关键词: 结直肠肿瘤, 肿瘤转移, 基因检测, 超声灰度值

Abstract: Objective To explore the key factors influencing progression-free survival (PFS) and overall survival (OS) after radical ablation therapy for metachronous liver metastases from colorectal cancer(CRC) from multiple dimensions, including genetic testing, clinical characteristics, and ultrasound indicators, to identify the primary prognostic risk factors for ablation treatment of colorectal liver metastases (CRLM) and to establish a predictive model. 
Methods A retrospective, multicenter analysis was conducted, including 298 eligible patients with CRLM. These patients were randomly divided into a model training cohort (n=208) and a validation cohort (n=90). Clinical data, clinical characteristics of tumor, imaging, and routine laboratory indicators were collected for all patients. Univariate and multivariate Cox regression analyses were used to identify the main factors affecting prognosis in patients with CRLM. Based on the Cox model′s risk results, independent factors were assigned scores, and risk prediction nomograms were constructed to predict 1-year PFS and 3-year OS, followed by curve validation. 
Results Univariate analysis identified patient gender, age, ultrasound grayscale values ≥100, and targeted therapy as factors associated with poorer PFS (P＜0.05). Multivariate analysis confirmed that male gender, age ≥60, value of gray-scale ultrasound ≥100, and targeted therapy were influencing factors of poorer PFS (P＜0.05). Similarly, univariate analysis showed that BRAF gene mutation and value of gray-scale ultrasound ≥100 were associated with poorer OS (P＜0.05). Multivariate analysis confirmed BRAF gene mutation and value of gray-scale ultrasound ≥100 as significant factors influencing OS (P＜0.05). Based on the model analysis of the training set data, several highly correlated variables were included in the prognostic model for radical ablation. Nomograms were developed to predict 1-year PFS and 3-year OS, where the overall probability of 1-year PFS was calculated as follows: 27.5×(female)+ 31×(age＜60)+100×(value of gray-scale ultrasound＜100)+ 22.5×(no targeted therapy). The 3-year OS probability was calculated as: 25×(BRAF wild-type) + 100×(value of gray-scale ultrasound＜100). The area under the ROC curve (AUC) of the 1-year PFS prediction model was 0.88 (95%CI: 0.76-1.00) in the training cohort and 0.81 (95%CI: 0.71-0.90) in the validation cohort. For the 3-year OS prediction model, the AUC value was 0.80 (95%CI: 0.66-0.94) in the training cohort and 0.54 (95%CI: 0.36-0.72) in the validation cohort. 
Conclusion Value of gray-scale ultrasound, genetic mutations, and treatment modalities are influencing factors of PFS and OS in patients with metachronous liver metastases from CRC. By integrating these prognostic risk factors into a comprehensive model, it may effectively predict the medium-and long-term outcomes for these patients. 

Key words: colorectal neoplasms, neoplasm metastasis, genetic testing, value of grayscale ultrasound