Fragmentomics of cell-free DNA from targeted panels in genitourinary malignancies.
Bibliographic record
Abstract
199 Background: The detection of genomic alterations in cancer is critical for identifying clinically actionable alterations for treatment decisions. Tumor samples historically have been required, but obtaining tissue for molecular profiling is not always feasible, especially in the metastatic setting. The isolation and analysis of cell-free DNA (cfDNA) including circulating tumor DNA (ctDNA) via blood-based “liquid” biopsies offers non-invasive sampling of the tumor. Recently, fragmentation patterns of cfDNA (i.e. “fragmentomics”) have emerged as a method for inferring epigenomic and transcriptomic information. However, these analyses use whole-genome sequencing, which lacks the necessary depth to cost-effectively assess genomic alterations, limiting the application of these techniques clinically. Methods: We developed a novel cfDNA fragmentomics machine learning approach for standard targeted cancer gene panels in order to identify genitourinary (GU) and non-GU cancers from two independent metastatic cancer cohorts: a published cohort from GRAIL (prostate cancer, non-GU cancers, normal samples, N=198) and an expanded institutional cohort from the University of Wisconsin (prostate adenocarcinoma, neuroendocrine prostate cancer (NEPC), renal cell carcinoma (RCC), bladder cancer, non-GU cancers, normal samples, n = 431). Results: In the GRAIL cohort, 10-fold cross-validation AUCs were 0.987 for identifying prostate cancer, 1.00 for normal samples, and ranged from 0.922-0.958 for the non-GU cancers. In the UW cohort, 10-fold cross-validation AUCs were 0.950 for bladder cancer, 0.982 for prostate cancer, 0.993 for NEPC, 0.925 for RCC, and 0.980 for normal samples, and ranged from 0.874-0.954 for non-GU cancers. Our assay can sensitively detect and accurately distinguish GU and non-GU malignancies despite a median ctDNA fraction of only 0.076 in the GRAIL cohort, and a median ctDNA fraction of only 0.022 in the UW cohort. Conclusions: Our enhanced machine learning approach for examining fragmentomics in standard cancer gene cfDNA panels unlocks the potential for these established assays to be utilized to answer critical clinical questions beyond somatic variant identification. The excellent performance of our innovative approach even in samples with low ctDNA fractions suggest potential applications such as multi-cancer early detection and minimal residual disease monitoring. This framework could dramatically expand the potential of already existing clinically used assays and minimizes barriers for continued biomarker development.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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".