Abstract 1639: Predictive modeling of cancer-type in Li-Fraumeni syndrome
Bibliographic record
Abstract
Abstract Objective: Li-Fraumeni syndrome (LFS) is an autosomal dominant cancer predisposition syndrome associated with a germline TP53 mutation. Individuals with LFS are prone to developing a wide spectrum of tumors. This heterogeneity in tumor type makes it difficult to provide patient specific surveillance protocols. As a result, there is an immediate need to develop robust, evidence-based stratification strategies to tailor surveillance protocols to an individual patient’s risk of cancer. Germline TP53 mutations themselves do not explain specific cancer phenotypes in LFS, nor the collective clinical heterogeneity. Hence, it is important to consider genomic level data in the development of predictive algorithms in these patients. The proposed study is two-fold—to identify germline modifiers predictive of tumor type in LFS, and to use these modifiers to develop a predictive model for the early detection of cancer type. Methods: This study consists of blood-derived DNA methylation and whole genome sequencing (WGS) from a cohort of LFS patients (n=134); a subset of this cohort was held out as a test set. The first objective was to identify germline modifiers: single nucleotide variants (SNVs), insertions and deletions (indels), structural variants (SVs), copy number variation (CNV) and differentially methylated regions (DMRs) predictive of cancer type. Blood DNA methylation was generated using Illumina HumanMethylation450 BeadChip array. The top statistically significant DMRs were identified between cancer types by performing pairwise comparisons using a linear model. CNV, SNVs, indels, and SVs were detected from WGS using benchmarked tools: CNVnator, ERDS, GATK, and Delly. Custom filtering pipelines, and a curated list of cancer genes were established to determine high quality, biologically relevant modifiers. The second objective was to develop a model using the identified modifiers to predict cancer type in LFS patients. A generalized linear model was implemented using elastic net regularization to estimate the probability of getting a particular cancer type. Results: The classifier can accurately identify the cancer type of all the individuals in the test set. The model can determine with > 90% accuracy whether an individual has cancer and can differentiate between cancer types among affected individuals with 42%-92% probability. Pathway analysis of the predictors highlight the TGF-β signaling pathway as a possible therapeutic target for personalized treatment in LFS. Conclusion: This project is the first comprehensive molecular analysis of LFS and uses the largest LFS cohort to-date. It illustrates the contribution of genetic changes in LFS beyond TP53 and highlights the existence of molecular differences within LFS patients that contribute to phenotypic differences. Ultimately, this study will allow for the early detection of tumor-onset in LFS patients to assist clinicians in developing personalized surveillance protocols. Citation Format: Vallijah Subasri, Nicholas Light, Benjamin Brew, Nathaniel Anderson, Adam Shlien, Anna Goldenberg, David Malkin. Predictive modeling of cancer-type in Li-Fraumeni syndrome [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 1639.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".