Projects

Polarized Spatial Frequency Domain Imaging (pSFDI)

pSFDI vs SALS - fiber angle
A native heart valve leaflet imaged with pSFDI and small angle light scattering (SALS) showing agreement in fiber orientation trends. pSFDI also requires no tissue preparation, has much higher resolution, and can capture large FOVs in seconds, as compared to the ~3 hours of tissue prep and 1 hour of point-scanning required for the SALS imaging technique.

In my primary thesis work, I am using polarized light imaging combined with spatial frequency domain imaging (SFDI) to study tissue microstructure (specifically, collagen fiber structure). Polarized light imaging provides wide-field maps of fiber directionality and degree of alignment, while SFDI allows us to optically section the top layers of the tissue. The technique is large field of view (5-25cm^2), high resolution (2MP+), rapid (<5s acquisition time), and requires no tissue preparation, allowing multiple time-point imaging of tissues undergoing dynamic testing. The main application currently is to study the structure of native and bioprosthetic heart valve tissues during mechanical testing to understand the role that fibers play in the mechanical behavior of the tissues. This will ultimately aid in the development of more appropriate materials for building bioprosthetic heart valves with longer lifetimes.

References:

Goth, Will, et al. “Interpreting fiber structure from polarization dependent optical anisotropy.” Imaging, Manipulation, and Analysis of Biomolecules, Cells, and Tissues XV. Vol. 10068. International Society for Optics and Photonics, 2017.

Goth, Will, et al.. Wide-field mapping of collagen fiber orientation and orientation distribution in soft tissues. Imaging Techniques in Biomechanics. Biomedical Engineering Society Annual Conference, Minneapolis, MN 2016.

Goth, Will, et al. “Optical-based analysis of soft tissue structures.” Annual review of biomedical engineering 18 (2016): 357-385.

Goth, Will, et al. “Polarized spatial frequency domain imaging of heart valve fiber structure.” Optical Elastography and Tissue Biomechanics III. Vol. 9710. International Society for Optics and Photonics, 2016.

Yang, Bin, et al. “Polarized light spatial frequency domain imaging for non-destructive quantification of soft tissue fibrous structures.” Biomedical optics express 6.4 (2015): 1520-1533.


Sub-diffuse Spatial Frequency Domain Imaging (sdSFDI)

Prelminary sdSFDI work
A Mohs surgery skin cancer sample with basal cell carcinoma (BCC) is shown imaged with it’s accompanying histology image. The sub-diffuse optical properties (bottom left) show changes in the BCC tumor region, which can then be used to extract approximate tumor boundaries (bottom right)

By imaging skin cancer samples at high spatial frequencies, we are able to detect changes in the cellular structure. Changes in the density, size, and shape of cells (hallmarks of cancer tumors) result in changes to the way those cells scatter light. By detecting this change in scattering, we are able to identify areas of dysplasia and aid physicians in identifying tumor boundaries. This technique is non-contact, wide-field (5-25cm^2), requires no tissue preparation, and can image bulk tissue biopsy samples. The speed of the image acquisition (<10 seconds) and processing time (<10 seconds) could provide a much faster workflow in scenarios where histological processing and examination is the main bottleneck, such as Mohs micrographic surgery.

References:

Goth, Will, et al. “Rapid empirical characterization of sub-diffuse reflectance imaging from Mohs surgery skin samples.” Lasers in Surgery and Medicine. Vol. 50. 111 River St, Hoboken 07030-5774, NJ USA: Wiley, 2018.


Optical Phantom Development

Sub-diffuse tissue phantoms
Optical properties of phantoms with different polymer micro-bead concentrations were developed using Mie calculations of desired diffuse and sub-diffuse scattering properties. sdSFDI at high spatial frequencies reveal these differences in the diffuse and sub-diffuse regimes.

To validate my imaging systems, I have developed several types of scattering tissue phantoms. These include solid and liquid scattering bead phantoms and TiO2/silicone phantoms with tune-able diffuse and sub-diffuse optical properties. Additionally, I have created or imaged electrospun fiber phantoms with various fiber alignment properties in collaboration with Dr. Alicia Allen from the Zoldan Lab and Dr. Siliang Wu from the Cosgriff-Hernandez Lab.

References:

Goth, Will, et al. “Rapid empirical characterization of sub-diffuse reflectance imaging from Mohs surgery skin samples.” Lasers in Surgery and Medicine. Vol. 50. 111 River St, Hoboken 07030-5774, NJ USA: Wiley, 2018.

Goth, Will, et al. “Predictive model of probe-dependent sampling depth in diffuse reflectance spectroscopy.” Optical Imaging and Microscopy Platform Session. Biomedical Engineering Society, San Antonio, TX 2014.

Hennessy, Ricky, et al. “Effect of probe geometry and optical properties on the sampling depth for diffuse reflectance spectroscopy.” Journal of biomedical optics 19.10 (2014): 107002.

Hennessy, Ricky, et al. “Sampling Depth of Diffuse Reflectance Spectroscopy Probes: Computational and Experimental Analysis.” Biomedical Optics. Optical Society of America, 2014.


Diffuse Reflectance Spectroscopy (DRS)

DRS with different Source-Detector Separations
By changing the source-detector separation (SDS) of the illumination and detection fibers of a diffuse reflectance spectroscopy probe, different sampling depths/volumes can be achieved. Shown is the ‘photon banana’ of expected light propagation paths through tissue using different SDS.

My first project in my graduate studies was in probe-based diffuse reflectance spectroscopy (DRS), which samples single points of tissue at a time, as opposed to wide-field imaging. This was part of a larger project which has shown high sensitivity and specificity of the DRS technique in identifying cancerous tissue in several clinical studies. My work in this project was experimentally validating the impact of using different source-detector separations (SDS) in the DRS probe, which ultimately allows the probe the be sensitive to different tissue depths.

References:

Goth, Will, et al. “Predictive model of probe-dependent sampling depth in diffuse reflectance spectroscopy.” Optical Imaging and Microscopy Platform Session. Biomedical Engineering Society, San Antonio, TX 2014.

Hennessy, Ricky, et al. “Effect of probe geometry and optical properties on the sampling depth for diffuse reflectance spectroscopy.” Journal of biomedical optics 19.10 (2014): 107002.

Hennessy, Ricky, et al. “Sampling Depth of Diffuse Reflectance Spectroscopy Probes: Computational and Experimental Analysis.” Biomedical Optics. Optical Society of America, 2014.


Phase-shifting Surface Profilometry (PSP)

Phase-shifting Surface Profilimetry
Example of the the Phase-shifting Surface Profilimetry (PSP) technique to extract 3D image information from a 2D image set. The ‘warping’ of the lines projected on the sample (visible in the structured illumination image, C) is due to the changes in height of the sample. By quantifying the ‘warping’ over the entire sample, a height-map of the object can be extracted.

As part of my thesis work, we are investigating the ability to add a further modality, Phase-shifting Surface Profilimetry (PSP), to our pSFDI technique. PSP allows extraction of surface topology from samples using essentially the same illumination technique as SFDI. We hope to image the 3D fiber structure of intact heart valves (rather than excised heart valve leaflets) using this multi-modal technique.


At-home polarized-light dermoscopy

PLI
Early prototype polarized light imaging attachment for a standard iPhone 5S.

One of the major limitations in skin cancer screening is the time that a physician has to examine a patient, which is often inadequate to properly evaluate all suspicious lesions. An at-home imaging device which allows consistent polarized light imaging, paired with an image-processing algorithm which ranks the lesions in terms of their relative likelihood for cancer, could aid the process considerably. Allowing easy tracking of changes in suspicious lesions over time and ‘ugly duckling’ lesions which are considerably different from other lesions on a patient are two powerful screening methods this device could permit to be adopted into the screening process. The overall goal is decrease the number of in-person dermatologist visits and unnecessary biopsies that are common in skin cancer screening. My role in this project was to develop initial device prototypes and supervise senior design teams in developing robust devices for a clinical study.

References:

Zhang, Yao, et al. “An automated mole ranking sysmem (MoleList) for skin cancer screening.” Lasers in Surgery and Medicine. Vol. 50. No. 4. 111 River St, Hoboken 07030-5774, NJ USA: Wiley, 2018.


Electrospun Fiber Mats

esf

In collaboration with Dr. Alicia Allen of the Zoldan lab, I designed and built a high-speed electrospinning mandrel to facilitate fabrication of electrospun fiber scaffolds. The device is capable of collection speeds from 0 to 3000RPM, operates at high voltages (~2kV), and is a workhorse instrument for producing fiber mats used to induce preferential alignment in cardiac cell cultures. This system was built for a fraction of the cost ($10,000).

Additionally, we are beginning a collaboration to image electrospun fibers fabricated in the Cosgriff-Hernandez group.

References:

Allen, Alicia CB, et al. “Electrospun poly (N-isopropyl acrylamide)/poly (caprolactone) fibers for the generation of anisotropic cell sheets.” Biomaterials science 5.8 (2017): 1661-1669.


Tissue Mechanical Analysis Device

Mountable Biaxial Stretching Device
The design process for my custom-made biaxial stretching device to mechanically test corneal tissue while simultaneously imaging with Brillouin microscopy.

During my summer ‘externship’ in my first year of graduate school and as part of the NIH T-32 Training program, I worked with Dr. Giuliano Scarcelli at the Wellman Center for Photomedecine (now at the University of Maryland) to create a custom biaxial tissue stretching device to mount on his Brillouin microscope. The goal was to determine mechanical properties (elastic moduli) from stress/strain experiments and see how well they corresponded to the properties found using the Brillouin microscopy technique.


Optical Coherence Tomography (OCT) Forward-directed Beam Scanning

bouma

The summer after I completed my undergraduate degree, I worked in Dr. Brett Bouma’s OCT research lab at the Wellman Center for Photomedicine to develop a technique for forward-direct, probe based OCT scanning. The constraints were that the system needed to be low-profile (to fit within a 3-5mm probe width) and allow rapid scanning for video-rate capture. My design incorporated a conical mirror to redirect the beam off the optical axis, using a combination of axial translation and rotation, which resulted in translating the beam in a spiral pattern the lateral (imaging) plane. This scanning method required reshaping of the beam, as the concical mirror warped the wavefront considerably. I modeled the system in Zemax and designed a lens system that would counteract the majority of the wavefront distortion.


Photo-acoustic Schlieren Elastography (PhASE)

PhASE2
Photo-acoustic Schlieren elastography (PhASE) device. This protoype system combined photo-acoustic stimulation of pressure waves in clear samples (such as cornea) which were subsequently visualized using Schlieren imaging. The speed of the wave-fronts could then be used to deduce mechanical properties of the sample.

During a directed study as an undergraduate, myself, Dr. Charles DiMarzio, and (now Dr.) Evan Perillo devised a non-contact system to detect mechanical properties of clear tissues, such as the cornea. I subsequently proposed the project for my Senior Design Capstone course, in which we design and validated the system instrumentation.