Bulletin of the American Physical Society
APS April Meeting 2022
Volume 67, Number 6
Saturday–Tuesday, April 9–12, 2022; New York
Session D08: Physics in MedicineRecordings Available
|
Hide Abstracts |
Sponsoring Units: GMED Chair: Wojciech Zbijewski, Johns Hopkins University Room: Salon 4 |
Saturday, April 9, 2022 1:30PM - 1:42PM |
D08.00001: In Silico Prototyping for Intranasally Administered Agents for COVID-19 and Other Respiratory Pathogens Zachary E Silfen, Mohammad M.H. Akash, Mark G Cherepashensky, Arijit Chakravarty, Saikat Basu, Diane Joseph-McCarthy For respiratory pathogens, such as SARS-CoV-2, a dominant early infection site is the nasopharynx. Antivirals administered directly to this site are likely to have a broader therapeutic window than systemically administered agents. To account for the dearth of patient data from various demographics, we propose a machine learning-enabled protocol to identify optimal formulation design parameters that can be matched to nasal spray device parameters for effective drug delivery. For that, we have measured 11 anatomical parameters (e.g., nasopharyngeal volume, nostril heights) for ten representative CT-based nasal geometries. We have also performed 160 CFD simulations of drug delivery for a range of breathing conditions (by applying varying pressure gradients driving the inhaled air transport) on the same geometries to determine drug deposition at the nasopharynx for nasal inhalers. With this test set, a proof-of-concept machine learning model is being developed to quantify targeted drug delivery in a wider demographic, as a correlative function of upper airway geometric variations. This work contributes to the design of a personalized, efficient intranasal delivery modality for prophylactics, therapeutics, and vaccines; the results will find use in a variety of respiratory diseases. |
Saturday, April 9, 2022 1:42PM - 1:54PM |
D08.00002: Clinical Telemedical Measurement of Grip Strength Thresholds for Work-Related Injury Permanent Impairment Rating Jerry L Artz, Bruce Bolon, Zach Griebel, John Alchemy Neuromuscular grip strength evaluation is challenging with the growing importance of telemedicine. Objective documentation of this hand function requires a simple procedure using an expensive dynamometer (e.g., a JAMAR grip dynamometer) traditionally used in an in-person medical office setting. This pilot study tests a method of measuring grip strength via telemedicine to determine if an individual, with a work-related injury, meets American Medical Association Guides criteria for permanent impairment to be eligible for insurance benefits. The procedure involves griping a specified water bottle, commonly found in one’s home, that is elevated and horizontal. The procedure requires the individual to suddenly squeeze hard to project the water. The height of the water bottle and maximum distance of the water stream are measured. An individual who does not meet a minimum distance of projection may qualify for compensation due to loss of grip strength. Thirty-nine volunteers in this pilot study provide calibration for this procedure, each doing five repeated trials according to the established protocol. The many subtle details and results of this pilot study are presented in this paper. |
Saturday, April 9, 2022 1:54PM - 2:06PM |
D08.00003: Dosimetric Analysis of Breast Swelling During External Beam Radiotherapy using Biomechanical Deformation Modeling David M McClatchy Breast cancer is the most common cancer in women worldwide. External beam radiotherapy with MV photons remains a central treatment for breast cancer, with treatments lasting up to five weeks. Over this timeframe, breast swelling due to radiation inhibiting normal fluid drainage poses a dosimetric challenge, as the size, shape, and water equivalent depth of the treatment area can change. In current clinical practice for whole breast radiation therapy, radiation fields are created with extra margin around the tissue to ensure the breast is still inside the treatment field if swelling presents, but no estimates are made of the dosimetric consequences of the swelling. Here we will present a dosimetric study where patient CT scans used for radiation planning are deformed using a biomechanical algorithm to simulate breast swelling in-silico. These deformed reference simulation CT scans will be compared to actual replan CT scans for patients with clinically significant swelling, to validate the accuracy of the biomechanical model. Then radiation dose will be estimated using the original treatment plan on the deformed CTs for a diverse set of patients. Finally, a multiparametric model will be built to quantitatively demonstrate the dosimetric impact of varying magnitudes of swelling as a function of tumor depth, breast volume, treatment modality (static versus modulated arc therapy), fractionation scheme, and remaining fractions. |
Saturday, April 9, 2022 2:06PM - 2:18PM |
D08.00004: Cost-Effective Depth-Encoding Methods for Time-of-Flight PET Scanners William J Matava, Kyle T Klein, Firas Abouzahr, Christopher Layden, Akhil Sadam, John Cesar, Shawn Park, Trang Do, Victoria Koptelova, Tri Truong, Stefaan Tavernier, Marek Proga, Karol Lang Positron Emission Tomography (PET) is a non-invasive medical imaging technique with a unique utility toward the diagnosis and location of cancer, and with growing applications to the study of neurodegenerative diseases. In order to improve position resolution and reduce parallax error during image reconstruction, state-of-the-art scanners feature both time-of-flight (TOF) capability and depth-of-interaction (DOI) sensitivity. Using Monte Carlo simulations in Geant4 and benchtop experiments, we have explored low-cost methods for achieving DOI sensitivity in lutetium–yttrium oxyorthosilicate (LYSO) PET scanners based on light-sharing between scintillator pixels. We have also explored the impact that LYSO's surface polish has on its TOF and DOI resolutions. |
Saturday, April 9, 2022 2:18PM - 2:30PM |
D08.00005: An Inexpensive Polyvinyltoulene Barrel PET Scanner Design Akhil Sadam, Christopher Layden, Kyle T Klein, William J Matava, Karol Lang We present a novel barrel positron-emission tomography (PET) scanner design that provides an inexpensive alternative to current clinical scanners without compromising resolution or sensitivity, using extruded polyvinyltoluene scintillators. While lower efficiency, these scintillators provide a sufficiently fast time resolution at much lower cost, allowing quick exploratory scans for tumor metastases and pathogenic substance identification. The complete detector is comprised of a 24-array barrel, each array 3x4 units of 4 scintillators 1m long wrapped in reflective film and coupled on either end to silicon photomultipliers (SiPMs). Preliminary Monte Carlo simulation via the Geant4 toolkit predicts that this scanner attains an energy resolution of about 16.1 to a central point source, and via our k-nearest neighbor algorithm, a depth of interaction resolution of about 15.5mm full-width-half-maximum (FWHM). Assuming a SiPM time resolution of 100ps, a coincidence time resolution of about 270ps FWHM is attained. In our talk we present more details of the design and performance of this PET scanner. |
Saturday, April 9, 2022 2:30PM - 2:42PM |
D08.00006: DEEP LEARNING TECHNIQUES FOR KNEE MR IMAGES RECONSTRUCTION María Margarita López-Titla, Héctor Gómez-Morales, Kelvin Lin, Sarmad Malik, Zheng Cheng MRI scanners acquire data samples in the spatial frequency domain (k-space) and classical or deep learning imaging techniques are applied to reconstruct the final image. A limitation of using classic reconstruction techniques is that it requires to acquire the full set of Fourier domain data. Recent deep learning techniques use subsampled Fourier data to produce diagnostic images. The aim of this project is to maximize the accuracy of the reconstructed images while minimizing the amount of data needed, the size of the model, and the training time. We used the Knee MR images from NYU fastMRI database to perform advance machine learning algorithms to reconstruct MR images. The quality assurance of these images was evaluated by NMSE, PSNR, SSIM and by expert opinions. A: We varied the size of fast MRI´s UNet B. We modified the UNet to use loss functions other than L1-loss. C: We implemented smaller variations of Resnet, CS-Net, DCCNN, CDCNN. D: We considered the expert evaluation for the feedback of the final outputs of the tuned neural networks. The CS-Net was the most performant. The CS-Net might have been judged the best bacause tempts to emulate a traditional MRI reconstruction method. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700