Session: Imaging and Imaging in RT - Salle de bal et foyer

June 23, 2022 from 3:00pm EST to 4:00pm EST

Scientific Session – Imaging and Imaging in RT
Thursday, June 23, 2022, 15:00-16:00


Scientific Session Imaging and Imaging in RT – Presentation 1 

Navigation and non‑navigation CT scan of the sinuses: comparison of the effective doses of radiation in children and adults

Mario Chretien, Noémie Villemure-Poliquin, Jacques E. Leclerc
CHU de Québec - Université Laval, Université Laval

Purpose: The advent of 3D navigation imaging has gained in popularity among otolaryngologists for endoscopic sinus and skull base surgeries, in both adults and children. However, the increased tissue radiation with 3D navigation protocols computed tomography (CT) scans is a source of concern because of its potential health hazards. We aimed to compare the effective doses of radiation between 3D navigation protocols and standard protocols for sinus CT scans for adult and pediatric population.

Methods: We performed a retrospective cohort study of patients undergoing sinus CT scans using a Siemens Drive CT scanner. The effective dose of radiation was calculated in mSv. Average irradiation doses were compared using a Student’s T-Test or a Kruskall–Wallis test when appropriate.

Results: A total of 115 CT scans were selected for analysis : 47 standard and 68 3D navigation; 31 exams on children and 84 exams on adults. For the total population, mean effective dose in the non-navigation CT scans was 0.37 mSv (SD: 0.16, N = 47) and in the 3D navigation sinus CT group was 2.33 mSv (SD: 0.45, N = 68). The mean difference between the two groups was statistically significant 1.97 mSv (CI 95% − 2.1 to − 1.83; P < 0.0001).

Conclusion: There was a sixfold increase in radiation with utilization of 3D navigation. The ratio was identical with the pediatric and the adult subset. Otolaryngologists should be aware of this significant increase dose and should attempt to decrease the radiation exposure of their patients.


Scientific Session Imaging and Imaging in RT – Presentation 2

Tumour Coverage Evaluation for Percutaneous Liver Tumour Ablations:  2D US vs 3D US

Shuwei Xing, Derek W. Cool, Elvis C.S. Chen, Terry M. Peters, Aaron Fenster
Robarts Research Institute - Western University, Department of Medical Imaging,

Purpose: Ultrasound (US)-guided percutaneous thermal ablation is a promising curative treatment technique for focal liver tumours. Assessing the tumour coverage intra-procedurally is required to ensure that the tumour can be covered entirely by the ablation zone plus a 5 or 10 mm safety margin. Since 2D US images may not provide sufficient volumetric evaluation, we aim to compare 2D with 3D US-based tumour coverage evaluation and investigate whether 3D US-based evaluation is a more efficacious approach.

Methods: We conducted a prospective patient trial to collect data (2D and corresponding 3D US images) from 12 consecutive cases. To evaluate the tumour coverage, we measured the ablation margin defined as the minimum signed Euclidean distance from the estimated ablation zone to the tumour. Then we validated our estimation by comparing them (2D and 3D) with corresponding oncological outcomes.

Results: The minimum estimated ablation margin evaluated by the 2D US was generally larger than the 3D US-based one. 11 of 12 cases were correctly assessed by 3D US-based tumour coverage evaluation and the remaining one was a false positive case. However, two cases that had residual tumours were not identified by the conventional 2D US-based approach.

Conclusion: The smaller estimated ablation margin evaluated using 3D US indicated the risk of insufficient evaluation from 2D US images. Thus, our validation results demonstrated that 3D US-based tumour coverage assessment is more efficacious.


Scientific Session Imaging and Imaging in RT – Presentation 3

ExacTrac imaging dose for real-time tumor tracking in prostate SBRT patients

Ruwan Abeywardhana, Alan Spurway, Mike Sattarivand
Nova Scotia Health Authority

Purpose: To quantify three-dimensional (3D) dose for planning target volume (PTV) and critical organs at risk (OAR) during ExacTrac real-time tumor monitoring for prostate stereotactic body radiotin therapy (SBRT).

Materials and Methods: Thirty prostate SBRT patients were retrospectively selected and divided into three (large, medium, and small) categories using body mass index with 10 patients in each category. EGS phantoms were created from patients’ CT data and 3D dose distribution from the ExacTrac system was computed in DoseXYZ using a previously validated ExacTrac Monte Carlo (MC) model. Imaging protocols were set to 130kVp, 120kVp, and 110kVp (with 25mAs) for large, medium, and small patients as per vendor recommendation.

Results: For all the patients, mean D50 values for PTV, bladder, femoral head left, femoral head right, rectum, bone, and skin were 41 ±(11), 25.3 ±(6.7), 25.9 ±(11.8), 21.7 ±(7.3), 82.7 ±(23.4), 64.9 ±(14.5), 5.1 ±(1.7) mGy, respectively; while mean D2 values were 77 ±(21), 56.4 ±(20.7), 141.5 ±(52.2), 123 ±(35.5), 143 ±(34.8), 750 ±(159), 398 ±(57) mGy, respectively. The skin D2 values were statistically significant and increased with increasing patient sizes. Contradictorily, internal OARs and PTV of larger patients have received significantly lower doses (D2 and D50). The highest reported D2 as a percentage of the prescribed dose was 2.67% and 1.42% for bone and skin respectively.

Conclusions: 3D imaging dose distributions from ExacTrac real-time stereoscopic system was quantified. Imaging dose from real-time IGRT in prostate SBRT patients using ExacTrac is only a few percentages of prescribed treatment dose.


Scientific Session Imaging and Imaging in RT – Presentation 4 

Comparison of 4DCT and 4DCBCT in measuring motion trajectory of moving targets in phantom

Satyapal Rathee, Bhumika Handa
University of Alberta

Purpose: To study the discrepancy in motion trajectories measured by 4DCT and 4DCBCT, and the motion effects in 3DCBCT.

Methods: A 2.5 cm diameter spherical target undergoing cos4 motion in 3D was imaged using 4DCT (pitch as per breathing period), 4DCBCT (basic and advanced reconstructions), and 3DCBCT. For 4D images, the center of target in each phase determined the motion envelope. The measured motion envelope in each direction was fit with cos4function to determine the peak-to-peak motion amplitude and compared with the programmed motion. For 3DCBCT, enlargement of target due to motion was measured as full width at 95% of maxima of profiles through target center in three cardinal directions. Subtracting actual target size from measured enlargement indirectly estimated the peak-to-peak target motion. This analysis was done for several combinations of motion amplitudes in 3D and breathing periods. 

Results: The mean difference between programmed and measured peak-to-peak motion amplitude was (4DCT, 4DCBCT basic, and 4DCBCT advanced) -0.3 mm, -0.2 mm, and -0.9 mm in SI; -0.1 mm, -0.1 mm, and -0.2 mm in LR; -0.1 mm, 0.3 mm and -0.2 mm in AP directions.  The mean difference (standard deviation) between programmed and indirectly measured peak-to-peak motion amplitude by 3DCBCT was 0.2mm (0.7mm), 0.0mm (0.3mm) and -0.5mm (0.7mm) in SI, AP and LR directions respectively.

Conclusions: Both 4DCT and 4DCBCT are able to measure the position of regularly moving target accurately with some underestimation in advanced 4DCBCT reconstruction. Apparent enlargement of moving target in 3DCBCT is as per motion in 3D.


Scientific Session Imaging and Imaging in RT – Presentation 5

Automated measurements of patient positioning using EPID images in DIBH Breast Hybrid IMRT

Jonathan Redekopp, Jorge Alpuche, Ryan Rivest, David Sasaki, Stephen Pistorius
University of Manitoba, CancerCare Manitoba

Purpose: The purpose of this work is to develop a methodology to automatically measure intrafraction motion, interfraction motion, and setup error in breast DIBH EPID images. 

Methods : An algorithm was developed to measure chest wall distance (CWD) on cine EPID images. The images were acquired for fields without MLCs in DIBH breast hybrid IMRT treatments. Intrafraction motion, interfraction motion and setup errors were measured for all fractions of 10 patients. Intrafraction motion was quantified as the change in CWD compared to the first frame while interfraction motion as the change in average CWD from the first fraction. Setup error was measured for the first fraction as the difference in CWD between the first EPID frame and a MV DRR. Manual measurements were made for validation. 

Results: Validation of the algorithm against manual measurement of the chest wall agreed on average by 0.7 ± 0.3 mm with the EPID images and 1.1 ± 0.7 mm with the MV DRR images. The largest intrafraction motion found was 2.45 mm and agreed within manual measurements by 0.05 mm. The largest interfraction change was 6.6 mm and was validated within 0.5 mm. Setup errors were manually validated and agreed within 1.4 mm ± 1 mm.

Conclusions: An algorithm was developed to automatically detect the CWD on EPID images. Our results suggest that it can be used to measure intrafraction motion, interfraction motion and setup errors within 1.4 mm on average. Future work will aim to expand its scope to fields with MLCs.


Scientific Session Imaging and Imaging in RT – Presentation 6

One Stop Shot – Personal protective equipment Pb equivalency assessment tool

Thorarin Bjarnason, Kevin Hammerstrom
Medical Imaging - Interior Health Kelowna, Department of Radiology - University of British Columbia, Department of Physics - University of British Columbia

We developed a rugged tool for personal protective equipment (PPE) Pb equivalency assessment using radiographic detectors. A strip of Pb foils of various thicknesses were sandwiched between acrylic large enough to fully cover radiographic detectors, with the PPE placed on top of the tool and beside the Pb strip. Using a single x-ray shot per kVp, both thorough and quick Pb equivalency assessments were made. After the exposure, analysis was performed on PACS by placing ROIs within each Pb foil and within the PPE. A full Medical Physics assessment of PPE was performed using linear lookup tables (LUTs) from digital radiographic equipment. Calibration curves were created by mapping grey scale value to Pb thickness using ROIs from the Pb strip. The PPE grey scale measurement was then compared to the calibration curve to determine precise Pb equivalency of the PPE. For spot-checking PPE that may have had their labels worn or torn off, a Technologist assessment was performed using Abdomen LUTs. In this scenario, the technologist simply needed to compare the grey scale value with those on the tool’s Pb strip to get a close estimate of the PPE Pb equivalency. These measurements were all verified using Philips and Siemens DR systems.


Scientific Session Imaging and Imaging in RT – Presentation 7

Non-invasive measurement of the arterial input function for dynamic positron emission tomography: Simulation of clinical workflow

Liam Carroll, Shirin A. Enger
McGill University

Purpose: Dynamic positron emission tomography (dPET) is a functional imaging technique that has many uses throughout the radiotherapy treatment cycle. dPET requires the measurement of the time course arterial radioactivity concentration, called the arterial input function (AIF), which is normally acquired using arterial catherization. We present simulations of a proposed non-invasive detector (NID) designed to measure the AIF. 

Methods: A Geant4 Monte Carlo user-code using PENELOPE cross-sections was developed to simulate the NID. The detector consisted of 59 plastic scintillating fibers (1 mm diameter, 10 cm length) arranged in two layers on a polyethylene wrist phantom (64.13 mm diameter, 10 cm length). Two water filled cylinders (2.30 mm diameter, 6 mm apart), representing the radial artery and vein, were placed in the phantom at varying depths (2.0 to 3.0 mm). 18F, 11C, 15O, 68Ga radioisotopes were randomly distributed between the artery and vein and were allowed to decay 100 million times per simulation (44 radial artery/vein depth and radioisotope combinations). Energy deposition per event per scintillator channel were tracked separately for positrons and annihilation photons. A post processing algorithm considering the biodistribution of the injected bolus and using an expectation maximization maximum likelihood algorithm was tested.

Results: The average percent error for each radioisotope were 0.035 %, 0.031 %, 0.09 % and 0.07 % for 18F, 11C, 15O, 68Ga respectively. 6 data-sets had an error greater than 1 % and these were all for 18F data sets.

Conclusion: The proposed clinical workflow was able to separate arterial and venous counts.