Drug Response Prediction of Liver Cancer Cell Line Using Deep Learning

Cancer is the second deadliest human disease worldwide with high mortality rate. Rehabilitation and treatment of this requires precise automatic assessment effective drug response control system. Prediction treated untreated cancerous cell line one most challenging problems for targeted delivery response. A novel approach proposed prediction cancer automatically by employing modified Deep neural networks. Human hepatocellular carcinoma (HepG2) cells are exposed to anticancer functionalized CFO@BTO nanoparticles developed our lab. models modifying ResNet101 exploiting transfer learning concept. Last three layers re-trained identification cells. Transfer in an appropriate choice especially when there limited amount annotated data. The technique validated on acquired 203 fluorescent microscopy images HepG2 cobalt ferrite@barium titanate (CFO@BTO) magnetoelectric vitro. achieved accuracy 97.5% sensitivity 100% outperformed other approaches. performance reveals effectiveness approach. It scalable fully which can be extended similar diseases such as lung, brain tumor breast cancer.