Cancer kills more Americans under the age of 75 than any other disease. Although most cancers occur in epithelial surfaces that can be directly visualized, the majority of cases are detected at an advanced stage. Optical imaging and spectroscopy may provide a solution to the need for non-invasive and effective early detection tools. These technologies are capable of examining tissue over a wide range of spatial scales, with widefield macroscopic imaging typically spanning several square-centimeters, and high resolution in vivo microscopy techniques enabling cellular and subcellular features to be visualized. This work presents novel technologies in two important areas of optical imaging: high resolution imaging and widefield imaging. For subcellular imaging applications, new high resolution endomicroscope techniques are presented with improved lateral resolution, larger field-of-view, increased contrast, decreased background signal, and reduced cost compared to existing devices. A new widefield optical technology called multi-modal spectral imaging is also developed. This technique provides real-time in vivo spectral data over a large field-of-view, which is useful for detecting biochemical alterations associated with neoplasia. The described devices are compared to existing technologies, tested using ex vivo tissue specimens, and evaluated for diagnostic potential in a multi-patient oral cancer clinical trial.

Thesis (M.A.)--Boston University, 2011. / OBJECTIVE: Thyroid cancer is the most common endocrine malignancy and patients presenting with thyroid nodules often undergo surgery solely for diagnostic purposes. The goal of our study was to examine the accuracy of Elastic Scattering Spectroscopy (ESS) in distinguishing between benign and malignant thyroid nodules in fresh ex vivo specimens and to design an in vivo ESS probe and device, manufacture it and conduct a clinical trial. METHODS: Patients already undergoing thyroidectomy surgery were consented for the ex vivo study. ESS data was obtained from ex vivo specimens by recording 5 readings per nodule with five repetitive readings per each site. Final pathology reports were used to confirm the diagnosis. The spectra were analyzed using principal component analysis, linear discriminant analysis and leave one out technique. The in vivo ESS study was conceptually designed and IRB approval from Boston Medical Campus was obtained. RESULTS: The ex vivo study showed that ESS could predict the difference between benign and malignant tumors with a sensitivity of 74%, specificity of 90%, positive predictive value of 82% and negative predictive value of 85%. 193 spectra were analyzed from 64 patients, 120 spectra were from benign nodules and 73 from malignant nodules. Subanalysis examined only indeterminate nodules showed sensitivity of 65%, specificity of 79%, PPV 77% and NPV 67%. The in vivo ESS probe was designed and 12 identical instruments were manufactured. Initial experimental readings were taken and parameters were adjusted for the in vivo tissue environment. The clinical trial is underway. CONCLUSIONS: ESS is a practical tool that can accurately identify malignancy in ex vivo thyroid specimens with high specificity and sensitivity. Initial in vivo experimental trials have been conducted and show promise for similar results.

Surface acoustic waves (SAWs) have a large number of applications and the majority of them are in the sensor and actuator fields targeted to satisfy market needs. Recently, researchers have focused on optimizing and improving device functions, sensitivity, power consumption, etc. However, SAW actuators and sensors still cannot replace their conventional counterparts in some mechanical and biomedical areas, such as actuators for liquid pumping under microfluidic channels and sensors for real-time cell culture monitoring. The two objectives of this dissertation are to explore the potential of piezoelectric materials and surface acoustic waves for research on actuators and sensors in the mechanical pump and biosensor areas. Manipulation of liquids in microfluidic channels is important for many mechanical, chemical and biomedical applications. In this dissertation, we first introduced a novel integrated surface acoustic wave based pump for liquid delivery and precise manipulation within a microchannel. The device employed a hydrophobic surface coating (Cytop) in the device design to decrease the friction force and increase the bonding. Contrary to previous surface acoustic wave based pumps which were mostly based on the filling and sucking process, we demonstrated long distance media delivery (up to 8mm) and a high pumping velocity, which increased the device’s application space and mass production potential. Additionally, the device design didn’t need precise layers of water and glass between substrate and channel, which simplified the design significantly. In this study, we conducted extensive parametric studies to quantify the effects of the liquid volume pumped, microchannel size, and input applied power as well as the existence of hydrophobic surface coating on the pumping velocity and pump performance. Our results indicated that the pumping velocity for a constant liquid volume with the same applied input power could be increased by over 130% (2.31 mm/min vs 0.99 mm/min) by employing a hydrophobic surface coating (Cytop) in a thinner microchannel (250 µm vs 500 µm) design. This device could be used in circulation, dosing, metering and drug delivery applications which necessitated small-scale precise liquid control and delivery. This dissertation also introduced a novel SAW-based sensor designed and employed for detecting changes in cell concentration. Before conducting cell concentration experiments, preliminary experiments were conducted on weight concentration differentiation of microfluidic particles based on a polydimethylsiloxane (PDMS) channel and surface acoustic wave resonator design. The results confirmed that our device exerted an ultra-stable status to detect liquid properties by monitoring continuous fluids. An improved design was carried out by depositing a 200 nm ZnO layer on top of the lithium tantalate substrate surface increased the sensitivity and enabled cell concentration detection in a microfluidic system. Comprehensive studies on cell viability were carried out to investigate the effect of shear horizontal (SH) SAWs on both a cancerous (A549 lung adenocarcinoma) and a non-cancerous (RAW264.7 macrophage) cell line. Two pairs of resonators consisting of interdigital transducers (IDTs) and reflecting fingers were used to quantify mass loading by the cells in suspension media as well as within a 3-dimensional cell culture model. In order to predict the characteristics and optimize the design of the SH-SAW biosensor, a 3D COMSOL model was built to simulate the mass loading response of the cell suspensions. These results were compared to experimental data generated by pipetting cell concentrations of 3.125K, 6.25K 12.5K, 25K and 50K cells per 100µL into the PDMS well and measuring to obtain the relative frequency shift from the two oscillatory circuit systems (one of which functioned as a control). Frequency shift measurements were also collected from A549 cells cultured on a 3D nanofiber scaffold produced by electrospinning to evaluate the device’s ability to detect changes in cell density as the cells proliferated in culture over the course of eight days. The device’s ability to detect changes in cell density over time in a 3D model along with its biocompatibility reveal great potential for this device to be incorporated into 3D in vitro cancer research applications.

A microwave imaging system processes scattered electromagnetic fields in the microwave region to create images. It is an alternative or complementary imaging tool that can be used in breast cancer imaging. It employs non-ionising radiation and during measurement, compression of the scanned body part is avoided. These benefits potentially lead to safer and more comfortable examinations. It also has the potential to be both sensitive and specific to detect small tumors, whilst being much lower cost than current methods, such as magnetic resonant imaging, mammography and ultrasound. This thesis reports a multi-layer active antenna array for breast imaging using microwaves from 3 GHz to 8 GHz. The proposed structure resolves the outstanding problem in the design of large active antenna arrays for tissue imaging, namely, the isolation of the antennas from the electronic circuits. A ground plane within the multi-layer design separates the antenna array from the electronics array while providing shielding to the antennas from the back and improved power coupling into the tissue. The possibility of a high-speed vertical connector to provide interconnection between the antenna array and the mixer array is investigated and measurements show that it could be utilized for the frequency range from 3 GHz to 8 GHz. / Thesis / Master of Applied Science (MASc)

No description available.

No description available.

The abundance of next-generation sequencing (NGS) data has encouraged the adoption of machine learning methods to aid in the diagnosis and treatment of human disease. In particular, the last decade has shown the extensive use of predictive analytics in cancer research due to the prevalence of rich cellular descriptions of genetic and transcriptomic profiles of cancer cells. Despite the availability of wide-ranging forms of genomic data, few predictive models are designed to leverage multidimensional data sources. In this paper, we introduce a deep learning approach using neural network based information fusion to facilitate the integration of multi-platform genomic data, and the prediction of cancer cell sub-class. We propose the dGMU (deep gated multimodal unit), a series of multiplicative gates that can learn intermediate representations between multi-platform genomic data and improve cancer cell stratification. We also provide a framework for interpretable dimensionality reduction and assess several methods that visualize and explain the decisions of the underlying model. Experimental results on nine cancer types and four forms of NGS data (copy number variation, simple nucleotide variation, RNA expression, and miRNA expression) showed that the dGMU model improved the classification agreement of unimodal approaches and outperformed other fusion strategies in class accuracy. The results indicate that deep learning architectures based on multiplicative gates have the potential to expedite representation learning and knowledge integration in the study of cancer pathogenesis. / Thesis / Master of Science (MSc)

Conventional methods of cancer therapy have been severely limited by inefficient delivery of therapeutic doses without incidence of harsh and toxic side effects in normal tissues. Consequently, countless new methods for early detection and drug delivery have been investigated in the area of nanoparticles and hydrogels. Although many of these methods are promising, the complex nature of cancer increases the difficultly for the development of the perfect system. Cellulose nanocrystals (CNCs) have been studied widely for a variety of applications. Despite their advantages, investigations of their abilities in the biomedical field have not been explored. The goal of this project is to delve into the potential uses of CNCs in detection, targeted drug delivery, and potentiation of irreversible electroporation (IRE)-induced cell death in folate receptor (FR)-positive cancers. To accomplish this task we have prepared stable and reproducible CNCs from wood pulp via sulfuric acid hydrolysis. Furthermore, we have functionalized the surface of these nanoparticles and conjugated them with the targeting ligand folic acid (FA) and the fluorescent imaging agent fluorescein-5\'-isothiocyanate (FITC) to create FITC-CNC-FA; CNCs have also been conjugated with doxorubicin (DOX), a potent chemotherapeutic (DOX-ALAL-CNC-FA). We have determined FITC-CNC-FA's and DOX-ALAL-CNC-FA's ability to specifically target FR-positive cancer cells in vitro; meanwhile non-targeted CNCs (FITC-CNC) were shown unable to bind to these cell types. In addition, we have investigated FITC-CNC-FA's pharmacokinetic activity in vivo. To properly model the CNC conjugate's activity in vivo, a physiologically based pharmacokinetic (PBPK) model has been constructed. We have also examined CNCs' ability to potentiate a new technique for tumor ablation, IRE. Pre-incubation with FA-conjugated CNCs (CNC-FA) have shown an increase in cytotoxicity in FR-positive cancer cells induced by IRE. In addition, CNC-FA did not potentiate IRE-induced cytotoxicity in a FR-negative cancer cell type. For a more comprehensive understanding of CNC-FA's ability to potentiate IRE induced cytotoxicity, we optimized a 3D in vitro hydrogel system. Preliminary data suggest this method of experimentation will be more realistic to in vivo studies to be completed in the future. Together, these studies showcase CNCs as efficient and effective nano-carriers in tumor detection and treatment. / Ph. D.

We characterized the thermoacoustic and photoacoustic properties of large-area, few-layer graphene by pulsed microwave and optical excitations. Due to its high electric conductivity and low heat capacity per unit area, graphene lends itself to excellent microwave and optical energy absorption and acoustic signal emanation due to the thermoacoustic effect. When exposed to pulsed microwave or optical radiation, distinct thermoacoustic and photoacoustic signals generated by the few-layer graphene are obtained due to microwave and laser absorption of the graphene, respectively. Clear thermoacoustic and photoacoustic images of large-area graphene sample are achieved. A numerical model is developed and the simulated results are in good accordance with the measured ones. This characterization work may find applications in ultrasound generator and detectors for microwave and optical radiation. It may also become an alternative characterization approach for graphene and other types of two-dimensional materials. (C) 2016 AIP Publishing LLC.

Early detection of cancer is paramount for improved patient survival. This dissertation presents work developing imaging techniques to improve cancer diagnostics and detection utilizing light and microwave induced thermoacoustic imaging. In the second chapter, the well-established pre-clinical mouse window chamber model is interfaced with simultaneously acquired high-resolution pulse echo (PE) ultrasound and photoacoustic (PA) imaging. Co-registered PE and PA imaging, coupled with developed image segmentation algorithms, are used to quantitatively track and monitor the size, shape, heterogeneity, and neovasculature of the tumor microenvironment during a month long study. Average volumetric growth was 5.35 mm³/day, which correlated well with two dimensional results from fluorescent imaging (R = 0.97, p < 0.01). Spectroscopic PA imaging is also employed to probe the assumed oxygenation status of the tumor vasculature. The window chamber model combined with high-resolution PE and PA imaging could form a powerful testbed for characterizing cancers and evaluating new contrast and therapeutic agents. The third chapter utilizes a clinical ultrasound array to facilitate fast volumetric spectroscopic PA imaging to detect and discriminate endogenous absorbers (i.e. oxy/deoxygenated hemoglobin) as well as exogenous PA contrast agents (i.e. gold nanorods, fluorophores). In vivo spatiotemporal tracking of administered gold nanorods is presented, with the contrast agent augmenting the PA signal 18 dB. Furthermore, through the use of spectral unmixing algorithms, the relative concentrations of multiple endogenous and exogenous co-localized absorbers were reconstructed in tumor bearing mice. The concentration of Alexaflour647 was calculated to increase nearly 20 dB in the center of a prostate tumor after a tail-vein injection of the contrast agent. Additionally, after direct subcutaneous injections of two different gold nanorods into a breast tumor, the concentration of each nanoparticle was discriminated in vivo with a signal-to-noise ratio of greater than 25 dB. This technique has great potential for improved early cancer detection and individualized cancer treatment through advanced pharmacokinetic monitoring of therapeutic agents. Finally, the fourth chapter presents significant improvements made to enhance breast cancer detection with thermoacoustic (TA) imaging. In a breast cancer simulating phantom, the initial demonstration of TA spectroscopy (TAS) is used to detect and discriminate relative water / fat composition based solely on the sample's intrinsic spectral absorption. The slope of the TA signal was highly correlated with that of the absorption coefficient (R² = 0.98, p < 0.01), indicating TAS can distinguish materials based on their dielectric properties. Furthermore, the use of carbon nanotubes as a potential TA contrast agent is explored. These nanoparticles significantly enhance the magnitude of the TA signal (8 dB larger than water), and also demonstrate unique absorption spectra. Finally, short microwave pulses (Δt ≥ 10 ns) are achieved through novel microwave hardware, and used to generate high-frequency TA signals. In conclusion, this section presents advancements made to the sensitivity, contrast, and resolution of TA imaging. Overall, this dissertation presents enhancements made to the diagnostic capabilities of PA and TA imaging for improved detection and characterization of cancer.