The goal of this research work is to investigate and develop a framework for medical visualization, answering the needs of both end users (clinicians), working in their specific domain, and application designers, who build applications and visualizations for them. Most flexible visualization systems currently available are based on the dataflow paradigm. Dataflow-based systems suffer from a poor management of the screen real estate, which is a problem for end users. In order to deal with this problem, spreadsheet-based visualization systems –which are a generalization of conventional spreadsheets– were introduced. In comparison with dataflow-based visualization systems, spreadsheet based systems spend their screen space on operands rather than operators, which are usually more interesting to the user. We proposed a spreadsheetbased framework for interactive medical visualization. This framework is analyzed with regard to its usability and validated with concrete medical use cases. We have also investigated specific visualization methods and interaction techniques to answer the needs of clinicians. We used a multilevel color-mapping of scalar values on models of organ for effectively comparing bones even if the data set contain errors. We proposed also a method for organizing high-dimensional data sets that are difficult to explore on a two dimensional layout. A haptic rendering application is also presented for investigating the use of haptics for providing the clinicians with tools that allow them to "touch" and "feel" the virtually simulated tissues in addition to their visual rendering.