Convolutional Neural Networks for Chagas' Parasite Detection in Histopathological Images.

Chagas disease is a widely spreaded illness caused by the parasite Trypanosoma cruzi (T. cruzi). Most cases go unnoticed until the accumulated myocardial damage affect the patient. The endomyocardium biopsy is a tool to evaluate sustained myocardial damage, but analyzing histopathological images takes a lot of time and its prone to human error, given its subjective nature. The following work presents a deep learning method to detect T. cruzi amastigotes on histopathological images taken from a endomyocardium biopsy during an experimental murine model. A U-Net convolutional neural network architecture was implemented and trained from the ground up. An accuracy of 99.19% and Jaccard index of 49.43% were achieved. The obtained results suggest that the proposed approach can be useful for amastigotes detection in histopathological images.Clinical relevance- The proposed method can be incorporated as automatic detection tool of amastigotes nests, it can be useful for the Chagas disease analysis and diagnosis.

Machine Learning based Classification of Local Robotic Surgical Skills in a Training Tasks Set.

During surgical training, it is important for the surgeon develops good motor skills throughout his training. For this reason, various surgical training systems have been developed to enhance these skills. However, one of the great challenges in these training systems is being able to objectively measure the ability and performance of the main surgical tasks, where currently only a global measurement is obtained once the task is completed. In this work, a temporal evaluation scheme is proposed, that is, an evaluation of local surgical performance at different time intervals during the training of typical tasks (knot-tying, needle-passing and suturing). The goal is to automatically classify expert (experience >100 hrs) and non-expert (experience <10 hrs) surgeons according to their performance during training, based on three classifiers: K-Nearest Neighborhood, Random Forest, and Support Vector Machine Unlike other previously reported methods, this work proposes a new evaluation scheme based on segments or time intervals, which can be an indicator of the surgeon's local performance during a robotic surgical task, without the need for direct labeling of the data at the segment level. The classification performance from obtained results was in accuracy 83% to 100%, 88% to 100% of AUC-ROC, and 88% to 100% of F1-Score in the final test between experts and non-experts surgeons, where the Support Vector Machine classifier presented the best performance. These results suggest that this proposed method by time intervals could be used in various surgical trainers to evaluate the local performance of a surgeon during trainingand thus be able to provide a tool for the quantitative visualization of opportunities to improve surgical skills.Clinical relevance- We consider that the proposed method to carry out a local performance evaluation during surgical training can provide useful information in the learning and improvement of surgical skills.

Tortuosidad discreta como medida morfométrica en tumores cerebrales / Discrete tortuosity as a morphometric measure in brain tumors

Resumen: Los tumores cerebrales pueden presentar cambios morfológicos dependiendo de su grado de malignidad. El objetivo de este trabajo es poder detectar y cuantificar estos cambios morfológicos a partir imágenes de resonancia magnética, ya que puede representar una ventaja importante para el diagnóstico no invasivo de los pacientes. Una forma de identificar dichos cambios morfológicos es a través de la medición de su tortuosidad. La tortuosidad discreta es un descriptor que caracteriza curvas bidimensionales, como el contorno de una región. En este trabajo se propone una variante para calcular la tortuosidad de superficies volumétricas. Esta técnica se basa en el uso del código cadena de cambios de pendientes del contorno superficial de un volumen y la denominamos como tortuosidad discreta tridimensional. Este descriptor se utiliza como un índice morfométrico para estudiar la tortuosidad de tumores cerebrales. Para ello se analizan imágenes de resonancia magnética de 20 pacientes con presencia de gliomas de bajo y alto grado de malignidad, considerando cuatro regiones de interés: edema, tumor entero, región activa y necrosis. Como resultado, se muestran los distintos grados de tortuosidad de las diversas regiones, encontrándose solo en algunas de ellas diferencias significativas. Cabe señalar que una desventaja que se tiene presente, es la dependencia de la medición a la utilización de un método robusto de segmentación de las regiones, sin embargo la propuesta de la tortuosidad discreta para superficies volumétricas es satisfactoria.

Abstract: A decision tree based system with heuristic weight factors oriented to diagnosis by Morphological changes in brain tumors may be related to their malignancy. The objective of this work is to be able to detect and quantify these changes in a magnetic resonance imaging, since it can represent an important advantage for the noninvasive diagnosis in patients. One way to identify such morphological changes can be through the measurement of their tortuosity. The discrete tortuosity is a descriptor that characterizes bi-dimensional curves, as the contour of a region. In this work an alternative procedure for calculating the volumetric tortuosity of a surface is proposed. This technique is based in the slope chain code of the surface contour of a volume, and here we call it tridimensional discrete tortuosity. This descriptor is used as a morphometric index to study the tortuosity of brain tumors. For this, magnetic resonance images from 20 patients with low and high malignancy levels were analyzed, considering four regions: edema, whole tumor, enhancing region, and necrotic region. As a result, the tortuosities of the different regions are presented, with significant differences only in some of them. It should be noted that a disadvantage that is present, is the dependence of the measurement to the use of a robust method of segmentation, nevertheless the proposal of the discrete tortuosity for volumetric surfaces is satisfactory.

Hyperglycemia and the fate of apparent diffusion coefficient-defined ischemic penumbra.

BACKGROUND AND PURPOSE: Previous data have shown the feasibility of identifying ischemic penumbra in patients with acute stroke by using a semiautomated analysis of ADC maps. Here, we investigated whether the fate of ADC-defined penumbra was altered by HG. We also examined the interaction between HG and arterial recanalization on infarct growth. MATERIALS AND METHODS: We examined 94 patients by using MR imaging within 6 hours of stroke onset and a follow-up MR imaging within 7 days. The ADC-defined tissue-at-risk was calculated from the early MR imaging. Patients were classified according to high (>7 mmol/L; n = 34/94, HG) or normal (n = 60/94) baseline SGL. The impact of HG status on infarct growth was assessed by using multiple regression models and analysis of the slopes of regression lines for each group. Interaction between HG status and arterial recanalization on infarct growth was investigated by using multiple regression analysis. RESULTS: The slope of the predicted versus observed infarct growth regression line was steeper in HG than non-HG patients (P = .0008), suggesting that infarct growth within ADC-defined tissue-at-risk was increased in HG patients. The effect was 2.8 times more severe in nonrecanalized patients (P = .01) than in patients with recanalization (P = .001). CONCLUSIONS: ADC-defined tissue-at-risk may represent ischemic penumbra because part of this area may be salvaged in normal SGL patients. The toxicity in HG patients seems to be more related to penumbra-infarction transition than reperfusion injury in humans because the effect was larger in nonrecanalized than in recanalized patients.