Oscar Alfonso Jiménez del Toro

Creating a Large-Scale Silver Corpus from Multiple Algorithmic Segmentations

Currently, increasingly large medical imaging data sets become available for research and are analysed by a range of algorithms segmenting anatomical structures automatically and interactively. While they provide segmentations on a much larger scale than possible to achieve with expert annotators, they are typically less accurate than experts. We present and compare approaches to estimate segmentations on large imaging data sets based on a small number of expert annotated examples, and algorithmic segmentations on a much larger data set. Results demonstrate that combining algorithmic segmentations is reliably outperforming the average individual algorithm. Furthermore, injecting organ specific reliability assessments of algorithms based on expert annotations improves accuracy compared to standard label fusion algorithms. The proposed methods are particularly relevant in putting the results of large image analysis algorithm benchmarks to long-term use.

Hierarchic Multi–atlas Based Segmentation for Anatomical Structures: Evaluation in the VISCERAL Anatomy Benchmarks

Computer–based medical image analysis is often initialized with the localization of anatomical structures in clinical scans. Many methods have been proposed for segmenting single and multiple anatomical structures. However, it is uncommon to compare different approaches with the same test set, namely a publicly available one. The comparison of these methods objectively defines the advantages and limitations for each method. A hierarchic multi–atlas based segmentation approach was proposed for the segmentation of multiple anatomical structures in computed tomography scans. The method relies on an anatomical hierarchy that exploits the inherent spatial and anatomical variability of medical images using image registration techniques. It was submitted and tested in the VISCERAL project Anatomy benchmarks. In this paper, the results are analyzed and compared to the results of the other segmentation methods submitted in the benchmark. Various anatomical structures in both un–enhanced and contrast–enhanced CT scans resulted in the highest overlap with the proposed method compared to the other evaluated approaches. Although the method was trained with a small training set it generated accurate output segmentations for liver, lungs and other anatomical structures.

Epileptogenic lesion quantification in MRI using contralateral 3D texture comparisons.

Epilepsy is a disorder of the brain that can lead to acute crisis and temporary loss of brain functions. Surgery is used to remove focal lesions that remain resistant to treatment. An accurate localization of epileptogenic lesions has a strong influence on the outcome of epilepsy surgery. Magnetic resonance imaging (MRI) is clinically used for lesion detection and treatment planning, mainly through simple visual analysis. However, visual inspection in MRI can be highly subjective and subtle 3D structural abnormalities are not always entirely removed during surgery. In this paper, we introduce a lesion abnormality score based on computerized comparison of the 3D texture properties between brain hemispheres in T1 MRI. Overlapping cubic texture blocks extracted from user-defined 3D regions of interest (ROI) are expressed in terms of energies of 3D steerable Riesz wavelets. The abnormality score is defined as the Hausdorff distance between the ROI and its corresponding contralateral region in the brain, both expressed as ensembles of blocks in the feature space. A classification based on the proposed score allowed an accuracy of 85% with 10 control subjects and 8 patients with epileptogenic lesions. The approach therefore constitutes a valuable tool for the objective pre-surgical evaluation of patients undergoing epilepsy surgery.

Multi-structure Atlas-Based Segmentation Using Anatomical Regions of Interest

The Visceral project organizes a benchmark on multiple anatomical structure segmentation. A training set is provided to the participants that includes a sample of the manual annotations of these structures. To evaluate different segmentation approaches a testing set of volumes must be segmented automatically in a limited period of time. A multi-atlas based segmentation approach is proposed. This technique can be implemented automatically and applied to different anatomical structures with a large enough training set. The addition of a hierarchical local alignment based on anatomical knowledge and local contrast is explained in the approach. An initial experiment to evaluate the impact of using a local alignment and its results show a higher overlap (\({>}9.7\,\%\)) of the structures measured with the Jaccard coefficient. The approach is an effective and easy to implement method that adjusts well to the Visceral benchmark.

Convolutional neural networks for an automatic classification of prostate tissue slides with high-grade Gleason score

The Gleason grading system was developed for assessing prostate histopathology slides. It is correlated to the outcome and incidence of relapse in prostate cancer. Although this grading is part of a standard protocol performed by pathologists, visual inspection of whole slide images (WSIs) has an inherent subjectivity when evaluated by different pathologists. Computer aided pathology has been proposed to generate an objective and reproducible assessment that can help pathologists in their evaluation of new tissue samples. Deep convolutional neural networks are a promising approach for the automatic classification of histopathology images and can hierarchically learn subtle visual features from the data. However, a large number of manual annotations from pathologists are commonly required to obtain sufficient statistical generalization when training new models that can evaluate the daily generated large amounts of pathology data. A fully automatic approach that detects prostatectomy WSIs with high–grade Gleason score is proposed. We evaluate the performance of various deep learning architectures training them with patches extracted from automatically generated regions–of–interest rather than from manually segmented ones. Relevant parameters for training the deep learning model such as size and number of patches as well as the inclusion or not of data augmentation are compared between the tested deep learning architectures. 235 prostate tissue WSIs with their pathology report from the publicly available TCGA data set were used. An accuracy of 78% was obtained in a balanced set of 46 unseen test images with different Gleason grades in a 2–class decision: high vs. low Gleason grade. Grades 7–8, which represent the boundary decision of the proposed task, were particularly well classified. The method is scalable to larger data sets with straightforward re–training of the model to include data from multiple sources, scanners and acquisition techniques. Automatically generated heatmaps for theWSIs could be useful for improving the selection of patches when training networks for big data sets and to guide the visual inspection of these images.

Processing Diabetes Mellitus Composite Events in MAGPIE

The focus of this research is in the definition of programmable expert Personal Health Systems (PHS) to monitor patients affected by chronic diseases using agent oriented programming and mobile computing to represent the interactions happening amongst the components of the system. The paper also discusses issues of knowledge representation within the medical domain when dealing with temporal patterns concerning the physiological values of the patient. In the presented agent based PHS the doctors can personalize for each patient monitoring rules that can be defined in a graphical way. Furthermore, to achieve better scalability, the computations for monitoring the patients are distributed among their devices rather than being performed in a centralized server. The system is evaluated using data of 21 diabetic patients to detect temporal patterns according to a set of monitoring rules defined. The system’s scalability is evaluated by comparing it with a centralized approach. The evaluation concerning the detection of temporal patterns highlights the system’s ability to monitor chronic patients affected by diabetes. Regarding the scalability, the results show the fact that an approach exploiting the use of mobile computing is more scalable than a centralized approach. Therefore, more likely to satisfy the needs of next generation PHSs. PHSs are becoming an adopted technology to deal with the surge of patients affected by chronic illnesses. This paper discusses architectural choices to make an agent based PHS more scalable by using a distributed mobile computing approach. It also discusses how to model the medical knowledge in the PHS in such a way that it is modifiable at run time. The evaluation highlights the necessity of distributing the reasoning to the mobile part of the system and that modifiable rules are able to deal with the change in lifestyle of the patients affected by chronic illnesses.

A formal method for selecting evaluation metrics for image segmentation

Evaluating the quality of segmentations is an important process in image processing, especially in the medical domain. Many evaluation metrics have been used in evaluating segmentation. There exists no formal way to choose the most suitable metric(s) for a particular segmentation task and/or particular data. In this paper we propose a formal method for choosing the most suitable metrics for evaluating the quality of segmentations with respect to ground truth segmentations. The proposed method depends on measuring the bias of metrics towards/against the properties of the the segmentations being evaluated. We firstly demonstrate how metrics can have bias towards/against particular properties and then we propose a general method for ranking metrics according to their overall bias. We finally demonstrate for 3D medical image segmentations that ranking produced using metrics with low overall bias strongly correlate with manual rankings done by an expert.

Benefits of texture analysis of dual energy CT for Computer-Aided pulmonary embolism detection

Pulmonary embolism is an avoidable cause of death if treated immediately but delays in diagnosis and treatment lead to an increased risk. Computer-assisted image analysis of both unenhanced and contrast-enhanced computed tomography (CT) have proven useful for diagnosis of pulmonary embolism. Dual energy CT provides additional information over the standard single energy scan by generating four-dimensional (4D) data, in our case with 11 energy levels in 3D. In this paper a 4D texture analysis method capable of detecting pulmonary embolism in dual energy CT is presented. The method uses wavelet-based visual words together with an automatic geodesic-based region of interest detection algorithm to characterize the texture properties of each lung lobe. Results show an increase in performance with respect to the single energy CT analysis, as well as an accuracy gain compared to preliminary work on a small dataset.

Multi atlas-based segmentation with data driven refinement

Anatomical structure segmentation is the basis for further image analysis processes. Although there are many available segmentation methods there is still the need to improve the accuracy and speed of them to be used in a clinical environment. The VISCERAL project organizes a benchmark to compare approaches for organ segmentation in big data. A fully–automatic segmentation method using the VISCERAL data set is proposed in this paper. It incorporates both the local contrast of the image using an intensity feature as well as atlas probabilistic information to compute the definite labelling of the structure of interest. The usefulness of the new intensity feature is evaluated using contrast–enhanced CT images of the trunk. An overall average increase is computed in the overlap of the segmentations with an improvement of up to 33% for several anatomical structures when compared to only using an atlas based segmentation method. Qualitative results are also shown for MR images supporting the inclusion of this contrast feature in atlas–based segmentation methods for several modalities.

Using Probability Maps for Multi–organ Automatic Segmentation

Organ segmentation is a vital task in diagnostic medicine. The ability to perform it automatically can save clinicians time and labor. In this paper, a method to achieve automatic segmentation of organs in three–dimensional (3D), non–annotated, full–body magnetic resonance (MR), and computed tomography (CT) volumes is proposed.