Poster Presentations

 

Morning Session

   

    Matthew Frederick Bruce

Harmonic doppler imaging of kidney stones

           Abstract: TBD.

 

    Adam Glaser 

Rapid pathology of fresh surgical specimens using an inverted scanning light sheet microscope

Abstract: For microscopic inspection of fresh surgical oncology specimens, pathologists must currently cut thick excised tissues into thin sections, a time-consuming and expensive practice. While various optical-sectioning microscopy methods have been proposed as an alternative means to physical tissue sectioning, they have typically suffered from a narrow depth of focus, making it logistically difficult to image the irregular surface of these fresh tissue specimens. Additional challenges include insufficient resolution, contrast, field of view, and/or imaging speed. To overcome these limitations, we have designed an inverted scanning light sheet microscope with an extended depth of focus (to image irregular tissue surfaces) that is able to provide rapid (<10 minutes), wide-area (up to 10x10 cm), high-resolution (1-2 microns), high contrast imaging of fresh tissues. The system design, characterization, and example feasibility results from human prostatectomy specimens will be presented.

 

    Efren Lee

Statistical distributions of ultra-low dose CT sinograms in the data processing stream

 Abstract: Low dose CT imaging is typically constrained to be diagnostic. However, there are applications for even lower-dose CT imaging, including image registration across multi-frame CT images and attenuation correction for PET/CT imaging. We define this as the ultra-low-dose (ULD) CT regime where the exposure level is a factor of 10 lower than current low-dose CT technique levels. In the ULD regime it is possible to use statistically-principled image reconstruction methods that make full use of the raw data (sinogram) information. However, clinical CT scanners have a data processing stream that uses the standard approach of a negative logarithm transformation in as well as pre-log and post-log corrections (most importantly a pre-log non-positivity correction). Our goal is to understand the statistical distribution of ULD CT data through the different data processing steps to understand when analytic or iterative image reconstruction methods may be effective in producing images that are useful for image registration or attenuation correction in PET/CT imaging. We used calibrated simulation studies and the Kolmogorov–Smirnov statistic to evaluate the normality of processed sinogram data. In summary, our results indicate that there are three general regimes: (1) Diagnostic CT, where post-log data are well modeled by normal distribution. (2) Low-dose CT, where normal distribution remains a reasonable approximation and statistically-principled (post-log) methods that assume a normal distribution have an advantage. (3) An ULD regime that is photon-starved and the normal approximation is no longer effective. This leads to fundamental limits in the estimation of ULD CT data when using a standard data processing stream.

 

    Mengyuan Liu

Computational methods for automated analysis of human brain anatomy, function and connections

Abstract: In our group, we develop computational methods to study early human brain development using different imaging modalities. We acquire structual and functional MRI images of fetal and premature neonatal human brain and apply computer vision and machine learning techniques to obtain motion/bias corrected high-resolution images. With structural MRI, we built a spatio-temporal atlas to model the brain development during the third trimester. With functional MRI and DTI, we also learn the development of functional connections. With these tools, we aim to use medical images of fetal/neonatal human brain to predict neuro-developmental outcomes of newborns.

​

    Jingjiang Xu

Spectral-domain optical coherence tomography- based angiography for scalable wide-field vascular imaging

Abstract: We report a high-speed (147kHz), high sensitive (~105dB), long imaging depth (7mm) spectral-domain OCT-based angiography, which shows scalable ultra-wide field of view (up to 750mm2) with flexible resolution for functional vascular imaging.

 

    Yu Wang

Raman-encoded molecular imaging (REMI) for guiding breast cancer lumpectomy

Abstract: There is a need for intraoperative imaging technologies to guide breast cancer lumpectomy surgeries and to reduce the high rates of re-excision for patients (20%-60%) in which residual tumor is found at the surgical margins during post-operative pathology analyses. The molecular imaging of tumor biomarkers is potentially a highly specific method to identify residual tumors at the surgical margins. However, since the molecular phenotype of breast tumors vary greatly between patients as well as within a single tumor mass, multiplexed molecular imaging is necessary to ensure a high level of sensitivity of tumor detection. We are developing an intraoperative molecular imaging technique to enable surgeons to rapidly visualize residual tumors at the surgical margins of freshly resected breast tumors. In particular, this work employs topically applied surface-enhanced Raman scattering (SERS) nanoparticles (NPs), in conjunction with ratiometric imaging of targeted vs. untargeted NPs, to enable the sensitive and multiplexed detection of a large number of cell-surface biomarkers of breast cancer. By developing high-affinity targeted SERS NPs, a sensitive raster-scanned spectral-imaging device, and an optimized topical-delivery protocol and device, we have recently demonstrated the simultaneous quantification of four biomarkers at the surgical margins of freshly resected breast tumors in 10 min. Feasibility studies are being performed for the clinical translation of this Raman-encoded molecular imaging (REMI) technique to guide breast cancer lumpectomy procedures.

 

    Kristen Wangerin

A virtual clinical trial comparing static versus dynamic PET imaging to measure response to breast cancer therapy

Abstract: Background: Positron emission tomography (PET) imaging is used to monitor the response of cancer to therapy and to assess the effectiveness of therapy as early as possible. Both standardized uptake values (SUVs) and estimates of the kinetic rate parameters have been used in treatment evaluation. Objective: Our goal was to determine which quantitative imaging biomarker, SUV from a static scan or kinetic parameters from a dynamic scan, is better at determining tumor response to therapy as a function of model parameters. Methods: We conducted a virtual clinical trial that used a combination of measured and simulated data to account for all known sources of PET imaging variability in order to characterize static SUV measurements and kinetic parameter estimates. We generated parameter estimate histograms before and after therapy and corresponding receiver operating characteristic (ROC) curves. Results: For high uptake tumors, the simpler static SUV measurement effectively distinguished response to therapy. For lower uptake tumors, kinetic parameters better distinguished response. Therefore, SUV from a static PET scan determined response to therapy for higher uptake tumors, while kinetic parameters enabled earlier assessment of response for lower uptake tumors..

 

    QinQin Zhang

OCT-based angiography of choroidal neovascularization by removing projection artifacts

Abstract: To accurately visualize choroidal neovascularization (CNV) and its treatment response, a projection artifact removal (PAR) algorithm was performed on the outer retinal avascular slab (ORAS) of OCT-based angiography (OCTA) images to remove projection artifacts from the top retinal circulations. Optical microangiography (OMAG) as one of the leading techniques of OCTA utilizing the complex OCT signal with a high sensitivity to detect the capillaries was employed in the study. The application of the PAR algorithm on OMAG angiograms can allow for more accurate quantitative evaluation of CNVs before and after treatment and provide comparable or even better images to the current gold standard FA/ICGA. Thus this algorithm may be a useful tool in interpreting images of CNVs that can complement current standard imaging modalities such as FA/ICGA.

​

    David Li

Photoacoustic activation of nanoemulsions and their use as theranostic agents

Abstract: Photoacoustic imaging is a hybrid ultrasound imaging modality that can be used in conjunction with traditional ultrasound imaging. One of the major drawbacks of photoacoustic imaging is the limited imaging depth due to light scattering. An efficient light absorbing emulsion based nanoemulsion is presented to greatly improve the signal to noise ratio of photoacoustic imaging. The emulsions feature a low boiling point liquid perfluorocarbon oil core (on the order of 200 nm in diameter) coated by a layer of polypyrrole or gold nanoparticles, which have a high optical absorption efficiency. Heating of the oil core due to optical absorption causes the emulsion to reversibly vaporize and recondense back into the liquid phase. The emulsion based photoacoustic contrast agents not only provide a stronger photoacoustic response than nanoparticle or dye based photoacoustic agents, benchtops experiments have shown that the reversible vaporization can be also be used for therapy to mechanically break down blood clots and restore blood flow.

​

​

                                                                                                                                                                                                                                                       

Afternoon Session

 

    LinPeng Wei and ChengBo Yin

A handheld optical-sectioning microscope for cancer detection and surgical guidance.

Abstract: There is a need for miniature optical-sectioning microscopes to enable in vivo interrogation of tissues as a real-time and noninvasive alternative to gold-standard histopathology. In this study, we developed a handheld line-scanned (LS) dual-axis confocal (DAC) microscope for the early detection of cancer as well as for guiding tumor-resection procedures. The DAC architecture has demonstrated an advantage over the conventional single-axis confocal configuration for reducing background noise. The use of line scanning enables fast frame rates, which mitigates motion artifacts of a hand-held device during clinical use. This handheld microscope has been validated with both ex vivo imaging in tissues and in vivo imaging in animal models, and will be tested in human patients for oral-cancer detection and for guiding brain-tumor resections.

 

    Jin Liu

Motion Detection and Correction for Carotid Artery Wall Imaging using Structured Light

Abstract: Carotid artery wall MRI is often affected by complex neck motion. We aimed to separate different motion components and correct them for better carotid artery wall delineation using structured light system. A healthy volunteer was scanned for 2D carotid MRI. It was demonstrated that voluntary abrupt motion, unconscious bulk motion and involuntary respiration can all be detected effectively. Both abrupt motion and bulk neck shift can be corrected for better vessel wall delineation, but the duration of abrupt motion can affect motion correction effectiveness. Bulk neck shift distance optimization by maximizing sharpness can future reduce motion artifact.

​

    Geng Zeng

A simulation study of sensitivity and specificity of the PET/X scanner

Abstract: This is a brief introduction to the PET/X scanner that is being developed in our lab and my simulation study on its performance. The scanner is intended to evaluate the efficacy of clinical breast cancer treatments for confirmed patients.

 

    Ruoqiao Zhang

Impact of extended-duration cine CT on motion compensation for respiratory gated PET/CT: A preliminary study

Abstract: PET/CT scanners has become a standard component in many diagnostic imaging procedures such as oncology, surgical planning, and cancer staging. However, PET/CT imaging of the lung and abdomen is typically affected by patient respiratory motion, which may lead to major inaccuracy of organ uptake values and artifacts. While respiratory gated PET/CT with cine CT can be somewhat effective in reducing motion-induced errors, the traditional cine acquisition is usually insufficient to provide full motion compensation due to the short cine duration, especially for patients with irregular breathing patterns. In this study, we conducted phantom simulations to investigate the impact of increasing the cine CT duration on reducing CT artifacts induced by respiratory motion.

 

    Maria Zontak

Speeding up 3D speckle tracking using PatchMatch

Abstract: Echocardiography provides valuable information to diagnose heart dysfunction. A typical exam records several minutes of real-time cardiac images. To enable complete analysis of 3D cardiac strains, 4-D (3-D+t) echocardiography is used. This results in a huge dataset and requires effective automated analysis. Ultrasound speckle tracking is an effective method for tissue motion analysis. It involves correlation of a 3D kernel (block) around a voxel with kernels in later frames. The search region is usually confined to a local neighborhood, due to biomechanical and computational constraints. For high strains and moderate frame-rates, however, this search region will remain large, leading to a considerable computational burden. Moreover, speckle decorrelation (due to high strains) leads to errors in tracking. To solve this, spatial motion coherency between adjacent voxels should be imposed, e.g., by averaging their correlation functions. This requires storing correlation functions for neighboring voxels, thus increasing memory demands. In this work, we propose an efficient search using PatchMatch, 2 a powerful method to find correspondences between images. Here we adopt PatchMatch for 3D volumes and radio-frequency signals. As opposed to an exact search, PatchMatch performs random sampling of the search region and propagates successive matches among neighboring voxels. We show that: 1) Inherently smooth offset propagation in PatchMatch contributes to spatial motion coherence without any additional processing or memory demand. 2) For typical scenarios, PatchMatch is at least 20 times faster than the exact search, while maintaining comparable tracking accuracy.

 

    Shaozhen Song

Optical coherence elastography based on high speed imaging of single-shot laser-induced acoustic wave at 16 kHz frame rate

Abstract: The proposed ultra-high speed Shear Wave Imaging OCT (SWI-OCT) technique utilizes high speed, phase-stabilized swept source OCT to capture single-shot shear wave propagation at a real frame rate of 16 kHz. The shear wave is generated photothermal-acoustically, by ultra-violet pulsed laser, which requires no contact to OCE subjects, while launching high frequency shear waves that carries rich localized elasticity information. The image acquisition and processing can be performed at video-rate, which enables real-time 3D elastography. This approach opens up the feasibility to perform real-time 3D SW-OCE in clinical applications, to obtain high-resolution localized quantitative measurement of tissue biomechanical property.

 

    Chieh-Li Chen

Peripapillary retinal nerve fiber layer (RNFL) vascular microcirculation in glaucoma using optical coherence tomography-based microangiography (OMAG)

Abstract: Purpose: To investigate the vascular microcirculation changes in peripapillary RNFL in normal, glaucoma suspects (GS), and open angle glaucoma (OAG) subjects using OMAG. Methods: One eye from each subject was scanned using a montage scanning protocol with a 68 kHz Cirrus HD-OCT 5000 based OMAG prototype system (Zeiss, Dublin, CA) covering a 6.7x6.7 mm2 area centered at the optic nerve head (ONH) with a 9.8 µm sampling resolution. Blood flow signals were extracted using a complex OCT signal based OMAG method. Peripapillary RNFL was segmented by a proprietary semi-automatic segmentation program. RNFL vascular en face images were generated using blood flow signals with the highest flow intensity along the axial direction within the RNFL. RNFL vascular microcirculation was measured by calculating the overall blood flux and vessel density within an annulus region centered at the ONH excluding the big retinal vessels (2.5-mm and 3.7-mm as inner and outer diameters, Figure) and compared among groups. Blood flux was defined as the averaged flow signal intensity in the vessels; vessel density measured the percentage of the detected vessels and capillaries within the annulus region, and was defined as the ratio between areas of vessels and the annulus. One-way ANOVA and t-tests were used for statistical analysis. P<0.05 and <0.0167 were considered as statistically significant among three groups, and between each two groups after Bonferroni correction. Results: Seventeen eyes from 17 normal, 27 eyes from 27 GS, and 42 eyes from 42 OAG subjects were recruited. The mean age and RNFL thickness for normal, GS, and OAG groups were 68.8±11.5, 67.7±7.3, and 65.7±10.7 years (p=0.49), and 94.3±12.4, 88.1±8.4, and 69.8±9.8 µm, respectively (p<0.0001). Significant differences in RNFL thickness were detected between normal and OAG, and GS and OAG, but not between normal and GS (p=0.06). RNFL microcirculation measured as blood flux was significantly lower in OAG and GS compared to normal eyes (p≤0.0015). Significant differences in vessel density were detected between normal and OAG, and GS and OAG, but not between normal and GS eyes (p=0.58) (Figure). Conclusion: Peripapillary RNFL perfusion detected by OMAG showed significant differences among normal, GS, and OAG subjects. RNFL microcirculation measurement using OMAG may provide useful information for detection of glaucoma.

​

    Christina Keravnou

Microvascular injury and perfusion changes induced by ultrasound and microbubbles in a machine-perfused pig liver

Abstract: Localized drug delivery and uptake can benefit from the combined action of ultrasound and microbubbles at a specific site. Some of the possible mechanisms suggested are vessel and/or cell poration, but still the exact acoustic parameters that trigger those phenomena remain unknown. Ex-vivo machine perfusion of human-sized organs is a technique that provides an ideal environment for preclinical investigations with high physiological relevance not possible with in-vitro experiments. In this work, ex-vivo machine-perfused pig livers combined with an image-guided therapy system were used to investigate microvascular flow changes caused by the interaction of ultrasound driven microbubbles with the vasculature. The effect of acoustic pressure (1.7 – 4 MPa peak negative pressures) and the number of cycles (1000 or 20 cycles) were examined. Perfusion changes due to the action of ultrasound on microbubbles in the microcirculation were qualitatively and quantitatively assessed with Contrast Enhanced Ultrasound (CEUS) and were used as a metric of the extend of vessel perforation and thus extravasation. Areas that were exposed to peak negative pressure above 1.7 MPa showed a detectable and irreversible perfusion change. Complete devascularization of the area exposed to ultrasound was observed at much larger acoustic pressures (~4 MPa). Shorter acoustic pulses (20 cycles) produced markedly less perfusion changes than longer pulses (1000 cycles) under the same acoustic amplitude exposure.