Establishment and Validation of a Novel Promising Recurrence Prediction Molecular Subtype Classification for Patients with Stage II Colorectal Cancer

Objective: Stage II colorectal cancer (CRC) manifests various treatment responses and outcomes. To assess the recurrence risk, we established a molecular subtype classification for stage II CRC. Methods: We developed a novel molecular subtyping clustering algorithm CRC-MSC for determining molecular subtypes in stage II CRC using only whole exome sequencing (WES). It utilized a specified distance estimate, calculated according to the shared alterations among samples. CRC-MSC was qualified for classification and risk stratification on samples from National Cancer Center (NCC) and The Cancer Genome Atlas (TCGA). To facilitate application to individual management decision, this algorithm was simplified by a decision tree classifier (CRC-MDTC). Phylogenetic analysis was conducted to clarify the mutated genes constituting the decision tree. Results: Using CRC-MSC successfully subtyped 60 stage II CRC samples from NCC into three groups, favorable prognosis (S-I), unfavorable prognosis (S-II) and intermediate (S-III) groups, which was verified in TCGA dataset. These subtypes feature unique molecular, clinicopathological phenotypes and clinical outcomes, especially immune traits.A simplified version of decision tree classifier (CRC-MDTC), TMB, PRKDA and KRAS alterations are three top classification indications, could also reliably categorize samples into three groups. It was validated in an independent cohort with an accuracy of 74.7%. CRC-MSC and CRC-MDTC were available at https://bioinfo.org/crcmsc. Conclusions: In addition to the initial risk evaluation, we developed a framework for the classification of stage II CRC subtypes based on genome profiling, which may shed new light on the management of patients with stage II CRC, especially with regarded to individualized chemotherapy and immunotherapy. Funding: This work was supported by National Program Project for Precision Medicine in National Research and Development Plan of China (2016YFC0905300), and CAMS Innovation Fund for Medical Sciences (CIFMS) (2019-I2M-2-002). Declaration of Interest: None to declare. Ethical Approval: The ethical approval of this study was granted by the Institutional Review Board of Chinese Academy of Medical Sciences (CHCAMS, Beijing, China)/National Cancer Center (NCC, Beijing, China).