Upstate New York Association of Physicists in Medicine, Inc. (A Chapter of the AAPM)
Spring Meeting – Friday, April 21st, 2017

Swift Auditorium, Buffalo General Hospital

100 High Street, Buffalo, NY  14203



Business Meeting


Lunch and Vendor Exhibits (Lunch Sponsor: Landauer)


Meeting Introduction

Lalith Kumaraswamy, PhD

UNYAPM President


Invited Talk

MR IGRT: Realizing the Vision

Mr. Paul Naine



Young Investigator Award Symposium

Magnitude of motion-induced blur on High-Resolution Fluoroscopic (HRF) detector projection views due to gantry rotation in Cone-Beam Computed Tomography (CBCT)

Jordan Kerbs

University at Buffalo


Linearity Check and X Ray spectrum comparison with various modes in a Photon Counting Detector (PCD)

Alok Shankar

University at Buffalo


Coherent scatter imaging Monte Carlo simulation

Laila Hassan

University at Albany


Dosimetric effects of immobilization board and couch structures in prone breast radiation

Amy Lau

Roswell Park Cancer Institute


Coffee Break and Vendor Exhibits (Coffee Break Sponsor: Sun Nuclear)


Young Investigator Award Symposium

Implementation of basis material decomposition in neurovascular intervention

Alexander Podgorsak

University at Buffalo


Effect of using region-of- interest (ROI) attenuators on organ and

effective dose in CBCT scan and interventional fluoroscopy

Zhenyu Xiong

University at Buffalo


Impact of prostate focused alignment on planned pelvic lymph node dose

Joshua Kilian-Meneghin

Roswell Park Cancer Institute


Fiducial marker detectability improvement in EPID during

radiation therapy treatment

Tianjun Ma

Roswell Park Cancer Institute


Invited Talk


ABR MOC Update for Medical Physicists including the new OLA Program

Matthew Podgorsak, PhD, FAAPM

Trustee, Therapy Medical Physics, American Board of Radiology


Young Investigator Award Presentation


Meeting Adjourned



Meeting Abstracts


Magnitude of motion-induced blur on High-Resolution Fluoroscopic (HRF) detector projection views due to gantry rotation in Cone-Beam Computed Tomography (CBCT)


Jordan Krebs, Megan Russ, Alok Shankar Pookotte Alanchery, Daniel R. Bednarek, Stephen Rudin


Purpose: Source and detector movement induces image blur in CBCT projection views. As detector resolution increases the impact of image blur increases. A method of predicting blur was investigated and experimentally verified.


Methods: A high resolution detector with 75 um pixels was mounted on an angiographic C-Arm.

Suspended metal wire and lead beads were imaged as the C-arm rotated at a constant speed. Image blurring was predicted to be a function of three variables: rotational speed, object’s distance from isocenter, and frame exposure time. Rotational speed was kept constant for each run and the object’s vertical distance from isocenter was adjusted incrementally. Blurring of the beads was measured in images at a gantry angle of 0o and compared to images for projection views without motion. The object’s image width was corrected for magnification and plotted to show a linear dependence of blur on distance from isocenter. The plot’s slope was predicted to be the product of rotational speed and frame exposure time.


Results: Significant image blur was observed along the direction of motion in projections with gantry motion. Motion-induced blur would have a detrimental impact on spatial and contrast resolution for projection views acquired in tomosynthesis or CBCT. A mathematical model was derived for predicting motion blur in a projection view for objects anywhere in the FOV. The model was experimentally verified for the special case in which objects lie along the central axis.


Conclusion: The impact of motion-induced blur on current CBCT with pixels of about 400 um would be minor. For high definition detectors with pixels of 75 um or less, appropriate methods for reducing the effect of motion blur during CBCT acquisition with continuous gantry motion will be needed.



Linearity Check and X Ray spectrum comparison with various modes in a Photon Counting Detector (PCD).


A. Shankar, M. Russ, J. Krebs, D.R. Bednarek, S. Rudin


Purpose: To compare the linearity and sensitivity of the four operational modes in XCounter’s Actaeon PCD. And to evaluate performance of these modes in effectively measuring the spectrum from an X-Ray tube source.


Materials and Methods: In Energy Integrating Detectors (EID), the resultant image intensity is proportional to energy deposited by the photons, whereas in PCD the intensity is proportional to number of photons detected above a certain threshold. XCounter AB’s Actaeon imager is a direct detector with 0.75mm thick CdTe, 100µm pixel pitch, and 256 x 256 pixels. Various operational modes in Actaeon were used to image a Stenosis/Aneurysm Artery block (Nuclear Associates model 76-705) The exposure settings were varied from 50 to 400 mA and the total counts for each mode were evaluated using ImageJ for two separate thresholds (10 keV & 32 keV) to evaluate linearity. X-ray spectra at 70 kVp and 110 kVp from a Toshiba Rotanode DRX T744GFS tube were measured by summing 150 individual frames to find total counts at varying threshold settings. Initially dark field measurements were taken to check for background/scatter counts as a function of threshold settings up to 70 keV.


Results: High Sensitivity with Anti-Coincidence- On (HS-On) appears to result in overall gain of total counts compared to modes without charge sharing correction as it sums up counts below the threshold and attributes them to one pixel. However, for the HS-On mode at higher photon fluence rates the linearity breaks down due to delay in implementing the anti-coincidence charge sharing correction for neighboring pixels hence resulting in pulse pile up losses. High Sensitivity with Anti-Coincidence On (HS-On) mode outperformed the other modes in giving the spectrum with greatest counts for thresholds above 20 keV and 50 keV (for 70 kVp and 110kVp respectively) probably because the charge sharing circuit enables counts in adjacent pixels that would have had values below the threshold, now to be able to contribute to more accurate values above the threshold and hence counted unlike for other modes. This was evidenced when the charge sharing was turned off but high sensitivity maintained (HS - Off). The dark field/scatter counts were significant only up to a 4keV threshold after which instrumentation noise appears to be negligible.


Conclusion: Initial test of Actaeon PCD evidenced higher sensitivity but non-linear performance when charge sharing correction was implemented. For higher thresholds and where instrumentation noise was negligible, energy spectra were greatest in magnitude for the HS-ON PCD mode indicating the most potential for dual energy applications of the four available modes.



Coherent scatter imaging Monte Carlo simulation


Laila Hassan and Carolyn A. MacDonald


University at Albany, State University of New York, Department of Physics, 1400 Washington Avenue, Albany, New York 12222, United States


Conventional mammography can suffer from poor contrast between healthy and cancerous tissues due to the small difference in attenuation properties. Coherent scatter slot scan imaging is an imaging technique which provides additional information and is compatible with conventional mammography. A Monte Carlo simulation of coherent scatter slot scan imaging was performed to assess its performance and provide system optimization. Coherent scatter could be exploited using a system similar to conventional slot scan mammography system with antiscatter grids tilted at the characteristic angle of cancerous tissues. System optimization was performed across several parameters, including source voltage, tilt angle, grid distances, grid ratio, and shielding geometry. The simulated carcinomas were detectable for tumors as small as 5 mm in diameter, so coherent scatter analysis using a wide-slot setup could be promising as an enhancement for screening mammography. Employing coherent scatter information simultaneously with conventional mammography could yield a conventional high spatial resolution image with additional coherent scatter information.


Dosimetric Effects of Immobilization Board and Couch Structures in Prone Breast Radiation


A. Lau, K. Salerno, T. Ma, I. Wang


Purpose: The prone breast technique, initially introduced for treatment of pendulous breasts to minimize skin effects, may provide improved dose homogeneity and reduced doses to the heart and lung compared to supine in breast conservation patients. As the medial tangential beam may pass through the immobilization board and couch structures, there is concern regarding attenuation and potential bolus effects.


Methods: The ClearVue™ prone breast board used consisted of an insert on which the contralateral breast rests and a base indexed to the couch. Transmission factors (TF) were measured at various gantry angles using a Farmer chamber at 4cm depth in a solid water phantom. Treatment plans were also calculated in the Eclipse™ treatment planning system (TPS) for comparison. A parallel plate chamber at phantom surface was used to quantify the bolus effect.


Results: Measured insert TFs ranged 95.6-97.7% for 6 MV and 97.7-99.8% for 23 MV. TPS values were in agreement within 1.4%. Base TFs ranged 97.9-98.5% and 98.9-99.8% for 6 and 23 MV, respectively. TPS values agreed within 0.5%. Measured couch rail TFs were 96.3% and 97.9% for 6 and 23 MV, respectively. The insert increased the surface dose by 62-93% for 6 MV and 115-153% for 23 MV. The board base increased the surface dose by 39-55% for 6 MV and 65-83% for 23 MV. Compared to a 10cm air gap between the breast and base, close contact increased surface dose by 25-30% for both energies.


Conclusion: Results demonstrated possible reduced coverage by attenuating external structures and an increased surface dose from the breast board. Proper modelling of immobilization devices and couch structures in the TPS may allow more accurate dose computation in treatment planning. Further study is needed regarding the clinical impact and significance of these findings.



Implementation of basis material decomposition in neurovascular intervention


Alexander R. Podgorsak1, 2, Ashwin Venkataraman2, S.V. Setlur Nagesh2, Daniel Bednarek2, Stephen Rudin1, 2, and Ciprian N. Ionita1, 2


1 Department of Biomedical Engineering, University at Buffalo

2 Toshiba Stroke and Vascular Research Center, University at Buffalo


Purpose: To implement and assess material decomposition processing using dual-energy CBCT images taken of a 3D-printed patient specific neurovascular phantom following an endovascular intervention.


Methods: An endovascular intervention was carried out on a 3D-printed aneurysm model. A platinum coil was deployed into the aneurysm sac, and a cobalt-chromium stent was placed to span the aneurysm orifice. Imaging of the model was performed with a CMOS-based micro-CBCT system using dual-energy acquisition done for tube voltages of 35 and 70 kVp. Axial slices of the model were computed using the FDK algorithm. Basis material decomposition was performed using the reconstructed dual-energy image data, considering the platinum and cobalt-chromium as basis materials for decomposition. An additional trial was carried out following addition of iodine contrast agent to the model, considering the platinum, cobalt-chromium, and iodine as basis materials for decomposition. X-ray tube and reconstruction parameters were kept constant throughout.


Results: Examination of the output basis material slices from the intervention without iodine indicates that the algorithm was able to decompose between the platinum and cobalt-chromium. The decomposition was carried out with minimal error. With the addition of the iodine contrast agent, the decomposition between the platinum and cobalt-chromium was similarly successful. When a slice devoid of platinum and with the cobalt-chromium and iodine present was considered, the decomposition between the cobalt-chromium and iodine was accurate, and the platinum map correctly displayed no material.


Conclusion: This work indicates the potential clinical utility of basis material decomposition in the context of neurovascular intervention, as well as the usefulness of this method in the imaging and assessment of 3D-printed patient specific endovascular models.



Effect of using region-of- interest (ROI) attenuators on organ and effective dose in CBCT scan and interventional fluoroscopy


Zhenyu Xiong, Sarath Vijayan, Stephen Rudin, Daniel R. Bednarek


Toshiba Stroke and Vascular Research Center, University at Buffalo, Buffalo, NY


Purpose: In some medical imaging procedures using CBCT and fluoroscopy, only the center of the field of view may be needed to be visualized with optimal image quality. The aim of this study was to quantify the dose reduction benefit of the region of interest (ROI) imaging during a typical CBCT and interventional fluoroscopy procedures in the head and torso.


Methods: The Toshiba Infinix C-Arm System was modeled in BEAMnrc/EGSnrc with and without the ROI attenuator. Patient organ and effective doses were calculated in DOSXYZnrc/EGSnrc Monte Carlo software for CBCT and interventional procedures. The Monte Carlo model was first validated by comparing the surface dose distribution in a 20 cm thick solid water block phantom with measurement by Gafchromic film. The dependence of dose reduction on the ROI attenuator thickness, the opening size of the ROI, the axial beam position and the location of the different organs for both neuro and thoracic imaging was evaluated.


Result: The results showed that the entrance surface dose reduction in the solid water is about 85.7% outside the ROI opening and 10.5% in the ROI opening. For a ROI attenuator made with 0.7 mm thick Cu and a circular aperture which was 12% of the FOV, the results of Monte-Carlo calculations showed a reduction in the dose to most organs of 45%-50% and in effective dose of about 46% compared to the dose in standard neuro and thoracic CBCT scans without the ROI attenuator; a reduction in dose for most radiosensitive organs of 55%-60% (65%-70%) and in effective dose of 58% (66%) compared to the dose in a simulated neuro (cardiac) interventional fluoroscopic procedure without the ROI attenuator was determined.


Conclusion: This work provides evidence of substantial reduction of organ and effective doses when using an ROI attenuator during CBCT and fluoroscopic procedures.



Impact of Prostate Focused Alignment on Planned Pelvic Lymph Node Dose

J. Kilian-Meneghin, T. Ma, L. Kumaraswamy

Purpose: Prostate patients with positive lymph node margins receive an initial course of 45Gy to the PTV comprised of prostate, seminal vesicles, and lymph nodes with a 1cm margin.  The prostate is localized via implanted fiducial markers before each fraction is delivered using portal-imaging. However, the pelvic lymph nodes (PLN) are affixed to the bony anatomy and are not mobile in concert with the prostate. The aim of this study is to determine whether a significant difference in PLN coverage exists between planned and delivered external beam therapy treatments for these patients.

Methods: The recorded prostate motions were gathered for 20 patients; conjointly the PLN motions were determined by manual registration of the bony anatomy in the Kv-images. The difference between the prostate and the bony anatomy coordinates were input into Eclipse as field shifts to represent the deviation in planned vs delivered PLN coverage.

Results: Structure volume at D(100) was recorded for each patient for two structures: summed PLNs and PLNs+1cm margin to express their contribution to the PTV. For the PLNs+1cm, the average difference between the planned coverage and calculated delivered coverage was 4%, with a paired p-value of p=0.0006. Based upon bony anatomy registration, only 25% of patients received the intended dose coverage based upon D(100) dose criteria for PLN+1cm. Dose value differences between the two plans at minimum were 2.48±3.29Gy, at mean were 0.51±0.47Gy, and at maximum were 0.26±0.33Gy. For the PLNs, the average difference between the planned coverage and calculated delivered coverage was 0.8%, with a paired p-value of p=0.58.

Conclusions: The results indicate a significant difference exists between the planned coverage and calculated delivered coverage for the PLNs+1cm, with no significant difference for PLNs. We conclude that lymph node motion must be considered with the prostate motion when aligning patients before each fraction.


Fiducial marker detectability improvement in EPID during radiation therapy treatment

Tianjun Ma, Josh Kilian-Meneghin, Lalith K. Kumaraswamy

Purpose: Hypofractionated prostate radiotherapy has the advantage of maximizing tumor kill while minimizing toxicities to rectum and bladder due to the relatively low α/β ratio for prostate compared to rectum and bladder.  However, large prostate movements during treatment can reduce the therapeutic effectiveness of the hypofractionated treatments. Real time monitoring can help accurately locate the target and, potentially, narrow the tumor margin to alleviate post-treatment radio-toxicities. The aim of this study is to maximize detectability of the prostate during hypofractionated VMAT treatments with real time EPID imaging.

Materials and Methods: Patients routinely will have at least three fiducial markers (FMs) implanted inside the prostate before computed tomography scan are acquired for treatment planning. During VMAT delivery, the detectability of the FMs is a challenge due to the dynamic MLC movement. In-house software was developed to extract the marker locations and project them onto the EPID detector. Detecting radius and FM separation distance can be set to estimate the probability of FM detection in real-time. The overall detectability score and overall percentage time of FM detection for at least two separate FMs were obtained to evaluate the probability of FM detection. Radiation treatment plans were re-optimized to take the FMs detectability into account.

Results: The Detectability score (SD) studied in the four original treated plans has an averaged value of 0.41±0.13. While the total percentage time of fiducial marker detection (TD) was 0.38±0.17. After optimization for FM detection, SD increased by 11% and TD had a significant improvement to 0.49± 0.16, an increase of detection time by 22.4%.

Conclusions: The results indicated that incorporating the FM detectability into plan optimization greatly enhances the localization of the target in hypofractionated prostate treatments.