kratz et al., 2017), which factors in the age and sex of the individual. In addition, National Comprehensive Cancer Network (NCCN) guidelines recommend that individuals with LFS receive aggressive cancer screening, including: annual breast MRI starting at age 20; colonoscopy every 2-5 years starting at age 25; whole-body MRI; esophagogastroduodenoscopy; physical exam every 6-12 months; and a dermatologic exam starting at age 18 years (Daly et al., 2019). For women with LFS, risk-reducing mastectomy also should be considered.
Due to the high cancer risks conferred by germline TP53 PVs, hereditary pan-cancer gene panels typically include TP53. Recent work showed that apparent germline TP53 PVs were not uncommon among individuals who received hereditary cancer panel testing, accounting for 1% of germline PVs identified (Rosenthal, Bernhisel, Brown, Kidd, & Manley, 2017). Next-generation sequencing (NGS) is used to accommodate the large number of genes often included in hereditary cancer panels; however, NGS is more sensitive than traditional sequencing methods and can detect variants at very lowallele frequencies. Heterozygous germline variants will have NGS sequence reads with approximately a 1:1 ratio to those of the wildtype allele, producing an NGS allele frequency of about 50%. Allele frequencies that differ significantly from 50% in blood-derived samples suggest that the variant occurs in only a proportion of blood cells and therefore was acquired somatically. Somatic variants also can be detected in saliva-derived samples, although the allele frequencies in saliva may be different than in blood, as saliva samples often contain a mix of epithelial and white blood cells.
A previous study showed that likely somatic PVs accounted for 0.71% of all PVs identified during hereditary cancer panel testing, with somatic PVs occurring most commonly in TP53 and among older individuals (Coffee et al., 2017). This was consistent with previous work, which has shown that clonal expansion of blood cell subpopulations carrying somatic variants during hematopoiesis, also termed clonal hematopoiesis of indeterminate potential (CHIP), has been observed for a number of genes (Genovese et al., 2014;Jaiswal et al., 2014;Mitchell et al., 2018;Xie et al., 2014) and occurs with normal aging (Steensma et al., 2015). However, it was notable that in one case, the NGS allele frequency of a TP53 PV increased from less than 30% (likely somatic range) to 45% (apparent germline range) when retested after 3 months (Coffee et al., 2017). This led to the hypothesis that the NGS allele frequency range for somatically acquired TP53 PVs may overlap with that expected for heterozygous germline variant alleles, thus potentially confounding test result interpretation.
Unlike germline PVs, variants acquired somatically through CHIP do not carry hereditary cancer risk, though they are associated with other risks such as hematologic cancer, cardiovascular disease, and all-cause mortality (Genovese et al., 2014;Gillis et al., 2017;Jaiswal et al., 2014;Jaiswal et al., 2017;Takahashi et al., 2017). Given the severe clinical interventions for individuals with LFS, it is critical that apparent germline PVs in TP53 be investigated further to determine whether they are truly in the germline. For this reason, our laboratory offers follow-up testing to individuals with an apparent germline TP53 PV identified during the course of clinical hereditary cancer testing, to determine whether the PV was germline or acquired somatically. This report presents the results of follow-up testing.

| Editorial policies and ethical considerations
Only deidentified data collected as part of clinical testing were used in this analysis. All individuals provided informed consent for clinical testing. The analysis set did not include individuals from states with laws that prevent the use of deidentified genetic data. The corresponding and senior authors confirm that they had full access to all data used in the analysis and take responsibility for data integrity and the accuracy of the data analysis. All authors are employees of Myriad Genetic Laboratories, Inc., and receive salary and stock options as compensation.

| Cohort and genetic testing
This analysis assessed individuals who had hereditary pan-cancer panel testing (Myriad Genetic Laboratories, Inc., Salt Lake City, UT) as part of normal clinical care between September 2013 and February 2018. Testing was performed using blood or saliva samples.
Validation of the NGS panel has been described previously and included analysis of TP53 (Judkins et al., 2015). Sequence variants and large rearrangements (LRs) were analyzed relative to the wildtype TP53 gene sequence (RefSeq NM_000546.5). For sequence variants, apparent germline PVs were defined as those identified with an NGS allele frequency between 30% and 70% and which received a laboratory classification of Pathogenic or Likely Pathogenic. This threshold was assigned to minimize false negative results and was based on our laboratory's experience and expertise in clinical genetic testing. The broad range captures potential germline allele frequency imbalances that may occur due to technical or biological reasons. For LR variants, apparent germline PVs were identified through NGS dosage analysis, which used normalized read counts from sequencing amplicons to determine gene copy number.
While this approach did not allow assignment of precise allele frequencies for LRs, dosage increases or decreases relative to those expected for a heterozygous germline deletion or duplication could be evaluated. A germline LR typically presents as a 50% increase or decrease in dosage.
All sequence PVs were confirmed by Sanger sequencing. Pathogenic LRs were detected by NGS and confirmed by microarray, multiplex ligation-dependent probe amplification, and/or repeat NGS. Individuals who were found to have an apparent heterozygous germline PV in TP53 were offered follow-up testing at no additional cost to determine whether the variant was germline or somatic.
Clinical information, including age, personal cancer history, and family cancer history, was obtained from provider-completed test request forms and deidentified.

| Follow-up testing
Follow-up testing was performed either on blood or saliva samples from informative family members or on a cultured fibroblast sample from the proband (fibroblast cell culturing was performed by ARUP Laboratories, Salt Lake City, UT). DNA samples were extracted using standard methods. Follow-up fibroblast and family testing consisted of Sanger single-site analysis for the TP53 PV that was initially identified in the proband. Figure 1 depicts the interpretation of the PV identified in the proband based on the results of follow-up, confirmatory testing. When family member testing was performed, the TP53 PV was "confirmed germline" if it also was identified in a relative who was not a child or grandchild. In this scenario, the PV was considered constitutionally present in all tissues, consistent with a molecular diagnosis of LFS. The PV was considered "likely germline" when it was present in a child or grandchild of the proband. In this scenario, it is likely that the PV is constitutional in the proband, yet it remains possible that it was acquired during early embryonic development and segregated to germ cells and other tissues, but not all tissues, rendering the proband constitutionally mosaic for the TP53 PV. Family member testing was considered "uninformative" when the TP53 PV was not detected in the tested family member.
When fibroblast testing was performed, the TP53 PV was considered "somatic" if it was absent from the proband fibroblasts.
This demonstrates that the PV is not present in all tissues; however, it remains undetermined whether the PV is restricted to blood (CHIP) or is present in some but not all tissues (constitutional mosaic). The PV was considered "likely germline" if it was present in the fibroblast sample from the proband. While identification of the PV in skin fibroblasts is strongly supportive of a constitutional variant, it does not definitively confirm the presence in the germline.
Fibroblast testing results were "inconclusive" when initial NGS pancancer panel testing showed an allele frequency consistent with a germline variant, but the ratio of wild-type to PV allele in the fibroblast sample was significantly skewed from the expected 50:50 ratio. Such skewing might reflect the presence of more than one cell population in the skin biopsy, or it might result from technical or cell culture artifacts.

| Analysis
Analyses were performed for all individuals with an apparent germline PV in TP53, were stratified by variant origin (confirmed germline, likely germline, somatic, uninformative, inconclusive, and unknown), and included age at testing and age at cancer diagnosis (if applicable). For the subset of individuals who had follow-up testing, the results of follow-up testing and NGS allele frequency were also assessed.
For all individuals who carried an apparent germline PV in TP53, clinical criteria for LFS were evaluated using classic LFS (Li et al., 1988) and revised Chompret LFS criteria (Bougeard et al., 2015). Classic LFS criteria were: proband diagnosed with sarcoma before 45 years of age, with a first-degree relative diagnosed with any cancer before 45 years of age, and an additional first-or second-degree relative diagnosed with any cancer before 45 years of age or a sarcoma at any age. Chompret LFS criteria were: (a) proband diagnosed with an LFS tumor (soft tissue sarcoma, osteosarcoma, brain tumor, adrenocortical carcinoma, leukemia, F I G U R E 1 Follow-up testing schema. The possible interpretations of follow-up test results are provided for family testing and fibroblast testing in the proband COFFEE ET AL.

| 205
lung, or premenopausal breast [DCIS, invasive, or lobular invasive]) before 46 years of age, and at least one first-or second-degree relative diagnosed with one LFS tumor before 56 years of age or with multiple primaries at any age; or (b) a proband with multiple tumors (except multiple breast tumors), the first of which occurred before age 46, and two of which belong to the LFS tumor spectrum; or (c) a proband with adrenocortical carcinoma, choroid plexus carcinoma, or rhabdomyosarcoma of embryonal anaplastic subtype, regardless of family history; or (d) breast cancer before 31 years of age.

| RESULTS
A total of 221 individuals were identified as carrying an apparent TP53 PV (NGS allele frequency of 30-70%). The median age at testing was 41 years (range 15-88), and 28.5% (63/221) of individuals were older than 50 years ( Figure 2  When follow-up included fibroblast testing only, the TP53 PV was detected in fibroblasts in 56.0% (14/25) of cases and was designated as likely germline (Table 1). The PV was absent from fibroblasts and thus designated as somatic in 36.0% (9/25) of cases. In 8.0% (2/25) of cases, the variant status was determined to be inconclusive; while the variant appeared to be germline in blood, the ratio was significantly skewed in fibroblasts, indicating possible mosaicism in skin tissue.
When follow-up included both family member and fibroblast testing, the PV was absent in both fibroblasts and tested relatives in 28.6% (6/21) of cases and was considered somatic. In an additional 66.7% (14/21) of cases, the PV was present in both fibroblasts and tested relatives and was designated as confirmed germline or likely germline depending on the relationship of the tested relatives. For one case (4.8%; 1/21), the PV ratio was significantly skewed in fibroblasts and determined to be inconclusive.

| Confirmed and likely germline TP53 PV carriers
Overall, follow-up testing confirmed germline status for the original   (Table 2). None met the criteria for classic LFS.

| Somatic TP53 PV carriers
Among individuals with an apparent germline PV who had follow-up fibroblast testing, TP53 PVs were found to be somatic in 32.6% (15/46) of cases (Table 1). While confirmed germline/likely germline PVs shared similar NGS allele frequency ranges, grouping around 50%, the somatic PV allele frequencies were dispersed more evenly (Figure 3).
Eighty percent (12/15) of individuals with a somatic TP53 PV had a reported personal history of cancer, with a median age at diagnosis of 44 years (Figure 2). For six somatic PV carriers, early-onset breast cancer (ages 25, 28, 32, 33, 41, and 42)  *Fibroblast testing results were inconclusive when initial testing showed an allele frequency consistent with a germline variant, but the ratio of wild-type to PV allele in the fibroblast sample was significantly skewed from the expected 50:50 ratio. Family member testing was considered uninformative when the TP53 PV was not detected in the tested family member. COFFEE ET AL. On the other hand, fibroblast testing can determine that a PV is likely germline (present in fibroblasts) or that the PV was acquired somatically (absent in fibroblasts). Fibroblast testing, with or without additional family testing, showed that a substantial proportion (32.6%) of apparent germline TP53 PVs were acquired somatically. Although this proportion might reflect some self-selection among individuals who elected to undergo   (Genovese et al., 2014;Jaiswal et al., 2014;Xie et al., 2014).
CHIP has been shown to be present in about 10% of individuals older than 65 years, and its prevalence increases with age (Genovese et al., 2014;Jaiswal et al., 2014;Xie et al., 2014).
Association of somatically acquired, low-allele-frequency (10-30%) TP53 PVs with aging has been seen previously in individuals tested using NGS hereditary cancer gene panels (Coffee et al., 2017). It follows that further expansion of a hematopoietic cell subclone carrying a TP53 PV could occur over time, reaching an NGS allele frequency that overlaps with those seen for heterozygous germline variants. The observation that subclones with TP53 PVs in mouse bone marrow chimeras expand over time, specifically after exposure to chemotherapy, further supports this hypothesis (Wong et al., 2015).
Although it is likely that CHIP underlies the acquisition of somatic A primary strength of the analysis is the large clinical data set that afforded the broad analysis and follow-up testing in individuals tested for hereditary cancer risk. Personal and family cancer history entered on the test request form by the ordering provider further enhances data analysis and interpretation but comes with a limitation: the possibility that the information is less than complete. For some tested individuals, the family cancer history might be more extensive than was reported and could bring the family cancer history phenotype in closer alignment with LFS diagnostic criteria. Cytotoxic cancer therapy is known to be associated with CHIP (Coombs et al., 2017;Gillis et al., 2017;Slavin et al., 2019;Takahashi et al., 2017) and should be considered by ordering providers as a potential contributor to acquisition of somatic TP53 PVs. Unfortunately, cancer treatment information was not available in this clinical testing data set; this is an important area for investigation using study cohorts that are designed to capture such information.
In summary, this study demonstrates that TP53 PVs with NGS allele frequencies consistent with those expected for a germline PV can be either inherited or acquired somatically. Therefore, clinical reporting of TP53 PVs identified by NGS should include the possibility of either germline or somatic origin and the corresponding clinical implications. For individuals who carry germline TP53 PVs, confirmation of the test result is essential to providing appropriate medical management. Determining that a TP53 PV was acquired somatically may help prevent unnecessary medical intervention for the proband and undue anxiety for family members who otherwise may think that they are at risk of carrying the PV. These findings highlight the need for a conservative approach to the COFFEE ET AL.

| 209
initial reporting of TP53-positive NGS test results, combined with investment in follow-up testing to deliver clinically accurate results.