The development of metal-based anticancer drugs is mainly governed by the experience accumulated with cisplatin and its analogues. The synthesis is focused on adding appropriate leaving and non-leaving groups to a transition metal in order to get more favorable DNA binding properties, and the biological activity is tested in vitro, always in a second step, looking for the cell line that is killed at the lowest drug concentration. This strategy seems unproductive today for the area of new drug development where the knowledge on cancer genomics is suggesting the use of targets selectively expressed, or overexpressed by cancer cells. These targets almost always are proteins, constituting membrane receptors or components of crucial biochemical pathways. Some data indicate that the antitumor activity of cisplatin might also be due to the interaction with protein targets. This critical review examines the possibilities for metal-based drugs to challenge tumors with innovative strategies, based on genomic approaches, capitalizing on the chemical experiences with metals in medicine and focusing on the nature of the ligands which are added to a metal depending on the selected tumor cells and on their molecular targets.