Abstract
Bone marrow edema (BME) is the term given to the abnormal fluid signal seen within the bone marrow on magnetic resonance imaging (MRI). It usually indicates the presence of underlying pathology and is associated with a myriad of conditions/causes. However, it can be misleading, as in some cases, it may be associated with normal changes in the bone, especially during the growth period of childhood, and objective methods for assessment are lacking. In this work, learning models for BME detection were developed. Transfer learning was used to overcome the size limitations of the dataset, and two different regions of interest (ROI) were defined and compared to evaluate their impact on the performance of the model: bone segmention and intensity mask. The best model was obtained for the high intensity masking technique, which achieved a balanced accuracy of 0.792 +/- 0.034. This study represents a comparison of different models and data regularization techniques for BME detection and showed promising results, even in the most difficult range of ages: children and adolescents. The application of machine learning methods will help to decrease the dependence on the clinicians, providing an initial stratification of the patients based on the probability of edema presence and supporting their decisions on the diagnosis.

Abstract
Biomedical engineering has been targeted as a potential research candidate for machine learning applications, with the purpose of detecting or diagnosing pathologies. However, acquiring relevant, high-quality, and heterogeneous medical datasets is challenging due to privacy and security issues and the effort required to annotate the data. Generative models have recently gained a growing interest in the computer vision field due to their ability to increase dataset size by generating new high-quality samples from the initial set, which can be used as data augmentation of a training dataset. This study aimed to synthesize artificial lung images from corresponding positional and semantic annotations using two generative adversarial networks and databases of real computed tomography scans: the Pix2Pix approach that generates lung images from the lung segmentation maps; and the conditional generative adversarial network (cCGAN) approach that was implemented with additional semantic labels in the generation process. To evaluate the quality of the generated images, two quantitative measures were used: the domain-specific Frechet Inception Distance and Structural Similarity Index. Additionally, an expert assessment was performed to measure the capability to distinguish between real and generated images. The assessment performed shows the high quality of synthesized images, which was confirmed by the expert evaluation. This work represents an innovative application of GAN approaches for medical application taking into consideration the pathological findings in the CT images and the clinical evaluation to assess the realism of these features in the generated images.

Abstract
Emerging evidence of the relationship between the microbiome composition and the development of numerous diseases, including cancer, has led to an increasing interest in the study of the human microbiome. Technological breakthroughs regarding DNA sequencing methods propelled microbiome studies with a large number of samples, which called for the necessity of more sophisticated data-analytical tools to analyze this complex relationship. The aim of this work was to develop a machine learning-based approach to distinguish the type of cancer based on the analysis of the tissue-specific microbial information, assessing the human microbiome as valuable predictive information for cancer identification. For this purpose, Random Forest algorithms were trained for the classification of five types of cancer-head and neck, esophageal, stomach, colon, and rectum cancers-with samples provided by The Cancer Microbiome Atlas database. One versus all and multi-class classification studies were conducted to evaluate the discriminative capability of the microbial data across increasing levels of cancer site specificity, with results showing a progressive rise in difficulty for accurate sample classification. Random Forest models achieved promising performances when predicting head and neck, stomach, and colon cancer cases, with the latter returning accuracy scores above 90% across the different studies conducted. However, there was also an increased difficulty when discriminating esophageal and rectum cancers, failing to differentiate with adequate results rectum from colon cancer cases, and esophageal from head and neck and stomach cancers. These results point to the fact that anatomically adjacent cancers can be more complex to identify due to microbial similarities. Despite the limitations, microbiome data analysis using machine learning may advance novel strategies to improve cancer detection and prevention, and decrease disease burden.

Abstract
The clinical relevance of pulse wave velocity (PWV), as an indicator of cardiac risk associated to arterial stiffness, has gained clinical relevance over the last years. Optic sensors are an attractive instrumental solution for this type of measurement due to their truly non-contact operation capability, which has the potential of an interference free measurement. The nature of the optically originated signals, however, poses new challenges to the designer, either at the probe design level as at the signal processing required to extract the timing information that yields PWV. In this work we describe the construction of two prototype optical probes and discuss their evaluation using three algorithms for pulse transit time (PTT) evaluation. Results, obtained in a dedicated test bench, that is also described, demonstrate the possibility of measuring pulse transit times as short as 1ms with less than 1% error. © Springer-Verlag Berlin Heidelberg 2013.

Abstract
Sub-millimetre distension waveforms (0.7 mm, max) are assessed using two new optical probes. The probes differ on the type of photo-detector used: planar photodiodes (PPD), in one case, and avalanche photodiodes (APD), in the other. Performance of the probes is evaluated in an especially developed test setup and in vivo, at the carotid site of humans. In the latter case, distension (associated to the pressure wave generated by the left ventricle contraction that propagates through the arterial system) carries clinically relevant information that can be extracted if, as will be shown, the waveforms are accurate and have enough resolution. An ultrasound image system, Vivid" e, was used as source of reference data for comparison. Along with the probes, a set of software routines was also developed to extract artefact-free data and evaluate the error. Results from the test setup demonstrate the possibility of waveform distension measurements with less than 6% error for both optical probes in this study. In comparison with an ultrasound system, the optical sensors allow the reproduction of the arterial waveform with a higher resolution, adequate to feed feature extraction algorithms.

Abstract
The non-invasive assessment of hemodynamic parameters has been a permanent challenge posed to the scientific community. The literature shows many contributions to this quest expressed as algorithms dedicated to revealing some of its characteristics and as new probes or electronics, featuring some enhanced instrumental capability that can improve their insight. A test system capable of replicating some of the basic properties of the cardiovascular system, especially the ones related with the propagation of the arterial pressure wave (APW), is a powerful tool in the development of those probes and in the validation of the various algorithms that extract clinically relevant information from the data that they can collect. This work describes a test bench system, based on the combination of a new programmable pressure wave generator with a flexible tube, capable of emulating some of these properties. It discusses its main characterization issues and demonstrates the system in a relevant case study. Two versions of the system have been set up: one that generates a short duration pulse-like pressure wave from an actuator operated in a switched mode, appropriate to system characterization; a second one, using a long stroke actuator, linearly operated under program control, capable of generating complex, including cardiac-like, pressure waveforms. This configuration finds its main use in algorithm test and validation. Tests with a new piezoelectric probe, designed to collect the APW at the major artery sites are shown, demonstrating the possibility of non-invasive precise recovery of the pressure waveform.