Abstract
An early diagnosis of cancer is essential for a good prognosis, and the identification of differentially expressed genes can enable a better personalization of the treatment plan that can target those genes in therapy. This work proposes a pipeline that predicts the presence of lung cancer and the subtype allowing the identification of differentially expressed genes for lung cancer adenocarcinoma and squamous cell carcinoma subtypes. A gradient boosted tree model is used for the classification tasks based on RNA-seq data. The analysis of gene expressions that better differentiate cancerous from normal tissue, and features that distinguish between lung subtypes is the main focus of the present work. Differential expressed genes are analyzed by performing hierarchical clustering in order to identify gene signatures that are commonly regulated and biological signatures associated with a specific subtype. This analysis highlighted patterns of commonly regulated genes already known in the literature as cancer or subtype-specific genes, and others that are not yet documented in the literature.

Abstract

Abstract
The consumption of antibiotics, such as sulfonamides, by humans and animals has increased in recent decades, and with it their presence in aquatic environments. This contribute to the increasing of bacterial resistant genes, making the treatment of infectious diseases more difficult. These antibiotics are usually detected by taking a water sample to a laboratory and quantifying it using expensive methods. Recently, digital colorimetry, has emerged as a new method for detecting sulfonamides in water. When a reagent comes into contact water sample containing sulfonamides, a color is produced from which we can infer the concentration of sulfonamides. To ensure that the color is not affected by the illumination when taking a photograph, a color reference target is positioned next to the sample to correct the colors. This method has already been implemented in smartphones to provide a faster and more practical tool that can be used immediately when collecting water samples. Despite this improvement, the algorithms used can still be outperformed by the use of machine learning. In this work, we presented a machine learning approach and a mobile app to solve the problem of sulfonamides quantification. The machine learning approach was designed to run locally in the mobile device, while the mobile application is transversal to Android and iOS operation systems.

Abstract
Lung cancer is a highly prevalent pathology and a leading cause of cancer-related deaths. Most patients are diagnosed when the disease has manifested itself, which usually is a sign of lung cancer in an advanced stage and, as a consequence, the 5-year survival rates are low. To increase the chances of survival, improving the cancer early detection capacity is crucial, for which computed tomography (CT) scans represent a key role. The manual evaluation of the CTs is a time-consuming task and computer-aided diagnosis (CAD) systems can help relieve that burden. The segmentation of the lung is one of the first steps in these systems, yet it is very challenging given the heterogeneity of lung diseases usually present and associated with cancer development. In our previous work, a segmentation model based on a ResNet34 and U-Net combination was developed on a cross-cohort dataset that yielded good segmentation masks for multiple pathological conditions but misclassified some of the lung nodules. The multiple datasets used for the model development were originated from different annotation protocols, which generated inconsistencies for the learning process, and the annotations are usually not adequate for lung cancer studies since they did not comprise lung nodules. In addition, the initial datasets used for training presented a reduced number of nodules, which was showed not to be enough to allow the segmentation model to learn to include them as a lung part. In this work, an objective protocol for the lung mask's segmentation was defined and the previous annotations were carefully reviewed and corrected to create consistent and adequate ground-truth masks for the development of the segmentation model. Data augmentation with domain knowledge was used to create lung nodules in the cases used to train the model. The model developed achieved a Dice similarity coefficient (DSC) above 0.9350 for all test datasets and it showed an ability to cope, not only with a variety of lung patterns, but also with the presence of lung nodules as well. This study shows the importance of using consistent annotations for the supervised learning process, which is a very time-consuming task, but that has great importance to healthcare applications. Due to the lack of massive datasets in the medical field, which consequently brings a lack of wide representativity, data augmentation with domain knowledge could represent a promising help to overcome this limitation for learning models development.

Abstract
Lower limb amputation is a condition affecting millions of people worldwide. Patients are often prescribed with lower limb prostheses to aid their mobility, but these prostheses require frequent adjustments through an iterative and manual process, which heavily depends on patient feedback and on the prosthetist's experience. New computer-aided design and manufacturing technologies have been emerging as ways to improve the fitting process by creating virtual models of the prosthesis' interface component with the limb, the socket. Using Adversarial Autoencoders, a generative model describing both transtibial and transfemoral sockets was created. Two strategies were tested to counteract the small size of the dataset: transfer learning using the ModelNet dataset and data augmentation through a previously validated socket statistical shape model. The minimum reconstruction error was 0.00124 mm and was obtained for the model which combined the two approaches. A single-blind assessment conducted with prosthetists showed that, while generated and real shapes are distinguishable, most generated ones assume plausible shapes. Our results show that the use of transfer learning allowed for a correct training and regularization of the latent space, inducing in the model generative abilities with potential clinical applications.

Abstract
BackgroundTraumatic Brain Injury (TBI) is one of the leading causes of injury related mortality in the world, with severe cases reaching mortality rates of 30-40%. It is highly heterogeneous both in causes and consequences, complicating medical interpretation and prognosis. Gathering clinical, demographic, and laboratory data to perform a prognosis requires time and skill in several clinical specialties. Machine learning (ML) methods can take advantage of the data and guide physicians toward a better prognosis and, consequently, better healthcare. The objective of this study was to develop and test a wide range of machine learning models and evaluate their capability of predicting mortality of TBI, at hospital discharge, while assessing the similarity between the predictive value of the data and clinical significance. MethodsThe used dataset is the Hackathon Pediatric Traumatic Brain Injury (HPTBI) dataset, composed of electronic health records containing clinical annotations and demographic data of 300 patients. Four different classification models were tested, either with or without feature selection. For each combination of the classification model and feature selection method, the area under the receiver operator curve (ROC-AUC), balanced accuracy, precision, and recall were calculated. ResultsMethods based on decision trees perform better when using all features (Random Forest, AUC = 0.86 and XGBoost, AUC = 0.91) but other models require prior feature selection to obtain the best results (k-Nearest Neighbors, AUC = 0.90 and Artificial Neural Networks, AUC = 0.84). Additionally, Random Forest and XGBoost allow assessing the feature's importance, which could give insights for future strategies on the clinical routine. ConclusionPredictive capability depends greatly on the combination of model and feature selection methods used but, overall, ML models showed a very good performance in mortality prediction for TBI. The feature importance results indicate that predictive value is not directly related to clinical significance.