New Paradigms in the Clinical Management of Li-Fraumeni Syndrome

Approximately 8.5%-16.2% of childhood cancers are associated with a pathogenic/likely pathogenic germline variant-a prevalence that is likely to rise with improvements in phenotype recognition, sequencing, and variant validation. One highly informative, classical hereditary cancer predisposition syndrome is Li-Fraumeni syndrome (LFS), associated with germline variants in the TP53 tumor suppressor gene, and a >90% cumulative lifetime cancer risk. In seeking to improve outcomes for young LFS patients, we must improve the specificity and sensitivity of existing cancer surveillance programs and explore how to complement early detection strategies with pharmacology-based risk-reduction interventions. Here, we describe novel precision screening technologies and clinical strategies for cancer risk reduction. In particular, we summarize the biomarkers for early diagnosis and risk stratification of LFS patients from birth, noninvasive and machine learning-based cancer screening, and drugs that have shown the potential to be repurposed for cancer prevention.

Empirical data drift detection experiments on real-world medical imaging data

While it is common to monitor deployed clinical artificial intelligence (AI) models for performance degradation, it is less common for the input data to be monitored for data drift - systemic changes to input distributions. However, when real-time evaluation may not be practical (eg., labeling costs) or when gold-labels are automatically generated, we argue that tracking data drift becomes a vital addition for AI deployments. In this work, we perform empirical experiments on real-world medical imaging to evaluate three data drift detection methods' ability to detect data drift caused (a) naturally (emergence of COVID-19 in X-rays) and (b) synthetically. We find that monitoring performance alone is not a good proxy for detecting data drift and that drift-detection heavily depends on sample size and patient features. Our work discusses the need and utility of data drift detection in various scenarios and highlights gaps in knowledge for the practical application of existing methods.

Early Cancer Detection in Li-Fraumeni Syndrome with Cell-Free DNA

People with Li-Fraumeni syndrome (LFS) harbor a germline pathogenic variant in the TP53 tumor suppressor gene, face a near 100% lifetime risk of cancer, and routinely undergo intensive surveillance protocols. Liquid biopsy has become an attractive tool for a range of clinical applications, including early cancer detection. Here, we provide a proof-of-principle for a multimodal liquid biopsy assay that integrates a targeted gene panel, shallow whole-genome, and cell-free methylated DNA immunoprecipitation sequencing for the early detection of cancer in a longitudinal cohort of 89 LFS patients. Multimodal analysis increased our detection rate in patients with an active cancer diagnosis over uni-modal analysis and was able to detect cancer-associated signal(s) in carriers prior to diagnosis with conventional screening (positive predictive value = 67.6%, negative predictive value = 96.5%). Although adoption of liquid biopsy into current surveillance will require further clinical validation, this study provides a framework for individuals with LFS.

SIGNIFICANCE

By utilizing an integrated cell-free DNA approach, liquid biopsy shows earlier detection of cancer in patients with LFS compared with current clinical surveillance methods such as imaging. Liquid biopsy provides improved accessibility and sensitivity, complementing current clinical surveillance methods to provide better care for these patients. See related commentary by Latham et al., p. 23. This article is featured in Selected Articles from This Issue, p. 5.

Multiple Germline Events Contribute to Cancer Development in Patients with Li-Fraumeni Syndrome

Li-Fraumeni syndrome (LFS) is an autosomal dominant cancer-predisposition disorder. Approximately 70% of individuals who fit the clinical definition of LFS harbor a pathogenic germline variant in the TP53 tumor suppressor gene. However, the remaining 30% of patients lack a TP53 variant and even among variant TP53 carriers, approximately 20% remain cancer-free. Understanding the variable cancer penetrance and phenotypic variability in LFS is critical to developing rational approaches to accurate, early tumor detection and risk-reduction strategies. We leveraged family-based whole-genome sequencing and DNA methylation to evaluate the germline genomes of a large, multi-institutional cohort of patients with LFS (n = 396) with variant (n = 374) or wildtype TP53 (n = 22). We identified alternative cancer-associated genetic aberrations in 8/14 wildtype TP53 carriers who developed cancer. Among variant TP53 carriers, 19/49 who developed cancer harbored a pathogenic variant in another cancer gene. Modifier variants in the WNT signaling pathway were associated with decreased cancer incidence. Furthermore, we leveraged the noncoding genome and methylome to identify inherited epimutations in genes including ASXL1, ETV6, and LEF1 that confer increased cancer risk. Using these epimutations, we built a machine learning model that can predict cancer risk in patients with LFS with an area under the receiver operator characteristic curve (AUROC) of 0.725 (0.633-0.810).

Significance

Our study clarifies the genomic basis for the phenotypic variability in LFS and highlights the immense benefits of expanding genetic and epigenetic testing of patients with LFS beyond TP53. More broadly, it necessitates the dissociation of hereditary cancer syndromes as single gene disorders and emphasizes the importance of understanding these diseases in a holistic manner as opposed to through the lens of a single gene.

Abstract 6545: Investigating secondary findings in a pediatric cancer cohort: preliminary findings

Purpose: An expected outcome following germline genome sequencing in oncology is the discovery of ‘secondary findings’ (SFs). SFs comprise pathogenic(P)/likely P (LP) germline variants in cancer genes not typically associated with the presenting cancer, in addition to germline variants of uncertain significance (VUS) to the patient’s cancer. Due to the rarity of childhood cancers and a dearth of studies analyzing SFs, many pediatric SFs are categorized as VUS without clinical interpretation. Interpreting SFs poses significant challenges: VUSs and other SFs are frequently not included in clinical molecular reports, and even when reported (often through research), their clinical utility and long-term impact on patient health are unclear. However, we know VUSs can have clinical importance because some VUSs, when investigated thoroughly, have been reclassified as pathogenic predictors of significant health conditions in children. We hypothesize that an in-depth characterization of the landscape of germline SFs/VUSs across a diverse pediatric cancer cohort will reveal new roles of these genes and mutations in pediatric cancers.

Methods: To explore germline SFs in pediatric cancer patients, we analyzed germline whole-genome sequencing (WGS) data for patients with rare, relapsed, refractory, and metastatic childhood cancers enrolled in the SickKids Cancer Sequencing Program (KiCS). We developed a custom analysis pipeline to identify germline single-nucleotide variants and indels deemed SFs, auto-classify their pathogenicity (ex. P, LP, or VUS) using CharGer, filter for PanCanAtlas-indicated cancer predisposition genes, and sort the remaining variants by cancer and non-cancer associations.

Results: The KiCS cohort (n = 511) encompassed over 133 different tumor types; the median age of participants was 14 years (SD = 10.27) and 55% of patients were male. Ongoing work in our lab will catalogue the frequency and distribution of SFs in KiCS and analyze germline variants by subgroup (gene, tumor subtype, stage, demographics, gene function). We will also compare SF prevalence in KiCS to the general population using the gnomAD dataset. Results from preliminary analyses of this cohort will be presented.

Significance: SFs/VUSs are under-utilized in cancer management. This work advances the holistic understanding of germline genomics in pediatric oncology and the roles of SFs in disease. Future studies will evaluate SFs by patient ancestry and validate cancer associations through the evaluation of allelic imbalance/loss of heterozygosity in matched tumor genomes.

Citation Format: Safa Majeed, Stephenie Prokopec, Brianne Laverty, Vallijah Subasri, Michael Taylor, Yvonne Bombard, Trevor Pugh, Adam Shlien, Anita Villani, David Malkin. Investigating secondary findings in a pediatric cancer cohort: preliminary findings. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6545.

Abstract 3364: Diagnosing Li-Fraumeni syndrome from the somatic genome

Background: Li-Fraumeni syndrome (LFS) is a hereditary cancer predisposition syndrome caused by germline mutations in the tumor suppressor gene TP53. LFS is estimated to occur in 1:1000 - 1:5000 people and is associated with a 80% lifetime cancer risk. This syndrome is diagnosed using familial cancer history and germline TP53 sequencing; however, clinical heterogeneity and variants of unknown significance limit diagnostic precision. Accurate diagnosis is imperative to implement surveillance for secondary malignancies and familial genetic testing.

Methods: We hypothesized that LFS cancers evolve uniquely from sporadic cancers, implying that the somatic genomes of LFS patients exhibit distinct characteristics that can infer the predisposition syndrome. To investigate this, we interrogated mutational signatures, TP53 copy number, TP53 loss of heterozygosity, ploidy, and the incidence of chromothripsis in tumor compared to normal samples (blood or fibroblast) of individuals with germline TP53 mutations (n=27), somatic TP53 mutations (n=17) and WT for TP53 (n=158). We created a random forest model with 10-fold cross validation to determine if somatic features could diagnose LFS.

Results: No signatures were significantly associated with LFS (Wilcoxon rank-sum test, Benjamin-Hochberg FDR correction). As previously reported in the literature, LFS compared to non-LFS cancers in our cohort were more likely to: be hyperdiploid (odds ratio (OR) = 11.83, FDR < 0.0001, Fisher exact test), have undergone TP53 loss of heterozygosity (OR = 23.15, FDR < 0.0001, Fisher exact test) and experience chromothripsis (OR = 7.76, FDR < 0.001, Fisher exact test).

The area under the receiver operating curve (AUROC) for our random forest model with 10-fold cross validation was 0.90, the area under the precision recall curve (AUPRC) was 0.59, the positive predictive value (PPV) was 0.70, the negative predictive value (NPV) was 0.93 and the F1-score was 0.52. This implies that the somatic genomic features are reliable indicators of this germline syndrome. We have obtained access to a future 50 LFS samples from the Pediatric Cancer Genome Project dataset, which we hope will improve our model’s performance.

Conclusion: We have developed a machine learning tool that uses somatic features to identify LFS, a germline cancer predisposition syndrome. As the importance of precision oncology becomes apparent, a tool to identify LFS patients from the somatic genome will facilitate early diagnosis. This will allow individuals to enter a surveillance program for early detection of secondary tumors, leading to improved outcomes.

Citation Format: Brianne Laverty, Vallijah Subasri, Nicholas Light, David Malkin. Diagnosing Li-Fraumeni syndrome from the somatic genome [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3364.

Abstract 1428: DNA methylation predicts early onset of primary tumor in patients with Li-Fraumeni syndrome

Background: Li-Fraumeni syndrome (LFS) is an autosomal dominant cancer predisposition syndrome. Approximately 80% of individuals with LFS harbor a germline TP53 pathogenic variant rendering them susceptible to a wide spectrum of early-onset malignancies. A comprehensive surveillance regimen termed the ‘Toronto Protocol’, has recently been adopted for early tumor detection, demonstrating significant improvement in survival among TP53 pathogenic variant carriers. However, the protocol’s “one-size-fits-all” approach fails to consider an individual patient's risk of cancer. We built a machine learning model that predicts early-onset of primary tumors in LFS by estimating the probability of cancer onset before the age of six years, leveraging patient peripheral blood leukocyte methylation profiles.

Methods: We built a gradient-boosted tree model to predict the probability of cancer onset before the age of six using methylation data from 288 TP53 pathogenic variant carriers. An external test set of 82 TP53 pathogenic variant carriers was used to validate our model. To increase the signal-to-noise ratio, methylation probes associated with TP53 status were retained and probes associated with aging were removed. In our study, we were primarily interested in minimizing the false negative rate (i.e. to reduce the number of patients who developed cancer before the age of six but were undetected by our algorithm.

Findings: We correctly predicted whether the first tumor will occur before the age of six with an accuracy of 79% in our external test set. Importantly, our model classified 90% of the patients that developed cancer prior to the age of six correctly. In addition, 81% of the individuals without cancer in the external test set were predicted correctly.

Interpretation: Our tool provides additional value to clinicians in stratifying patients into low- or high-risk groups of developing early-onset malignancies, and helps inform rational use of clinical surveillance tools for early cancer detection, with the ultimate aim to improve overall patient outcomes.

Citation Format: Vallijah Subasri, Benjamin Brew, Lauren Erdman, Tanya Guha, Jordan R. Hansford, Elizabeth Cairney, Carol Portwine, Christine Elser, Jonathan L. Finlay, Kim E. Nichols, Wendy Kohlmann, Noa Alon, Ana Novokmet, Ledia Brunga, Anita Villani, Kelvin C. de Andrade, Payal P. Khincha, Sharon A. Savage, Joshua D. Schiffman, David Malkin, Anna Goldenberg. DNA methylation predicts early onset of primary tumor in patients with Li-Fraumeni syndrome [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1428.

Abstract 2114: Horvath clock as a predictor of cancer risk in LFS patients

Li-Fraumeni syndrome (LFS) is an autosomal dominantly inherited cancer predisposition syndrome associated with germline mutations of the TP53 tumor suppressor gene. TP53 mutation carriers are susceptible to a wide range of cancers that occur at strikingly earlier age of onset than their sporadic counterparts. The lifetime cancer risk in TP53 mutation carriers is estimated to be ~73% for males and approaching 100% in females. Although improved survival outcomes have been demonstrated for carriers undergoing intense clinical surveillance, there is continued interest in identifying new environmental, genetic, and epigenetic risk factors that could improve our ability to predict disease onset and outcome. A number of studies have demonstrated age-associated DNA methylation (DNAm) changes at specific CG dinucleotides and that these changes can be combined into epigenetic age predictors to estimate chronological age. Deviation of chronological and predicted age have been associated with age-associated illnesses such as metabolic disease and cancer. For a given chronological age, older epigenetic age is presumed to indicate poorer health.

An epigenetic profile defining the DNA methylation age (DNAm age) of an individual has been suggested to be a biomarker of aging, and thus possibly providing a tool for assessment of health and mortality. Our goal was to test whether DNAm age could be a possible predictor of cancer risk in LFS patients. We applied the DNA age calculator (http://dnamage.genetics.ucla.edu/) (Horvath 2013) to DNA methylation profiles derived from lymphocytes extracted from 157 LFS patients' blood samples using the Illumina HumanMethylation450 BeadChip. While a correlation of DNAm age and actual age was observed in both ‘normal' and LFS patients, the latter showed significant deviations (differences between DNAm age and chronological age). Moreover, the extent of deviations seems selectively associated with two distinct age groups (0-5 years and 20-50 years). Remarkably, this bimodal DNAm age profile shows striking resemblance to the epidemiologically characterized age dependency of LFS cancers (Amadou, 2018). Individuals with germline mutant or germline wild-type TP53 and no cancer showed no epigenetic age acceleration whereas individuals who were carriers of mutant TP53 who developed cancer showed accelerated epigenetic aging. Our preliminary results suggest that DNAm age is a dynamic, real-time correlate of patient-specific cancer risk in LFS. Further, the age dependent deviations suggest that the cancer risk profiles derived from Horvath signatures are dynamic and reflect the changes in cancer risk throughout an individual's lifetime, and could be used as a predictor of cancer onset in TP53 mutation carriers.

Citation Format: Malgorzata Pienkowska, Nardin Samuel, Sanaa Choufani, Vallijah Subasri, Nish Patel, Rosanna Weksberg, Ran Kafri, David Malkin. Horvath clock as a predictor of cancer risk in LFS patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2114.

Pathogenic Germline Variants in Cancer Susceptibility Genes in Children and Young Adults With Rhabdomyosarcoma

Rhabdomyosarcoma (RMS) is the most common pediatric soft-tissue sarcoma and accounts for 3% of all pediatric cancer. In this study, we investigated germline sequence and structural variation in a broad set of genes in two large, independent RMS cohorts.

MATERIALS AND METHODS

Genome sequencing of the discovery cohort (n = 273) and exome sequencing of the secondary cohort (n = 121) were conducted on germline DNA. Analyses were performed on 130 cancer susceptibility genes (CSG). Pathogenic or likely pathogenic (P/LP) variants were predicted using the American College of Medical Genetics and Genomics (ACMG) criteria. Structural variation and survival analyses were performed on the discovery cohort.

RESULTS

We found that 6.6%-7.7% of patients with RMS harbored P/LP variants in dominant-acting CSG. An additional approximately 1% have structural variants (ATM, CDKN1C) in CSGs. CSG variants did not influence survival, although there was a significant correlation with an earlier age of tumor onset. There was a nonsignificant excess of P/LP variants in dominant inheritance genes in the patients with FOXO1 fusion-negative RMS patients versus the patients with FOXO1 fusion-positive RMS. We identified pathogenic germline variants in CSGs previously (TP53, NF1, DICER1, mismatch repair genes), rarely (BRCA2, CBL, CHEK2, SMARCA4), or never (FGFR4) reported in RMS. Numerous genes (TP53, BRCA2, mismatch repair) were on the ACMG Secondary Findings 2.0 list.

CONCLUSION

In two cohorts of patients with RMS, we identified pathogenic germline variants for which gene-specific therapies and surveillance guidelines may be beneficial. In families with a proband with an RMS-risk P/LP variant, genetic counseling and cascade testing should be considered, especially for ACMG Secondary Findings genes and/or with gene-specific surveillance guidelines.

Abstract 3666: The genomic landscape and clonal evolution of tumours arising in TP53 mutation carriers

Li-Fraumeni syndrome (LFS) is a familial cancer predisposition syndrome (CPS) caused by germline mutations in TP53, associated with a high frequency of sarcomas, breast cancers, adrenocortical carcinomas and CNS tumours. Recently, our group and others have identified distinct mutational signatures, defined by the spectrum and context of somatic mutations, in tumours arising in individuals with another CPS, constitutional mismatch repair deficiency (CMMRD). CMMRD tumours frequently harbor somatic mutations disrupting the proofreading function of the DNA polymerases POLE and POLD1, which consequently leads to an ultra-hypermutant cancer genome, with early MMR mutational signatures and late POLE/POLD1 mutational signatures. This ultrahypermutant tumor phenotype is essentially diagnostic for the syndrome and provides a rational for immune checkpoint inhibitors which have shown success in this context. Based on our results in CMMRD tumors we hypothesize that LFS tumours might likewise harbor distinct mutational events and/or evolutionary dynamics from sporadic tumours of the same histiotype. Although several cancer genomics landscape studies have included a handful of tumours from LFS patients, to our knowledge no study to date has attempted to comprehensively characterize the cancer genomes of LFS patients. To investigate the somatic mutational events driving tumourigenesis in LFS we performed whole-genome sequencing (WGS) analysis of 22 tumours derived from patients with pathogenic germline TP53 mutations. Tumours from germline TP53 wildtype patients were analyzed by the same methods to serve as a control data set. For each tumour sample, where possible, we performed WGS on multiple spatially distinct micro-dissected tumour regions in order to reconstruct the evolutionary history of each cancer. Somatic variant calling was performed at high sensitivity using in silico reconstructed high-depth bulk WGS (80-120X coverage), with somatic mutations, structural variants, copy number alterations, mutational signatures and subclones identified using MuTect2, delly, battenberg, SigProfiler and phyloWGS respectively. High confidence variants were identified using in-house designed filtering pipelines. The resulting analyses reveals the life history of LFS cancers is marked by a high frequency of early catastrophic genomic rearrangement events, a diverse range of somatic driver events and in at least some cases marked intratumoural spatial heterogeneity of CNVs and SNVs.

Citation Format: Nicholas Light, Matthew Zatzman, Nathaniel Anderson, Vallijah Subasri, Mehdi Layeghifard, Ana Novokmet, James Tran, Richard de Borja, Fabio Fuligni, Joshua Schiffman, David Malkin, Adam Shlien. The genomic landscape and clonal evolution of tumours arising in TP53 mutation carriers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3666.