Deep learning based automated delineation of the intraprostatic gross tumour volume in PSMA-PET for patients with primary prostate cancer.

PURPOSE: With the increased use of focal radiation dose escalation for primary prostate cancer (PCa), accurate delineation of gross tumor volume (GTV) in prostate-specific membrane antigen PET (PSMA-PET) becomes crucial. Manual approaches are time-consuming and observer dependent. The purpose of this study was to create a deep learning model for the accurate delineation of the intraprostatic GTV in PSMA-PET. METHODS: A 3D U-Net was trained on 128 different 18F-PSMA-1007 PET images from three different institutions. Testing was done on 52 patients including one independent internal cohort (Freiburg: n = 19) and three independent external cohorts (Dresden: n = 14 18F-PSMA-1007, Boston: Massachusetts General Hospital (MGH): n = 9 18F-DCFPyL-PSMA and Dana-Farber Cancer Institute (DFCI): n = 10 68Ga-PSMA-11). Expert contours were generated in consensus using a validated technique. CNN predictions were compared to expert contours using Dice similarity coefficient (DSC). Co-registered whole-mount histology was used for the internal testing cohort to assess sensitivity/specificity. RESULTS: Median DSCs were Freiburg: 0.82 (IQR: 0.73-0.88), Dresden: 0.71 (IQR: 0.53-0.75), MGH: 0.80 (IQR: 0.64-0.83) and DFCI: 0.80 (IQR: 0.67-0.84), respectively. Median sensitivity for CNN and expert contours were 0.88 (IQR: 0.68-0.97) and 0.85 (IQR: 0.75-0.88) (p = 0.40), respectively. GTV volumes did not differ significantly (p > 0.1 for all comparisons). Median specificity of 0.83 (IQR: 0.57-0.97) and 0.88 (IQR: 0.69-0.98) were observed for CNN and expert contours (p = 0.014), respectively. CNN prediction took 3.81 seconds on average per patient. CONCLUSION: The CNN was trained and tested on internal and external datasets as well as histopathology reference, achieving a fast GTV segmentation for three PSMA-PET tracers with high diagnostic accuracy comparable to manual experts.

Investigation and benchmarking of U-Nets on prostate segmentation tasks.

In healthcare, a growing number of physicians and support staff are striving to facilitate personalized radiotherapy regimens for patients with prostate cancer. This is because individual patient biology is unique, and employing a single approach for all is inefficient. A crucial step for customizing radiotherapy planning and gaining fundamental information about the disease, is the identification and delineation of targeted structures. However, accurate biomedical image segmentation is time-consuming, requires considerable experience and is prone to observer variability. In the past decade, the use of deep learning models has significantly increased in the field of medical image segmentation. At present, a vast number of anatomical structures can be demarcated on a clinician's level with deep learning models. These models would not only unload work, but they can offer unbiased characterization of the disease. The main architectures used in segmentation are the U-Net and its variants, that exhibit outstanding performances. However, reproducing results or directly comparing methods is often limited by closed source of data and the large heterogeneity among medical images. With this in mind, our intention is to provide a reliable source for assessing deep learning models. As an example, we chose the challenging task of delineating the prostate gland in multi-modal images. First, this paper provides a comprehensive review of current state-of-the-art convolutional neural networks for 3D prostate segmentation. Second, utilizing public and in-house CT and MR datasets of varying properties, we created a framework for an objective comparison of automatic prostate segmentation algorithms. The framework was used for rigorous evaluations of the models, highlighting their strengths and weaknesses.

Radiomics in prostate cancer imaging for a personalized treatment approach - current aspects of methodology and a systematic review on validated studies.

Prostate cancer (PCa) is one of the most frequently diagnosed malignancies of men in the world. Due to a variety of treatment options in different risk groups, proper diagnostic and risk stratification is pivotal in treatment of PCa. The development of precise medical imaging procedures simultaneously to improvements in big data analysis has led to the establishment of radiomics - a computer-based method of extracting and analyzing image features quantitatively. This approach bears the potential to assess and improve PCa detection, tissue characterization and clinical outcome prediction. This article gives an overview on the current aspects of methodology and systematically reviews available literature on radiomics in PCa patients, showing its potential for personalized therapy approaches. The qualitative synthesis includes all imaging modalities and focuses on validated studies, putting forward future directions.

Intelligent Sea States Identification Based on Maximum Likelihood Evidential Reasoning Rule.

It is necessary to switch the control strategies for propulsion system frequently according to the changes of sea states in order to ensure the stability and safety of the navigation. Therefore, identifying the current sea state timely and effectively is of great significance to ensure ship safety. To this end, a reasoning model that is based on maximum likelihood evidential reasoning (MAKER) rule is developed to identify the propeller ventilation type, and the result is used as the basis for the sea states identification. Firstly, a data-driven MAKER model is constructed, which fully considers the interdependence between the input features. Secondly, the genetic algorithm (GA) is used to optimize the parameters of the MAKER model in order to improve the evaluation accuracy. Finally, a simulation is built to obtain experimental data to train the MAKER model, and the validity of the model is verified. The results show that the intelligent sea state identification model that is based on the MAKER rule can identify the propeller ventilation type more accurately, and finally realize intelligent identification of sea states.

Quantitative Regular Expressions for Arrhythmia Detection.

Implantable medical devices are safety-critical systems whose incorrect operation can jeopardize a patient's health, and whose algorithms must meet tight platform constraints like memory consumption and runtime. In particular, we consider here the case of implantable cardioverter defibrillators, where peak detection algorithms and various others discrimination algorithms serve to distinguish fatal from non-fatal arrhythmias in a cardiac signal. Motivated by the need for powerful formal methods to reason about the performance of arrhythmia detection algorithms, we show how to specify all these algorithms using Quantitative Regular Expressions (QREs). QRE is a formal language to express complex numerical queries over data streams, with provable runtime and memory consumption guarantees. We show that QREs are more suitable than classical temporal logics to express in a concise and easy way a range of peak detectors (in both the time and wavelet domains) and various discriminators at the heart of today's arrhythmia detection devices. The proposed formalization also opens the way to formal analysis and rigorous testing of these detectors' correctness and performance, alleviating the regulatory burden on device developers when modifying their algorithms. We demonstrate the effectiveness of our approach by executing QRE-based monitors on real patient data on which they yield results on par with the results reported in the medical literature.

c302: a multiscale framework for modelling the nervous system of Caenorhabditis elegans.

The OpenWorm project has the ambitious goal of producing a highly detailed in silico model of the nematode Caenorhabditis elegans A crucial part of this work will be a model of the nervous system encompassing all known cell types and connections. The appropriate level of biophysical detail required in the neuronal model to reproduce observed high-level behaviours in the worm has yet to be determined. For this reason, we have developed a framework, c302, that allows different instances of neuronal networks to be generated incorporating varying levels of anatomical and physiological detail, which can be investigated and refined independently or linked to other tools developed in the OpenWorm modelling toolchain.This article is part of a discussion meeting issue 'Connectome to behaviour: modelling C. elegans at cellular resolution'.

Quantitative monitoring of STL with edit distance.

In cyber-physical systems (CPS), physical behaviors are typically controlled by digital hardware. As a consequence, continuous behaviors are discretized by sampling and quantization prior to their processing. Quantifying the similarity between CPS behaviors and their specification is an important ingredient in evaluating correctness and quality of such systems. We propose a novel procedure for measuring robustness between digitized CPS signals and signal temporal logic (STL) specifications. We first equip STL with quantitative semantics based on the weighted edit distance, a metric that quantifies both space and time mismatches between digitized CPS behaviors. We then develop a dynamic programming algorithm for computing the robustness degree between digitized signals and STL specifications. In order to promote hardware-based monitors we implemented our approach in FPGA. We evaluated it on automotive benchmarks defined by research community, and also on realistic data obtained from magnetic sensor used in modern cars.

A Multi-Sensor Data Fusion Approach for Atrial Hypertrophy Disease Diagnosis Based on Characterized Support Vector Hyperspheres.

Disease diagnosis can be performed based on fusing the data acquired by multiple medical sensors from patients, and it is a crucial task in sensor-based e-healthcare systems. However, it is a challenging problem that there are few effective diagnosis methods based on sensor data fusion for atrial hypertrophy disease. In this article, we propose a novel multi-sensor data fusion method for atrial hypertrophy diagnosis, namely, characterized support vector hyperspheres (CSVH). Instead of constructing a hyperplane, as a traditional support vector machine does, the proposed method generates "hyperspheres" to collect the discriminative medical information, since a hypersphere is more powerful for data description than a hyperplane. In detail, CSVH constructs two characterized hyperspheres for the classes of patient and healthy subject, respectively. The hypersphere for the patient class is developed in a weighted version so as to take the diversity of patient instances into consideration. The hypersphere for the class of healthy people keeps furthest away from the patient class in order to achieve maximum separation from the patient class. A query is labelled by membership functions defined based on the two hyperspheres. If the query is rejected by the two classes, the angle information of the query to outliers and overlapping-region data is investigated to provide the final decision. The experimental results illustrate that the proposed method achieves the highest diagnosis accuracy among the state-of-the-art methods.

Guided search for hybrid systems based on coarse-grained space abstractions.

Hybrid systems represent an important and powerful formalism for modeling real-world applications such as embedded systems. A verification tool like SpaceEx is based on the exploration of a symbolic search space (the region space). As a verification tool, it is typically optimized towards proving the absence of errors. In some settings, e.g., when the verification tool is employed in a feedback-directed design cycle, one would like to have the option to call a version that is optimized towards finding an error trajectory in the region space. A recent approach in this direction is based on guided search. Guided search relies on a cost function that indicates which states are promising to be explored, and preferably explores more promising states first. In this paper, we propose an abstraction-based cost function based on coarse-grained space abstractions for guiding the reachability analysis. For this purpose, a suitable abstraction technique that exploits the flexible granularity of modern reachability analysis algorithms is introduced. The new cost function is an effective extension of pattern database approaches that have been successfully applied in other areas. The approach has been implemented in the SpaceEx model checker. The evaluation shows its practical potential.

Curvature analysis of cardiac excitation wavefronts.

We present the Spiral Classification Algorithm (SCA), a fast and accurate algorithm for classifying electrical spiral waves and their associated breakup in cardiac tissues. The classification performed by SCA is an essential component of the detection and analysis of various cardiac arrhythmic disorders, including ventricular tachycardia and fibrillation. Given a digitized frame of a propagating wave, SCA constructs a highly accurate representation of the front and the back of the wave, piecewise interpolates this representation with cubic splines, and subjects the result to an accurate curvature analysis. This analysis is more comprehensive than methods based on spiral-tip tracking, as it considers the entire wave front and back. To increase the smoothness of the resulting symbolic representation, the SCA uses weighted overlapping of adjacent segments which increases the smoothness at join points. SCA has been applied to a number of representative types of spiral waves, and, for each type, a distinct curvature evolution in time (signature) has been identified. Distinct signatures have also been identified for spiral breakup. These results represent a significant first step in automatically determining parameter ranges for which a computational cardiac-cell network accurately reproduces a particular kind of cardiac arrhythmia, such as ventricular fibrillation.

Teaching cardiac electrophysiology modeling to undergraduate students: laboratory exercises and GPU programming for the study of arrhythmias and spiral wave dynamics.

As part of a 3-wk intersession workshop funded by a National Science Foundation Expeditions in Computing award, 15 undergraduate students from the City University of New York(1) collaborated on a study aimed at characterizing the voltage dynamics and arrhythmogenic behavior of cardiac cells for a broad range of physiologically relevant conditions using an in silico model. The primary goal of the workshop was to cultivate student interest in computational modeling and analysis of complex systems by introducing them through lectures and laboratory activities to current research in cardiac modeling and by engaging them in a hands-on research experience. The success of the workshop lay in the exposure of the students to active researchers and experts in their fields, the use of hands-on activities to communicate important concepts, active engagement of the students in research, and explanations of the significance of results as the students generated them. The workshop content addressed how spiral waves of electrical activity are initiated in the heart and how different parameter values affect the dynamics of these reentrant waves. Spiral waves are clinically associated with tachycardia, when the waves remain stable, and with fibrillation, when the waves exhibit breakup. All in silico experiments were conducted by simulating a mathematical model of cardiac cells on graphics processing units instead of the standard central processing units of desktop computers. This approach decreased the run time for each simulation to almost real time, thereby allowing the students to quickly analyze and characterize the simulated arrhythmias. Results from these simulations, as well as some of the background and methodology taught during the workshop, is presented in this article along with the programming code and the explanations of simulation results in an effort to allow other teachers and students to perform their own demonstrations, simulations, and studies.

CellExcite: an efficient simulation environment for excitable cells.

BACKGROUND: Brain, heart and skeletal muscle share similar properties of excitable tissue, featuring both discrete behavior (all-or-nothing response to electrical activation) and continuous behavior (recovery to rest follows a temporal path, determined by multiple competing ion flows). Classical mathematical models of excitable cells involve complex systems of nonlinear differential equations. Such models not only impair formal analysis but also impose high computational demands on simulations, especially in large-scale 2-D and 3-D cell networks. In this paper, we show that by choosing Hybrid Automata as the modeling formalism, it is possible to construct a more abstract model of excitable cells that preserves the properties of interest while reducing the computational effort, thereby admitting the possibility of formal analysis and efficient simulation. RESULTS: We have developed CellExcite, a sophisticated simulation environment for excitable-cell networks. CellExcite allows the user to sketch a tissue of excitable cells, plan the stimuli to be applied during simulation, and customize the diffusion model. CellExcite adopts Hybrid Automata (HA) as the computational model in order to efficiently capture both discrete and continuous excitable-cell behavior. CONCLUSIONS: The CellExcite simulation framework for multicellular HA arrays exhibits significantly improved computational efficiency in large-scale simulations, thus opening the possibility for formal analysis based on HA theory. A demo of CellExcite is available at http://www.cs.sunysb.edu/~eha/.