Positron Emission Tomography-Based Radiomics Features May Be Prognostic Of Outcome In Patients With Esophageal Squamous Cell Carcinoma Treated With Chemoradiation Therapy

W. Yu,X.L. Fu,Z.Y. Xu,L. Du,H. Zhang

DOI: https://doi.org/10.1016/j.ijrobp.2016.06.956

2016-01-01

Abstract:To identify quantitative PET-based radiomics features with prognostic significance in patients with esophageal cancer. Patients with unresectable esophageal squamous cell carcinoma are enrolled to receive concurrent chemoradiation therapy (CRT) and 18FDG PET/CT scanning both before and 4 weeks after the beginning of CRT. Quantitative features are extracted from within the region of interests (ROIs) containing the primary tumor contoured by experienced physicians, quantifying SUV intensity distribution, texture, and associated changes resulting from CRT. Radiomics features and clinical parameters are combined to explore the association with survival using Kaplan-Meier analysis and LASSO (Least Absolute Shrinkage and Selection Operator) -Cox regression is performed to identify features with prognostic performance. From 2012 to 2015, 34 patients of median age of 62.5 years (range, 40∼75 years) were enrolled to receive concurrent CRT. Median radiation BED was 71.175 Gy (range, 67.2∼73.925 Gy) and 91.2% of the patients completed at least 2 courses of chemotherapy with 5-FU and cisplatin. With the median follow-up time of 13.3 months (range, 2.5∼28.4 months), 2-year OS and PFS were 60.6% and 41.8%, respectively. Using LASSO-Cox method, pre-treatment SUVmax and inverse difference moment and the ratios of post- and pre-treatment total lesion glycolysis (TLG) and coarseness were the most important prognostic features. To predict the local failure, pre-treatment SUV bias-corrected skewness and correlation and the ratio of post- and pre-treatment SUVmax exhibited significant performance. The original and relative changes of radiomics features extracted from PET images may have played an effective role in predicting the survival and local control of patients with esophageal cancer who underwent concurrent CRT, which promise to guide the individualization of clinical decision making and radiation therapy delivery.