Medical imaging informatics has revolutionized healthcare practice. Images are now stored on a central server and can be accessed remotely through web technologies and standard formats. These images are accompanied by patient-identifiable metadata, allowing systems to execute queries based on that data and present it alongside the pixel data. Despite these advantages, this data has privacy constraints when sharing and reusing it in different scenarios. In research, there is an urgent need for large volumes of data to develop new technologies for diagnosis and therapy purposes. We propose a cloud-ready architecture to build a collaborative environment with anonymized images that can be safely used in research. The solution is based entirely on open-source components, allowing the construction of anonymization pipelines that process and forward images to distinct repositories. The proposal is resilient and scalable and has been tested in a production setting for several clinical studies and use cases.

Computer-assisted diagnosis has been advancing significantly with the proliferation of intelligent image analysis algorithms. These tools enhance medical image screening and diagnostic workflow by highlighting relevant features for doctors, resulting in faster and more accurate diagnoses. However, developing these tools requires large annotated image datasets for achieving acceptable performance. Moreover, developing the analysis algorithms requires knowledge about these technologies, often limited to medical staff. Finally, the integration of those solutions in clinical environments is still a hard barrier due to interoperability issues. This work proposes a solution that extends a Vendor-Neutral Archive for supporting the integration of AI tools for research and production. It makes use of two open-source components, Dicoogle and MONAI, resulting in a Web and DICOM compliant platform that uses active learning strategies for automatic image annotation and inference services.

Clinical pathology has been adopting digital technologies in the review process of the slides. This paradigm shift is being instigated by the advances in technology that allow pathologists to remotely access data banks and visualize high-resolution images with the support of cutting-edge imaging tools. Computer-assisted diagnosis is one of such improvements that is proliferating due to the recent advancements in deep learning technologies. In this clinical area, these tools are especially relevant, due to the big spatial resolution of these images. They can identify regions of interest or diagnostically relevant features on the image, improving the screening times. This article presents the use of active learning strategies to train an object detection model for the mitotic annotation use case, developed in the iPATH research project. The model was successfully integrated into a Pathology-PACS and is being actively used by pathologists for research purposes. Results regarding the evolution and development of this model are also presented in this work.

Many medical studies have been conducted aiming for better understanding of the causes of diseases and to assist in treatments and protective factors. In some cases, these studies do not produce impactful findings due to the small number of participants. Some initiatives already invested efforts in conducting multicentre studies, which raises other technical challenges due to the heterogeneity of datasets. The analysis of such data sources implies dealing with different data structures, terminologies, concepts, languages, and most importantly, the knowledge behind the data. In this paper, we present a methodology to centralise different datasets into the tranSMART application, using a harmonising strategy based on standard data schema. This methodology can help researchers to generate evidence from a wider variety of data sources. This proposal was validated using Alzheimer's Disease cohorts from several countries, combining at the end 6,669 subjects and 172 clinical concepts. The harmonised datasets can provide multi-cohort queries and analysis. The software package is available, under the MIT license, at https://github.com/bioinformatics-ua/tranSMART-migrator.

The use of digital imaging in medicine has become a cornerstone of modern diagnosis and treatment processes. The new technologies available in this ecosystem allowed healthcare institutions to improve their workflows, data access, sharing, and visualization using standardized formats. The migration of these services to the cloud enables a remote diagnostic environment, where professionals can review the studies remotely and engage in collaborative sessions. Despite the advantages of cloud-ready environments, their adoption has been slowed down by the demanding scenario high-resolution medical images pose. Some studies can have several gigabytes of data that need to be managed and consumed in the network. In this context, performance constraints of the software platforms can result in severe denial of clinical service. This work proposes a highly scalable cloud platform for extreme medical imaging scenarios. It provides scalability with auto-scaling mechanisms that allow dynamic adjustment of computational resources according to the service load.

In the last decade, digital technologies have been transforming clinical pathology, dematerializing the processes, improving the workflow, and promoting the development of decision support tools. The glass slides can be digitized with high-quality scanners and sent to network archives, opening doors to data sharing and universal access. Despite these advancements, however, the adoption of digital pathology solutions in production environments has been slowed down by the lack of standardization in the available software. Among the improved visualization features in digital pathology is the ability to capture a series of images at the same spatial location with varying focal planes, also known as Z-stack. This article proposes a full-stack, standard solution that efficiently supports the visualization of these 4D images. It is a Cloud-based architecture that supports Z-stack image management and visualization. The solution is made vendor-neutral by providing a pipeline for the integration of different image vendors.

Recent advancements in the acquisition of digital imaging modalities with high-throughput technologies, such as confocal laser scanner microscopy (CLSM) and focused-ion beam scanning electron microscopy (FIB-SEM), are providing researchers with unprecedented opportunities to collect massive amounts of multidimensional datasets. This data can be used to visualize the internal structure of tiny particles (mostly cells and organisms) or to develop analytic algorithms. Visualizing newly obtained multidimensional microscope imaging data is beyond the capabilities of traditional 3D visualization packages, as it carries much information in the form of additional dimensions. Typically, these extra dimensions correspond to space, time, and channels, which has driven the development of new visualization applications. In this article, we describe the design and implementation of an interactive web-based multidimensional 3D visualization tool for CLSM and FIB-SEM microscope imaging modalities. The proposed 3D visualization application accepts DICOM files as input and provides a variety of visualization choices ranging from 3D volume/surface rendering to multiplanar reconstruction approaches. The solution performance was tested by uploading and rendering microscopy images of distinct modalities.

The use of medical imaging in clinical practice has increased dramatically in recent decades. The adoption and migration to the Cloud of medical imaging systems and services is an excellent opportunity for telemedicine, telework and collaborative work environments. However, the adoption of this paradigm has been slow in this scenario. While the migration has many advantages, it also introduces new challenges mainly related with data storage and management. One of most important open problems is with the efficient handling and transmission of large volumes of data. This issue is particularly critical if the service requests data regional redundancy and high scalability. In this context, this article proposes and describes a new architecture compliant with medical imaging requirements and following standard Cloud and medical imaging interfaces. The solution is based on Kubernetes, an open-source system to deploy, scale and manage containerized applications anywhere. The proposal includes an intelligent component for distributed management of the medical studies based on service policies. The result was a scalable Cloud-based medical imaging repository that can be deployed in a standard way in multiple providers that can achieve a better performance than single node.

Because of its high-resolution visibility across bone tissue, digital whole slide imaging (WSI) in pathology has become increasingly popular in biomedicine for diagnostic, educational, and research purposes. The technology, however, is not yet sufficiently mature to support remote clinical practice in pathology. Several technical challenges have hampered the gradual rollout of telematics platforms. Most notably, there is a lack of system interoperability because scanner manufacturers have yet to adopt a standard for data format and communications processes, specifically the DICOM standard, which already supports WSI. At the moment, it is uncommon to find an institutional repository capable of acquiring, storing, and visualizing samples scanned by different vendors’ equipment. Although some vendors offer DICOM interfaces, the vast majority still use proprietary solutions. This article proposes a framework for multi-vendor data integration through the dicomization of proprietary images, including metadata, and optimization of the visualization process, with a focus on the stacked gigapixel WSI produced by new generation scanners. These images are difficult to preprocess and visualize in most environments due to their size, which is typically in the hundreds of Gigabytes range, and there are no reliable packages that can load, preprocess, and visualize these stacked WSI on a single platform. In this work, we created a simple automated pipeline that can read the majority of proprietary WSI images and focal planes, convert them to DICOM for preprocessing and visualization, and make them compatible with modern Web PACS platforms. The solution was tested with eight distinct proprietary files, two of which were stacked images, from acquisition to Web visualization, demonstrating good performance and efficient use of computational resources.

Nowadays, the medical imaging laboratories are an important piece in the diagnosis within the healthcare institutions. Every day, medical imaging modalities generate new data and the growth over the last recent years has been exponential. The storage and distribution of such data is orchestrated by the Picture Archiving and Communication System (PACS) but the current technologies are facing new challenges. On one hand, the massive amount of data requires a huge local infrastructure to handle the load, which may be expensive. On the other hand, the outsourcing of the data to cloud services may be slow and not compliant with regulations. It becomes urgent to investigate new ways to handle the storage and distribution of the medical images in an healthcare institution. In this work, we discuss the current solutions for the problem of storage and management of medical images. Furthermore, we introduce a solution based on a private network of nodes that can overcome problems as privacy, redundancy and high availability. The developed system offers features that totally match with the Digital Imaging and Communications in Medicine (DICOM) standard in terms of data integrity, immutability, and tolerance to failure. The results show that the proposed architecture performance are very similar with the baseline solutions. So, it reveals itself as a good solution considering the advantages pointed out.

Breast cancer is the leading type of cancer in women and the second most common cancer overall. With thousands of breast screening exams being performed daily around the world, it is a time-consuming task for radiologists, that often find it difficult to analyze and classify them all, with most of exams turning out to be normal cases. The dedicated time to review negative cases could be applied to reviewing more complex cases that require additional care. In this work, the authors propose a support to diagnosis system, focused on mammography screening, based on deep Convolutional Neural Networks (CNNs) and an ensemble classifier, designed to relieve radiologists of normal cases. The architecture takes advantage of macroscopic full mammogram-level features, patch-level features, and patient metadata to output an image-level classification. The results translate the architectures’ high confidence on the exam being normal, relieving the radiologist from analyzing the exam, or suspicious, requiring a specialized radiologist’s attention for a final classification. The developed system was trained and validated using three public datasets (CBIS-DDSM, BCDR and INbreast) and achieved a final, exam-level AUC of 0.98, with a specificity of 96% and a sensitivity of 87%. A conclusion of this work is the possibility to reduce the radiologist’s workload, with potential to reduce the requirement of a second reader, creating further opportunities to review and analyze more complex cases.

In recent years, the quality of fluorescence microscopic imaging produced by a laser scanning microscope (LSM) system has been increased through the use of optical imaging techniques being used for small biological or nonbiological particles like instance, bacteria or cells in tissue samples. Currently, multiple manufactures provide LSM equipment able to capture single-layer or stacked images. However, the manufactures still using distinct data formats including proprietary ones, limiting the interoperability with third systems. In the clinical environment, the scanners need to be integrated with a vendor-neutral archive, storing images in a standard format and interface, making them accessible to healthcare professionals regardless of what proprietary system created it. Having distinct software solutions to manage the data is tiresome and time-consuming. This article proposes a normalization pipeline that can convert distinct vendor formats in a standard DICOM structure including pixel data and metadata. After conversion, the images are sent to a DICOM-compliant repository being able to be consumed in the network using normalized communication and visualization processes. The proposed solution is reliable and uses efficiently the less memory, a critical issue since the resolution of pathology whole-slide images can reach several gigapixels.

In the nineties, the adoption of the DICOM standard format in radiology departments brought numerous advantages to clinical practice. The setup of PACS with standard communication processes and data formats allowed the creation of central repositories, fast retrieval of images, visualization of images acquired with several modalities, and simultaneous access at distributed places. Nowadays, microscopy imagining faces the same normalization challenge with the proliferation of equipment that stores data in a proprietary format and provides dedicated visualization software. This reality severely limits the implementation of vendor-neutral archives with common visualization processes, conditioning the research work and its integration in clinical environments. This paper proposed a pipeline for the integration of multiple microscopy imaging modalities into the PACS-DICOM universe, including the numerous metadata elements. A proof-of-concept system was developed, for validation purposes, and integrated with the Dicoogle open-source PACS, providing image storage, metadata indexing and visualization.

Cohort studies have been playing a key role in helping our understanding of diseases, health conditions, and treatments. These cohorts are often composed of a small number of subjects, especially in rare diseases studies, which reduces the statistical power of the results. One solution that can strengthen the scientific findings is to combine distinct studies and perform then multi-cohort analysis. However, even studies conducted for the same purpose in distinct research groups can have different scopes and medical observations, which preclude across-cohort exploration. In this paper, we propose a recommendation system to automatically discover cohorts of interest. This methodology uses context-based retrieval techniques combined with collaborative filtering to find relevant cohorts and scientific literature about a specific clinical investigation. The system was validated in a community focused on the study of Alzheimer's diseases, which includes 62 cohorts.

Medical imaging laboratories are supported by information and communication systems commonly denominated as PACS, that encompasses technology for acquisition, archive, distribution and visualization of digital images in network. Concerning the data and workflow management, traditional solutions used in production provide a limited set of services usually configured at system installation. As result, healthcare institutions are not able to fully explore their infrastructure or adapt it to new operational requirements, either for clinical or research procedures. This article proposes a framework for services orchestration and workflow management in distributed medical imaging environments. It was designed for end-user usage and is accessible through a Web portal that allows to document, repeat and allocate procedures and tasks to correct resources, either from information systems or human interventions. It provides an abstraction layer for integration with distinct data sources through standard services, allows the creation of new services through orchestration of existent ones and the scheduling of tasks. Moreover, it includes a logging and alert mechanism integrated with email service. The solution was validated through the specification of two use cases that were deployed in production environment.

The general trend of information access anywhere and anytime is also leading to the emergence of innovative medical imaging systems, adapted to this new reality. The HTML5 standard led Web applications to another software level, providing a set of features that allows developing professional Web-based medical imaging applications. However, despite the visualization quality that is already possible in HTML5 browsers, the performance is still an issue, due to the typical size of image studies. In this paper, we present a caching and prefetching solution that enriches the end-user experience in a Web-based DICOM viewer, by reducing data access latency of examinations under revision. We deployed the system in a radiology center for mammography screening, at a national level, and the results show that this technique significantly reduces the average examination access latency, during the reviewing process.

Nowadays, medical imaging laboratories are supported by heterogeneous systems, including image repositories, acquisition devices, viewer workstations and other administrative information systems. They hold tremendous amounts of data resulting not only from imaging modalities, but also from patient diagnosis, treatment, and services management. Unfortunately, the interoperability between the different medical information systems is still a major limitation. Despite the existence of standards to support the distinct RIS and PACS applications, such as DICOM and HL7, the interoperability between them is, in most cases, limited to a few sets of information elements. As a result, the establishment of cooperative workflows or the integrated visualization of patient data is still compromised. Moreover, this scenario severely constraints the usage of these data for research and business analytics purposes, commonly referred as secondary uses of data. In this document, we propose a method for transforming echocardiography reports held by proprietary information systems into DICOM Structured Reports (SR), the gold standard for interoperability in medical imaging. As a result, reports, images, and associated metadata can be accessed and shared by all PACS applications in an integrated and structured manner. Furthermore, the large-scale federation of those elements has a tremendous interest for data analytics and secondary uses of data.

Web-based technologies have been increasingly used in picture archive and communication systems (PACS), in services related to storage, distribution, and visualization of medical images. Nowadays, many healthcare institutions are outsourcing their repositories to the cloud. However, managing communications between multiple geo-distributed locations is still challenging due to the complexity of dealing with huge volumes of data and bandwidth requirements. Moreover, standard methodologies still do not take full advantage of outsourced archives, namely because their integration with other in-house solutions is troublesome. In order to improve the performance of distributed medical imaging networks, a smart routing mechanism was developed. This includes an innovative cache system based on splitting and dynamic management of digital imaging and communications in medicine objects. The proposed solution was successfully deployed in a regional PACS archive. The results obtained proved that it is better than conventional approaches, as it reduces remote access latency and also the required cache storage space.

Healthcare institutions are increasingly taking advantage of information and computational systems to enhance the efficiency and quality of their services. These IT systems are normally able to handle huge amounts of digital data, and to extract relevant fingerprints useful to improve the quality of clinical practice. However, building automatized processes to achieve this over multiple and heterogeneous databases is still a challenge. This paper presents a new approach able to collect and index information from distinct medical data sources, allowing us to identify important metrics to evaluate the performance and quality of clinical services. A case study combining information from ultrasound medical images and echocardiogram clinical reports is also presented.