Recurrence risk stratification model for patients with stage I-III colorectal cancer based on clinicopathological and postoperative ctDNA predictors.
Jin Gu,Fuming Lei,Xiaodong Wang,Hui Chen,Wensheng Huang,Xiaomang He,Yuming Hong,Qingmin Zeng,Yanzhao Wang,Qingkun Gao,Zheping Yuan,Pengfei Niu,Dandan Huang,Zhaoya Gao,Changming Ding,Zhichao Zhai,Ke An,Shiqing Chen,Xiaochen Zhao,Yuezong Bai
DOI: https://doi.org/10.1200/jco.2022.40.16_suppl.e15511
IF: 45.3
2022-01-01
Journal of Clinical Oncology
Abstract:e15511 Background: The aim of this study was to propose a new kind of recurrence risk classification and further establish a prognostic model for resected stage I-III colorectal cancer (CRC). Methods: From 2017 to 2020, 142 patients diagnosed with stage I-III CRC at Peking University Shougang Hospital were recruited for this study. Tissues and paired white blood cells were analyzed with a 733-gene NGS panel, and peripheral plasma obtained 7-10 days after surgery was analyzed with a 127-gene ultra-deep target panel in a CLIA-certified laboratory. The relationship between the postoperative ctDNA status, clinicopathological features and recurrence-free survival (RFS) was identified by log-rank test and Cox regression analysis. Results: In total, 108 patients successfully completed both tissue and postoperative plasma NGS tests, of which 17 patients were postoperatively positive for ctDNA (tumor-informed, ≥1 somatic mutation). Postoperative ctDNA positivity was independently associated with poorer RFS (HR = 8.4, P < 0.0001). Postoperative recurrence rates were significantly higher in ctDNA-positive patients than ctDNA-negative patients (76.5% vs 16.5%, P < 0.0001). Univariate analysis showed that, in addition to ctDNA status, stage, vascular invasion, neural invasion, and preoperative CEA were all significantly associated with worse RFS (P < 0.05). Furthermore, in a COX regression model, the contribution of these clinicopathological factors to RFS has been adjusted. The model showed good discrimination [the concordance index (c-index): 0.811] of recurrence. Clinically high-risk groups according to clinicopathological risk model had significantly poorer RFS than low-risk groups (HR = 10.6, P < 0.001). In the clinicopathological high-risk group, ctDNA-positive patients can still further differentiate the higher risk of recurrence (HR = 3.7, P = 0.0012). There were similar trends in the clinicopathological low-risk group (HR = 6.1, P = 0.066). Combining the clinicopathological model and ctDNA, patients with either clinicopathological high-risk or ctDNA-positive had a significantly poorer RFS, compared with both clinicopathological low-risk and ctDNA-negative (HR = 12.9, P < 0.0001). Conclusions: Postoperative ctDNA can independently predict the risk of recurrence in stage I-III CRC. Combining clinicopathological factors and ctDNA showed superiority for recurrence risk stratification over either alone. This recurrence risk stratification model may serve as a useful clinical tool to help clinicians determine individualized treatment plans for stage I-III CRC patients. However, the model still needs further external cohort validation.