A human behaviour imitation in the computing industry can be obtained using Neural Network architectures. One of the popular and influential neural network architecture is multilayer back propagation network (MLBPN). This paper aims to design and implement the MLBPN model for analyzing medical images. Many earlier research works have focused on implementing machine and deep learning models for analyzing medical data. But the computational speed is less and requires more epochs to obtain accurate results. Some researchers have overcome the challenges by using different integrated circuit techniques which are not cost-effective. This paper implements the multi layer perceptron (MLP) architecture using a double gate metal oxide semiconductor field effect transistor (DGMOSFET) model. The physical and logical functionalities of the MOSFET are integrated with MLP to provide the similar electrical behaviour of the transistor to control the current and voltage. The proposed MLP with DG-MOSFET model is trained with datasets having mixed of disease cases which were taken from FDA-NCI Clinical Proteomics Program Databank and the testing process is evaluated. The simulation is carried out with MATLAB and Cadence software, and then the analysed values are compared with actual values. The output proves that the speed and accuracy are improved compared with single gate devices.