Leventouri, Theodora

Stereotactic Body Radiation Therapy (SBRT) is a modern precision radiation therapy to deliver the dose in 1 to 5 fractions with high target dose conformity, and steep dose gradient towards healthy tissues. The dose delivered is influenced by the leaf width of the MLC, especially in case of SBRT. Treatment plans with high definition (HD) MLC having leaf-width 2.5 mm and normal MLC having leaf-width 5 mm, were compared to quantify dosimetric and radiobiological parameters. Dosimetric parameters conformity indices (CI), gradient indices (GI) and heterogeneity indices (HI) were compared. The radiobiological parameters were evaluated by normal tissue complication probability (NTCP) and tumor control probability (TCP) based on the equivalent uniform dose (EUD). The results show that there is dosimetric and radiobiological merit of the HD MLC over the normal MLC. However, the improvement is not consistent with all the plans and thus further research is required prior to conclusion.

Quality Assurance (QA) for medical linear accelerators (linac) is the primary concern in
external beam radiation therapy. In this research, we have developed a MATLAB-based software
named Quality Assurance for Linacs (QALMA), which is unique, due to cost-effectiveness, user
friendly interface, and customizability. It includes five modules to perform different QA tests: Star
Shot analysis, Picket Fence test, Winston-Lutz test, MLC log file analysis, and verification of light
& radiation field coincidence. We also pay attention to quality assurance of 6DOF treatment couch
that plays a very important role in radiation therapy. We developed an Arduino based 3D printed
6DOF robotic phantom to check the accuracy of the treatment couch. This robotic phantom was
experimentally validated under clinical standards, and customizable upon requirements of the
quality assurance Task. The current features of this robotic phantom open development opportunities
beyond the specific couch application, such as organs motion simulation.

Intracavitary high-dose-rate brachytherapy (HDRBT) is a treatment option for
endometrial cancer, depending on the cancer stage. Because of the steep high dose gradient
of HDRBT, very small differences in the treatment plans, surrounding organ’s anatomy, or
procedures during the treatment could potentially cause significant dose variation to the
tumor, as well as organs at risks (OAR) nearby the treatment area, which could result in
unwanted radiobiological side effects. In this retrospective study, the radiobiological plan
evaluation parameters Equivalent Uniform Dose (EUD), Normal Tissue Complication
Probability (NTCP) are used as assessment tools to evaluate HDRBT plans. Furthermore,
gynecological applicator position in the coordinate system, and possible dose variations to
the tumor and critical organs from the initial fraction in comparison with subsequent
fractions over the entire multi fractionated treatment are studied. The evaluations were
performed for 118 HDR treatment plans for 30 patients by registration of the subsequent treatment plans into the initial CT-image guided plan. Dose fractionation regimens varied
from 4Gy to 7Gy per fraction, 1 or 2 fractions per week, depending on the cancer stage.
Our results demonstrate no significant radiobiological impacts on organs at risks (OAR).
In addition, the results of the applicator positions’ study indicate that improvement of
immobilization and localization devices are recommended.

It is hypothesized that the uncertainty of the Synchrony® model from the rotation of a
geometrically asymmetrical single fiducial shall be non-zero during the motion tracking.
To validate this hypothesis, the uncertainty was measured for a Synchrony® model built
for a respiratory motion phantom oriented at different yaw angles on a Cyberknife®
treatment table.
A Mini-ball Cube with three cylindrical GoldMark™ (1mmx5mm Au) numbered
fiducials was placed inside a respiratory phantom and used for all tests. The fiducial with
the least artifact interference was selected for the motion tracking. A 2cm periodic,
longitudinal, linear motion of the Mini-ball cube was executed and tested for yaw
rotational angles, 0° – 90°. The test was repeated over 3 nonconsecutive days. The
uncertainty increased with the yaw angle with the most noticeable changes seen
between20° and 60° yaw, where uncertainty increased from 23.5% to 57.9%. A similar test was performed using a spherical Gold Anchor™ fiducial. The uncertainties found
when using the Gold Anchor™ were statistically lower than those found when using the
GoldMark™ fiducial for all angles of rotation.
For the first time, it is found that Synchrony® model uncertainty depends on fiducial
geometry. In addition, this research has shown that tracking target rotation using a single
fiducial can be accomplished with the Synchrony® model uncertainty as it is displayed on
the treatment console.
The results of this research could lead to decreased acute toxicity effects related to
multiple fiducials.

The purpose of this study is to verify and validate the dose at various points of
interest in accelerated partial breast irradiation (APBI) treated with the Strut Adjusted
Volume Implant (SAVI) applicator using Thermoluminescent Dosimeters (TLDs). A set
of CT images were selected from a patient’s data who had received APBI using the SAVI
applicator. The images were used to make 3D models. TLDs were calibrated for
Brachytherapy. Various points of interest were marked out and slots were carved in the 3D
models to fit the TLDs. CT scans were taken of the 3D models with expanded SAVI
applicator inserted. A plan was made following B-39 protocol. The TLDs were read and
the absorbed doses were calculated and compared to the delivered doses. The results of this
study show that the overall average reading of the TLDs is within expected value. The TPS
shows overestimated dose calculations for brachytherapy.

Purpose: To highlight the importance of provide proper solid immobilization initially and in every
treatment in ABPI with brachytherapy. Materials and Methods: 100 patients receiving brachytherapy
treatments in ABPI using the Savi applicators were considered for the study. The CT scans used in the
initial plan was compared with every scan obtained before each treatment. Each of these scans were exported in the planning system and registered with the initial plan. Dosimetric evaluations were
performed with their consequences to the ribs and the skin surface. Results: Making the dosimetric
comparison for the critical points on the ribs and skin due to very small changes in the interfractionation
position revealed dramatic differences in the maximum dose to these critical organs. The
cavity's volume manifested changes between fractions as well as the distances to the two critical
organs. Therefore the maximum dose manifested variance between 10 and 32 in both of the sites
Conclusions: This study demonstrates that using CT scan before each treatment will minimize the risk
of delivering undesired high doses to the critical organs. This reveals the urgent need of increasing and
improving the immobilization methods when treating ABPI with Savi. In 30 of the cases re-planning was
necessary between fractions therefore we conclude that in each case the treatment and planning teams
must be prepared for re-plan as needed.

Simulated Annealing algorithm is utilized for Intensity Modulated Radiation Therapy IMRT optimization.
The goal in IMRT is to give the prescribed radiation dose to the tumor while minimizing the dose given
to normal organs.

Purpose: The purpose of this study is to compare the dose-volumetric results of intensity modulated
radiation therapy (IMRT) with RapidArc (RA Varian Medical Systems, Palo Alto, CA) for whole breast
irradiation. Methods: 25 patients previously treated for whole left breast (either RapidArc plan or IMRT)
were the subjects of this planning study. Eclipse v 11.0.47 was used to make all retrospective plans using
the same contours, energy, machine and normalization. Prescription dose to the planning target volume
was 5000 Gy in 25 fractions. All plans were normalized such that 100% covered 95% of planning target
volume (PTV). Results: V10, V20 and Dmean Gy of left lung significantly differed between the two plans
(p-value <0.0001, =0.0473 and <0.0001 respectively), but V30 Gy did not (p-value 0.463). V25, D33 and
Dmean Gy of heart significantly differed between the two plans (p-value =0.034, <0.0001 and 0.01
respectively), but V10 Gy did not (p-value 0.058). V5 of both right breast and right lung significantly
differed between the two plans (p-value <0.0007 and =0.0112, respectively). Also Dmean of both right
breast and right lung significantly differed between the two plans (p-value <0.0001 for both). The mean
conformity index did not significantly differ, p-value 0.142. There was a significant difference between the
mean MUs of the two plans as well, p-value <0.0001.

There are many available options today for treating small lesion cancer cells. Two of the
most used options are the planning systems BrainLab and Eclipse. The main difference between
the two is the algorithms that are used to calculate the dose distribution of external beam
radiation therapy. BrainLab offers a Monte Carlo based algorithm while Eclipse utilizes the
Anisotropic Analytical Algorithm. An investigative study on the quality of the planning system
is done for cases in lung, head and neck and prostate. In particular, lung cases are highly
heterogeneous which can lead to problems in the calculation. The ability to be able to plan on the
best system for individual cases can lead to better and more consistent treatments for cancer
patients.

Purpose: To explore offset values in dose optimization with pencil beam (PB) algorithm
to minimize dosimetric differences with plans calculated with Monte Carlo (MC) for lung cancer
treatment with Stereotactic Body Radiotherapy (SBRT). Methods: 20 cases of Non-Small Cell
Lung Cancer, treated with gated full motion range SBRT were selected. According to the
proximity of the Gross Tumor Volume (GTV) to the chest wall, two groups are defined.
Treatment plans were created on 4D average intensity projection (AIP) CT set with Brainlab
iPlanDose® 4.1.2 planning system. The D97 of PTV was normalized to 50Gy using the fast PB
and compared with MC. The optimized plan was then recomputed over each 4D respiratory
phase, and compared with MC. Results: The mean difference in the PB and MC D97 of the ITV
was 10.5% (±0.8%) of the prescription dose (50Gy). PB algorithm showed 2.3-2.4% less
overestimation to the D97 of the ITV, when comparing to MC, in the maximum exhalation phase
than in the maximal inhalation phase. Significantly smaller dose difference between PB and MC
is also shown in plans for peripheral lesions (7.7 ± 0.7%) versus for central lesions (12.7±0.8%)
(p< 0.01). Conclusions: The dosimetric differences between PB and MC can be reasonably
predicted depending on the location of lesion in the lung, and may be used as offset value in dose
optimization with PB. Caution is suggested when using the maximum inhalation phase for
treatment planning.