Prognostic value of postoperative circulating tumor DNA and tumor markers in resected pancreatic adenocarcinoma (PAAD): An interim analysis of a prospective observational study.

4192 Background: Despite the benefits of postoperative adjuvant chemotherapy, recurrence rates of PAAD remain as high as 60% within the first year, underscoring the need for improved strategies to optimize treatment plans. This prospective observational study aims to evaluate the potential of postoperative ctDNA-based minimal residual disease (MRD) as an early predictor of disease relapse in resected PAAD using next-generation sequencing. Methods: Pathologically confirmed stage I-III PAAD patients who underwent surgical resection and carried KRAS mutations were enrolled. Eligible criteria included negative surgical margins and no metastasis prior to adjuvant therapy. Patients who did not receive adjuvant therapy were excluded. Tumor tissue samples collected during surgery were analyzed using a 769-gene NGS panel, while plasma samples were obtained at landmark (4–8 weeks post-operation and pre-therapy), 12-, 24-, 36-, and 48-weeks post-operation, and 2 weeks post-adjuvant therapy. Plasma samples were assessed for MRD using MinerVa (Genecast Biotechnology). Concurrently, tumor biomarkers such as CA19-9 were measured. The primary endpoint was overall survival (OS), while the secondary endpoint was disease-free survival (DFS). The trial is designed for a total follow-up period of three years. Results: As of November 27, 2024, a total of 133 patients underwent MRD analysis. KRAS mutations were distributed as follows: p.G12D in 55%, p.G12V in 27%, p.G12R in 13% and other subtypes in 5%. Nine patients were excluded due to loss of follow-up, and 12 were excluded for less than six months of follow-up, leaving 112 patients for this interim analysis. MRD positivity rates were 16%, 9%, 10%, 15%, 11%, and 11% at landmark, 12, 24, 36, 48 weeks, and 2 weeks post-therapy, respectively. Landmark MRD positivity was a significant predictor of disease recurrence (HR = 2.39, 95%CI:1.14-5.01, p = 0.017). Combining landmark MRD with biomarker CA19-9 improved prognostic accuracy (HR = 2.70, 95%CI:1.4-5.5, p = 0.002). Patients with longitudinal MRD-positive (detected at any time point excluding landmark or relapse) had significant worse DFS (HR = 3.13, p = 0.001) and worse OS (HR = 0.73, p = 0.001) compared to those with negative MRD. Analysis of MRD status at landmark and week 36 revealed that patients transitioning from MRD-positive to MRD-negative (clearance) had significantly better DFS compared to those with persistent MRD positivity (p = 0.023). Notably, no patients deceased from the MRD clearance group. Conclusions: These findings underscore the clinical utility of integrating landmark and longitudinal MRD assessments with tumor markers for comprehensive risk stratification and prognostication in resected PAAD patients. This approach could potentially guide personalized treatment strategies and improve patient outcomes. Clinical trial information: NCT05479708 .