Getting medicinal substances to the right place is a big issue while treating various illnesses. The shortcomings of conventional chemotherapy include low selectivity and poor biodistribution. Lower dosages of medications are therefore needed in order to potentially overcome these restrictions by delivering pharmaceuticals to the site of action and preventing rapid drug breakdown or removal to improve drug concentration in target tissues. Active targeting, passive targeting and dual targeting are a few types of smart medication delivery systems. Drug delivery vehicles, therapeutic medicines, and targeting moieties are the three main parts of a smart drug delivery system that is intended to transport the drug to target tissue and facilitate simple biodistribution of a targeted drug. Drug delivery vehicles such as liposomes, dendrimers and viral vectors are utilized to successfully convey the loaded drug. A targeting moiety, such as an antibody, polyethylene glycol, or other proteins, helps release the medication into the intended organs, tissues, or cells by successfully and selectively recognizing the target cells. Even though a smart drug delivery system has the ability to treat a variety of ailments, its efficacy is influenced by a number of variables, including extracellular matrix, pH, glucose, low oxygen content, ions, enzymes, biological membranes and target site pH. Therefore, research should be done to appropriately choose, design and further alter drug delivery vehicles, therapeutic medicines and targeting moieties in order to overcome the constraints and improve the effectiveness and clinical applicability of smart drug delivery systems.

DL and AI techniques have led to substantial advancements in the field of picture analysis. Radiologists and doctors will both profit from artificial intelligence-based procedures, even though they won't completely replace them. Rather than being used for decision-making, these methods will probably be employed for consultation and decision support. Radiologists, however, need to understand these tools and how the medical industry employs them. Studies have shown that AI algorithms are very accurate, resilient, fast and useful in medical imaging; nevertheless, the majority of these algorithms, especially DL algorithms, are still in the experimental stage and have not yet been developed further or applied to clinical settings. There are many barriers that prevent modern AI technology from being widely used in therapeutic settings. First off, using very large data sets to train these algorithms is frequently impractical. Furthermore, all institutions must use the same study methods prior to the implementation of DL algorithms. Improving algorithm accuracy and performance is a difficult, complex problem.