The Applications Of Radiomics For Predicting Survival Rate Of Patients With Head And Neck Squamous Cell Carcinoma (Hnscc)

The radiomics based on positron emission tomography (PET) data and random forest can predict overall survival rate of head and neck squamous cell carcinoma (HNSCC). We used the texture features extracted from PET and clinical information from patients with HNSCC (n = 138). The Spearman's correlation analysis, Kaplan-Meier log rank test and random forest were used for survival significance and to predict survival rate of patients with HNSCC. Zone Length Non-Uniformity (ZLNU) was defined as a new key radiomics feature to predict survival rate. For stage N2 group, predicted survival rate was 76.2 % and actual survival rate was 73.3 %. For stage WA group, predicted survival rate was 74.7 % and actual survival rate was 73.8 %. The result of this study that applications of F-18-(FDG)-PET images using radiomics features was validated and could be expected to be used as the basis for future research using MRI images with more distinct structures.