Body composition analysis for predicting response to first-line bevacizumab in patients with unresectable colorectal liver metastases: A multicenter study.

OBJECTIVES

To investigate the association between baseline body composition and objective response to first-line bevacizumab-based chemotherapy in patients with unresectable colorectal liver metastases (CRLM), and to develop a model for response prediction in this treatment setting.

METHODS

This multicenter retrospective cohort included 185 patients with newly diagnosed unresectable CRLM from three hospitals who received first-line bevacizumab therapy. Pre-treatment body composition parameters were quantified at the L3 level on computed tomography and compared between objective response and non-response groups. Sex-specific median cutoffs were applied to evaluate associations with response. Subgroup analyses explored interactions with tumor burden. Logistic regression identified independent predictors. Model performance was assessed using receiver operating characteristic, calibration, and decision curve analyses.

RESULTS

The objective response group showed significantly higher visceral fat area, skeletal muscle area, and skeletal muscle density, but lower intermuscular fat area. Multivariable analysis confirmed higher visceral fat area and skeletal muscle area, and lower intermuscular fat area as independent predictors. The combined model incorporating age, sex, visceral fat area, skeletal muscle area, and intermuscular fat area demonstrated excellent discrimination, with areas under the curve of 0.883 in the training cohort and 0.870 in the external validation cohort.

CONCLUSION

Baseline CT-derived body composition is significantly associated with objective response to first-line bevacizumab in unresectable CRLM. The proposed model may serve as a practical, noninvasive tool for predicting objective response.