Back

Comprehensive baseline ctDNA characterization in two biomarker-selected Phase 1/2 studies using genomic and methylation profiling

Silverman, I. M.; Schonhoft, J. D.; Herzberg, B.; Yablonovitch, A.; Lagow, E.; Sethuraman, S.; Ulanet, D.; Yang, J.; Kim, I.; Basciano, P.; Cecchini, M.; Lee, E.; Lheureux, S.; Fontana, E.; Carneiro, B. A.; Reis-Filho, J. S.; Yap, T. A.; Zinda, M.; Rosen, E. Y.; Rimkunas, V.

2024-11-11 oncology
10.1101/2024.11.11.24316049 medRxiv
Show abstract

The development of DNA damage response (DDR)-directed therapies is a major area of clinical investigation, yet to date Poly (ADP-ribose) polymerase (PARP) inhibitors remain the only approved therapy in this space. Major challenges to DDR-targeted therapies in the post-PARPi era are the context dependency of DDR alterations and the presence of pre-existing resistance in this heavily pre-treated population. To that end, we used a contemporary platform to analyze pre-treatment circulating tumor DNA (ctDNA) samples from 173 patients enrolled onto two Phase 1/2 trials harboring pathogenic variants (PVs) in DDR genes. Baseline ctDNA analysis revealed a wealth of insights, including circulating tumor fraction estimation, impact of clonal hematopoiesis, PV allelic status, homologous recombination deficiency (HRD) signatures and presence of pre-existing resistance. HRD reversions were detected in 44% of evaluable patients and included large genomic rearrangements leading to deletion of whole or partial exons. We also discovered reversion of ATM in two patients previously treated with platinum chemotherapy, which has not previously been described. This study showcases the genomic complexity of DDR-altered tumors, revealed through baseline ctDNA profiling, an understanding of which is crucial for the future clinical development of novel DDR-directed therapies and combinations.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.