Article: Deep learning-based assessment of body composition and liver tumour burden for survival modelling in advanced colorectal cancer
Personalised therapy is a challenge in advanced colorectal cancer care. Much research has been carried out on prognostic and predictive markers of this disease, and a strong correlation was found between sarcopenia and survival in such patients. Currently, selecting personalised strategies for patients is based on very few parameters, not making sufficient use of all available clinical information. Therefore, this paper suggests that it is possible to use body composition and liver tumour burden through automated extraction from CT images. Such automated segmentation would allow one to extract prognostic parameters from the routine imaging data which is collected from patients. This could provide personalised survival modelling for colorectal cancer patients. Specifically, the inclusion of body composition as a factor holds great promise in improving current strategy making for patient care.
This review by Keyl J et al. aimed to explore automated assessment of body composition and liver metastases from CT images can improve personalised risk assessment.
Key learnings
Improving personalised therapy strategies for advanced colorectal cancer patients may be possible by using information gathered from routine CT scans. Increasing individualised treatment can help improve management and quality of life of these patients.
Reviewed by: Z. Beketova
Authors: Keyl J, Hosch R, Berger A, et al.
Published in: JCSM 2023
