Nowadays, diagnosing patient diseases remotely is a routine and digital medical images analysis as a part of Telecare Medical Information Systems (TMIS) play a fundamental role in early diagnosing and treating most common and serious diseases such as breast cancer. In this context, altering or distorting even a single pixel of a medical image during its transmission over an unsure channel could lead to a wrong diagnosis and harm patient health, induce damaging delays. Therefore the security and privacy of the transmitted medical images must be addressed most seriously. Several techniques are proposed in the literature to address these issues using different techniques such as chaos theory, more recently, elliptic curves cryptography (ECC) or improved classical methods such as linear cryptography. In this paper, we address the security level concern of an image encryption technique combining ECC with Hill cipher (ECCHC) which has been recently proposed by Dawahdeh et al [13]. Our study rises concerns about some weaknesses and flaws of the analyzed encryption scheme against some plain-text and known plain-text attacks. In addition, and not least issue, it is found that the key length used in Dawahdeh et al. scheme is not sufficiently large to be robust against brute force attack. To fix the found flaws and to improve the encryption scheme, a generalized cryptosystem is suggested. In the enhanced version, the key matrix negotiation is redefined to a cipher that combines a modified EC Integrated Encryption Scheme (ECIES) and the linear multiplication matrix is generalized to key matrix of (2n × 2n), n > 2 to counter efficiently the exhaustive search attack. The effectiveness of the proposed version is evaluated and verified through extensive experimentation and most recent available security tools. Compared with the state-of-the art techniques, the proposed version exhibits excellent security features and can resist to various knowing attacks.