Securing Neural Network-Based Personalized Medicine for Advanced Liver Cancer Detection in Healthcare

This research is aimed at discussing the critical meeting point of healthcare and artificial intelligence, including the need to protect the required treatment solution. Investigation into the utility of various deep learning techniques such as VGG 16, VGG 19, Convolutional Neural Networks (CNN), and Recurrent Neural Network (RNN) for the prediction of liver cancer stages and their respective treatments has been completed. The generation of 2340 of medical images from multiple sources, including the internet and records of hospitals, as a dataset was performed in a secure cloud. The security measures have allowed only authorized personnel to access the data and ensured the privacy and secrecy of data. The created dataset was then separated into two parts: training, which consisted of training the model on 70% of images, and testing, which implied that 30% of images were used to test the generated model. For each of the run models, performance evaluation measurements such as accuracy, precision, recall, and F1 score were calculated. The results have shown the highest efficiency of VGG 19 model, which resulted in 98.44% of accuracy. For the generated model, a fictitious ROC curve and specificity rate were also provided to showcase the discrimination abilities of each of the submitted models. The results show that the utility of neural network-based personalized medicine is beneficial for the further progression of the early prediction stage of liver cancer and its treatment. The result showed that the proposed models has successfully contributed to the field of healthcare informatics and further opportunities for artificial intelligence application to increase patient outcomes in the form of personalized medicine that may be performed.