Program Information
Prediction of Cervical Cancer Treatment Response Using Radiomics Features Based On F18-FDG Uptake in PET Images
B Altazi*, D Fernandez , G Zhang , M Biagioli , E Moros , H. Lee Moffitt Cancer Center, Tampa, FL, University of South Florida, Tampa, FL
Presentations
SU-E-J-258 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose:Radiomics have shown potential for predicting treatment outcomes in several body sites. This study investigated the correlation between PET Radiomics features and treatment response of cervical cancer outcomes.
Methods:our dataset consisted of a cohort of 79 patients diagnosed with cervical cancer, FIGO stage IB-IVA, age range 25-86 years, (median age at diagnosis: 50 years) all treated between: 2009–14 with external beam radiation therapy to a dose range between: 45–50.4 Gy (median= 45 Gy), concurrent cisplatin chemotherapy and MRI-based brachytherapy to a dose of 20–30 Gy (median= 28 Gy). Metabolic Tumor Volume (MTV) in patient’s primary site was delineated on pretreatment PET/CT by two board certified Radiation Oncologists. The features extracted from each patient’s volume were: 26 Co-occurrence matrix (COM) Feature, 11 Run-Length Matrix (RLM), 11 Gray Level Size Zone Matrix (GLSZM) and 33 Intensity-based features (IBF). The treatment outcome was divided based on the last follow up status into three classes: No Evidence of Disease (NED), Alive with Disease (AWD) and Dead of Disease (DOD). The ability for the radiomics features to differentiate between the 3 treatments outcome categories were assessed by One-Way ANOVA test with p-value < 0.05 was to be statistically significant. The results from the analysis were compared with the ones obtained previously for standard Uptake Value (SUV).
Results:Based on patients last clinical follow-up; 52 showed NED, 17 AWD and 10 DOD. Radiomics Features were able to classify the patients based on their treatment response. A parallel analysis was done for SUV measurements for comparison.
Conclusion:Radiomics features were able to differentiate between the three different classes of treatment outcomes. However, most of the features were only able to differentiate between NED and DOD class. Also, The ability or radiomics features to differentiate types of response were more significant than SUV.
Contact Email: