O11: Mutational spectra of >1000 pediatric acute leukemias and their clinical utility in prognosis, risk stratification, and precision medicine therapeutics

Introduction: Cell-free DNA (cfDNA) fragments are released by a variety of types of cells in the body into the blood (eg, healthy, placental, and cancer cells).Genetic and epigenetic analysis of cfDNA can be used to differentiate cfDNA tissue-of-origin for the early detection of cancer but has not been applied to BRCA1/2 heterozygotes.People who harbor a BRCA1/2 pathogenic variant are at high risk of developing breast, ovarian, prostate, pancreas cancer, and other types of cancer.Early detection in BRCA1/2 carriers includes intensive surveillance with breast imaging and prostate-specific antigen (PSA) test for prostate cancer.However, many BRCA1/2-related cancers do not have effective early detection surveillance (eg, ovarian and pancreas cancer).Prophylactic mastectomy and salpingo-oophorectomy are preventative options but have associated risks.Developing screening tools that are less invasive, more accessible, and can screen for multiple cancer types in BRCA1/2 patients is crucial.To address this, we have developed a cell-free DNA (cfDNA) blood test to predict cancer development in BRCA1/2 carriers for early detection surveillance.Methods: We collected 181 blood samples from 93 patients with a germline pathogenic or likely pathogenic variant in BRCA1 or BRCA2.We collected 77 samples from patients with active cancer and 104 samples from patients without cancer.We performed plasma Targeted Gene Panel Sequencing (TS) (n = 181) and shallow Genome Sequencing (pGS) (n = 177) to detect somatic variants, cancer-associated copy number changes, and fragmentation profiles in cfDNA.We additionally performed shallow pGS on 26 healthy individuals.Results: Using somatic variant identification, we were able to identify mutations in 21/77 samples from patients with cancer (27%) and 21/104 from patients without cancer (20%).We further used pGS to investigate copy number alterations (CNA) and tumor fraction (TF) prediction.In CNA analysis, 6 samples had a detectable TF (TF >= 10%).Of the samples that underwent both TS and pGS (samples = 162, patients = 73), 44 samples (from 32 patients) had a detectable TF or somatic mutation (44/162; 27.1% detection rate) from 21 and 23 cancer positive and negative samples, respectively.Fragment ratio (FR) analysis revealed significant differences in correlation to healthy individuals (FR to healthy correlation < 0.8) in both cancer positive (25/79; 31.6%detection rate) and negative (11/98; 11.2% detection rate) cases.We are currently exploring further fragmentomic features, including the use of fragment length, end motif, and nucleosome accessibility, to improve the cancer detection rate of our cfDNA assay.Conclusion: cfDNA sequencing is a promising tool to improve early cancer detection and bridge the gaps in current screening programs for BRCA1/2 carriers.Our work demonstrates that multimodal analysis of cfDNA increased the cancer detection rate over unimodal analysis.We are continuing to explore additional features of cfDNA that are most predictive of cancer, with the goal of incorporating them into our multimodal analysis.