Improvements in Small Animal Dosimetry: CIX3 Irradiator Characterization, Novel Phantom Investigation, and Shepherd Cs-137 Irradiator Dose Uniformity Analysis

Thumbnail Image




Journal Title

Journal ISSN

Volume Title

Repository Usage Stats



Project 1 (Chapter 2): X-Strahl CIX3 CharacterizationPurpose: In Fall 2021 Duke University purchased an Xstrahl CIX3 Cabinet Xray Irradiator. A characterization of this machine was performed to determine its dosimetry characteristics and best practices for radiobiological studies at Duke. Basic irradiator acceptance tests were expanded on to fully characterize this new machine. Two unique aspects of this machine were of particular interest. First a unique (non-uniform) filter design was investigated to determine if it has unintended side effects on field conformity. Second the lack of accessible cable ports made thermoluminescent dosimeters (TLDs) the primary dosimeters for experiments. For this reason additional investigations were conducted characterizing TLDs. Materials and Methods: To assess the dosimetric properties of this new machine the following parameters were investigated: output consistency, beam quality, field uniformity, and exposure rates. Output consistency was measured by comparing expected and observed max energy in kVp using a Piranha X-ray multimeter. Beam quality was measured as half value layer of aluminum and compared to expected results from Spekcalc. Spekcalc was used to determine energy fluence and mean energy of the spectrum. Field uniformity was assessed using Gafchromic ™ EBT3 film on an Epson Expression 10000 XL scanner with lateral response artifact correction factors and film calibrated using a NIST traceable 0.18 cc ion chamber. Exposure rates were characterized using a NIST traceable 0.18cc ion chamber and varying filtration, tube energy (kVp), and tray position for all available configurations. TLD energy response, positional response, and batch correction factor techniques were characterized on this machine. Energy response was determined by irradiating TLD’s to a range of energies (70-300 kVp) and the charge response reliv ative to the exposure (50 R) received was determined. Positional response in the field was investigated using the Duke Radiation Dosimetry Laboratory (DRDL) TLD holder and Gafchromic ™ EBT3 film. The relative exposure each TLD received was determined and compared to the ion chamber exposure. Results: The CIX3 had consistent energy output as measured by max energy conformance. The average difference from input voltage to output voltage was 0.84 % with the worst being 1.6 % (150 kVp, 1 mm copper Filtration). Theoretical estimations (Spekcalc) of the beam quality had good agreement with measured half value layers (using Piranha) with an average difference of 1.36 % and worst error of 2.99 % across the energy ranges sampled (50-150 kVp). Field uniformity results indicated general conformance to machine data (90 % within 25.9 cm diameter field) but some non-uniformities were identified. Areas of higher dose (105-110%) relative to the center were observed in the upper right quadrant of the field (from beam eye perspective). TLD energy response followed expected over-response in lower energy ranges, reducing to a normal response as energy increases. The highest over response was at 70 kVp, a 15 % over response compared to 300 kVp (the max energy of the CIX3). The film study to determine TLD positional response determined there were unequal exposures to the 50 TLDs. The trend observed an increase in exposure consistent with the field uniformity results in which the dose relative to the center of the field increased by 5-10 % towards the upper right quadrant. This effect was more pronounced at the lower energy level sampled (90 kVp). Conclusions: The full characterization of the CIX3 was very important to understand the nuances of the new machine. Conformance in beam quality results gave good indications that the machine is operating as designed and radiobiological studies expect to have consistent results between experiments under like conditions. The unique field uniformity results observed could help inform future experimental planning. More importantly it is an extremely important finding for TLD dose calibration. Since TLDs experience at most 6-7 % over exposure compared to the ion chamber it is important to use these findings in future calibrations. Project 2 (Chapter 3): Dose Depth of Small AnimalWater Phantom Purpose: Much is understood about the midpoint dose estimation of small animal phantoms and it has been the focus of DRDL to conduct dosimetry using this value. However, little data was available on the dose depth profile of small animal phantoms. This investigation sought to fill in that dose depth data, compare doses under varying experimental conditions in order to fully understand how the dose is distributed along the beam axis for small animal phantoms. Materials and Methods: A small animal water phantom (50 cc water vial) was characterized on the CIX3 using Gafchromic ™ EBT3 film. Film was calibrated using a NIST traceable 0.18 cc ion chamber. The phantom was irradiated with strips of film (15 total) varying filtration and whether a backscatter plate (uniform piece of acrylic) was included. The film was scanned and analyzed using Film QA software and aggregated dose depth profiles were determined using R Studio and Excel. Results: The dose depth of the phantom was characterized with a coefficient of variation of about 2-3 % across all depths and configurations. The inclusion of the backscatter plate improved the dose uniformity by an average of 2.14 % with most improvements coming from the bottom half of the phantom (closest to the backscatter plate). Average dose rates under each configuration were determined. The midpoint dose rate was found to have good conformance (within 0.5%) to the averaged dose rate across the depth. The dose rate increased by 33 % when using the backscatter plate due to the increased backscatter spectrum and the inverse square law effects. Conclusions: This data gave increased certainty in using the midpoint dose as a surrogate measure for whole body dose averages in small animal phantoms. The improvement in dose uniformity when using the backscatter plate seems like a promising addition to future experimental configurations. However in application the size of the backscatter plate makes it unusable for large experiment samples and it should be implemented only in specific studies as determined during the dosimetry consult with DRDL. Project 3 (Chapter 4): Novel Mouse Phantom Investigation Purpose: In recent years DRDL developed a novel mouse phantom based on feedback from researchers. This novel phantom flattened the mouse to mimic a mouse laying on an irradiation platform, especially when sedated. This investigation sought to determine if the novel ’flat’ phantom made of polymethyl methacrylate (acrylic) (PMMA) had a significant difference in dosimetry when compared with a standard cylindrical phantom made of soft tissue equivalent material. Materials and Methods: To compare the dosimetric differences each phantom a FLUKA Monte Carlo simulation was compared with experimental results from TLDs on the CIX3. In the simulation the dose was compared using a midpoint dose volume. In the experimental design TLDs were used to determine the dose at the midpoint of each phantom. The dose rates were analyzed and compared to determine if there was a significant difference. Results: The Monte Carlo results indicated there were very slight differences between phantom rates. The cylinder phantom had a dose rate of 1.83±0.038 Gy/min while the flat phantom had a midpoint dose of 1.81 ± 0.045 Gy/min, 1.1 percent lower than the cylindrical phantom. An unpaired t-test was performed to determine if the samples were different and was found to give a p-value of 0.71 which gives a high probability that the sampled data are not significantly different. The experimental results were found to be similar. The cylinder phantom had a dose rate of 1.77 ± 0.05 Gy/min while the flat phantom had a midpoint dose of 1.81 ± 0.045 Gy/min, 2.25 percent higher than the cylindrical phantom. Again a t-test was performed and these were determined to not be significantly different (p = 0.31). Conclusions: The flat phantom therefore is very similar to the dosimetry found in the cylinder phantom. Small variations due to material properties, height of phantom, scatter material, and attenuating material all balanced out to provide dosimetry properties that are similar. This means that the cheaper to manufacture flat phantom is just as good as the much more expensive cylinder phantom and both can be used for small animal dosimetry. Project 4 (Chapter 5): Shepherd Mk I - 68A Dose Uniformity Purpose: In 2009 a study was conducted to determine the dose uniformity in a Shepherd Mk I-68A Cs-137 irradiator. Since then vast improvements have been made on film design and software to analyze scanned film with improved accuracy. A follow up study was designed to revisit this previous characterization and update the dose uniformity of the irradiator cavity using these new dosimeters and techniques. Materials and Methods: Gafchromic ™ EBT3 film was calibrated using a NIST traceable 0.18 cc ion chamber on the Shepherd irradiator in each of the three positions available. Large film sheets were then irradiated in a 2 mm acrylic holder in all positions under rotating and non-rotating configurations. The scanned film was analyzed using FilmQA software, R studio, and Excel to determine the dose uniformity relative to the dose in the center. Results: An ion chamber sample was compared to the film results and found to be in good agreement (within 1 %) which indicated the film was appropriately irradiated, scanned, and calibrated. Rotating dose distributions in positions 2 and 3 were nearly equivalent to manufacturer predicted isodose distributions with noted discrepancies at the edges of the field. At a height of 15 cm to achieve dose uniformity of 100 % ± 5 % the rotating tray has a usable radius of 7 cm from the center in position 2 and 8 cm in position 3. Stationary dose distributions were compared to previous uniformity and found to be in general agreement in the center. However the isodose mapping previously characterized did not include a scanner lateral response artifact correction factor which indicated a better uniformity than what was found in this experiment. Position 1 results were similar to previous dose distributions and most importantly confirmed the positioning of the source in the chamber. Conclusions: The updated dose uniformity data provides QA feedback as part of a larger dosimetry program for Duke University. These results indicated that this irradiator was and still is performing as expected and no mechanical failures have caused a source to become misaligned or any major changes to the expected dosimetry.





Filip, Kevin T (2023). Improvements in Small Animal Dosimetry: CIX3 Irradiator Characterization, Novel Phantom Investigation, and Shepherd Cs-137 Irradiator Dose Uniformity Analysis. Master's thesis, Duke University. Retrieved from


Dukes student scholarship is made available to the public using a Creative Commons Attribution / Non-commercial / No derivative (CC-BY-NC-ND) license.