The paramount significance of security in medical image communication lies in its quality of the patient’s personal information. Safeguarding this data is crucial when digital images and patient information are transmitted over public networks. Medical images contain high sensitive and critical data. Each pixel in these images holds significant diagnostic value, and any alteration could lead to inaccurate diagnoses. Ensuring robust security for these images is imperative, and any compromise in security could have severe consequences, as the margin for redundancy is minimal. The embedding capacity of medical images is limited, making it challenging to employ traditional security measures. Some researchers say that encrypted data and hidden data are examples of data protection technologies that can be used to guarantee the database. These two may cause stagnation in the processing time and inadequacy, in situations with medical images. This paper suggests applying the Fragmented based Elliptical Curve Cryptography combined with Convolutional Neural Network algorithms to design a system that enables secure disease diagnosis using medical images. The experimental results further prove the efficiency of the system applied to Lung CT scan images collected from open medical data sources. The implemented security measures provide a high level of protection for sensitive medical data during transmission.