Most techniques in segmental spinal fixation surgery rely on the ident
ification of predefined targets with the help of anatomical landmarks
and on intraoperative use of image intensifiers. However, because ther
e is no direct link between the image information, the accessible spin
al anatomy, and the action of surgical instruments several potential p
roblems and possible complications are still involved. A novel system
for spinal surgery has been designed allowing for the real-time, intra
operative localization of surgical instruments in medical images. In p
ractice this was achieved by combining image-guided stereotaxis with a
dvanced optoelectronic position sensing techniques. Modules were devel
oped for image data processing, surgical planning and simulation, and
various intraoperative procedures. A detailed validation of the system
was performed indicating an overall accuracy to be better than the sl
ice distance of the spinal image used. In an in-vitro setting 20 pilot
holes for pedicle screws were prepared in human cadaveric lumbar spin
es. An analysis in 77 histological cuts showed an ideal location in 70
and only minor cortex engagement in seven sections. In vivo the syste
m has been successfully applied in three posterior low lumbar stabiliz
ations with overall 15 transpedicular screws. This article focuses on
the clinical evaluation of a computer-assisted surgery system and its
application to the operating theatre for transpedicular fixation of th
e spine. The given approach effectively keeps the surgeon 'in the loop
' and requires only minor modifications of the established surgical te
chniques and associated instruments. The results of this study indicat
e that advanced computer-assisted tech niques may significantly improv
e the accuracy and safety of surgical interventions of the spine. The
proposed technique may in future be adapted to other applications in o
rthopaedic surgery.