Discovery of a novel non-negative matrix factorization (NMF)-based homologous recombination deficiency (HRD) score and subsequent exploration in TALAPRO-2 (TP-2), a phase 3 study of talazoparib (TALA) + enzalutamide (ENZA) vs placebo (PBO) + ENZA as first-line treatment in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC).
Bibliographic record
Abstract
5021 Background: TP-2 (NCT03395197) demonstrated significantly improved radiographic progression-free survival (rPFS) in pts with mCRPC who received TALA + ENZA (n=402) vs PBO + ENZA (n=403). We discovered a novel NMF-based HRD score and used it to explore potential associations of HRD with efficacy in TP-2. Methods: The Cancer Genome Atlas Prostate Adenocarcinoma (TCGA-PRAD) dataset was used to train a novel NMF-based HRD predictive score incorporating gene expression (RNA seq) and homologous recombination repair (HRR)12 genomic features (mutations possibly or probably damaging by PolyPhen, deleterious by SIFT, or shallow/deep deletion; HRR12 genes: BRCA1, BRCA2, PALB2, ATM, ATR, CHEK2, FANCA, RAD51C, NBN, MLH1, MRE11A, CDK12). A previously published composite HRD score incorporating genomic loss of heterozygosity, large scale transitions, and telomeric allelic imbalances (Knijnenburg et al. Cell Rep. 2018;23:239-254.e6) was used as the “original” HRD reference score. HRR12 genomic features in TCGA-PRAD were associated with higher original and predicted HRD scores. Two datasets were used to generate the HRD score for TP-2: a FoundationOneLiquid CDx (F1LCDx) dataset (Azad et al. ASCO 2023, #5056) of prospectively collected/retrospectively analyzed plasma samples (n=681) and a tumor transcriptomic dataset generated via HTG’s Oncology Biomarker Panel (with 10 additional genes implicated in PARPi sensitivity; n=304). This NMF-based HRD score was then applied to the evaluable TP-2 safety population (n=285). Predicted HRD scores for TP-2 were categorized as high (≥0.46 [median]) or low (
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 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".