Necrotising enterocolitis (NEC) is a potentially devastating intestinal inflammation of multifactorial aetiology in premature or otherwise vulnerable neonates. Because of the broad spectrum of presentations, diagnosis and timing of surgical intervention may be challenging, and imaging needs to be an integrated part of management. The first four studies included in this thesis used routinely collected, nationwide register data to describe the incidence of NEC in Sweden 1987‒2009, its variation with time, seasonality, space-time clustering, and associations with maternal, gestational, and perinatal factors, and the risk of intestinal failure in the aftermath of the disease. Early infant survival increased dramatically during the study period. The incidence rate of NEC was 0.34 per 1,000 live births, rising from 0.26 per 1,000 live births in the first six years of the study period to 0.57 in the last five. The incidence rates in the lowest birth weights were 100‒160 times those of the entire birth cohort. Seasonal variation was found, as well as space-time clustering in association with delivery hospitals but not with maternal residential municipalities. Comparing NEC cases with matched controls, some factors, positively associated with NEC, were isoimmunisation, fetal distress, caesarean section, persistent ductus arteriosus, cardiac and gastrointestinal malformations, and chromosomal abnormalities. Negative associations included maternal pre-eclampsia, maternal urinary infection, and premature rupture of the membranes. Intestinal failure occurred in 6% of NEC cases and 0.4% of controls, with the highest incidence towards the end of the study period. The last study investigated current practices and perceptions of imaging in the management of NEC, as reported by involved specialists. There was great consensus on most issues. Areas in need of further study seem mainly related to imaging routines, the use of ultrasound, and indications for surgery. Developing alongside the progress of neonatal care, NEC is a complex, multifactorial disease, with shifting patterns of predisposing and precipitating causes, and potentially serious long-term complications. The findings of seasonal variation, spacetime clustering, and negative associations with antenatal exposure to infectious agents, fit into the growing understanding of the central role of bacteria and immunological processes in normal maturation of the intestinal canal as well as in the pathogenesis of NEC. Imaging in the management of NEC may be developed through future studies combining multiple diagnostic parameters in relation to clinical outcome.

There is a need for quantitative CT data in radiation therapy. Currently there are only few algorithms that address this issue, for instance the commercial DirectDensity algorithm. In scientific literature, an example of such an algorithm is DIRA. DIRA is an iterative model-based reconstruction method for dual-energy CT whose goal is to determine the material composition of the patient from accurate linear attenuation coefficients. It has been implemented in a two dimensional geometry, i.e., it could process axial scans only.  There was a need to extend DIRA so that it could process projection data generated in helical scanning geometries. The newly developed algorithm (DIRA-3D) implemented (i) polyenergetic semi-parallel projection generation, (ii) mono-energetic parallel projection generation and (iii) the PI-method for image reconstruction. The computation experiments showed that the accuracies of the resulting LAC and mass fractions were comparable to the ones of the original DIRA. The results converged after 10 iterations.

SAMMANFATTNING Bakgrund Antalet personer som drabbas av kolorektalcancer ökar varje år och är den tredje vanligaste cancersjukdomen i Sverige. Det gör att efterfrågan av undersökningsmetoderna blir större för att man ska kunna ställa en korrekt diagnos och påbörja behandling. DT-kolon är en av undersökningsmetoderna och används ofta som en andrahandsmetod. Syfte Syftet med studien är att undersöka vad DT-kolon har för betydelse vid utredningen av kolorektalcancer. Metod Studien är en systematisk litteraturstudie. Insamling av data har gjorts via databaserna PubMed och CINAHL där vetenskapliga artiklar som uppfyllt inklusionskriterierna och besvarat frågeställningen har kvalitetsgranskats och analyserats. Resultat DT-kolon kan påvisa både koloniska- och extrakoloniska fynd och är en särskilt fördelaktig metod vid lokalisering av extrakoloniska fynd. Vid den preoperativa bedömningen anses DT-kolon vara en betydelsefull metod vid exakt lokalisering av patologier i kolon och rektum och är ett bra komplement till koloskopiundersökningar. DT-kolon ses som en tillförlitlig metod och har en sensitivitet på 94–99.9 %. När CAD används som granskningsmetod vid DT-kolon så ökar sensitiviteten, speciellt vid detektion av patologier som är mindre än 10 mm. Slutsats DT-kolon är en betydelsefull undersökningsmetod vid utredning av kolorektalcancer och av andra patologier i buken, kolon och rektum. Undersökningsmetoden är tillförlitlig med hög sensitivitet och med CAD som granskningsmetod ökar sensitiviteten. En stark fördel är att man får med hela bukområdet vid bildtagning och kan därmed även lokalisera extrakoloniska fynd så som exempelvis metastaser. Man kan även se tarmväggens tjocklek och utvändig kontur av kolon och rektum. Nyckelord Kolorektalcancer, Datortomografi, Kolografi / ABSTRACT Background The number of people suffering from colorectal cancer increases every year and is the third most common cancer disease in Sweden. Therefore the demand increases for the examination methods in order to be able to make a correct diagnosis and start treatment. CT Colonography is one of the methods often used as a secondary method. Aim The purpose of this study is to examine what significance CT Colonography has in the investigation of colorectal cancer. Method This study is a systematic review. The collection of data were made from the databases PubMed and CINAHL. Studies which fulfil the inclusive criteria and answered the aim of this study were quality reviewed and analyized. Result CT Colonography can show colonic findings and is particularly useful to locate extracolonic findings. The method is important in the preoperative evaluation because of its exact localisation rate and is a good compliment to colonoscopy. CT Colonography is a reliable method with a sensitivity at 94–99.9 % which increases when CAD is used as a reading method, especially in the detection of pathology >10 mm. Conclusion CT Colonography is a valuable method in the investigation of colorectal cancer and other pathologies in the abdomen, colon and rectum. It’s a reliable method with a high sensitivity which increases with CAD as a reading method. A strong advantage of this method is that the whole abdomen is visible in the images so the localisation of extracolonic findings can be detected. Keywords Colorectal neoplasms, Computed Tomography, Colonograph

Cancer is a widespread type of diseases that each year affects millions of people. It is mainly treated by chemotherapy, surgery or radiation therapy, or a combination of them. One modality of radiation therapy is high dose-rate brachytherapy, used in treatment of for example prostate cancer and gynecologic cancer. Brachytherapy is an invasive treatment in which catheters (hollow needles) or applicators are used to place the highly active radiation source close to or within a tumour. The treatment planning problem, which can be modelled as a mathematical optimization problem, is the topic of this thesis. The treatment planning includes decisions on how many catheters to use and where to place them as well as the dwell times for the radiation source. There are multiple aims with the treatment and these are primarily to give the tumour a radiation dose that is sufficiently high and to give the surrounding healthy tissue and organs (organs at risk) a dose that is sufficiently low. Because these aims are in conflict, modelling the treatment planning gives optimization problems which essentially are multiobjective. To evaluate treatment plans, a concept called dosimetric indices is commonly used and they constitute an essential part of the clinical treatment guidelines. For the tumour, the portion of the volume that receives at least a specified dose is of interest while for an organ at risk it is rather the portion of the volume that receives at most a specified dose. The dosimetric indices are derived from the dose-volume histogram, which for each dose level shows the corresponding dosimetric index. Dose-volume histograms are commonly used to visualise the three-dimensional dose distribution. The research focus of this thesis is mathematical modelling of the treatment planning and properties of optimization models explicitly including dosimetric indices, which the clinical treatment guidelines are based on. Modelling dosimetric indices explicitly yields mixedinteger programs which are computationally demanding to solve. The computing time of the treatment planning is of clinical relevance as the planning is typically conducted while the patient is under anaesthesia. Research topics in this thesis include both studying properties of models, extending and improving models, and developing new optimization models to be able to take more aspects into account in the treatment planning. There are several advantages of using mathematical optimization for treatment planning in comparison to manual planning. First, the treatment planning phase can be shortened compared to the time consuming manual planning. Secondly, also the quality of treatment plans can be improved by using optimization models and algorithms, for example by considering more of the clinically relevant aspects. Finally, with the use of optimization algorithms the requirements of experience and skill level for the planners are lower. This thesis summary contains a literature review over optimization models for treatment planning, including the catheter placement problem. How optimization models consider the multiobjective nature of the treatment planning problem is also discussed.

This thesis investigated whether a deep learning model could learn features of Chaos,from the Chaos & Clues evaluation protocol, in a given dermatoscopic image data set. Asuccessful result could be of use in a future decision-support system for when dermatologists examine skin lesions for traces of melanoma (type of skin cancer). The chosen deep learning model (Inception V3) was trained to recognise four classesrelated to Chaos. Anonymous patient data was used, provided by the partnering companyGnosco. The data was partitioned into one or two classes depending on the symmetryproperties found in the corresponding image annotation. More than twenty differentmodel configurations was run to obtain the results in this thesis. The results indicate that the chosen model was not capable of learning features of Chaosfrom the dermatoscopical image data-set. Training the model to recognise features ofChaos resulted in an overfit system with low validation accuracy (close to 30%). The prediction target was changed to contrast the negative results from the Chaos classification. The chosen model was therefore configured to learn two classes, ’melanoma’ and’nevus’. This prediction target yielded a more positive result as the validation accuracywas close to 85%. However, the corresponding confusion matrix showed that these resultsare not trustworthy. It is inconclusive whether the negative results from the Chaos classification stem from thechosen approach or if the data set was insufficient for the task-difficulty. We propose adjustments to the data set for future work which could disclose if the outlined approach isviable or not.

Over 30 000 liver, abdomen, and thigh slices have been acquired by computed tomography for the SCAPIS and IGT study. To utilise the full potential of the large cohort and enable statistical pixel-wise body composition analysis and visualisation of associations with other biomarkers, a point-to-point correspondence between the scans is needed. For this purpose, an inter-subject image registration pipeline that combines the low-level information from CT images with high-level information from segmentation masks have been developed. It uses tissue-specific regularisation and processes images efficiently. The pipeline was used to deform 4603 CT scans of each slice into a respective common reference space in less than 30 hours. All but the ribs in the liver slices and the intra abdominal region of the abdomen were generally registered correctly. Vector and intensity magnitude errors indicating inverse consistency were on average less than 2.5 mm and 40 Hounsfield units respectively. The method may serve as a starting point for statistical pixel-wise body composition analysis, its association with non-imaging data, as well as saliency mapping analysis of the three-slice CT scans from the large SCAPIS and IGT cohorts.

This thesis describes a comparison of several state-of-the-art methods used for re-identification of a person between several non-overlapping views captured by surveillance cameras. Since 2014, the focus of the area of person re-identification has been heavily oriented towards approaches employing the use of neural network due to the increase in performance shown from this approach. Three different methods employing convolutional neural networks as a means of attempting automatic person re-identification have mainly been evaluated in this thesis. These three methods are named Spatial-Temporal Person Re-identification (ST-reID), Top DropBlock Network (Top-DB-Net), and Adaptive L2 Regularization. A fourth method known as Multiple Expert Brainstorming Network (MEB-Net) using domain adaptation is used for comparison to the results of applying the trained models from the other three methods on an unseen environment. As an attempt at improving the results of applying the models on an unseen environment, two different approaches have been taken. The first of these is an attempt at segmenting the person from the background by creating a mask that encapsulates the person while disregarding the background, as opposed to using a rectangular cropped image for training and evaluating the methods. To do this, Mask-RCNN which is a framework for object instance segmentation is used. The second approach explored in this thesis is attempting automatic white balancing as a means of removing the effect of the illumination source of the scenes before the person images are extracted. Both approaches show positive results when the model is applied on an unseen environment as opposed to using the unchanged person images, although the results have not been able to match those that have been obtained using domain adaptation. / <p>Examensarbetet är utfört vid Institutionen för teknik och naturvetenskap (ITN) vid Tekniska fakulteten, Linköpings universitet</p>

Malignant melanoma is the deadliest form of skin cancer. If correctly diagnosed in time, the expected five-year survival rate can increase up to 97 %. Therefore, exploring various methods for early detection can contribute with tools which can be used to improve detection of disease and finally to make sure that help is given in time. The purpose of this work was to investigate the performance and behavior of different convolutional neural network (CNN) architectures and to explore whether presegmenting clinical images would improve the prediction results on a binary classifier system. For the purposes of this paper, the two selected CNNs were Inception v3 and DenseNet201. The networks were pretrained on ImageNet and transfer learning techniques such as feature extraction and fine-tuning were used to extract the features of the training set. Batch size was varied and five-fold cross-validation was applied during training to find the optimal number of epochs for training. Evaluation was done on the ISIC test set, the PH2 dataset and a combined set of images from Karolinska University Hospital and FirstDerm, where the latter was also cropped to evaluate presegmentation. The achieved results for the ISIC test set were AUCs of 0.66 for Inception v3 and 0.71 for DenseNet201. For the PH2 test set, the AUCs were 0.82 and 0.73. The results for the Karolinska and FirstDerm set were 0.49 and 0.42. Presegmenting the latter test set resulted in AUCs of 0.58 and 0.51. In conclusion, quality of images could have a big impact on the classification performance. Batch size seems to affect the performance and could thus be an important hyperparameter to tune. Ultimately, the Inception v3 architecture seems to be less affected by different variability why selecting this architecture for a real-world clinical image application could be more suitable. However, the networks performed much worse than state of the art results in previous papers and the conclusions are based on rather inconclusive results. Therefore more research has to be done to verify the conclusions. / Malignt melanom är den dödligaste formen av hudcancer. Om en korrekt diagnos sätts tillräckligt tidigt kan den femåriga överlevnadsgraden uppgå till 97 %. Detta gör att forskningen efter metoder för tidigarelagd upptäckt kan bidra med verktyg som i sin tur kan användas till att upptäcka sjukdom och slutligen bidra till att hjälp sätts in i tid. Målet med detta arbete var att undersöka prestanda och beteende för olika faltningsbara neurala nätverk (CNN) och att undersöka ifall försegmentering av kliniska bilder kunde förbättra resultaten i ett binärt klassificeringssystem. De utvalda faltningsbara neurala nätverksarkitekturerna var Inception v3 och DenseNet201. Nätverken var förträanade på ImageNet och "Transfer-learning"-metoder som feature extraction och fine-tuning användes för att extrahera features från träningsuppsättningen. Batch size varierades och femtalig korsvalidering användes för att hitta det optimala antalet träningsepoker. Utvärderingen gjordes med bilder i testset från ISIC, PH2 och Karolinska och FirstDerm. Bilderna i den senare datamängden beskärdes för att utvärdera försegmenteringen av kliniska bilder. De uppnådda resultaten för ISIC testmängden var AUC-värden på 0.66 för Inception v3 och 0.71 för DenseNet201. För PH2 låg AUC-värdena på 0.82 respektive 0.73. Resultaten för testmängden med bilder frön Karolinska och FirstDerm var 0.40 och 0.42. Försegmenteringen av den sistnämnda testmängden gav AUC-värden på 0.58 och 0.51. Sammanfattningsvis kan bildkvalitet ha en stor inverkan på ett nätverks klassificeringsprestanda. Batch size verkar också påverka resultaten ochkan därför vara en viktig hyperparameter att stämma. Slutligen verkar Inception v3 vara mindre känslig för olika typer av variabiltet vilket görvalet av denna arkitektur mer lämplig ifall en riktig applikation ska byggas för detektion av exempelvis kliniska bilder. Det som bör understrykas i detta arbete är att resultaten var mycket sämre än det som bäst uppvisats i föregående rapporter och att slutasatserna är baserade på relativt ickeövertygande värden. Därför efterkrävs mer forskning för att styrka slutsatserna.

Magnetic resonance imaging (MRI) is a powerful imaging tool for diagnostics, which AMRA uses to segment and quantify certain anatomical regions. This thesis investigate the possibilities of using deep learning for the particular task of AMRAs segmentation, both for ordinary regions (fat and muscle regions) and injured muscles.The main approach performs muscle and fat segmentation separately, and compares results for three approaches; a full resolution approach, a down-sample approach (trained on down-sampled images) and an ensemble approach (uses voting among the 7 best networks).The results shows that deep learning segmentation is possible for the task, with satisfactory results. The down-sampled approach works best for fat segmentation, which can be related to the inconsistently over-segmented ground truth fat masks. It is therefore unnecessary with the additional resolution, which might only impair the performance. The down-sampled approach achieves better results also for muscle segmentation. Ensemble learning does in general not improve the neither the segmentation dice score nor the biomarker predictions. Injured muscles are more difficult to predict due to smaller muscles in the particular used dataset, and an increased data versatility. As a summary, deep learning shows great potential for the task. The results are overall satisfactory (mostly for a down-sampled approach), but further work needs to be done for injured muscles in order to make it clinically useful.

Studien utvärderade bildkvaliteten hos digitala panoramaröntgenbilder och korrelerade denna med synbarhet av normalanatomiska strukturer i bilderna. Studien undersökte även påverkan av subjektiv bildbehandling med avsikt att förbättra synbarheten av strukturerna.500 panoramaröntgenbilder (DICOM-format) framtagna med ett digitalt bildplattesystem utvärderades kvalitativt. Bildkvaliteten och synbarheten av utvalda normalanatomiska strukturer utvärderades i samtliga bilder. Bristande bildkvalitet medförde en subjektiv bildbehandling varefter en ny utvärdering gjordes.Enbart 10% av samtliga bilder var optimala. Felen bland de resterande bilderna dominerades av att patienten inte höll tungan mot gommen och positioneringsfel. Att patienterna ej höll tungan mot gommen påverkade i hög grad synbarheten av strukturer i maxilla. 80.5% (70 bilder) av alla bilder med horisontella positioneringsfel var till vänster, varav 8.6% (6 bilder) resulterade i en icke avbildad vänster käkled. Den mest effektiva kombinationen vid bildbehandling i studien visade sig vara en ökad kontrast och en minskad ljusstyrka vilket förbättrade bilder tagna utan tungan i gommen. Bilder med positioneringsfel var dock opåverkade av bildbehandling.