In medical images intratumour heterogeneity well translates specific cancer cells properties in a fast and non-invasive way. Medical community and researchers have developed proper tools (texture features, shape features…) permitting the quantification and the analysis of this tumour heterogeneity in particular in PET images. Yet so far, both doctors and researchers have mainly focused on the use of those tools before assessing their relevance and robustness. In this study, a new heterogeneous PET phantom allowing an extensive understanding of the key concept of heterogeneity quantification in PET images is designed and partially developed. This phantom aims at representing a complex enough environment mimicking clinical conditions to properly challenge PET heterogeneity data extraction and quantification methods further than commercial phantom already do. This study focuses, first, on defining sound specifications, design methodology and manufacturing method for this new object. Second, the study focuses on designing and developing the defined solution. Third, a proof-of-concept analysis is conducted within the study to test and validate the developed PET phantom prototype. It can be concluded that the produced PET phantom was successfully designed and developed and can be used by other operators; yet, there is still room for improvement. Finally, besides developing a novel PET phantom, this study yields recommendations to improve the work done toward a more handy, flexible and realistic tool.

Pencil beam scanning is becoming a more common treatment modality. However, its ability to deal with moving targets is known to be limited, as beam motion and target motion can reinforce each other, deteriorating the planned dose distribution in what is called the interplay effect. Literature concerning breathing motion usually investigates regular patterns. This work aims to investigate the magnitude of the interplay effect when considering irregular breathing signals. In silico calculations of dose distributions were made in the treatment planning system RayStation (version 7.99), using an XCAT phantom with 50 CT phases to model moving patient anatomy. An interplay calculator was included in RayStation, allowing calculation of disturbed doses based on a treatment plan and an irradiation time model for a proton therapy accelerator. The target investigated was a spherical liver tumour with 5cm diameter, irradiated with two beams delivering a prescription dose of 63 Gy. Plans without and with 5x layered repainting were created. Clinically realistic regular breathing patterns were generated to establish a baseline, after which irregularities were introduced. The basic form for all patterns was a sin^4 signal, with regular signal amplitudes ranging from 6 to 18 mm, period ranging from 3 to 4 s and phase between 0 and 2π rad. Considered irregularities were baseline shifts up to 34 mm, changing amplitudes between 6 and 18 mm, changing periods between 1.6 and 5.2 s and combinations. Evaluation was done by looking at dose homogeneity HI_5 and the fraction of the CTV volume that received a dose outside of the clinical limits of 95% and 107%, V_107/95. For the regular patterns, both a systematic and a randomised analysis were carried out. For irregular patterns, only a systematic analysis was carried out. The mean HI_5 was found to be 31% for regular patterns; the means of all irregular patterns stay below this, even though the size of the irregularities for some breathing patterns was very large. The mean V_107/95 was found to be 0.7 for regular patterns. Irregularities did not cause further deterioration. Five times layered repainting causes a statistically significant decrease in magnitude of the interplay effect across all breathing patterns by 50-80%, but is approximately 50% less effective against baseline shift than against other types of breathing. Interplay effect size correlates strongly with amplitude, but this correlation can be obscured because period and phase introduce very large variance. The interplay effect in general is large for investigated target size, prescription dose, beam configuration and machine performance. It can cause up to 100% of the CTV to receive a clinically unacceptable dose and lead to large inhomogeneities. Irregular breathing was not found to be notably worse. Repainting is effective, even against irregular breathing, but baseline shifts can undermine its effectiveness. Separately considering breathing irregularities for tumours similar to that investigated here is deemed of low importance; it is more important to properly model the magnitude of the interplay effect using an accurate, individualised breathing pattern.

Background and purpose: One of the main challenges in external beam radiotherapy treatment of locally advanced cervical cancer patients is dealing with bladder and rectum filling. Organ filling causes inter­fraction motion of the uterus, requiring large treatment planning volumes, or a plan library. Current assessment of tumor position is mainly done by visual inspection of a Cone Beam Computed Tomography (CBCT) scan. Eventually, this can lead to inter-­ and intra-­observer variability when choosing the best treatment plan from the plan library based on bladder filling. The incoming introduction of auto­contouring tools to obtain automatically-­generated (AG) contours of the bladder and the rectum on CBCT scans, allows the easier identification of these organs at risk and consequently, faster localization of the tumor region. However, to rely on these AG contours in the decision of plan selection, it is necessary to know if they have been reliably segmented. The goal of this project is to develop a strategy based on quantitative image features, to evaluate the quality of the AG contours to know if they are suitable for plan selection assessment. Materials & Methods: 140 LACC patients from Erasmus MC were included. For each patient, bladder and rectum contours were obtained from each of the CBCT scans done throughout the treatment (five fractions (CBCT scans) per patient). These contours were automatically­generated using a deep learning­-based auto­segmentation algorithm. Gold­-standard contours were manually delineated in some CBCT scans, but the rest of the automatically­generated contours did not have the corresponding ground-­truth contour, hence they were labeled with a score between 1 (bad quality) and 5 (good quality). For consistency, gold­-standard contours were included in the dataset with the class label 5. The contours were relabeled to have a binary classification problem, and those with label 3 were removed. Each contour volume was divided into three subregions: core region, inner and outer shell. This contour data was used for a comparison study between two supervised machine learning (ML) methodologies: Random forest (RF) networks and Logistic Regression (LR). For both strategies, feature extraction and selection were implemented. In RF methodology, a prior step of dimensionality reduction using principal component analysis (PCA) was performed. In LR, univariate feature selection followed by a multivariate logistic regression analysis was done. Before implementing the classifiers, the dataset was split into a training set and a test set. The ML models were trained using the training set, and they were tested on new unseen data. Predictions on the test data were obtained and used for evaluation of the model's performance using evaluation metrics: accuracy, sensitivity, specificity, confusion matrix, ROC curve, and AUC. Results: The RF classifier performed on the bladder test data with an AUC value of 0.87, while for the LR model, the value obtained was 0.77. The trained RF model identified the accurate and inaccurate bladder contours with a sensitivity of 94% and a specificity of 54%. The trained LR model resulted in a sensitivity of 91% and a specificity of 42%. In the case of the rectum, the RF classifier performance is indicated with the AUC value of 0.89, while the LR model obtained a value of 0.84. In the case of sensitivity and specificity, the RF model got 96% and 38%, and the LR classifier 95% and 38%, respectively. Conclusion: Random forest classifiers give the best results in terms of performance and classification skills for the OARs considered, especially for the bladder. It has been demonstrated that quantitative image features, paired with the corresponding contour class label, can be used for deriving statistical relationships from the data. This allows the identification of contouring errors and classifying the contours based on their quality. With the increasing automation of different steps in the radiotherapy treatment workflows, the automatic contour QA tool developed would be a key step in the process to ensure a faster, more feasible, and consistent plan selection. The tool could act as a support tool for radiotherapy technicians when choosing the plan from the plan library that best fits the daily anatomy of the patient.

Particle therapy (PT), including proton therapy, has important advantages compared to external beam photon therapy (section 1.1). This is because most of the therapeutic effect of a proton beam is localized at the endpoint, where most of its energy is imparted to the medium (Bragg peak), with nearly no dose deposited beyond that point. However, the highly localized dose deposition makes proton therapy more sensitive to (1) patient morphological alterations, including tumor progression / regression, (2) organ motion, (3) patient setup errors, (4) tissue lateral heterogeneities that render the results obtained with non-Monte-Carlo-based treatment planning algorithms unreliable to some degree, (5) beam characteristics utilized for treatment planning, and (6) the conversion of Hounsfield units (computed tomography data), to tissue density and stoichiometry. In addition, uncertainties in the mean excitation potential I, necessary to calculate the stopping power of the penetrating ions, further contribute to potential beam range inaccuracies. Given the aforementioned sources of treatment error, an imaging technique capable of providing feedback proportional to the quality of the treatment being delivered is highly desired and a very active field of research in proton therapy (section 1.2). Specifically, it is of utmost importance to develop an imaging technique capable of providing feedback with respect to the in vivo beam range, especially when highly-heterogeneous beam paths are crossed by a pencil beam. Such a imaging technique can make use of secondary gamma (γ) radiation emitted by the patient, as a result of nuclear interactions between the projectiles and the nuclei of the irradiated medium. These techniques are mainly divided into two categories, according to the type of secondary γ rays probed: (1) positron emission tomography (PET), which makes use of delayed emission, namely pairs of 511 keV annihilation photons, resulting from β+-decay; and (2) prompt gamma (PG) imaging, which makes use of the emission of single photons typically on a sub-nanosecond timescale…@en

Photon-counting detectors (PCD) for medical X-ray computed tomography (CT) are designed to measure the number of X-ray photons incident on a detector pixel as well as the energy of the individual X-rays. They are expected to yield improvements in image quality for a given radiation dose, and to offer opportunities for spectral imaging beyond dual-energy techniques. However, the fluence rate incident on the detector can exceed 108 mm-2 s-1 in CT, so that the detector pulses generated by the X-rays likely pile up on each other, which distorts the measurement. The semiconductors CdTe and Cd1-xZnxTe (CZT, x ≈ 0.1-0.2) are commonly considered efficient X-ray absorbers that provide sufficient rate capability (fast pulses in the order of 101 ns and a high pixel density ≥ 4 mm-2) and energy resolution (8-20% FWHM at 60 keV). In such detectors, an X-ray is converted into electron-hole pairs, which travel to (pixelated) electrodes, on which they induce a current pulse. However, the cost-effective synthesis of material of sufficient quality to make this a stable and reliable detection process appears to remain an issue. Thus, the aim of this thesis is to explore the photon-counting performance, e.g., the rate capability and energy resolution, of an alternative detector concept based on a scintillator, which converts an X-ray into a light pulse, and a silicon photomultiplier (SiPM), which detects the light. Since such a detector relies on light rather than charge transport, it may enable cost-effective manufacturing of stable and reliable PCDs... @en

The need of image (frame) acquisitions within short time intervals is of major importance for preclinical SPECT imaging. The short frame times enable higher temporal resolution which is required in bio-distribution and pharmacokinetic studies where fast dynamic imaging is performed. The present study evaluates and compares the performance of two different preclinical multipinhole SPECT systems (NanoScan, VECTor) for short frames acquisitions. Prior to the systems comparison, the comparison and selection of the best performing fast imaging mode provided by NanoScan system (Mediso) was performed. The fast imaging modes of this system provide acquisitions with 1,2 and 3 detector position around the animal bed. This comparison was performed by using uniform phantoms (syringes) and the rods of the NEMA NU4IQ phantom (frames: 6-30s). The down-sized version of NU4IQ phantom (SPECTIQ phantom) was used in this study to compare the performance of VECTor (MILabs) and NanoScan when performing acquisitions with short frame times (18s-600s, whole body scans). The quality of the acquired images was assessed in terms of absolute quantification (recovery coefficient), noise levels and visual evaluation. The quantification with the NanoScan was accurate (±5%) regardless of times frames duration and activity concentrations when imaging large structures. The increase in number of detector positions yielded images with lower noise levels. In the case of small structures, acquisition with 3 detector positions (Semi-3 mode) appeared to provide more accurate activity recovery compared to acquisitions with 1 (stationary mode) and 2 detector positions (Semi-2 mode). Especially in the case of the 2mm diameter rod of the NU4IQ phantom, the Semi-3 mode appears to provide significantly more accurate activity recovery (30s frame). The systems comparison showed activity recovery with up to 5% deviation from the dose calibrator measurement when imaging the uniform region of SPECTIQ phantom (d = 21mm). Both systems could recover the three largest rods (d = 1.5, 1.0, 0.75mm) for the longest frames used(180,360,600s). None of the systems could recover the two smallest rods of the phantom (d=0.5,0.35mm). As the frame time decreased, both systems could recover less number of rods. VECTor appeared to provide higher activity recovery than NanoScan for the three largest rods of the phantom. However, as the frame time decreased the differences became less significant. Furthermore, VECTor provided and 22.2% and 46.6% less spillover in airand water-filled phantom regions (after reaching convergence) than NanoScan did. The performances of two preclinical SPECT systems (NanoScan, VECTor) for short time acquisitionswere compared. The conducted experiments showed that the systems perform equally when conducting short frames imaging. Furthermore, the fast imaging mode of NanoScan employing three detector positions showed better performance than the other two fast imaging modes provided by this system.

The goal of this research was to perform a 3D end-to-end test on a MR-linac to check the whole workflow using a clinical treatment plan. Dosimetric gel was used to obtain 3D spatial information, with the phantom in the same position for irradiation and scanning. In order to achieve this, fundamental elements of gel dosimetry needed to be investigated. In the MR-linac, irradiation is delivered in the presence of a permanent magnetic field. Therefore, the dosimetric response within a 1.5 T magnetic field should be validated. It is also important to investigate the time-dependence of the gel. It is preferable to read-out the gels within approximately one hour, so that the phantom does not have to be moved. Ideally, scanning and irradiation would be done at the same time, to see the dynamical dose delivery. The VIPAR gel was used for this research. The experiments demonstrated that R2 values for doses irradiated with magnetic field were the same as R2 values for the same dose irradiated without magnetic field. R2 values are still proportional to the dose. It was also shown that it is possible to scan the phantom within 20 minutes after irradiation. Sensitivity is at its highest after approximately 8 hours and stays stable afterwards, so scanning after 8 hours will improve the read-out accuracy. It was also possible to make a fit for the R2 versus time plots, which makes it possible to correct for change over time. The fit can be divided in two linear parts if time is plotted on a logarithmic scale, one fit for the time points before 7 hours, one for the time points after 7 hours. The partial doses acquired by the gel during radiation delivery were estimated. The equivalent R2 values then agreed with the extrapolated fit to within 4%. This is a good indication that dynamic gel (4D) dosimetry may be achievable. A protocol for a relative end-to-end test was also developed. From the preliminary results, it appeared that a relative end-to-end test can be performed with the read-out of gel within 1 hour. A new MR sequence needs to be developed. For this end-to-end test, the sequence needs to scan a larger volume with a higher resolution, therefore, the scan time will increase and real-time dosimetry will not be possible. Changing the MR sequence might also change the optimal irradiation-scanning interval and the R2 versus time curve. To perform absolute dosimetry, an extra calibration would be required.