<p>Addressing the information transmission security challenges posed by the rapid advancement of electronic healthcare, this paper proposes a multi-image medical image encryption algorithm capable of encrypting any number of medical images. Tailored to the characteristics of medical images, this paper designs a multi-image encryption algorithm based on three-dimensional spiral transformation and adaptive multi-directional diffusion. First, due to the presence of extensive non-interesting regions in medical images, preprocessing is required before encryption. An adaptive thresholding method based on the average image grayscale value is designed to perform clipping on the images. During the key generation phase, a four-dimensional Hamiltonian conservative chaotic system is employed for subsequent operations. In the scrambling phase, multiple medical images are aggregated into a three-dimensional vector, segmented into sub-blocks using three-dimensional sub-blocking, and subjected to three-dimensional spiral transformations within each sub-block alongside inter-sub-block scrambling operations. During the diffusion phase, multiple rounds of diffusion are performed based on the distinct grayscale values within each subblock. Testing and performance analysis demonstrate that this algorithm achieves effective encryption, resists common attacks, and offers high security levels with fast processing speeds.</p>