Click here to


Are you sure ?

Yes, do it No, cancel

Integration of Ex Vivo Optical Reporter Imaging with in Vivo Radiologic Imaging: A Practical Multiscale Method for Measuring Whole Brain Disease Burden

M Scarpelli*, D Healey, S Mehta, C Quarles, Barrow Neurological Institute, Phoenix, AZ


(Thursday, 7/16/2020) 2:00 PM - 3:00 PM [Eastern Time (GMT-4)]

Room: Track 4

Purpose: A translational barrier for many molecular/functional imaging techniques is lack of a clear link between imaging parameters and underlying biologic characteristics. To overcome these limitations, we integrate in vivo radiologic imaging with ex vivo optical reporter imaging. Illustrative examples demonstrate how our methodology enables cross-validation and localized comparison of preclinical imaging techniques for assessing hypoxia, microvasculature, and tumor growth.

Methods: ¹8F-FMISO PET and T2w MRI data was acquired in mice and rats with orthotopically implanted brain tumors. These animals were then sacrificed, and whole brains were removed. Each excised brain was secured within a pathology slice block and scanned ex vivo with MRI. This was followed by slicing the brain into 1-mm pathology slices. These 1-mm slices were optically cleared. Fluorescence imaging of the 1-mm slices was then performed using an IVIS. All images were registered to the ex vivo MRI so that they were within the same spatial frame of reference.

Results: To evaluate the accuracy of in vivo MRI for identifying tumor margins, we used ex vivo fluorescence images of labelled tumor cells as a reference standard for comparison. After registration, the median surface distance between tumor ROIs segmented on in vivo MRIs and ROIs segmented on ex vivo fluorescence images was 34µm (interquartile range of 0-340µm; n=5 rats). The largest surface distance values arise because the MRI has lower sensitivity for detecting tumor regions than the fluorescence imaging. Additionally, brain tumor habitats with regions of hypoxic viable and nonviable tumor tissue were identified after registration of ex vivo fluorescence images with in vivo PET images.

Conclusion: These results demonstrate how the developed methodology facilitates the spatial validation of preclinical imaging techniques. This method can be used to establish the fidelity of in vivo imaging methods for detecting whole brain disease burden and inform their clinical utility.

Funding Support, Disclosures, and Conflict of Interest: This works was sponsored by the Arizona Biomedical Research Centre (ADHS18-198850), Blue Earth Diagnostics Ltd, Barrow Neurological Foundation, Dignity Health, ASU Collaborative Strategic Initiatives, and Students Supporting Brain Tumor Research.


Optical Imaging, Functional Imaging, Brain


IM- Multi-modality imaging systems: Development (new technology and techniques)

Contact Email