Two New Radiomics Features Predict Lung Cancer Survival in Low-Dose CT Screening

June 3, 2020 | Radiomics

Jaileene Perez Morales, MD, of the Moffitt Cancer Center in Tampa, Florida and colleagues identified two predictive radiomic features in non-small cell lung cancer. Statistical root mean square, one of the features, was correlated with occurrence of the genes LOC285403 and FOXF2, the latter of which has been associated with poor outcomes in non-small cell lung cancer. The second identified feature was Neighborhood Grey Tone Difference Matrix (NGTDM) Busyness. Based on these features, the researchers divided patients into low-, moderate- and high-risk groups. Patients identified as high-risk had 0% 5-year overall survival while patients who were identified as low-risk had 78% 5-year overall survival.