Circulating Tumor DNA as a Novel Prognostic Biomarker of Therapeutic Efficacy in Patients with Unresectable Colorectal Liver Metastases Treated with Hepatic Arterial Infusion
- PMID: 40634807
- DOI: 10.1245/s10434-025-17752-5
Circulating Tumor DNA as a Novel Prognostic Biomarker of Therapeutic Efficacy in Patients with Unresectable Colorectal Liver Metastases Treated with Hepatic Arterial Infusion
Abstract
Background: Hepatic arterial infusion (HAI) chemotherapy is an expanding treatment for patients with unresectable colorectal liver metastases (uCRLM). This study examined circulating tumor DNA (ctDNA) as a novel biomarker to prognosticate oncologic outcomes after HAI.
Methods: The study examined HAI outcomes for patients with uCRLM at Duke University from February 2022-December 2024. Signatera was used to determine circulating tumor DNA (ctDNA) quantity as mean tumor molecules per mL (MTM/mL). Survival outcomes were calculated from the date of HAI initiation and compared using Kaplan-Meier and Cox proportional hazards methods.
Results: In this study, 37 patients had ctDNA collected with 20 months of median follow-up time.. The patients with pre-HAI ctDNA ≤ 100 MTM/mL had a significantly better overall survival (OS) (hazard ratio [HR] of 5.5 (95% confidence interval [CI] 1.3-22.7; p = 0.05). After 3 months of HAI, the patients whose ctDNA decreased by less than 75% had significantly worse progression-free survival (PFS) (< 75% vs 75-99% vs 100% decrease; median PFS, 3 vs 10 vs 11 months; p < 0.01, log rank test) and OS (< 75% vs 75-99% vs 100% change; median, 10 months vs not reached [NR] vs NR; p = 0.01, log rank test). Although the patients with initial carcinoembryonic antigen (CEA) ≤ 10 ng/mL had significantly better OS than those with a CEA > 10 ng/mL (HR 2.5; 95% CI 1-6; p = 0.04), the magnitude of CEA change after 3 months of HAI was not associated with OS or PFS.
Conclusions: Preoperative ctDNA and ctDNA dynamics were associated with oncologic outcomes for patients who had uCRLM treated with HAI. These hypothesis-generating findings require validation before quantitative ctDNA is incorporated into patient selection algorithms and ctDNA dynamics are used to guide HAI therapy for uCLRM.
© 2025. Society of Surgical Oncology.
Conflict of interest statement
Disclosure: John H. Strickler has a consultant or advisory role at Abbvie, Agenus, Astellas, AstraZeneca, Bayer, Beigene, Daiichi-Sankyo, Eli Lilly, GSK, Johnson and Johnson, Jazz Pharmaceuticals, Merck, Natera, Pfizer, Roche/Genentech, Regeneron, Sanofi, Taiho, Takeda, Xilio Therapeutics, and Triumvira Immunologics (stock options). He has received research funding or contracted research from Abbvie, Amgen, AStar D3, Bayer, Beigene, Curegenix, Daiichi-Sankyo, Eli Lilly, Erasca, GSK, Leap Therapeutics, Novartis, Pfizer, Quanta Therapeutics, Revolution Medicines, and Roche/Genentech.
References
-
- Morris VK, et al. Treatment of metastatic colorectal cancer: ASCO guideline. JCO. 2023;41:678–700. - DOI
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Medical
