An Automated 3D Image-Processing Strategy for Small Animal Bioluminescence Cancer Studies

Nikolai V. Slavine, Roderick W. McColl, Edmond Richer, Ralph P. Mason and Peter P. Antich



Small-animal bioluminescence tomographic (BLT) imaging can detect cellular or molecular events in vivo and non-invasively. The specific goal of this project is to develop and validate optimal methodologies for imaging light emission at optical wavelengths in order to study small-animal tumors in vivo. Our BLT System with multiple rotating, high sensitivity CCD cameras offers imaging of bioluminescent light sources within living animals and phantoms with appropriate spatial resolution. To identify bioluminescence source location and strength we use the surface light flux detected on the surface of imaged object by CCD cameras. The quantitative 3D reconstruction of bioluminescence sources can be strengthened by incorporation of a priori knowledge on the small-animal anatomy and optical properties. An important part of BLT is an automation of image processing from the data acquisition to obtaining 3D reconstructions. In order to optimize this procedure Linux based software with multi-modality image handling environment was developed in user-derived formats, such as Finite Element Meshes or Color maps are also easily incorporated. Serial graphical applications which operate in an object-oriented or data-driven fashion are also available. A further objective of this work was to develop and test a strategy behind automated methods for BLT image processing to decrease the time of volumetric imaging and quantitative assessment.

Index Terms Bioluminescent imaging, 3D reconstruction, automation.

Full Text (PDF)