His presentation on Computer-Aided Oral Implantology can

Technical Issues of
Image Guided Surgery
W. Birkfellner
Digital Image Processing Laboratory,
Center for Biomedical Engineering and Physics
Medical University Vienna
W. Birkfellner

Image­Guided
CMF­Surgery
W. Birkfellner
An incomplete list
of systems...
� ARTMA Virtual Patient
� RoboDent System
� DenX IGI System
� Drill Guides
(Materialise, T. Fortin et
al.)
� Modification of
commercial systems
like the BrainLAB suite
or the Medtronic
TREON
� ... and VISIT

Reviving old VHS ­Tapes:
VISIT in action
W. Birkfellner et al., IEEE Trans
Med Imaging. 19(6):616­620,
2000
W. Birkfellner et al., Clin Oral
Implants Res. 12(1):69­78, 2001
F. Watzinger et al., Plast Reconstr
Surg. 107(3):659­667, 2001
A. Wagner et al., Clin Oral
Implants Res. 14(3):340­348,
2003

Challenges from
a technical point of view
• Imaging – CT protocols & handling of artefacts
• Patient – to ­ image registration
• Image processing issues – visualization &
segmentation
• Hardware for image­guided CMF surgery
• Workflow optimization
• What's about to come ?
• Non­technical challenges ...
W. Birkfellner

Imaging
• CT is the modality of choice for dental and
CMF surgery
• Streak artefacts stem from the reconstruction
process and are difficult to avoid or to be
removed
• Pantomography is strongly dependent on
patient positioning due to the strong perspective
and the tomographic distortion of the image
• CBCT introduces other artefacts such as
truncation and ring artefacts
W. Birkfellner

Artefacts in CBCT
Truncation­ and ring­
artefacts are common ...
W. Birkfellner

Quantitative CT
W. Birkfellner
Homolka P, Beer A,
Birkfellner W et al.
Radiology, 224(1): 247­
52, (2002)

Planning in VISIT
W. Birkfellner
• Planning of implant
channels based on
available bone
volume and ­density
is rather
straightforward
• For extraoral
implants, a more
sophisticated
interface is
necessary

Tracking & Registrationa
Link to the Real World
• Registration: Finding a
transformation between the
coordinate system of a
volume image and the real
world
• Tracking: Maintaining that
registration by measuring
motion in 3D by
• optical tracking
• electromagnetic probe
• alternative techniques
W. Birkfellner
W. Birkfellner et al., Med Phys 25(11):2242­2248,1998
W. Birkfellner et al., IEEE Trans Med Imaging.
17(5):737­742, 1998
J. Hummel et al., Med Phys 29(10):2205­2212, 2002

Key issues in Tracking
• Accuracy and resolution: Optical trackers are
rather accurate and still superior to EM probes
• Reliability: A free line of sight for an optical tracker
always be maintained
• Handling: Trackers are cumbersome in general;
passive systems have a higher acceptance rate than
active (=wired) systems, but are less robust.
• Sterilizability: Autoclavation of electronics is difficult,
ETO is expensive and time­consuming
W. Birkfellner

Patient to Image
Registration
Point to point registration (Horn) using
at least three markers and a point probe
error 0.7 +/­ 0.15 mm
Birkfellner W, Solar P, Gahleitner A et al, Clin Oral Implants Res 12(1):69­78, (2001)
W. Birkfellner

Some experiences
­ Minimally invasive insertion of titanium
microscrews is feasible; the markers stay stable
for several weeks
­ Increasing the number of markers beyond 3
screws does not influence registration accuracy
­ Exact localisation of screwhead centroids is
necessary
­ Mobility of mucosa in atrophic patients caused
inaccuracies when using splint­type solutions,
even if residual dentition is available
W. Birkfellner

Image Processing:
Segmentation
Segmentation is not exact –
Identification of surfaces
depends on algorithms used
W. Birkfellner

Alternative techniques
Surface registration: Highly dependent on the
quality of surface segmentation and the
presence of outliers
­ Surface scanning by laser
ranging is hampered by the
small visible area of the
mucosa
­ Stab incisions are
cumbersome, but give
good results
W. Birkfellner
© by Tactile Technologies, IL
Schicho K, Seemann R, Cohen V et al. J Clin
Periodontol. 34(11):991­7, (2007)

Splint techniques
© by DMFR, from Eggers G, Kress B, Rohde S, Mühling J.
Dentomaxillofac Radiol. 38(1):28­33, (2009)
W. Birkfellner
­ Problematic in
completely
edentulous patients
­ Residual elasticity
of 2­component
molding materials is
recognizable
­ Segmentation of
molding material
and mucosa
sometimes difficult

Surface Scanning
Result from the Medtronic FAZER surface
digitizer – from Schicho K, Figl M,
Seemann R et al. J Neurosurg, 106(4):704­9, (2007)
Colchester AC, Zhao J, Holton­Tainter
KS et al. Med Image Anal, 1(1):73­90,
(1996)
­ Has a long history in frameless stereotaxy
­ Complicated for fully intraoral interventions
W. Birkfellner

Tracking technologies
for CMF & oral surgery
­ 32 implants,
5 patients after
tumor resection and
primary reconstruction
­ Accuracy:
top 1.0 +/­ 0.5 mm
bottom 1.3 +/­ 0.9 mm
Wagner A, Wanschitz F, Birkfellner W et al. Clin Oral Implants Res. 14(3):340­8, (2003)
W. Birkfellner

Challenges in
tool design ­Sterilizable electronics are
cumbersome, autoclavation is
difficult.
­ Passive markers somewhat
solve the problem and are handy,
but provide less accurate results
when being conatminated with
West JB, Maurer CR Jr. IEEE
Trans Med Imaging. 23(5):533­
45, (2004)
blood
­ Tool geometry (and size) is
crucial
W. Birkfellner

An early academic effort
Hybrid Tracking
­ Combination of
electromagnetic and
optical sensors to
overcome the drawbacks
of both technologies
W. Birkfellner
Birkfellner W, Watzinger F, Wanschitz F et al.
IEEE Trans Med Imaging. 17(5):737­42,
(1998)

Looking back ­
Implant insertion ­ 2000
W. Birkfellner

Implant insertion after
hemimaxillectomy Sealing of the nasopharynx
with an obturator
prosthesis after radical
resection of squamous
cell carcinoma
W. Birkfellner

Restoration after ablative
tumor resection
­ Surgical removal
of an osteosarcoma
in the
mandible
­ Implant insertion
in a scapula flap
after reconstruction
Wagner A, Wanschitz F, Birkfellner W et al. Clin Oral
Implants Res. 14(3):340­8, (2003)
W. Birkfellner

Minimally invasive
implant insertion
Comparison of VISIT and the Medtronic
TREON system. Accuracy ranges from
0.6 +/­ 0.5 to 1.0 +/­ 0.5 mm for both
systems
W. Birkfellner
Wittwer G, Adeyemo WL, Schicho
K et al.Int J Oral Maxillofac
Implants. 22(5):785­90, (2007)

Medical Augmented
Reality
W. Birkfellner et al., IEEE Trans Med Imaging. 21(8):991­997, 2002
F. Wanschitz et al., Clin Oral Implants Res. 13(6):610­616, 2002
W. Birkfellner et al., Phys Med Biol. 48(3):N49­57, 2003
M. Figl et al., IEEE Trans Med Imaging. 24(11):1492­1499, 2005
W. Birkfellner

Technology of image
injection
Birkfellner W, Figl M, Huber K et al. IEEE Trans Med Imaging.21(8):991­7, (2002)
W. Birkfellner

Linking navigation and
augmented reality
W. Birkfellner

Conclusions &
Outlook
Research directions for Computer Aided Surgery
in the CMF­area include
� Technical developments concerning TRE …
� Developments for optimized workflow and cost
including socioeconomic research
�Evaluation of suitable clinical applications
W. Birkfellner

Acknowledgments
Dept. of Cranio­Maxillofacial Surgery
Dental School Vienna
Dept. of Diagnostic Radiology
My Colleagues at the CBMTP
… and R. Schulze and the IADMFR for inviting me!
wolfgang.birkfellner@meduniwien.ac.at
http://www.zbmtp.meduniwien.ac.at/index.php?id=79
W. Birkfellner