| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
|
|||||||||
|
|
Instrumentation & Data Analysis: InstrumentationInstrumentation Posters |
1 Chair for Computer Aided Medical Procedures (CAMP); 2 Department of Nuclear Medicine, TUM, Munich, Germany
1720
Objectives: The concept of SLNB for head and neck tumors has not yet become standard due to the complex anatomy and the difficulty of finding SLN when in close proximity to the injection site. Recently we introduced flexible intra-operative 3D tomographic imaging using hand-held gamma-probes and spatial localization systems. Here we present preliminary experiments to confirm the usefulness of this technology in intra-operative image-guidance for SLNB in the neck.
Methods: The readings of a gamma probe were acquired simultaneously with its position and orientation from a spatial localization system. Based on that a system matrix was computed on-the-fly using an ad-hoc forward model of the acquisition process. A foam phantom with four Tc-99m (50 [kBq]) hot spots simulating SLNs was scanned 4 times each by 3 test persons. After reconstruction using ML-EM the generated images were compared with a SPECT scan.
Results: Our reconstruction was evaluated in contrast to a SPECT scan used as ground truth. The average NCC was 0.512 (std 0.077). The mean value of deviation for the centroids of the hotspots in x direction was 0.78mm, in y direction 3.50mm and in z direction 2.72mm. Deviations in the y-axis (dorsal direction) were higher since readings were mainly acquired on top of the phantom and thus lacking projections from below. The mean acquisition time was 5 minutes; the voxel size was 5.3 x 5.3 x 6 mm3 for a volume of 10.6 x 3.7 x 10.2 cm3.
Conclusions: The introduced system proved to be able to reconstruct 3D tomographic images. Due to its flexibility and its intra-operative nature, it has high potential as an intra-operative image support system for SLNB.
| ||||||||||||||||||||||||||||||||||||||
