In radiosurgery a moving beam of
radiation is used as an ablative surgical instrument to destroy brain
tumors.  Classical radiosurgical systems rely on rigid skeletal
fixation of the anatomic region to be treated. This fixation procedure
is very painful for the patient and limits radiosurgical procedures to
brain lesions.  Furthermore, due to the necessity of rigid fixation,
radiosurgical treatment with classical systems cannot be fractionated.
A new 

{camera-guided radiosurgical system}  \* link to neuro.ps *\

capable of tracking patient motion during
treatment has been built to overcome these problems.  The radiation
source is moved by a six degree-of-freedom robotic arm.  In addition
to offering a more cost effective, less invasive, and less painful
treatment, the robotic gantry allows for arbitrary spatial motion of
the radiation source. Based on this feature we can treat non-spherical
lesions with accuracies unachievable with classical radiosurgical
systems.  The system introduces a new class of radiosurgical
procedures, called non-stereotactic, or image-guided
radiosurgery.  At the heart of these procedures are algorithms for
planning both a treatment and the corresponding beam motion, given the
geometric description of the tumor shape and relative locations in the
particular case.


Project Group: 

Achim Schweikard
Rhea Tombropoulos
Jean-Claude Latombe
John Adler

Departments of Neurosurgery and Robotics Laboratory, Department of
Computer Science, Stanford University.


Recent Publications

Treatment Planning for a Radiosurgical System with General Kinematics,
Proc. IEEE Conf. Robotics & Automation, 1994.

Planning for Image Guided Radiosurgery
AAAI Spring Symp. Medical Applications of Computer Vision, 1994.