Clinical imaging using magnetic resonance imaging (MRI), computed tomography (CT), positron emission tomography (PET), and/or single-photon emission computed tomography (SPECT) is the standard of care for the detection and staging of human cancer. However, clinical imaging is a macroscopic process, with up to 106 malignant cells filling every 1 mm3 voxel. At present, it is extremely difficult to correlate clinical imaging findings with the cellular genotype and/or phenotype leading to the finding. Hence, "biomarkers," such as dynamic contrast enhancement (DCE) on MRI, or a "hot" voxel on PET seen after injection of a targeted radiotracer, are difficult to validate since the fusion of macroscopic clinical imaging findings with microscopic histological findings is fraught with technical challenges. To solve this problem, our laboratory has developed new near-infrared (NIR) fluorescence technology that permits simultaneous (same slide) immunostaining and hematoxylin/eosin (H&E) staining of any pathological specimen. Hence, the chronic problem of co-registering tissue slices at the cellular level is eliminated, while the "gold-standard" of H&E histology is preserved. This technology lays the foundation for an integrated platform that permits high accuracy co-registration of macroscopic and microscopic data sets from an individual patient's cancer. We have also developed an automated microscope that acquires H&E and NIR fluorescence simultaneously, and which permits up to 28 2"x3" whole-mount slides (or 56 1"x3" slides) to be scanned without human intervention. Using this technology, and several other innovations described in the application, 3-D data sets of clinical pathological specimens, at microscopic resolution, can now be generated. However, to bridge the gap between the microscopic and macroscopic domains, we have formed an academic-industrial partnership with the Imaging Department of Siemens Corporate Research (SCR) in Princeton, NJ. SCR is expert in the co-registration of 3-D volumes that have undergone non-linear deformations, in image segmentation, and in pattern recognition. Using algorithms developed by SCR for this study, we present an automated and integrated platform for cancer biomarker validation. Our study also leverages a unique clinical resource at the Beth Israel Deaconess Medical Center, the Hershey Prostate Cancer Tissue Bank. Through the Hershey Tissue Bank, men undergoing radical prostatectomy for prostate cancer receive a preoperative endorectal coil 3T MRI (pre- and post-Gd DCE), and at prostatectomy, the entire gland is available for whole-mount preparation. The paired data sets of clinical imaging (DCE-MRI) and whole mount histology will provide proof of principle for our biomarker validation platform, and will also be used to determine the mechanism of DCE-MRI at the cellular level. Completion of the specific aims by academic and industrial teams with complementary skill sets will permit virtually any proposed biomarker, for any type of cancer, to be validated rapidly and efficiently.
Biomarkers for cancer imaging are extremely difficult to validate, since clinical imaging is performed on a macroscopic scale and histological evaluation of resected tissue is performed on a microscopic scale. We have formed an academic-industrial partnership between the Frangioni Laboratory at the Beth Israel Deaconess Medical Center in Boston, MA and Siemens Corporate Research in Princeton, NJ aimed at developing an integrated platform for biomarker validation using new near-infrared fluorescence and image fusion technology. This study also leverages a unique prostate cancer tissue bank that provides paired DCE-MRI and histological whole mounts from individual prostate cancer patients.
|Gibbs, Summer L; Genega, Elizabeth; Salemi, Jeffery et al. (2015) Near-infrared fluorescent digital pathology for the automation of disease diagnosis and biomarker assessment. Mol Imaging 14:|
|Inoue, Kazumasa; Gibbs, Summer L; Liu, Fangbing et al. (2014) Microscopic validation of macroscopic in vivo images enabled by same-slide optical and nuclear fusion. J Nucl Med 55:1899-904|
|Khurd, Parmeshwar; Bahlmann, Claus; Maday, Peter et al. (2010) COMPUTER-AIDED GLEASON GRADING OF PROSTATE CANCER HISTOPATHOLOGICAL IMAGES USING TEXTON FORESTS. Proc IEEE Int Symp Biomed Imaging 14-17 April 201:636-639|