The prognostic value of CT-measured body composition combined with radiomics in predicting the survival of patients with resectable colon cancer.
1/5 보강
[OBJECTIVE] To explore the prognostic value of body compositions and radiomics in patients with resectable colon cancer, and to develop and validate a clinical-radiomics model for predicting the posto
- p-value P = 0.035
- p-value P < 0.001
- HR 0.398
APA
Zhi X, Nie T, et al. (2026). The prognostic value of CT-measured body composition combined with radiomics in predicting the survival of patients with resectable colon cancer.. La Radiologia medica, 131(3), 350-362. https://doi.org/10.1007/s11547-025-02135-3
MLA
Zhi X, et al.. "The prognostic value of CT-measured body composition combined with radiomics in predicting the survival of patients with resectable colon cancer.." La Radiologia medica, vol. 131, no. 3, 2026, pp. 350-362.
PMID
41201567
Abstract
[OBJECTIVE] To explore the prognostic value of body compositions and radiomics in patients with resectable colon cancer, and to develop and validate a clinical-radiomics model for predicting the postoperative overall survival of patients with resectable colon cancer.
[METHODS] This study included 296 patients (43 months of median follow-up) with resectable colon cancer. Non-contrast CT images were used to quantify the body composition at the level of the third lumbar vertebra. Radiomics features were extracted from portal venous-phase CT scans. The recursive feature elimination and the least absolute shrinkage and selection operator regression were used for feature selection and construction of radiomic signatures. Univariate and multivariate Cox regression analysis were used to identify body composition. Combined with radiomics features, clinical-radiomics prediction model was constructed and plotted by nomogram, with performance metrics including the area under the receiver operating characteristic curve, calibration curves, decision curve analysis, and integrated discrimination improvement index.
[RESULT] Low skeletal muscle density (HR = 0.398, 95%CI = 0.168-0.939, P = 0.035) and low visceral fat area (HR = 0.238, 95%CI = 0.108-0.524, P < 0.001) were significantly associated with poor OS. The integrated clinical-radiomics model achieved C-index of 0.802 and 0.786 in the training and test cohorts, with superior 3-year OS AUC values of 0.804 and 0.828. Furthermore, clinical-radiomics model has a significant improvement in performance compared with radiomics model (IDI: 23.2%, P < 0.001) and clinical model (IDI:5.2%, P = 0.008).
[CONCLUSION] Nomogram combining body composition and tumor radiomics features can help predict the long-term prognosis of patients with resectable colon cancer and may serve as an effective tool to promote individualized treatment.
[METHODS] This study included 296 patients (43 months of median follow-up) with resectable colon cancer. Non-contrast CT images were used to quantify the body composition at the level of the third lumbar vertebra. Radiomics features were extracted from portal venous-phase CT scans. The recursive feature elimination and the least absolute shrinkage and selection operator regression were used for feature selection and construction of radiomic signatures. Univariate and multivariate Cox regression analysis were used to identify body composition. Combined with radiomics features, clinical-radiomics prediction model was constructed and plotted by nomogram, with performance metrics including the area under the receiver operating characteristic curve, calibration curves, decision curve analysis, and integrated discrimination improvement index.
[RESULT] Low skeletal muscle density (HR = 0.398, 95%CI = 0.168-0.939, P = 0.035) and low visceral fat area (HR = 0.238, 95%CI = 0.108-0.524, P < 0.001) were significantly associated with poor OS. The integrated clinical-radiomics model achieved C-index of 0.802 and 0.786 in the training and test cohorts, with superior 3-year OS AUC values of 0.804 and 0.828. Furthermore, clinical-radiomics model has a significant improvement in performance compared with radiomics model (IDI: 23.2%, P < 0.001) and clinical model (IDI:5.2%, P = 0.008).
[CONCLUSION] Nomogram combining body composition and tumor radiomics features can help predict the long-term prognosis of patients with resectable colon cancer and may serve as an effective tool to promote individualized treatment.