This paper introduces a novel method for region-based encryption of medical images, employing a 4D hyperchaotic system. The hash value of the image determines the initial value of the state variables of the 4D hyperchaotic system, resulting in a unique and secure encryption method. The method begins by isolating the region of interest (ROI) from the background through binary thresholding, and the ROI is defined with the help of a binary mask, and in this work this mask is public. Extract the pixels from the ROI and store them in a 2D matrix. The approach involves multiple encryption stages, including scrambling, XOR operations, bitplane scrambling, and polyalphabetic substitution, where these operations are executed utilizing the sequences produced by the proposed chaotic system, with the initial values of the system being derived from the hash value of the actual image. This approach ensures the secure encryption of the ROI while preserving the integrity of the background, offering a robust solution for protecting sensitive medical imagery. The simulation and comparative analysis demonstrate that the suggested method effectively resists cryptographic attacks while maintaining strong encryption.