Machine Learning and specifically Deep Learning systems have received much interest recently in many fields including the domain of Pattern Recognition. However, these systems usually relay on centralized data for training, which causes vulnerability and security issues. In addition, the complexity of the learning algorithms and the fast growth of the amounts of data in the recent years is another challenge for these systems. Blockchain has the ability to solve the data security problems due to its decentralized nature. In the other hand, parallel optimizations can reduce the training time of the learning algorithms. In this paper, we’ll be discovering the benefits of integrating Blockchain technology and Parallelism with the Pattern Recognition systems. We’ll be discussing papers that implemented the Blockchain technology, Learning systems, and also the Parallel Optimization methods in the medicine field. Most of these papers focus on the prediction, early diagnosis, and diagnosis of certain diseases such as Cancer, COVID-19...etc. We’ll make a comparative study between these works based on several criteria: Blockchain technology, Learning algorithms, Parallel optimization methods, Datasets, and the Accuracy. Furthermore, We will present in-depth analyses of these papers outlining the advantages and limits of each one.