The mutational factors influencing the therapeutic response and prognosis of stage IV CRC patients undergoing simultaneous resection of primary and liver metastatic lesions and subsequent adjuvant therapy---A pilot study on the prognosis of stage IV CRC.

[BACKGROUND] Patients with stage IV colorectal cancer (CRC) with liver metastasis represent a specific group that can be treated with surgery. However, the influence of genomic alterations on the therapeutic response and prognosis was not clear in CRC patients underwent simultaneous surgery of primary and metastatic lesions.

[METHODS] Fifty-two patients underwent simultaneous surgery on primary and metastatic lesions were retrospectively recruited. The mutational landscape of primary lesion was established by whole-exome sequencing(WES). Non-parametric test, Fisher's exact test, multivariate analyses, Kaplan-Meier analyses were performed to identify risk factors for response and prognosis, and a Nomogram model was established. Analyses were performed and figures were plotted using the Graphpad PRISM 5.0 and the R software.

[RESULTS] Several top mutated genes were identified from the mutational landscape of primary lesions, including APC, TP53, KRAS and TTN, and many co-mutations, mutually exclusive mutations and aberrant functions or pathways were revealed. KRAS(P = 0.047) and TTN(P < 0.001) exhibited significant differences in tumor mutational burden(TMB) between mutant and wild-type groups. TP53(P = 0.045), MUC12(P = 0.012) and CEL(P = 0.032) mutational status significantly stratified the patient therapeutic response, in which MUT12 was an independent risk factor(P = 0.02). CRC location(P = 0.014), patient therapeutic response(P < 0.001), and the mutational status of ANKRD20A4(P = 0.006), EVC(P = 0.05), FHOD3(P = 0.05), MYO15A(P = 0.008) and POTEE(P < 0.001) showed significant stratification on patient prognosis, in which cancer location, response to therapy and ANKRD20A4 and EVC mutational status were independent risk factors. These factors were used to establish a Nomogram model to predict the individual patient prognosis. Internal and external validation verified the effectiveness of the model.

[CONCLUSIONS] Independent risk factors including cancer location, response to therapy and ANKRD20A4 and EVC mutational status were identified and were used in the establishment of a Nomogram model for patient prognosis prediction.