A quadruple isotope SPECT/CT imaging acquisition was performed on a rat. 450uCi (17MBq) of Tc99m-MDP, 1.4mCi (52MBq) of I123, 600uCi (22MBq) of In111, and 1mCi (37MBq) of Tl201-Chloride were administered to the rat. A 40-minute image acquisition was performed at 3 hours post-injection of the Tc99m/I123 and 1 hour post-injection of the In111/Tl201.
The color bars from left to right are: CT, Tc99m, I123, In111 and Tl201.
A three-part animation displaying the results of an automated lung segmentation of a rat using the newly developed inviCRO digital scalpel powered by Definiens. The layered model or "onion" model of the lung is then generated via topographical thinning or "peeling" of the segmented lung. Application of this model to the image data allows for automated, quantifiable deposition of the relevant radiopharmaceutical in the lung. The three scenes are comprised of an X-Ray CT/Onion model fusion, SPECT/CT and lastly SPECT/Onion model.
Exemplary segmentation of a mouse brain using the CT component of a NanoSPECT/CT image only. On the left hand side, the original CT slices are shown, while the right images displays a fusion with the four different classifications of this image (background=gray, body=red, bone=white, brain=yellow).
Such classified anatomical data can subsequently be used to isolate organs from other modalities and thus achieve very precise and reliable information on, for instance, uptake. See [141] for further details.
InviCRO offers tools for automated segmentation for this and many other applications.
Example of an automated tumor segmentation in a mouse using Definiens' Developer XD software. On the left, a volume rendering of the mouse torso (CT in grayscale, SPECT in NIH Fire 2) is shown, while on the right hand side, the tumor was isolated using inviCRO's digital scalpel tool.
The segmentation rule set provided by the Boston-based contract research organization finds the tumor based on its characteristics in the anatomical CT only. This approach not only allows a very precise volume measurement (including necrotic parts of the tumor), but also delivers highly consistent estimates of the tumor uptake from the SPECT image. SPECT and CT images generated on Bioscan's NanoSPECT/CT.
Please note, that different SPECT color scales where used in the left and right movie.
Special thanks to Dr. Hesterman (Bioscan) for his support in data pre-processing and image generation. The mouse image is courtesy of Ben Gershman (Univ. New Mexico).
To analyze the aerosol images shown in [131] and [133], inviCRO developed a segmentation rule set for Definiens' Developer XD software. This does not only allow for very precise and consistent segmentation of a single image, but can easily be used for batch-processing of an array of images. The CT-only based segmentation delivers valuable anatomical information to analyze the information gained from the SPECT scans. To isolate the lungs from the rest of the image, inviCRO used their digital scalpel.
Special thanks to Ky Harlin (inviCRO, LLC) for the image processing. For additional details on the mouse data, please see above links.
A tri-modality (grayscale MRI, green SPECT, NIH Fire2 CT) MIP of a rat brain acquired by inviCRO at the Center for Translational Neuroimaging in the newly created imaging facility at Northeastern University.
A rule set (written in the Definiens Developer package) was applied to the CT data to automatically segment the brain and bones. The resulting segmentation was used in inviCRO's digital scalpel to isolate brain from the MR and SPECT data with anatomical reference supplied by the bone CT data.
The MRI (100umx100umx500um, 30min acq.) was acquired on a 7T Bruker magnet using a spin-echo sequence (RARE) comprised of 50 slices with 6 averages. The SPECT (600um voxels, 30min acq.) and CT (400um voxels, 4.5min acq.) images were generated on a NanoSPECT/CT. 400 uCi of I125-beta-CIT (a dopamine and serotonin transporter imaging agent) was injected 4.5 hours prior to the SPECT study.
After this process, the image data is now ready for futher analysis, such as estimating uptake in different regions of the brain.
The unsegmented data can be found in [139] and [138].
Special thanks to Dr. Hesterman (Bioscan) for the data processing.
The transverse slices of rat brain data with MRI (grayscale), CT (NIH Fire 2), and SPECT (green) acquired by inviCRO at the Center for Translational Neuroimaging in the newly created imaging facility at Northeastern University (using a 7R Bruker magnet and Bioscan's NanoSPECT/CT).
Please see entry [141] for further details. The unsegmented data can be found in [139] and [138].
Transverse slices from a tri-modality study (MRI, SPECT, CT) of a rat brain acquired at the Center for Translational Neuroimaging in the newly created imaging facility at Northeastern University. The MRI (grayscale, 100umx100umx500um, 30min acq.) was acquired on a 7T Bruker magnet using a spin-echo sequence (RARE) comprised of 50 slices with 6 averages. The SPECT (green, 600um voxels, 30min acq.) and CT (NIH Fire2, 400um voxels, 4.5min acq.) images were generated on a NanoSPECT/CT. 400 uCi of I125-beta-CIT (a dopamine and serotonin transporter imaging agent) was injected 4.5 hours prior to the SPECT study. Tracer localized in eyes, striatum and thyroid. Animal was imaged on a Minerve bed under isoflurane. Data registration and presentation supported by inviCRO.
Displayed: Image rendering of an automated 3D segmentation of a dual-modality study from the NanoSPECT/CT (Tc99m-Glucarate on a tumor bearing mouse; early stage) using the Definiens Developer XD software platform. The relevant color-coded organs shown in the animation were segmented and classified using a step-wise Rule Set specific to 3D SPECT-CT image acquisitions. The above movie was generated as a collaborative effort between Definiens, inviCRO and Bioscan.
Displayed: On the left-hand side is a standard maximum intensity projection (MIP) of a study performed on a NanoSPECT/CT of Tc99m-Glucarate in a mouse with a very early stage tumor. The image on the right-hand side is also a MIP, however instead of a standard SPECT image the data set co-registered with the CT was generated by assigning integer values to organs automatically segmented using the Definiens Developer XD software platform. The data set was classified using a step-wise Rule Set specific to 3D SPECT-CT image acquisitions. The above movie was generated as a collaborative effort between Definiens, inviCRO and Bioscan.