Improved Stratification for Risk of Early Metastases by Baseline Circulating Tumor Cell Counts for Locally Advanced Rectal Cancer in Neoadjuvant Setting: an Exploratory Analysis from a Phase III trial

Abstract

In this study, we recruited 119 patients with LARC receiving nCRT. 595 serial plasma samples were collected at d0, d15, d25 of radiotherapy as well before and 7 days post-surgery. The level of ctDNA was calculated by dynamic monitoring the mutant allele frequency of somatic mutations in plasma. Plasma and tissue samples were subjected to targeted-NGS using a 422 cancer-related genes panel. We followed up patients with concomitant CT until disease progression or death. Results: Serial samples' data analysis showed detection of pre-treatment mutations after completion of nCRT was significantly associated with worse disease-free survival (DFS) (P<0.05). Patients with ctDNA mutations clearance during nCRT significantly have a better DFS compare with patients with ctDNA non-clearance (PZ0.02). The presence of TP53 mutations is predictive of treatment outcome and shows a statistically significant worse DFS (2 years) (54.3% VS 87.0%, PZ0.0005). Predictive model based on support vector machine was developed for prediction of pCR achieving a mean AUC of 0.85 assessed by repeated cross validation. Through tracking clonal extinction, persistence and emergence, patients were grouped into four evolutionary subtypes with distinct TRG and DFS. Conclusion: Our data showed the prognostic value of ctDNA on DFS. In patients treated with nCRT, the presence of ctDNA after completion of nCRT was associated with an inferior DFS (PZ0.02). Dynamic monitoring of ctDNA can be used to predict TRG and prognosis in LARC patients receiving nCRT. ctDNA sequencing depicts the evolutionary trajectories of sensitive and resistant clones during nCRT in LARC. CtDNA could potentially be used to guide patient selection for nCRT.