Prediction of the therapeutic response after target-combined chemotherapy treatment for patients with liver metastasis from colorectal cancer using computed tomography texture analysis.
10.7507/1001-5515.201801062
- Author:
Jing ZHANG
1
;
Yuwen ZHOU
2
;
Meng QIU
2
;
Bing WU
3
Author Information
1. Department of Radiology, West China Hospital, Sichuan University, Chengdu 610041, P.R.China.
2. Cancer Center, West China Hospital, Sichuan University, Chengdu 610041, P.R.China.
3. Department of Radiology, West China Hospital, Sichuan University, Chengdu 610041, P.R.China.bingwu69@163.com.
- Publication Type:Journal Article
- Keywords:
colorectal cancer liver metastasis;
computed tomography;
target-combined chemotherapy;
texture analysis;
therapeutic response
- From:
Journal of Biomedical Engineering
2018;35(6):914-920
- CountryChina
- Language:Chinese
-
Abstract:
This study aims to investigate the value of pre-treatment computed tomography (CT) texture analysis in predicting therapeutic response of liver metastasis from colorectal cancer after combined targeting chemotherapy. A total of 82 patients with colorectal cancer liver metastases who underwent chemotherapy combined with targeted therapy (cetuximab) between March 2011 and October 2017 comprised this retrospective study population. According to the RECIST1.1, the best curative effect evaluation of patients was recorded. Complete response (CR) and partial response (PR) were assigned to the response group, and the stable disease (SD) and progressive disease (PD) were assigned to the non-response group. The CT texture analysis was based on the Omini-Kinetics software, and the three-dimensional (3D) texture analysis was performed on the marked lesion on portal phase. The differences of texture parameters between the response group and the non-response group were compared. The receiver operating characteristic (ROC) curves were depicted on the parameters which with statistically difference, to characterize value in predicting the response to target-combined chemotherapy. The differences of Entropy, Energy, Variance, std. Deviation, Quantile95 and sumEntropy between the two groups in pre-treatment lesions were significant ( < 0.05). And lesions with higher Entropy, lower Energy, higher Variance, higher std Deviation and higher sumEntropy seemed to indicate a better therapeutic response. When sumEntropy > 0.867, good diagnostic efficiency could be obtained, with sensitivity of 60.5% and specificity of 79.5%, respectively. In conclusion, texture parameters derived from baseline CT images of colorectal cancer liver metastasis have the potential value acting as imaging biomarkers in predicting tumor response to combined target chemotherapy.
