The FeatureCloud Platform for Federated Learning in Biomedicine: Unified Approach.

BACKGROUND: Machine learning and artificial intelligence have shown promising results in many areas and are driven by the increasing amount of available data. However, these data are often distributed across different institutions and cannot be easily shared owing to strict privacy regulations. Federated learning (FL) allows the training of distributed machine learning models without sharing sensitive data. In addition, the implementation is time-consuming and requires advanced programming skills and complex technical infrastructures. OBJECTIVE: Various tools and frameworks have been developed to simplify the development of FL algorithms and provide the necessary technical infrastructure. Although there are many high-quality frameworks, most focus only on a single application case or method. To our knowledge, there are no generic frameworks, meaning that the existing solutions are restricted to a particular type of algorithm or application field. Furthermore, most of these frameworks provide an application programming interface that needs programming knowledge. There is no collection of ready-to-use FL algorithms that are extendable and allow users (eg, researchers) without programming knowledge to apply FL. A central FL platform for both FL algorithm developers and users does not exist. This study aimed to address this gap and make FL available to everyone by developing FeatureCloud, an all-in-one platform for FL in biomedicine and beyond. METHODS: The FeatureCloud platform consists of 3 main components: a global frontend, a global backend, and a local controller. Our platform uses a Docker to separate the local acting components of the platform from the sensitive data systems. We evaluated our platform using 4 different algorithms on 5 data sets for both accuracy and runtime. RESULTS: FeatureCloud removes the complexity of distributed systems for developers and end users by providing a comprehensive platform for executing multi-institutional FL analyses and implementing FL algorithms. Through its integrated artificial intelligence store, federated algorithms can easily be published and reused by the community. To secure sensitive raw data, FeatureCloud supports privacy-enhancing technologies to secure the shared local models and assures high standards in data privacy to comply with the strict General Data Protection Regulation. Our evaluation shows that applications developed in FeatureCloud can produce highly similar results compared with centralized approaches and scale well for an increasing number of participating sites. CONCLUSIONS: FeatureCloud provides a ready-to-use platform that integrates the development and execution of FL algorithms while reducing the complexity to a minimum and removing the hurdles of federated infrastructure. Thus, we believe that it has the potential to greatly increase the accessibility of privacy-preserving and distributed data analyses in biomedicine and beyond.

Parea: Multi-view ensemble clustering for cancer subtype discovery.

Multi-view clustering methods are essential for the stratification of patients into sub-groups of similar molecular characteristics. In recent years, a wide range of methods have been developed for this purpose. However, due to the high diversity of cancer-related data, a single method may not perform sufficiently well in all cases. We present Parea, a multi-view hierarchical ensemble clustering approach for disease subtype discovery. We demonstrate its performance on several machine learning benchmark datasets. We apply and validate our methodology on real-world multi-view patient data, comprising seven types of cancer. Parea outperforms the current state-of-the-art on six out of seven analysed cancer types. We have integrated the Parea method into our Python package Pyrea (https://github.com/mdbloice/Pyrea), which enables the effortless and flexible design of ensemble workflows while incorporating a wide range of fusion and clustering algorithms.

Biomedical image augmentation using Augmentor.

MOTIVATION: Image augmentation is a frequently used technique in computer vision and has been seeing increased interest since the popularity of deep learning. Its usefulness is becoming more and more recognized due to deep neural networks requiring larger amounts of data to train, and because in certain fields, such as biomedical imaging, large amounts of labelled data are difficult to come by or expensive to produce. In biomedical imaging, features specific to this domain need to be addressed. RESULTS: Here we present the Augmentor software package for image augmentation. It provides a stochastic, pipeline-based approach to image augmentation with a number of features that are relevant to biomedical imaging, such as z-stack augmentation and randomized elastic distortions. The software has been designed to be highly extensible meaning an operation that might be specific to a highly specialized task can easily be added to the library, even at runtime. Although it has been designed as a general software library, it has features that are particularly relevant to biomedical imaging and the techniques required for this domain. AVAILABILITY AND IMPLEMENTATION: Augmentor is a Python package made available under the terms of the MIT licence. Source code can be found on GitHub under https://github.com/mdbloice/Augmentor and installation is via the pip package manager (A Julia version of the package, developed in parallel by Christof Stocker, is also available under https://github.com/Evizero/Augmentor.jl).

Casebook: a virtual patient iPad application for teaching decision-making through the use of electronic health records.

BACKGROUND: Virtual Patients are a well-known and widely used form of interactive software used to simulate aspects of patient care that students are increasingly less likely to encounter during their studies. However, to take full advantage of the benefits of using Virtual Patients, students should have access to multitudes of cases. In order to promote the creation of collections of cases, a tablet application was developed which makes use of electronic health records as material for Virtual Patient cases. Because electronic health records are abundantly available on hospital information systems, this results in much material for the basis of case creation. RESULTS: An iPad-based Virtual Patient interactive software system was developed entitled Casebook. The application has been designed to read specially formatted patient cases that have been created using electronic health records, in the form of X-ray images, electrocardiograms, lab reports, and physician notes, and present these to the medical student. These health records are organised into a timeline, and the student navigates the case while answering questions regarding the patient along the way. Each health record can also be annotated with meta-information by the case designer, such as insight into the thought processes and the decision-making rationale of the physician who originally worked with the patient. Students learn decision-making skills by observing and interacting with real patient cases in this simulated environment. This paper discusses our approach in detail. CONCLUSIONS: Our group is of the opinion that Virtual Patient cases, targeted at undergraduate students, should concern patients who exhibit prototypical symptoms of the kind students may encounter when beginning their first medical jobs. Learning theory research has shown that students learn decision-making skills best when they have access to multitudes of patient cases and it is this plurality that allows students to develop their illness scripts effectively. Casebook emphasises the use of pre-existing electronic health record data as the basis for case creation, thus, it is hoped, making it easier to produce cases in larger numbers. By creating a Virtual Patient system where cases are built from abundantly available electronic health records, collections of cases can be accumulated by institutions.

On the usage of health records for the design of Virtual Patients: a systematic review.

BACKGROUND: The process of creating and designing Virtual Patients for teaching students of medicine is an expensive and time-consuming task. In order to explore potential methods of mitigating these costs, our group began exploring the possibility of creating Virtual Patients based on electronic health records. This review assesses the usage of electronic health records in the creation of interactive Virtual Patients for teaching clinical decision-making. METHODS: The PubMed database was accessed programmatically to find papers relating to Virtual Patients. The returned citations were classified and the relevant full text articles were reviewed to find Virtual Patient systems that used electronic health records to create learning modalities. RESULTS: A total of n = 362 citations were found on PubMed and subsequently classified, of which n = 28 full-text articles were reviewed. Few articles used unformatted electronic health records other than patient CT or MRI scans. The use of patient data, extracted from electronic health records or otherwise, is widespread. The use of unformatted electronic health records in their raw form is less frequent. Patient data use is broad and spans several areas, such as teaching, training, 3D visualisation, and assessment. CONCLUSIONS: Virtual Patients that are based on real patient data are widespread, yet the use of unformatted electronic health records, abundant in hospital information systems, is reported less often. The majority of teaching systems use reformatted patient data gathered from electronic health records, and do not use these electronic health records directly. Furthermore, many systems were found that used patient data in the form of CT or MRI scans. Much potential research exists regarding the use of unformatted electronic health records for the creation of Virtual Patients.

Computer-aided diagnosis of melanocytic skin tumors by use of confocal laser scanning microscopy images.

OBJECTIVE: To check the applicability of machine learning algorithms for the computer-aided diagnosis of confocal laser scanning microscopy (CLSM) views of skin lesions. STUDY DESIGN: Features, based on spectral properties of the wavelet transform, are very suitable for the automatic analysis because architectural structures at different scales play an important role in diagnosis of CLSM views. The images are discriminated by several machine learning algorithms, based on Bayes-, tree-, rule-, function (numeric)-, and lazy-classifiers. RESULTS: The function and lazy classifiers delivered best classification results. However, these algorithms deliver no information about the inference mechanism leading to the classification. The tree classifiers provided better results than the rule classifiers. To obtain more insight into the inference process, and to compare it with the diagnostic guidelines of the dermopathologists, we combined the advantages of tree, numerical, and rule classifiers and choose the classification and regression trees (CART) algorithm, which automatically generates accurate inferring rules. The classification results were relocated to the images by use of the inferring rules as diagnostic aid. CONCLUSION: The discriminated elements of the skin lesions images show tissue with features in good accordance with typical diagnostic CLSM features.