Research
Molecular Photoacoustic Imaging for Cancer Diagnostics
Emergence of new molecular biomarkers for personalized therapy requires efficient evaluation of biomarker-targeted drug delivery. Unfortunately, there is a dearth of preclinical imaging modalities that can provide molecularly specific imaging with high sensitivity and resolution at depth. High-resolution optical imaging is severely limited by sub-millimeter penetration depth and an extremely small imaging field, while bioluminescence imaging has poor spatial resolution. Therefore, there is an urgent need to develop reliable, reproducible, validated, and affordable molecular imaging tools for preclinical research that can accelerate the drug discovery process and allow thorough evaluation of promising therapeutics. Recently, photoacoustic imaging (PAI) has emerged as a powerful tool in preclinical imaging – with multiple commercial systems available and a growing number of users – where optical contrast can be detected at significant depth (up to 5 cm) with good resolution. However, these end-users lack a reliable, reproducible, and validated molecular PAI platform to complement their translational research. Currently available contrast agents either do not have adequate photostability under the pulsed illumination that is required for PAI, lack sufficient PAI-signal-generation ability for deep imaging, or their absorbance spectra significantly overlap with those of hemoglobin, which reduces imaging contrast.
Our hypothesis is that a commercial-grade PAI platform capable of simultaneous anatomical, functional, and molecular visualization of pathology in small animals will significantly enhance the outcome of fundamental and preclinical research. The central theme of this research project is development, optimization, and validation of a novel class of molecularly specific PAI contrast agents and signal/image processing algorithms to enable reproducible, quantitative, longitudinal, and tomographic imaging of disease and its response to therapy. Our contrast agents are based on antibody-targeted liposomes loaded with J-aggregates of indocyanine green (ICG) dye (Lipo-JICG). Encapsulation of ICG molecules in a liposomal compartment in a form of J-aggregates is associated with highly advantageous properties for in vivo PAI: (i) a strong, narrow absorbance at ~890 nm, where it can be readily unmixed from hemoglobin spectra; (ii) enhancement of PAI signal due to dye-aggregation-mediated increases in thermal gradients and absorbance; (iii) the ability to implement robust, semi-quantitative PAI analysis that does not interfere with imaging of important physiological parameters such as blood oxygen saturation. Our preliminary data show that Lipo-JICG dramatically improves stability and PAI-signal intensity compared to monomeric ICG and silica-coated gold nanorods, with the latter considered the best commercial PAI contrast agent.
Collaborators: Konstantin Sokolov, Ph.D., Jason Cook, Ph.D., Stacy Moulder, M.D., Gaiane Rauch, M.D., Anil Sood, M.D.
Target Phase-change Nanodroplets for Translatable Molecular Ultrasound Imaging of Micromesatses
Determination of malignant spread is the most important prognostic factor for oral-cavity cancer and is critical for the development of a comprehensive treatment plan. Such metastases are the primary driver behind over 6,400 annual deaths and a 66% five-year survival rate for oral-cavity cancer in the United States. As is common with other cancers, oral-cavity cancer cells initially spread to a sentinel lymph node (SLN), which tends to be proximal to the primary cancer site. Thus, accurate and timely assessment of the presence of micrometastases (mMets) in SLNs is critical for correct staging and therapy planning. There is currently no effective preoperative method to assess lymph node involvement from oral-cavity cancers. Despite its significant potential morbidity, elective neck dissection (END) is the gold standard for assessing the presence or absence of lymphatic disease in these patients. In an effort to minimize patient burden, SLN biopsies (SLNB) have been advocated as a less invasive means of achieving accurate detection of mMets. However, the current standard of care for guidance of SLNB procedures, lymphoscintigraphy, is plagued by a number of significant limitations, resulting in reported false-negative rates as high as 32%. Lymphoscintigraphy’s use of a radiotracer requires radiation safety compliance throughout the clinical workflow, which adds additional cost and logistical overhead, while the inherently poor resolution of the gamma camera used to localize this radiotracer during the procedure produces high false-negative rates when the SLN site is too close to the tumor and becomes masked. Because of these significant shortcomings, there is an urgent clinical need for a point-of-care imaging technology that can be safely and noninvasively implemented to reliably detect and adequately visualize SLN mMets in real-time and to obviate the need for highly invasive END procedures.
The goal of this research project is to develop a targeted contrast agent and a noninvasive imaging technique to accurately identify mMets in SLNs with high spatiotemporal resolution. Our hypothesis is that targeted activatable nanodroplets (TANs) can specifically label cancer cells in SLN mMets and can be imaged in real-time with high contrast using a conventional ultrasound (US) imaging system, yielding immediate diagnostic information to facilitate timely and less-invasive treatment (e.g., eliminate END procedures) and/or to intraoperatively guide surgical/biopsy interventions. Our imaging approach, TAN-molecular ultrasound lymphatic (MUSL) imaging, promises to deliver a uniquely versatile contrast agent with a size profile that is remotely changed from nano-scale (e.g., nanodroplet), which is required for molecularly targeted delivery, to micro-scale (i.e., microbubble) for high-contrast, high-resolution imaging with conventional US. This platform has potential to address all disadvantages of the current standard of care by providing a single, cancer-specific, non-radioactive contrast agent that can be noninvasively and clearly visualized using standard clinical US prior to or during surgery. Such technological advancements promise to make “leap” improvements to staging, treatment planning, and therapy guidance of oral-cavity cancer.
Collaborators: Konstantin Sokolov, Ph.D. & Stephen Lai, M.D., Ph.D.
Direct 3D printing of anatomically and functionally mimicking photoacoustic-ultrasound imaging phantoms
Quantitative Imaging Biomarkers (QIB) offer tremendous potential in providing more effective, patient-specific, and rational clinical care. However, translating QIB methods from research tools to clinical practice has proven challenging, in large part because the imaging phantoms needed to support robust quality assurance (QA) programs for these modalities are insufficient. Given the strong dependence of QIBs on the specific functional (e.g., blood flow) and anatomical (e.g., vessel topology) aspects of an interrogated biological system, current imaging phantoms, due to their overly “simplistic design,” do not adequately simulate the wide range of structural complexities and biological variation inherent in a human subject, thus leading to overestimated measurements for precision, particularly reproducibility, and inaccurate assessment of bias. Studies have noted that many physiological (e.g., vessel permeability) and anatomical (e.g., vessel scale and dimensionality) characteristics are not accurately replicated in a current phantoms. The RSNA QIB Alliance (QIBA) Metrology Working Group recently warned that “phantoms do not represent the complexity of human targets,” often resulting in precision overestimation; thus, they claimed that “improved realism of phantoms” is an area “worthy of further investment.” An improved phantom platform is critical for validation and optimization of more established ultrasound-based QIB methods, such as volume blood flow (VBF) imaging, contrast-enhanced US (CEUS), and shear wave speed (SWS) elasticity imaging, while such a platform would also be of tremendous value in the development of newer QIB approaches, such as super-resolution imaging, acoustic angiography, or photoacoustic-ultrasonic oxygen saturation imaging. To meet these needs, the next imaging phantom platform should be capable of emulating the scale, tortuosity, density, and functionality of vasculature and tissue backscatter heterogeneity, viscoelasticity, and anisotropy that is characteristic of human biology.
The overall goal of this research project is to develop a 3D printing platform that is capable of making more anatomically and functionally realistic phantoms for ultrasound-mediated imaging. Our hypothesis is that our 3D printing platform can deliver more robust validation of existing QIBs, provide powerful developmental tools for optimizing new such approaches, and offer biologically realistic environments for more effective clinician training. To this end, we propose using projection-based stereolithography with biocompatible tartrazine as a photoabsorber to print poly(ethylene glycol) (PEG)-based hydrogels to create stable phantoms for ultrasound-mediated imaging with locally controlled elasticity and backscatter and that contain realistic/tortuous, flow-supporting vessels at the scale of arterioles/venules that support oxygen transport to extraluminal cells.
Collaborator: Jordan Miller, Ph.D.