I would recommend to build on an existing open-source application framework instead of redeveloping all basic infrastructure again.
For medical applications, you can use 3D Slicer (that I’m one of the core developers of). Slicer uses ITK for image I/O and processing, VTK for visualization, and Qt for GUI. It already has all the features that you may need for stereotactic surgery planning and guidance, such as DICOM import, rich multimodality image and tool visualization, intensity-based multimodality image registration, patient registration, tool calibration, automatic stereotactic frame registration. All comes with open-source, restriction-free (BSD-type) license. It is used clinically (under IRB approval) for many procedures. It is used in several commercial products - you can apply your own branding, style sheet, etc. and use simplified, application-specific user interface.
If you have any questions you can post it on Slicer forum.