Browsing by Subject "Applied physics"

Purpose: This project aimed to establish a method for computing 3D cardiac motion given continuous 2D-Cine MRI frames as the inputs. This approach would be useful for continuously monitoring cardiac and respiratory motion during MR-guided cardiac radiation therapy, and thus supporting radiation delivery guidance and gating.Methods: 5D-MRI datasets of seven patients, with each consisting of 3D spatial volumes of the cardiac cycle and respiratory cycle, were used for quantitative evaluation of the heart motion due to respiratory and cardiac movements. This was achieved through deformable image registration (DIR). Subsequently, principal component analysis (PCA) was performed on the computed deformation vector fields (DVF) to extract scores that effectively represent the characteristics of the DVFs. A deep learning model was then trained to predict the cardiac motion PCA scores given the inputs of 2D-Cine MRI. The predicted PCA scores were then transformed into 3D DVFs, which were then used to track 3D target motion. Results: The model’s performance was quantitatively evaluated on ground truth data that were withheld from model training. Across all 7 subjects, the average 3D DVF prediction errors for the heart region consistently remained around 0.3 ± 0.1mm. The predicted target motion, computed from the predicted DVFs, was visually evaluated, and found to be satisfactory. Conclusion: The developed method demonstrated promising potential in accurately computing and tracking real-time 3D cardiac motion given 2D-Cine MRI inputs. This approach presents a viable solution for continuously monitoring the 3D cardiac and respiratory motion of the heart during MR-guided cardiac radiation therapy.

Introduction: VMAT treatment planning requires time-consuming DVH-based inverse optimization process, which impedes its application in time-sensitive situations. This work aims to develop a deep-learning based algorithm, Automatic VMAT Planning via MLC Dynamic Sequence Prediction (AVP-DSP), for rapid prostate VMAT treatment planning.

Methods: AVP-DSP utilizes a series of 2D projections of a patient’s dose prediction and contour structures to generate a single 360º dynamic MLC sequence in a VMAT plan. The backbone of AVP-DSP is a novel U-net implementation which has a 4-resolution-step analysis path and a 4-resolution-step synthesis path. AVP-DSP was developed based on 131 previous prostate patients who received simultaneously-integrated-boost (SIB) treatment (58.8Gy/70Gy to PTV58.8/PTV70 in 28fx). All patients were planned by a 360º single-arc VMAT technique using an in-house intelligent planning tool in a commercial treatment planning system (TPS). 120 plans were used in AVP-DSP training/validation, and 11 plans were used as independent tests. Key dosimetric metrics achieved by AVP-DSP were compared against the ones planned by the commercial TPS.

Results: After dose normalization (PTV70 V70Gy=95%), all 11 AVP-DSP test plans met institutional clinic guidelines of dose distribution outside PTV. Bladder (V70Gy=6.8±3.6cc, V40Gy=19.4±9.2%) and rectum (V70Gy=2.8±1.8cc, V40Gy=26.3±5.9%) results in AVP-DSP plans were comparable with the commercial TPS plan results (bladder V70Gy=4.1±2.0cc, V40Gy=17.7±8.9%; rectum V70Gy=1.4±0.7cc, V40Gy=24.0±5.0%). 3D max dose results in AVP-DSP plans(D1cc=118.9±4.1%) were higher than the commercial TPS plans results(D1cc=106.7±0.8%). On average, AVP-DSP used 30 seconds for a plan generation in contrast to the current clinical practice (>20 minutes).

Conclusion: Results suggest that AVP-DSP can generate a prostate VMAT plan with clinically-acceptable dosimetric quality. With its high efficiency, AVP-DSP may hold great potentials of real-time planning application after further validation.

There always exists a demand to accelerate the time-consuming MRI acquisition process. Many methods have been proposed to achieve this goal, including deep learning method which appears to be a robust tool compared to conventional methods. While many works have been done to evaluate the performance of neural networks on standard anatomical MR images, few attentions have been paid to accelerating other less conventional MR image acquisitions. This work aims to evaluate the feasibility of neural networks on accelerating Brain DTI and Gynecological Brachytherapy MRI. Three neural networks including U-net, Cascade-net and PD-net were evaluated. Brain DTI data was acquired from public database RIDER NEURO MRI while cervix gynecological MRI data was acquired from Duke University Hospital clinic data. A 25% Cartesian undersampling strategy was applied to all the training and test data. Diffusion weighted images and quantitative functional maps in Brain DTI, T1-spgr and T2 images in GYN studies were reconstructed. The performance of the neural networks was evaluated by quantitatively calculating the similarity between the reconstructed images and the reference images, using the metric Total Relative Error (TRE). Results showed that with the architectures and parameters set in this work, all three neural networks could accelerate Brain DTI and GYN T2 MR imaging. Generally, PD-net slightly outperformed Cascade-net, and they both outperformed U-net with respect to image reconstruction performance. While this was also true for reconstruction of quantitative functional diffusion weighted maps and GYN T1-spgr images, the overall performance of the three neural networks on these two tasks needed further improvement. To be concluded, PD-net is very promising on accelerating T2-weighted-based MR imaging. Future work can focus on adjusting the parameters and architectures of the neural networks to improve the performance on accelerating GYN T1-spgr MR imaging and adopting more robust undersampling strategy such as radial undersampling strategy to further improve the overall acceleration performance.

Purpose: To validate commissioning results associated with the Xstrahl Small Animal Radiation Research Platform (SARRP) installed at Duke University in October 2021, verify the accuracy of the Xstrahl Point Dose Calculator (PDC) and MuriPlan dose calculation in simple geometries, and design and characterize a novel in-house kV mini-GRID capability on the SARRP.Methods: After installation at Duke University, Xstrahl SARRP TG-61 output was measured for independent verification using a Farmer ion chamber at reference conditions (33 cm SSD, 2 cm depth, open field, 220 kVp, 13 mA). Half-value layer was measured using the same ion chamber, with copper sheets to vary thickness. The accuracy of the PDC was investigated in simple water and bolus stack phantoms using EBT3 film. A range of field sizes (10x10, 20x20, 30x30, 40x40, 10x20, 20x10, 15x40, 10x40, 30x70 mm2) and depths (1 cm, 2 cm) were spot-checked. MuriPlan simulations of treatment delivery to the bolus phantom and water phantom were compared to results of EBT3 film measurements. Metal Oxide Semiconductor Field Effect Transistor (MOSFET) detectors were also used for independent verification, with detectors being embedded within a tissue-equivalent mouse phantom at 1 cm depth. GRID irradiations were performed with the SARRP, using a 220 kVp beam, 13 mA, and a 40 mm x 40 mm field size. Pencil and bar GRIDs with beamlet spacings of 1 mm and 1.25 mm were characterized by first inserting GRID into a 3D-printed mount and positioning the mount on top of a PLA plastic block, surrounded by distilled water. EBT3 films were infixed to the top level of the PLA block and positioned at isocenter. PDC was utilized to determine irradiation time. The beam was turned on for 102 seconds—sufficient time to administer 6 Gy with a 40 mm x 40 mm field size to the surface film at isocenter with no GRID blocking the beam. EBT3 film results were analyzed to determine the output factors, peak-to-valley ratios, integral dose relative to open field, relative dose maps, as well as to produce dose volume histograms for each GRID. Results were compared to GRID characterizations in Johnson et al [18]. GRID characterizations were used to inform experimental plan for pre-clinical trial evaluating treatment efficacy of GRID therapy with PD-L1 checkpoint inhibition compared to conventional radiation therapy. Results: TG-61 dose rate and half-value layer measured during on-site commissioning showed excellent agreement with Xstrahl factory commissioning results (≈1% difference). The PDC and MuriPlan dose calculation predicted results for field sizes and depths demonstrated acceptable agreement with actual results measured by EBT3 film (.2% to 12%), with exception of several outliers. Using EBT3 film dosimetry for verification (tissue-equivalent bolus medium), MuriPlan simulations were within 2% and 12% difference from the film measured dose for 5/7 field sizes in the bolus phantom and within 3% and 13% for 4/5 field sizes in the water phantom. MOSFET detector measurements using the mouse phantom demonstrated improved agreement with the PDC-predicted dose, with percent errors ranging from .12% to 5.97% (with a single outlier at 18.3%). Measured output factors using the SARRP for the 20 mm x 20 mm pencil GRIDs were .77 ± .03 and .74 ± .02 (1 mm and 1.25 mm beamlet spacing, respectively). For the bar GRIDs, these values were evaluated to be .83 ± .03 and .80 ± .03 (1 mm and 1.25 mm beamlet spacing, respectively). Peak:valley ratios for the 1 mm and 1.25 mm beamlet spacing pencil GRIDs were determined to be 24.5 ± 0.6 and 25.1 ± 1.3, respectively. Peak:valley ratios for the 1 mm and 1.25 mm beamlet spacing bar GRIDs were found to be lower than for pencil GRIDs with equivalent beamlet spacing, being evaluated to be 13.2 ± 1.1 and 18.5 ± 1.2, respectively. Output factors, peak:valley ratios, integral dose relative to open field, and dose volume histograms for the pencil and bar GRIDs using the SARRP largely corroborated the results of Johnson et al in terms of experimental trends (peak:valley ratios being higher for pencil GRIDs and increasing with increasing beamlet spacing, output factors decreasing with increasing beamlet spacing for both GRID types, and decreasing integral dose with increasing beamlet spacing for pencil GRIDs and increasing integral dose with increasing beamlet spacing for bar GRIDs). 4.67% to 30.5% difference was observed for experimentally measured peak:valley ratios relative to the results for the same GRIDs in Johnson et al.. Better agreement was demonstrated in GRID output factor measurements (≈0% to 14%). Integral dose experimental measurements demonstrated exceptional agreement with Johnson et al.., with percent differences ranging from 1% to 2.1%. These measured differences are likely a result of using the SARRP versus the XRAD 225Cx used in Johnson et al, but lend significant credence to reproducibility of results found using the XRAD 225Cx. Conclusions: The PDC and MuriPlan computations provide an effective estimate of the exposure time necessary to deliver dose for corresponding MVC field sizes and depths (within 6% error using the MOSFET for verification). EBT3 film was determined to be an unreliable measure of SARRP dose delivery; MOSFET detectors demonstrated more consistency and effectiveness for treatment planning verification. Xstrahl’s SARRP was able to replicate the kV mini-GRID capabilities of the XRAD 225Cx used in Johnson et al. and can be used for mini-GRID characterizations and preclinical mouse trials.

In urban, dense, and/or environmentally sensitive regions, underground high-pressure fluid-filled (HPFF) transmission cables are used to transport electricity at high voltages to prevent electrical losses. With age and use, these HPFF cables degrade and can leak petroleum-based dielectric fluid (DF) into the surrounding environment. Maintaining adequate DF in these cables is required for safe and reliable operation. Currently, detecting and locating underground DF leaks is challenging, time intensive, and inconsistent. One leak location method exists utilizing perfluorocarbon (PFC) tracer molecules and a mobile gas chromatograph. This instrument is sensitive enough to detect the atmospheric background levels of PFC in the ppqv (parts per quadrillion by volume, 10-12) range; however, this instrument is prone to analyte saturation, is not fully portable, nor does it produce real-time results.

An improved mobile and highly sensitive PFC detection method is required. A cycloidal coded aperture miniature mass spectrometers (C-CAMMS) could be such a method due to its integration of three core technologies that help to overcome the throughput versus resolution tradeoff that has historically hampered the miniaturization of mass spectrometers. The C-CAMMS prototype combines: a cycloidal mass analyzer, aperture coding, and a focal plane array detector, to enable a mobile instrument capable of detecting PFC tracers with high resolution, sensitivity, and selectivity. The cycloidal mass analyzer utilizes perpendicularly-oriented overlapping electric and magnetic fields to linearly separate ions by mass-to-charge ratio (m/q). The C-CAMMS platform uses aperture coding to increase throughput without sacrificing resolution which commonly occurs during miniaturization. Finally, a capacitive transimpedance amplifier (CTIA) array ion detector offers sensitive, simultaneous ion detection across a wide mass range.

A detailed understanding of the detection process for this instrument was obtained through extensive simulations and experiments. From this knowledge, a set of design considerations and a six-step design approach have been established for reproducibly developing a cycloidal mass analyzer that uses a focal plane array detector. This design knowledge relates the magnetic and electric field homogeneity of the cycloidal mass analyzer to the performance of the complete mass spectrometer. The efficacy of this design roadmap is demonstrated by designing the PFC-CAMMS instrument for the specific use case of PFC tracer detection and location.

The design knowledge was implemented to create a PFC-CAMMS instrument with high resolution, high sensitivity, a large mass range, and a small form factor. PFC-CAMMS achieved a resolution of 0.071 u at m/q 69 and 0.27 u at m/q 331 and demonstrated an overall mass range of m/q 29 – 502. Additionally, using PMCH (C7F14) as the test molecule, the PFC-CAMMS instrument achieved a detection limit of 20 ppbv (parts per billion by volume, 10-9) with a response time of less than 5 s from sample introduction using a capillary inlet. There is no evidence of a hysteresis effect when detecting PFCs for this benchtop (66 x 46 x 30 cm3 and ~50 kg) laboratory prototype. This work describes not only the understanding that was generated about designing electromagnetic components for a PFC-CAMMS, but also explains the fabrication, alignment, and assembly details that enable this improved baseline performance. The reproducible procedure for developing a cycloidal mass analyzer with an array detector facilitates improved hardware designs that produce a more consistent system response function across the intended mass range. Following reconstruction of the coded mass spectrum, the PFC-CAMMS instrument can overcome the throughput vs. resolution tradeoff due to its stronger yet more uniform field.

Project 1 (Chapter 2): Dose Verification andMonte Carlo Modeling of an Image-GuidedSmall Animal Radiotherapy Irradiator

Purpose: Preclinical trials play a crucial role in advancing the understanding of cancer biology and developing effective therapeutic interventions. The purpose of this project is to simulate and validate beam output of a Small Animal Radiotherapy Research Platform (SARRP, xStrahl) with both physical dosimetry and Monte Carlo simulation models.

Materials and Methods: The SARRP console was set up to deliver an intended dose of 8Gy (4 Gy anterior-posteriorly (AP), 4 Gy posterior-anteriorly (PA)) in 142 seconds to a flat mouse phantom. The x-ray irradiation parameters were set to 13 mA, 220 kVp, with a 33.725 cm source to surface distance. Beam filtration included 0.8 mm Be (inherent) and 0.15 mm Cu (added), with collimation set to 40x30 mm. Dose verification was conducted through two methods: utilizing an energy-calibrated MOSFET dosimeter and employing Monte Carlo Simulations using Monte Carlo N- Particle Transport (MCNP). MOSFET Calibration encompassed four setups to ensure precision. The first two involved calibrating the MOSFET with an ion chamber in air at 0 degrees (Setup 1) and 180 degrees (Setup 2). The subsequent two setups calibrated the MOSFET positioned inside the phantom (Setups 3 and 4) with an ion chamber in air. After the calibrations, the MOSFET, placed inside the phantom, received the intended 4 Gy dose for verification. The MCNP simulation comprised two stages: a point source simulation and a simulation of the x-ray tube. For the point source, the SARRP geometry was replicated, with the x-ray tube modeled as a collimated point source. The x-ray tube simulation entailed modeling components of the xray tube. Validation methods included comparing energy spectra, Half Value Layer (HVL) testing, and film analysis of the anode heel effect.

Results: In the dose verification, Setup 1 exceeds the intended 4 Gy dose by 7.72%, while Setup 2 underdoses by 2.96%, resulting in a cumulative overdose of 2.38% for Setups 1 and 2. Setup 3 aligns with the intended 4 Gy dose, underdosing slightly by 0.14%, while Setup 4 underdoses by 2.31%. The cumulative dose for Setups 3 and 4 totals 7.90 Gy, indicating a 1.27% underdose. The two calibration techniques demonstrate a difference of 3.6%. Calibration in air is the preferred method due to the ionization chamber also being present in air. Point Source Simulation yielded doses of (4.27 ± 0.02) Gy (AP) and(3.77 ± 0.02) Gy (PA). X-ray Tube Simulation resulted in (3.95 ± 0.02) Gy (AP). Energy spectrum of the MCNP model showed good agreement with the manufacturer model in key spectral characteristics (peaks, mean energies). HVL comparison showed good agreement with only a 0.5% difference between simulated and experimental half value thicknesses. The anode-heel effect analysis was inconclusive.

Conclusions: The dose verification processes establish the SARRP’s efficacy in delivering the intended radiation dose. The integration of advanced measurement techniques set a benchmark for small animal dosimetry and ultimately strengthens the reliability of radiation doses in preclinical studies.

Project 2 (Chapter 3): Investigation of Occupational Radiation Exposure to InterventionalRadiologists from Use of Fluoroscopic Imaging

Purpose: The purpose of this project is to investigate radiation exposure among Interventional Radiology (IR) physicians using fluoroscopic imaging through experimental data collection and retrospective analysis, with objectives to understand Automatic Exposure Rate Control mechanisms, assess exposure rates to operators, and identify trends amongIR physicians.

Materials and Methods: Three interventional fluoroscopes were investigated: Philips AlluraClarity Xper FD 20/15, Philips Allura Xper FD20, and GE Discovery IGS 740. Two phantoms were employed to replicate patient and operator. The “patient” phantom was comprised of water-equivalent slabs (5cm thick, 30cm x 30cm). The “operator” phantom (Atom Dosimetry Labs Adult Male phantom) was placed beside the table and covered with a 0.25 mm lead apron. A Ludlum 9DP pressurized ion chamber was positioned at collar level of the operator phantom. Parameters varied included patient thickness (20cm-40cm), collimation, and fluoroscopy and acquisition modes. Exposure (mR) to “operator”were measured and normalized to number of x-ray pulses. Retrospective analysis used Radiation Dose Structured Report data for an 8-month period. Physician caseload, averageCumulative Air Kerma (CAK) by physician, and total CAK by physician were determined.

Results: Based on measured data, acquisition mode exhibits longer pulse widths (7.20x-31.3x) and higher tube current (1.48x-7.58x) compared to fluoroscopy for all units. Tube voltage increases with phantom thickness in both fluoroscopy and acquisition mode for all units. Increasing phantom size and collimated field size elevate the operator exposure rate for all C-arms. Larger field sizes contribute to higher exposure rates compared to small field sizes (4.91x-6.27x fluoroscopy, 5.51x-8.23x acquisition). Additionally, acquisition mode contributes to higher exposure rates (23.4x–107.1x) than fluoroscopy. Analysis of proceduraldata identified trends in case distribution and dose to patients across physicians.

Conclusions: Overall, patient size, collimation settings, and fluoroscopy vs. acquisition mode were identified as significant contributors to operator exposure rates. Outliers among IR physicians highlighted the need for targeted interventions to mitigate excessive radiation exposure.

Oropharyngeal Squamous Cell Carcinoma (OPSCC) is by far the most predominant form of head and neck cancer in the United States. The survival rate for OPSCC is very high, which, while fortunate, yields many patients who are left with the late term toxicities consequent of their treatment. This project aimed to use patient-reported outcome (PRO) data from two sources – the PRO-CTCAE and the QLQ-C30 – along with the dosimetric data of patients that have already been treated, in order to characterize retrospectively a relationship between patient dosimetric data and the severity of response of PRO data. In particular, PRO data was used as a way to characterize the severity of patient-experienced xerostomia and dysphagia. Additionally, this data was used to fit the radiobiological parameters for two normal tissue complication probability (NTCP) models: the Lyman-Kutcher-Burman (LKB) model, and the Relative Seriality (RS) model. Overall, it was found that the PRO-CTCAE data was more robust than the QLQ-C30 data in its characterization. Based on the PRO-CTCAE data, the V52 (volume which receives at least 52 Gy) of the combined constrictors and the V59 of the superior pharyngeal constrictor show the strongest relationship with patient-reported dysphagia. Additionally, the V27 of the contralaterals and the V12 of the contralateral parotid show the strongest relationship with patient-reported xerostomia. Furthermore, it was found that the dose response curves for both NTCP models fit the data with similar accuracy.

Trapped ions provide a promising platform to build a practical quantum computer. Scaling the high performance of small systems to longer ion chains is a technical endeavor that benefits from both better hardware system design and gate-level control techniques. In this thesis, I discuss our work on building a small-scale trapped-ion quantum computing system that features stable laser beam control, low-crosstalk individual addressing and capability to implement high-fidelity multi-qubit gates.

We develop control techniques to extend the pack-leading fidelity of entangling gates in two-ion systems to longer chains. A major error source limiting entangling gate fidelities in ion chains is crosstalk between target and neighboring spectator qubits. We propose and demonstrate a crosstalk suppression scheme that eliminates all first-order crosstalk utilizing only local control of target qubits, as opposed to an existing scheme which requires control over all neighboring qubits. Using the scheme, we achieve a $99.5\%$ gate fidelity in a 5-ion chain. Complex quantum circuits can benefit from native multi-qubit gates such as the $N$-Toffoli gate, which substantially reduce the overhead cost from performing universal decomposition into single- and two-qubit gates. We take advantage of novel performance benefits of long ion chains to realize scalable Cirac-Zoller gates, which uses a simple pulse sequence to efficiently implement $N$-Toffoli gates. We demonstrate the Cirac-Zoller 3- and 4-Toffoli gates in a five-ion chain with higher fidelities than previous results using trapped ions. We also present the first experimental realization of a 5-Toffoli gate.

Purpose: FLASH irradiation shows strong potential for clinical applications, offering tumor control comparable to conventional irradiation with lower levels of normal tissue toxicity. This combination of effects, the FLASH effect, could widen the therapeutic gap and improve the effectiveness of radiation therapy treatments of cancer and other diseases. While the FLASH effect is typically seen at mean dose rates (MDRs) above 40 Gy/s, the exact conditions for it are unknown. Furthermore, the underlying mechanisms are unclear, with current theories suggesting that FLASH irradiation induces transient hypoxia in normal tissue and causes reduced DNA damage. Duke University is in a unique position to investigate both the conditions and mechanisms behind the FLASH effect through the High Intensity Gamma-ray Source (HIGS) linear accelerator at the Triangle Universities Nuclear Laboratory (TUNL). This FLASH source has been combined with a unique ex vivo rat brain slice organotypic model to create a novel FLASH experimental platform. This platform has been demonstrated to reduce tumor burden, a key component of the FLASH effect. However, the normal tissue sparing portion of the FLASH effect has not been explored. The purpose of this work is to assess independent methods for measuring normal tissue health in the rat brain slice organotypic model and determine if a normal tissue sparing effect is present in this experimental setup.

Methods: Two main experiments were conducted: an experiment using the HIGS linac, and an experiment using a Varian 2100EX clinical linac to replicate the HIGS experiment at a conventional clinical dose rate. For the HIGS irradiation, nine well-plates, each with eight 350-micron thick rat brain slices, were divided into one unirradiated (No IR) and two experimental arms. Plates for each of the experimental arms were irradiated with 4 pulses of 35 MeV electrons from the HIGS linac. One experimental arm, the FLASH arm, was irradiated with 0.15 seconds in-between pulses; the other arm, the non-FLASH arm, was irradiated with 10 seconds in-between pulses. EBT-XD film was scanned using an EPSON 11000XL scanner to determine the dose delivered to each slice. Each treatment arm had a plate dedicated to one of three independent normal tissue health assays: yellow fluorescent protein (YFP), immunofluorescence, or cytokines. Depending on the assay, the slices in the plate either underwent YFP transfection of neurons pre-irradiation, were fixed and underwent immunofluorescence staining three days post-irradiation, or were fixed and underwent cytokine collection three days post-irradiation. Stereoscope images of YFP slices were taken over five days post-irradiation. Healthy neurons were manually tracked to determine surviving fractions for each arm, and YFP intensity was measured for each arm. Confocal images of immunofluorescent slices were taken, with microglia morphology and intensity measurements made with a custom CellProfiler pipeline. Morphology measurements of microglia included: area, perimeter, circularity, eccentricity, mean radius, median radius, major and minor axis lengths, and maximum and minimum Feret diameters. Intensity measurements included integrated intensity, mean intensity, and median intensity. The mean for each measurement was determined for each arm, and ANOVA tests were used to determine statistically significant differences between treatment arm means. Astrocyte branches were also segmented using a separate CellProfiler pipeline to measure total intensity, total intensity per unit area, mean intensity, and mean intensity per unit area for the segmented regions. ANOVA tests were performed to determine statistical significance between treatment arm means. Cytokine profiles were analyzed by Eve Technologies (Alberta, CA), and statistical significance determined using ANOVA tests. The conventional experiment replicated the dose delivered to the FLASH arm for two cytokine plates and an immunofluorescent plate and followed similar analysis methods.Results: Both the YFP-based surviving fraction measurements and the YFP signal intensity measurements could not effectively distinguish between different treatment arms. Both FLASH and CONV irradiation resulted in an increase in microglia size between the irradiated and non-irradiated arms based on area, perimeter, major and minor axis lengths, and maximum and minimum Feret diameter measurements (p < 0.0001). Microglia from the FLASH arm were larger compared with all other irradiated arms based off area, perimeter, major and minor axis lengths, and maximum and minimum Feret diameters (p < 0.05 to p < 0.0001), suggesting increased activation. This is further supported by a higher integrated intensity compared with the non-FLASH and HIGS No IR arms (p < 0.0001). No statistically significant difference was determined between treatment arms with astrocyte analysis. Increased levels of TNF-alpha in the FLASH arm compared with all other arms suggested activation of microglia into a pro-inflammatory M1 state (p < 0.01 to p < 0.0001). Increased levels of fractalkine in the FLASH arm compared with all other arms suggested the transition of microglia from the pro-inflammatory M1 state into an anti-inflammatory, restorative M2 state (p < 0.01 to p < 0.0001).

Conclusions: Any differences present in the YFP assay were below the sensitivity detection threshold of the assay. Microglia immunofluorescence and cytokine profile assays proved effective in detecting differences between treatment arms. Astrocyte analysis was not sensitive enough to distinguish between the different treatment arms. The increased size of microglia, TNF-alpha levels, and fractalkine levels in the FLASH arm compared with other arms suggest a stronger transition from a short-term pro-inflammatory state to an anti-inflammatory state compared with other treatment arms. Together, these results indicate differences in normal tissue response between treatment arms and suggests the possible presence of a normal tissue sparing effect.

As rapid developments in technology merge acoustics and microfluidics, acoustofluidic technology has been increasingly employed in biophysical and biomedical research to address various challenges, especially in the fields of tissue engineering, liquid biopsy, clinical diagnostics and therapeutics. Acoustofluidic technologies offer highly biocompatible, label-free and contact-less manipulation of objects based on differential effects including acoustic streaming and acoustic radiation force. However, acoustofluidic technologies have not been fully implemented into model animal studies to simplify the manipulation, lower the cost, and increase the throughput. In addition, despite the expansion of the scope of acoustic-based particle manipulation technologies from the micro to the nanoscale over the past decade, limitations continue to pose challenges in manipulating sub-100 nm particles using acoustic waves. As a result, developing an acoustic technique capable of manipulating sub-100 nm particles would strengthen the capabilities of acoustic manipulation and fulfill needs in areas such as biomedicine, biophysics, optics, electronics, and materials science.

First, with the nematode Caenorhabditis elegans (C. elegans) employed as a model animal in the field of developmental biology, neuroscience, human diseases, aging and drug screening for more than 50 years, we sought to extend acoustofluidic technologies into C. elegans research to address the key drawbacks of current C. elegans analysis procedures. An acoustofluidic chip capable of rotating C. elegans in both static and continuous flow in a controllable, precise, high-throughput and stable manner was then developed. Rotational manipulation was achieved by exposing C. elegans to a surface acoustic wave (SAW) field that generated a vortex inside a microchannel. By controlling the propagation of the SAW, we achieved bidirectional and stepwise rotation of C. elegans. Using this chip, we have clearly imaged the dopaminergic neurons, as well as the vulval muscles and muscle fibers of the C. elegans in different orientations. These achievements are difficult to realize through conventional microscopy. After that, another tool for effectively isolating and categorizing large quantities of C. elegans based on different phenotypes was developed as an integrated acoustofluidic chip. This chip was capable of identifying worms of interest based on expression of a fluorescent protein in a continuous flow and then separating them accordingly in a high-throughput manner. For example, L3 worms can be processed at a throughput of around 70 worms/min with a sample purity over 99%, which remains over 90% when the throughput is increased to around 115 worms/min. In our acoustofluidic chip, the time period to complete the detection and sorting of one worm is only 50 ms, which outperforms nearly all existing microfluidics-based worm sorting devices and may be further reduced to achieve higher throughput.

Moving forward with the experience we gained through manipulating C. elegans via acoustofluidics, we are aiming to solve a critical issue in current acoustofluidic technologies. Although acoustic fields have been increasingly used to pattern, focus, and separate micro- and nanometer-sized particles for biomedical applications, the acoustic-based separation of nanoscale bioparticles in sub-100 nm range remains a significant challenge. To address this problem, we present Acoustic Nanoscale Separation via Wave-pillar Excitation Resonance (ANSWER), allowing for the high-resolution, tunable separation of nanoscale bioparticles ranging from 50 nm to 1000 nm. We created virtual acoustic pillars that enable iterative deflection of particles for precision separation via an excitation resonance. Controlling the cut-off diameter is optimized by acoustic frequency, power, and microchannel dimensions in separating sub-100 nm particles. To demonstrate the potential of our ANSWER platform in biomedical applications, we have shown its ability to fractionate small extracellular vesicle (sEV) subpopulations. For the first time, sEV subpopulations can be rapidly separated (<10 minutes) directly from human plasma without sample preprocessing or complex nanofabrication. Due to its high separation purity (>96% small exosomes, >80% exomeres), ANSWER shows promise as a powerful tool that will enable more in-depth studies into the complexity, heterogeneity, and functionality of sEV subpopulations. To simplify the operation and keep the biological components in their native environment, separation without sheathflow was then discussed with the ANSWER platform. The same sheathless separation concept was then extended to the microscale for the isolation of plasma directly from human whole blood.

The work in this dissertation presents a comprehensive investigation and exploration of both the mechanism of specialized acoustic field generation and modulation, as well as the application of highly controllable manipulation of model animals and nanoscale (< 100 nm) biological particles. We hope our work can benefit and enable new possibilities in the relevant research fields.

Metasurface apertures provide an alternative approach to the very commonly used phased arrays or electronic scanned antennas (ESA) for wireless power transfer (WPT) and imaging applications. Array antennas use radiating elements which are often spaced at half-wavelength and uses active phase-shifter at each module to control the phase. However, in a metasurface antenna, the required phase is obtained from the sampled reference wave which propagates over the aperture providing an advance phase to each radiating elements. Metasurface apertures have very low manufacturing costs, planar form factor making them a suitable candidate for applications involving beamforming and wavefront shaping.

The thesis is divided mainly into two parts consisting of designing metasurface apertures for WPT applications and computational imaging purposes. In the first part, the proposed WPT system operating by focusing fields in the Fresnel region is presented with two proof of concept demonstrations. The first demonstration includes patch array antennas as the transmit and receive aperture and a half-wave rectifier to convert the RF to DC. The frequency of operation is 5.8 GHz (C-band) and the design of the patch array antenna is tedious and not suitable for a dynamic aperture which is possible by making use of metasurface apertures. The second demonstration consists of a metasurface aperture which uses a holographic technique to achieve focusing of microwaves at a particular focus distance for the transmit aperture. The receiving aperture is a metamaterial absorber which is connected to a rectifying circuit to harvest the power, thereby completing the WPT system. A LED connected as the load is illuminated which indicates the basic functionality of the WPT system. The RF-DC power transfer efficiency is in good agreement between simulations and experiments. The proposed system consisting of metasurfaces for both the transmit (focused aperture) and receive aperture (absorber) operates at 20 GHz (in K-band) has not been demonstrated in the literature and is a suitable candidate for higher frequencies (W-Band).

The second part consists of designing metasurface apertures demonstrating monostatic and bi-static microwave imaging systems where the metasurface apertures are frequency-diverse and operating at K-band frequencies (18-26.5 GHz). The metasurface apertures consist of radiating irises distributed over the sub-apertures in a periodic pattern. This frequency-diverse aperture produces distinct radiation patterns as a function of frequency that encode scene information onto a set of measurements; images are subsequently reconstructed using computational imaging approaches. In the monostatic case, the metasurface aperture is used as the transit aperture and 4 open-ended waveguides are used as receive aperture. In the case of bistatic case, both the transmit and receive apertures are metasurface apertures which result in increased mode diversity resulting in improved image reconstructions.

Exponentially localized Wannier functions (ELWFs) are an orthogonal basis for the low energy states of a material consisting of functions which decay exponentially quickly in space. When a material is insulating and periodic, conditions which guarantee the existence of ELWFs in dimensions one, two, and three are well-known and methods for constructing ELWFs numerically are well-developed. In this dissertation, we consider the case where the material is insulating but not necessarily periodic and develop an algorithm for calculating ELWFs.

In Chapter 3, we propose an optimization-free algorithm for constructing Wannier functions in both periodic and non-periodic insulating systems. In this chapter, we rigorously prove that under the assumption of ``uniform spectral gaps'', a technical assumption we introduce, that our algorithm constructs ELWFs.

While the uniform spectral gaps assumption is not always met in practice, in Chapter 4, we prove that for a wide class of systems (both periodic and non-periodic) it is always possible to modify our algorithm so that the uniform spectral gaps assumption holds. As a consequence of this result, we conclude that for both periodic and non-periodic systems our algorithm can construct ELWFs whenever they exist.

The results in this dissertation open the door for extending the theory of topological insulators, a recently discovered class of materials, to fully non-periodic systems.

The thesis has two parts. The first project relates to real time dosimetry in Gynecological (GYN) High Dose Rate (HDR) Brachytherapy Treatments and the second part investigates a new workflow for Image Guided Brachytherapy (IGBT) for centers that do not have access to MRI with applicator in-situ.

Project 1: Real-time Dosimetry in HDR Brachytherapy using the Duke designed NanoFOD dosimeter: Limits of Error Detection in Clinical Applications

Purpose: Previously, an optical fiber radiation detector system was shown to be capable of identifying potential high dose rate (HDR) brachytherapy delivery errors in real time. The purpose of this work is to determine this detector’s limits of error detection in vaginal cylinder and tandem and ovoid-type HDR gynecological brachytherapy treatments.

Method and Materials: The system consists of a scintillating nanoparticle-terminated fiber-optic dosimeter (NanoFOD) and a LabView platform (Versions 2015 and 2017; National Instruments Corporation, Austin, TX) which displays the real-time voltages measured by the NanoFOD during HDR treatment delivery. The platform allows for the measured voltage to be overlaid on the expected detector signal. To test the limitation of error detectability in vaginal cylinder brachytherapy, the NanoFOD was taped 1.5-cm from the tip of a 3-cm diameter Varian manufactured stump cylinder. This setup was

imaged using CT to localize the NanoFOD, and a plan was generated based on one of the institutional 3-cm stump cylinder templates with 9 dwell positions. For each dwell position, the expected voltage and doses were calculated using the dose distribution exported from the treatment planning system (TPS), a previously obtained distance-based calibration curve for the NanoFOD and TG-229 along and away table. The voltage values were imported to the LabView for real time monitoring. With a known location of the NanoFOD tip, the expected doses from the NanoFOD at each dwell position were calculated for all applicators for: 1) the clinical plan; 2) wrong source guide tube (SGT); 3) wrong cylinder; 4) wrong treatment template; 5) wrong step size, as well as 6) incorrect positioning of the cylinder insert. Measurements of voltage from each delivered plan were converted to dose per dwell and compared to the expected dose values.

In addition to experiments, a simulation-based limit of error detection with cylinder was studied and the results obtained from simulation were compared to experimental results. Simulation was done by modifying the applicator parameters and the incorrect plans were generated for 1) wrong SGT lengths 5-mm to 70-mm longer than correct length with a 5-mm interval; 2) wrong cylinder; 3) wrong treatment template; 4) wrong step size, as well as 5) incorrect positioning of the cylinder insert. Doses per dwell were also calculated and compared to the doses from correct plan.

Furthermore, simulations-only for T&O plans were also performed. Two representative clinical T&O-based plans with extra needles were used to establish the simulation-based limits of error detectability of the NanoFOD system. With a known location of the NanoFOD tip, the expected doses from the NanoFOD at each dwell position were calculated for all applicators for: 1) the clinical plan; 2) wrong treatment lengths (TL); 3) wrong connection to afterloader; 4) wrong digitization direction; and 5) wrong step size.

From previous work, it has been determined that the overall uncertainty in dose of this system in HDR brachytherapy is 20%. The percent difference (PD) between expected and measured doses and, between the simulated doses from correct and incorrect plans was calculated and compared to 20% to determine if a certain potential error can be detected with the NanoFOD system.

Results: For the cylinder experiments, when the correct treatment was delivered, the median PD over 9 dwells between measured and expected dose from each dwell was 11%. When the wrong SGT, wrong cylinder size, or wrong clinical template was used, the system detected PD values of -91.6%, -71.4% and -29.6% at the first dwell. With wrong step, the system showed large discrepancies starting at the third dwell position (58.6%). When incorrect cylinder insert length was 0.5cm, 1.0cm and 1.5cm, the significant PD values captured were -29.4%, -28.1% and -22.5% at 3rd, 2nd and 1st dwell, respectively.

In TPS simulations for the cylinder applicator, PD from wrong SGT, wrong cylinder size and wrong template and were -93.6%, -90.8% and -57.8% at first dwell. For wrong step size, the simulation predicted the PD to be greater than 20% (47.2%) starting at first dwell. For all incorrect insert length, simulation predicted greater than 20% PD (-38.6%, -34.9%, -38.6%) at first dwell. For incorrect SGT lengths range from 5mm to 70mm, simulation-based PD was from -28.0% to -82.3% at the first dwell, indicating the NanoFOD caught this error when the SGT was only 5-mm longer.

In TPS simulations with two T&O-based plans, with incorrect treatment length (TL, 5mm-70mm), the PD ranges for the two patients were 26.4%-26.8%, 20.4%-30.9%, and 30.9%-57% in 10mm TL tandem, 5mm TL ovoid and 10mm TL needle at first dwell. PD values were -44.1% to -36.4%, -75.3% to 96.4%, and -39.6% to 298.1% respectively at the first dwell when connections between tandem and left ovoid, left needle and left ovoid, left needle and tandem were swapped. When the two needles were switched, PD values were 43.1% and 63.6% at the 1st and 3rd dwell for patient A and patient B. Switching connection between the two ovoids did not produce significant PD for patient A because the two ovoids were identically loaded. For patient B, the PD was 87.5% when ovoids were swapped. Incorrect direction of digitization made the signal too low to be detected, an indication that treatment should be stopped. With wrong step size in tandem, ovoid and needle, simulation-based PD range in dose for two patients were 25.8%-34.9%, 34.3%-37.4%, and 20.7%-81.5%, respectively at 3rd, 2nd and 1st dwell for both patients.

Conclusion:

Using the 20% error detection threshold, the NanoFOD system was able to detect errors in real time cylinder tests when the wrong cylinder size, wrong SGT and wrong template were used, all at the first dwell position. For subtle errors such as small incorrect catheter insertions, the real-time system can detect errors starting with the second or third dwell position. The Labview interface is a good tool to use for real time tracking of the delivery.

TPS simulation predicted comparable discrepancies to PD obtained from experimental measurements when wrong cylinder size and wrong source guide tube was used. The disagreement between simulation-based PD and experiment-based PD is due to uncertainty in positioning during the experiment, as the nanoFOD in current form does not have a radio opaque marker to help with easy identification. TPS simulation was thus used for more complex applicators, knowing that simulations and experiments match.

With the 20% PD as an indication of wrong treatment in T&O-based HDR cases, the NanoFOD can capture all simulated gross errors within the first few dwells into the treatment, indicating it is capable of real-time verification of T&O HDR brachytherapy. Although the two T&O plans were different, the NanoFOD was able to capture the errors at similar dwell positions regardless of the difference in planning. Overall, the NanoFOD can catch, in real time, potential gross errors in clinical HDR.

Project 2: Investigation of Contour-based Deformable Image Registration (cbDIR) of pre-brachytherapy MRI (pbMRI) for target definition in HDR cervical cancer treatments

Purpose: While MRI-guided brachytherapy has been considered the gold standard for IGBT treatment for cervical cancer, practical factors such as cost, MRI access and clinical flow efficiency have limited the use of MRI for brachytherapy treatments. However, a pre-brachytherapy MRI (pbMRI) is usually taken before the brachytherapy to assess the tumor regression after external beam treatments. The purpose of this project is to investigate the role of contour-based deformable image registration of the pbMRI in target definition for high dose rate (HDR) cervical cancer treatments.

Method and Materials: Thirty-five patients with locally advanced cervical cancer treated at Duke University with Tandem and Ovoids (T&O) applicators were studied. A pbMRI was acquired for all patients. Each patient also had MRI and CT taken with applicator in situ at the time of the first fraction. The uterus structure contoured on pbMRI was deformed using contour-based (cb) and hybrid contour-based (hcb) DIR to the same structure contoured on the CT of first HDR fraction (MIM Software Inc., v 6.7.3 and v 6.8.11). The deformation matrix was used to deform the dHRCTV, which was then compared with the ground truth HRCTV (gtHRCTV) obtained on the MRI of 1st HDR fraction. Dice Similarity Coefficients (DSC) and Hausdorff Distance (med_HD, max_HD) were used to evaluate the overlap and mismatch at boundaries between the deformed HRCTV and gtHRCTV using both DIR methods. The process of generating dHRCTV requires user’s manual input. Therefore, inter- and intra-user variability using cbDIR, as well as the difference in the two DIR methods were investigated.

Results: The pbMRI scan was acquired on average 5.9+/-3.8 days before first HDR fraction. Median DSC for uterus was 0.9 (IQR 0.88-0.93) and 0.93 (IQR 0.92-0.94), and for HRCTV was 0.64 (IQR 0.52-0.71) and 0.68 (IQR 0.59-0.72) for cbDIR and hcbDIR, respectively. The median med_HD for uterus was 0.09cm (IQR 0.08-0.1cm) and 0.08cm (IQR 0.06-0.09cm). The median max_HD for HRCTV was 1.37cm (IQR 1.1-2.3cm) and 1.39cm (IQR 1.1-1.6cm) with cbDIR and hcbDIR respectively. Limited inter-user, intra-user, and DSC variability between DIR versions were found (p=0.53, p=0.77, p=0.18, respectively).

Conclusion: Contour-based DIR based on uterus/cervix structure is proven to be relatively user independent, improves when hybrid contoured based DIR algorithms are used, and is expected to serve as a good starting point for HRCTV definition for HDR planning in the absence of MRI with applicators in-situ.

Total skin irradiation (TSI) with electron beam has been very effective for patient with Mycosis fungoides. We recently developed and implemented a technique of laying down position for patients who are too frail for the standard standing position. In this study, we validated these measurements with Monte Carlo (MC) simulation which can provide more information on dose distributions and guidance on further optimization of the technique. The laydown technique consists of 6 equi-spaced beam directions relative to the patient cranial-caudal axis, similar to the standup technique. For the AP/PA directions (vertex fields), patient is placed directly under the gantry at 195cm source-to-skin distance (SSD) and 3 overlapping fields with gantry angles 60˚ apart are used. For the four oblique directions, patient is repositioned on the floor parallel to the gantry rotation axis at SSD of 305 cm with gantry at 300˚. A customized 0.25 mm Cu filter was placed in the linac interface mount to further broaden the beam. Each treatment fraction consists of 10 fields and 3 of them are unique. The Monte Carlo simulation was performed within the EGSnrc environment, using the phase space file provided by the linac vendor. The following quantities were studied and compared with the measurements: for each field/direction at the treatment SSDs, the percent depth dose (PDD), the profiles at the depth of maximum, and the absolute dosimetric output on the flat water phantom; the composite dose distribution on a cylindrical phantom of 30 cm diameter. Cu filter increases the beam FWHM by 44% but also reduces the output by 60%. The central regions within ±10% of the prescription dose were 170×70 cm2 for vertex fields and 140×80 cm2 for oblique fields. Profiles and output factors for both vertex fields and oblique fields agreed within 3% between MC and measurements. Vertex fields has dmax at (0.55: MC; 0.67: measurement)cm and R80 at (1.15; 1.40)cm, oblique field has dmax at (1.05; 0.86)cm and R80 at (1.55; 1.40)cm. When all fields are combined on the cylindrical phantom, the dmax shifted toward surface region. The composite dose distribution has the surface dose at (99.0; 95.2) %, dmax at (0.15; 0.15)cm, and R80 at (0.55; 0.75)cm. The maximum X-ray contamination at the central axis was (2.2; 2.1)%, and reduced to 0.2% at 40 cm off the central axis. Cylindrical phantom of 20 cm and 40 cm diameters for patient size simulation shows the surface dose of 93% and 103%, compared to 30 cm diameter. The Monte Carlo results in general agree well with the measurement, which provides secondary support in our commissioning procedure. In addition to those measurable quantities, the Monte Carlo simulation can provide further information such as the full dose distribution of the patient phantom, and the ability to investigate and optimize techniques such as different filter design, SSD and field size variations.