Radiomics for Estimating Recurrence Risk of Patients with Lung Cancer by Using Survival Analysis

Purpose: Because of the promotion cancer screening, number patients with lung detected at early stage has increased. However, it was reported that 30-40% I relapsed. If recurrence risk can be accurately predicted, is possible to give medical care for improving prognosis patients. The purpose this study develop a method prediction by using survival analysis radiomics approach. Method: A public database used in study. Fifty (25 recurrences and 25 censored cases) classified as or II were selected their pretreatment computed tomography (CT) images obtained. First, we one slice containing largest tumor area manually segmented regions. We subsequently calculated 367 radiomic features such size, shape, CT values, texture. Radiomic least absolute shrinkage selection (Lasso). Cox regression model random forest (RSF) estimating functions fifty Result: experimental result showed average under curve (AUC) values RSF accuracy 0.81 0.93, respectively. Conclusion: Since our scheme predict non-invasive image examinations, would useful treatment follow-up after treatment.