The aim of this work is to enhance the detection and classification pipelines of an artificial intelligence (AI)-based decision support system (DSS) for diagnosing acanthamoeba keratitis (AK), a vision-threatening disease. The images used are taken with the in vivo confocal microscopy (IVCM) technique, a complementary tool for clinical assessment of the cornea that requires manual human analysis to support diagnosis. The DSS facilitates automated image analysis and currently aids in diagnosing AK. However, the accuracy of AK detection needs improvements in order to use it in clinical practice. To address this challenge, we utilize image brightness processing through multiscale retinex (MSR), and develop a custom-built image processing pipeline with deep learning model and rule-based strategies. The proposed pipeline replaces two deep learning models in original DSS, resulting in an overall accuracy improvement of 10.23% on average. Additionally, our improved pipeline not only enhances the original system’s ability to aid AK diagnosis, but also provides a versatile set of functions that can be used to create pipelines for detecting similar keratitis diseases.

Modern medical imaging techniques have revolutionized health care over the last decades, providing clinicians with high-resolution 3D images of the inside of the patient's body without the need for invasive procedures. Detailed images of the vascular anatomy can be captured by angiography, providing a valuable source of information when deciding whether a vascular intervention is needed, for planning treatment, and for analyzing the success of therapy. However, increasing level of detail in the images, together with a wide availability of imaging devices, lead to an urgent need for automated techniques for image segmentation and analysis in order to assist the clinicians in performing a fast and accurate examination. To reduce the need for user interaction and increase the speed of vascular segmentation,  we propose a fast and fully automatic vascular skeleton extraction algorithm. This algorithm first analyzes the volume's intensity histogram in order to automatically adapt the internal parameters to each patient and then it produces an approximate skeleton of the patient's vasculature. The skeleton can serve as a seed region for subsequent surface extraction algorithms. Further improvements of the skeleton extraction algorithm include the expansion to detect the skeleton of diseased arteries and the design of a convolutional neural network classifier that reduces false positive detections of vascular cross-sections. In addition to the complete skeleton extraction algorithm, the thesis presents a segmentation algorithm based on modified onion-kernel region growing. It initiates the growing from the previously extracted skeleton and provides a rapid binary segmentation of tubular structures. To provide the possibility of extracting precise measurements from this segmentation we introduce a method for obtaining a segmentation with subpixel precision out of the binary segmentation and the original image. This method is especially suited for thin and elongated structures, such as vessels, since it does not shrink the long protrusions. The method supports both 2D and 3D image data. The methods were validated on real computed tomography datasets and are primarily intended for applications in vascular segmentation, however, they are robust enough to work with other anatomical tree structures after adequate parameter adjustment, which was demonstrated on an airway-tree segmentation.

Histopathological evaluation and Gleason grading on Hematoxylin and Eosin(H&E) stained specimens is the clinical standard in grading prostate cancer. Recently, deep learning models have been trained to assist pathologists in detecting prostate cancer. However, these predictions could be improved further regarding variations in morphology, staining and differences across scanners. An approach to tackle such problems is to employ conditional GANs for style transfer. A total of 52 prostatectomies from 48 patients were scanned with two different scanners. Data was split into 40 images for training and 12 images for testing and all images were divided into overlapping 256x256 patches. A segmentation model was trained using images from scanner A, and the model was tested on images from both scanner A and B. Next, GANs were trained to perform style transfer from scanner A to scanner B. The training was performed using unpaired training images and different types of Unsupervised Image to Image Translation GANs (CycleGAN and UNIT). Beside the common CycleGAN architecture, a modified version was also tested, adding Kullback Leibler (KL) divergence in the loss function. Then, the segmentation model was tested on the augmented images from scanner B.The models were evaluated on 2,000 randomly selected patches of 256x256 pixels from 10 prostatectomies. The resulting predictions were evaluated both qualitatively and quantitatively. All proposed methods outperformed in AUC, in the best case the improvement was of 16%. However, only CycleGAN trained on a large dataset demonstrated to be capable to improve the segmentation tool performance, preserving tissue morphology and obtaining higher results in all the evaluation measurements. All the models were analyzed and, finally, the significance of the difference between the segmentation model performance on style transferred images and on untransferred images was assessed, using statistical tests.

There are many clinical situations where magnetic resonance imaging (MRI) is preferable over other imaging modalities, while the major disadvantage is the relatively long scan time. Due to limited resources, this means that not all patients can be offered an MRI scan, even though it could provide crucial information. It can even be deemed unsafe for a critically ill patient to undergo the examination. In MRI, there is a trade-off between resolution, signal-to-noise ratio (SNR) and the time spent gathering data. When time is of utmost importance, we seek other methods to increase the resolution while preserving SNR and imaging time. In this work, I have studied one of the most promising methods for this task. Namely, constructing super-resolution algorithms to learn the mapping from a low resolution image to a high resolution image using convolutional neural networks. More specifically, I constructed networks capable of transferring high frequency (HF) content, responsible for details in an image, from one kind of image to another. In this context, contrast or weight is used to describe what kind of image we look at. This work only explores the possibility of transferring HF content from T1-weighted images, which can be obtained quite quickly, to T2-weighted images, which would take much longer for similar quality. By doing so, the hope is to contribute to increased efficacy of MRI, and reduce the problems associated with the long scan times. At first, a relatively simple network was implemented to show that transferring HF content between contrasts is possible, as a proof of concept. Next, a much more complex network was proposed, to successfully increase the resolution of MR images better than the commonly used bicubic interpolation method. This is a conclusion drawn from a test where 12 participants were asked to rate the two methods (p=0.0016) Both visual comparisons and quality measures, such as PSNR and SSIM, indicate that the proposed network outperforms a similar network that only utilizes images of one contrast. This suggests that HF content was successfully transferred between images of different contrasts, which improves the reconstruction process. Thus, it could be argued that the proposed multi-contrast model could decrease scan time even further than what its single-contrast counterpart would. Hence, this way of performing multi-contrast super-resolution has the potential to increase the efficacy of MRI.

The treatment plan for radiation therapy with protons is based on images from a computed tomography (CT) scanner. This is problematic since the photons in the x-ray beam from the CT scanner and the protons are affected differently by the tissue in the patient, which introduce an uncertainty in the track length of the protons. The hypothesis of this study is that a new generation of CT scanners (DECT), with the capacity to simultaneously scan the patient with two photon spectra of different mean energy, will improve the tissue characterisation and which in turn reduce the uncertainty in the track length of the protons. In this study, the accuracy and precision of a DECT-based method from the literature is compared to the conventional calibration method used today at the University clinics in Sweden to relate the attenuation of the photon beam to the slowing down of the protons. The methods are tested on CT images of a phantom, a plastic body containing tissue equivalent plastic inserts of known elemental composition. The results turned out to be inconclusive as there were large uncertainties in the measurements. The method has potential, as has been shown in the literature, but there are many questions that need to be answered before the method is ready to be implemented at the clinic. / En proton som färdas genom människokroppen deponerar endast en liten del av sin energi längs vägen innan den plötsligt deponerar allt i slutet på dess bana. Hur lång dess bana är beror på protonens ursprungliga energi och den atomära sammansättningen hos vävnaden den passerar igenom. Om sammansättningen är känd går det genom att justera den initiala energin bestämma banlängden. Denna egenskap gör protonen väldigt attraktiv för strålterpi, då det innbär möjligheten att behandla med hög precision samt bespara frisk vävnad onödig dos. Strålterapi med protoner planeras idag med bilder från en skiktröntgen (CT) som underlag. Ett problem med det är att röntgenstrålarna från CT-skannern påverkas annorlunda än protonerna av vävnaden, vilket introducerar en osäkerhet i protonernas banlängd. Hypotesen i denna studie är att en ny generation av CT-scanner (DECT), med möjlighet att simultant skanna patienten med två fotonspektran av olika medelenergi, på ett bättre sätt ska kunna bestämma den atomära sammansättningen för vävnaden och därmed reducera osäkerheten i protonernas banlängd. Noggrannhet och precision för en DECT-baserad metod från litteraturen jämförs med den SECT-baserade kalibreringsmetoden, som idag används på Universitetssjukhusen i Sverige för att relatera fotonstrålens dämpning i vävnaden till protonernas inbromsning. Metoderna testas på CT bilder av ett fantom, en plastkropp innehållandes olika cylindrar av vävnadsekvivalent plast med känd atomär sammansättning. Resultatet av den här studien är inte starkt nog för att bevisa hypotesen för studien. Det insamlade bildmaterialet innehåller höga brusnivåer jämfört med de som rapporteras i literaturen. Brusnivåer är så höga att det mesta av resultatet inte kan anses som statistiskt signifikant. Det är dessutom svårt att göra en direkt jämförelse av prestanda med befintlig teori för vävnadskaraktärisering, då bildmaterialet från de CT skanners som jämfördes är av olika typer. De resultat som publicerats i litteraturen visar att den DECT-baserade metoden har potential, men den här studien gör tydligt att det fortfarande finns frågor som måste besvaras innan metoden är redo att implementeras kliniskt.

The purpose of this project was to explore the performance of methods for estimating stopping power ratio (SPR) from Hounsfield units (HU) using dual energy CT scans, rather than the standard single energy CT scans, with the aim of finding a method which could outperform the current single energy stoichiometric method. Such a method could reduce the margin currently added to the target volume during treatment which is defined as 3.5 % of the range to the target volume + 1 mm . Three such methods, by Taasti, Zhu, and, Lalonde and Bouchard, were chosen and implemented in MATLAB. A phantom containing 10 tissue-like inserts was scanned and used as a basis for the SPR estimation. To investigate the variation of the SPR from day-to-day the phantom was scanned once a day for 12 days. The resulting SPR of all methods, including the stoichiometric method, were compared with theoretical SPR values which were calculated using known elemental weight fractions of the inserts and mean excitation energies from the National Institute of Standards and Technology (NIST). It was found that the best performing method was the Taasti method which had, at best, an average percentage difference from the theoretical values of only 2.5 %. The Zhu method had, at best, 4.8 % and Lalonde-Bouchard 15.6% including bone tissue or 6.3 % excluding bone. The best average percentage difference of the stoichiometric method was 3.1 %. As the Taasti method was the best performing method and shows much promise, future work should focus on further improving its performance by testing more scanning protocols and kernels to find the ones yielding the best performance. This should then be supplemented with testing different pairs of energies for the dual energy scans. The fact that the Zhu and Lalonde-Bouchard method performed poorly could indicate problems with the implementation of those methods in this project. Investigating and solving those problems is also an important goal for future projects. Lastly the Lalonde-Bouchard method should be tested with more than two energy spectra.

This thesis explores and explains a method using discrete curvature as a feature to find regions of vertices that can be classified as being likely to indicate the presence of an underlying tumor on a kidney surface mesh. Vertices are tagged based on curvature type and mathematical morphology is used to form regions on the mesh. The size and location of the tumor is approximated by fitting a sphere to this region. The method is intended to be employed in noninvasive radiotherapy with a dynamic soft tissue model. It could also provide an alternative to volumetric methods used to segment tumors. A validation is made using the images from which the kidney mesh was constructed, the tumor is visible as a comparison to the method result. The dynamic kidney model is validated using the Hausdorff distance and it is explained how this can be computed in an effective way using bounding volume hierarchies. Both the tumor finding method and the dynamic model show promising results since they lie within the limit used by practitioners during therapy.

Trauma care is the treatment of patients with injuries caused by external forces, for instance car crashes, assaults or fall accidents. These urgent patients typically arrive at the hospital’s Emergency Department, where they are treated by an interdisciplinary team of physicians and nurses, who collaborate to identify and address life-threatening injuries. In this thesis, the urgent phase of trauma care has been explored through observations of trauma calls and interviews with trauma care professionals, with the purpose of mapping the workflow and providing a basis for a discussion of IT systems within trauma radiology. The professionals, procedures and tools involved are collectively described as the initial trauma care system. There has been a focus on interaction between the units of this system, as well as on how decisions regarding treatment are made, often with the help of medical imaging. The initial trauma care system functions under significant time pressure, striving towards the well-defined objective of saving the life of the patient. To a great extent the system relies on standardized procedures, aiming for screening life-threatening injuries. The trauma team features a clear hierarchy and distinct roles, where the team leader role is considered vital for the team’s performance. Experience is valued and important for everyone, especially since the team often makes decisions, that may affect the future of the patient, based on incomplete information about the situation. Therefore, CT (computed tomography) images offer valuable decision-making support. The respondents are fairly satisfied with the current tools for viewing and manipulating radiological images. Little support for the need of improved or novel IT systems in trauma radiology is found, as is the use for 3D visualization of radiological images in this domain. Informants recognize communication failures and lacking teamwork as the major problems in trauma care. Difficulties like this may be decreased by education and training regarding these issues.

Dental radiography is one of the most frequent types of diagnostic radiological investigations performed. The equipment and techniques used are constantly evolving. However, dental healthcare has long been an area neglected by radiation safety legislation and the medical physicist community, and thus, the quality assurance (QA) regime needs an update. This project aimed to implement and evaluate objective tests of key image quality parameters for intra-oral (IO) X-ray images. The image quality parameters assessed were sensitivity, noise, uniformity, low-contrast resolution, and spatial resolution. These parameters were evaluated for repeatability at typical tube current, voltage, and exposure time settings by computing the coefficient of variation (CV) of the mean value of each parameter from multiple images. A further aim was to develop a semi-quantitative test for the correct alignment of the position indicating device (PID) with the primary collimator. The overall purpose of this thesis was to look at ways to improve the QA of IO X-rays systems by digitizing and automating part of the process. A single image receptor and an X-ray tube were used in this study. Incident doses at the receptor were measured using a radiation meter. The relationship between incident dose at the receptor and the output signal was used to determine the signal transfer curve for the receptor. The principal sources of noise in the practical exposure range of the system were investigated using a separation of noise sources based upon variance. The transfer curve of the receptor was found to be linear. Noise separation showed that quantum noise was the dominant noise. Repeatability of the image quality parameters assessed was found to be acceptable. The CV for sensitivity was less than 3%, while that for noise was less than 1%. For the uniformity measured at the center, the CV was less than 10%, while the CV was less than 5% for the uniformity measured at the edge. The low-contrast resolution varied the most at all exposure settings investigated with CV between 6 - 13%. Finally, the CV for the spatial resolution parameters was less than 5%. The method described to test for the correct alignment of the PID with the primary collimator was found to be practical and easy to interpret manually. The tests described here were implemented for a specific sensor and X-ray tube combination, but the methods could easily be adapted for different systems by simply adjusting certain parameters.

Osteoarthritis is a large problem affecting many patients globally, and diagnosis of osteoarthritis is often done using evidence from arthroscopic surgeries. Making a correct diagnosis is hard, and takes years of experience and training on thousands of images. Therefore, developing an automatic solution to perform the diagnosis would be extremely helpful to the medical field. Since machine learning has been proven to be useful and effective at classifying and segmenting medical images, this thesis aimed at solving the problem using machine learning methods. Multifractal analysis has also been used extensively for medical imaging segmentation. This study proposes two methods of automatic segmentation using neural networks and multifractal analysis. The thesis was performed using real arthroscopic images from surgeries. MultiResUnet architecture is shown to be well suited for pixel perfect segmentation. Classification of multifractal features using neural networks is also shown to perform well when compared to related studies.