Matheus Palhares Viana

Mapping road network communities for guiding disease surveillance and control strategies

Human mobility is increasing in its volume, speed and reach, leading to the movement and introduction of pathogens through infected travelers. An understanding of how areas are connected, the strength of these connections and how this translates into disease spread is valuable for planning surveillance and designing control and elimination strategies. While analyses have been undertaken to identify and map connectivity in global air, shipping and migration networks, such analyses have yet to be undertaken on the road networks that carry the vast majority of travellers in low and middle income settings. Here we present methods for identifying road connectivity communities, as well as mapping bridge areas between communities and key linkage routes. We apply these to Africa, and show how many highly-connected communities straddle national borders and when integrating malaria prevalence and population data as an example, the communities change, highlighting regions most strongly connected to areas of high burden. The approaches and results presented provide a flexible tool for supporting the design of disease surveillance and control strategies through mapping areas of high connectivity that form coherent units of intervention and key link routes between communities for targeting surveillance.

Wide residual networks for mitosis detection

One of the most important prognostic markers to assess proliferation activity of breast tumors is estimating the number of mitotic figures in H&E stained tissue. We propose the use of a recently published convolutional neural network architecture, Wide Residual Networks, for mitosis detection in breast histology images. The model is trained to classify each pixel of on an image using as context a patch centered on the pixel. We apply post-processing on the network output in order to filter out noise and select true mitosis. Finally, we combine the output of several networks using majority vote. Our approach ranked 2nd in the MICCAI TUPAC 2016 competition for mitosis detection, outperforming most other contestants by a significant margin.

Author Correction: Mapping road network communities for guiding disease surveillance and control strategies

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Lung Nodule Synthesis Using CNN-Based Latent Data Representation

Convolutional neural networks (CNNs) have been widely used to address various image analysis problems at the cost of intensive computational load and large amounts of annotated training data. When it comes to Medical Imaging, annotation is often complicated and/or expensive, and innovative methods for dealing with small or very imbalanced training sets are mostly welcome. In this context, this paper proposes a novel approach for efficiently synthesizing volumetric patch data from a small amount of samples using their latent data. Our method consists of two major steps. First, we train a 3D CNN auto-encoder for unsupervised learning of volumetric latent data by means of multivariate Gaussian mixture models (GMMs): while the encoder finds latent representations of volumes using GMMs, the decoder uses the estimated GMMs parameters to reconstruct the volume observed in the input. Then, we modify latent data of samples at training time to generate similar, but different, new samples: we run non-rigid registrations between patches decoded from real latent data and patches decoded from modified latent data, and warp the corresponding original image patches using the resulting displacement fields. We evaluated our method in the context of lung nodules synthesis using the publicly available LUNA challenge dataset, and generated new realistic samples out of real lung nodules, preserving their original texture and neighbouring anatomical structures. Our results demonstrate that 3D CNNs trained using our synthesis method were able to consistently deliver lower lung nodule false positive rates, which indicates an improvement in the networks discriminant power.

Deep positive-unlabeled learning for region of interest localization in breast tissue images.

Rapid digitization of whole-slide images (WSIs) with slide scanners, along with the advancements in deep learning strategies has empowered the development of computerized image analysis algorithms for automated diagnosis, prognosis, and prediction of various types of cancers in digital pathology. These analyses can be enhanced and expedited by confining them to relevant tumor region on the large-sized and multi-resolution WSIs. The detection of tumor-region-of-interest (TRoI) on WSIs can facilitate to automatically measure the tumor size as well as to compute the distance to the resection margin. It can also ease the process of identifying high-power-fields (HPFs), which are essential towards the grading of tumor proliferation scores. In practice, pathologists select these regions by visual inspection of WSIs, which is a cumbersome, time-consuming process and affected by inter- and intra- pathologist variability. State-of-the-art deep learning-based methods perform well on the TRoI detection task by using supervised algorithms, however, they require accurate TRoI and non-TRoI annotations to train the algorithms. Acquiring such annotations is a tedious task and incurs observational variability. In this work, we propose a positive and unlabeled learning approach that uses a few examples of HPF regions (positive annotations) to localize the invasive TRoIs on breast cancer WSIs. We use unsupervised deep autoencoders with Gaussian Mixture Model-based clustering to identify the TRoI in a patch-wise manner. The algorithm is developed using 90 HPF-annotated WSIs and is validated on 30 fully-annotated WSIs. It yielded a Dice coefficient of 75.21%, a true positive rate of 78.62% and a true negative rate of 97.48% in terms of pixel-bypixel evaluation compared to the pathologists annotations. Significant correspondence between the results of the proposed algorithm and the state-of-the-art supervised ConvNet indicates the efficacy of the proposed algorithm.

Methods for imaging mammalian mitochondrial morphology: A prospective on MitoGraph

Mitochondria are found in a variety of shapes, from small round punctate structures to a highly interconnected web. This morphological diversity is important for function, but complicates quantification. Consequently, early quantification efforts relied on various qualitative descriptors that understandably reduce the complexity of the network leading to challenges in consistency across the field. Recent application of state-of-the-art computational tools have resulted in more quantitative approaches. This prospective highlights the implementation of MitoGraph, an open-source image analysis platform for measuring mitochondrial morphology initially optimized for use with Saccharomyces cerevisiae. Here Mitograph was assessed on five different mammalian cells types, all of which were accurately segmented by MitoGraph analysis. MitoGraph also successfully differentiated between distinct mitochondrial morphologies that ranged from entirely fragmented to hyper-elongated. General recommendations are also provided for confocal imaging of labeled mitochondria (using mito-YFP, MitoTracker dyes and immunostaining parameters). Widespread adoption of MitoGraph will help achieve a long-sought goal of consistent and reproducible quantification of mitochondrial morphology.