Bringing AI to the clinic: blueprint for a vendor-neutral AI deployment infrastructure

Clinical Integration of Artificial Intelligence for Breast Imaging

This article describes an approach to planning and implementing artificial intelligence products in a breast screening service. It highlights the importance of an in-depth understanding of the end-to-end workflow and effective project planning by a multidisciplinary team. It discusses the need for monitoring to ensure that performance is stable and meets expectations, as well as focusing on the potential for inadvertantly generating inequality. New cross-discipline roles and expertise will be needed to enhance service delivery.

Biometric contrastive learning for data-efficient deep learning from electrocardiographic images

OBJECTIVE

Artificial intelligence (AI) detects heart disease from images of electrocardiograms (ECGs). However, traditional supervised learning is limited by the need for large amounts of labeled data. We report the development of Biometric Contrastive Learning (BCL), a self-supervised pretraining approach for label-efficient deep learning on ECG images.

MATERIALS AND METHODS

Using pairs of ECGs from 78 288 individuals from Yale (2000-2015), we trained a convolutional neural network to identify temporally separated ECG pairs that varied in layouts from the same patient. We fine-tuned BCL-pretrained models to detect atrial fibrillation (AF), gender, and LVEF < 40%, using ECGs from 2015 to 2021. We externally tested the models in cohorts from Germany and the United States. We compared BCL with ImageNet initialization and general-purpose self-supervised contrastive learning for images (simCLR).

RESULTS

While with 100% labeled training data, BCL performed similarly to other approaches for detecting AF/Gender/LVEF < 40% with an AUROC of 0.98/0.90/0.90 in the held-out test sets, it consistently outperformed other methods with smaller proportions of labeled data, reaching equivalent performance at 50% of data. With 0.1% data, BCL achieved AUROC of 0.88/0.79/0.75, compared with 0.51/0.52/0.60 (ImageNet) and 0.61/0.53/0.49 (simCLR). In external validation, BCL outperformed other methods even at 100% labeled training data, with an AUROC of 0.88/0.88 for Gender and LVEF < 40% compared with 0.83/0.83 (ImageNet) and 0.84/0.83 (simCLR).

DISCUSSION AND CONCLUSION

A pretraining strategy that leverages biometric signatures of different ECGs from the same patient enhances the efficiency of developing AI models for ECG images. This represents a major advance in detecting disorders from ECG images with limited labeled data.

A holistic approach to implementing artificial intelligence in radiology

OBJECTIVE

Despite the widespread recognition of the importance of artificial intelligence (AI) in healthcare, its implementation is often limited. This article aims to address this implementation gap by presenting insights from an in-depth case study of an organisation that approached AI implementation with a holistic approach.

MATERIALS AND METHODS

We conducted a longitudinal, qualitative case study of the implementation of AI in radiology at a large academic medical centre in the Netherlands for three years. Collected data consists of 43 days of work observations, 30 meeting observations, 18 interviews and 41 relevant documents. Abductive reasoning was used for systematic data analysis, which revealed three change initiative themes responding to specific AI implementation challenges.

RESULTS

This study identifies challenges of implementing AI in radiology at different levels and proposes a holistic approach to tackle those challenges. At the technology level, there is the issue of multiple narrow AI applications with no standard use interface; at the workflow level, AI results allow limited interaction with radiologists; at the people and organisational level, there are divergent expectations and limited experience with AI. The case of Southern illustrates that organisations can reap more benefits from AI implementation by investing in long-term initiatives that holistically align both social and technological aspects of clinical practice.

CONCLUSION

This study highlights the importance of a holistic approach to AI implementation that addresses challenges spanning technology, workflow, and organisational levels. Aligning change initiatives between these different levels has proven to be important to facilitate wide-scale implementation of AI in clinical practice.

CRITICAL RELEVANCE STATEMENT

Adoption of artificial intelligence is crucial for future-ready radiological care. This case study highlights the importance of a holistic approach that addresses technological, workflow, and organisational aspects, offering practical insights and solutions to facilitate successful AI adoption in clinical practice.

KEY POINTS

1. Practical and actionable insights into successful AI implementation in radiology are lacking. 2. Aligning technology, workflow, organisational aspects is crucial for a successful AI implementation 3. Holistic approach aids organisations to create sustainable value through AI implementation.

"KAIZEN" method realizing implementation of deep-learning models for COVID-19 CT diagnosis in real world hospitals

Numerous COVID-19 diagnostic imaging Artificial Intelligence (AI) studies exist. However, none of their models were of potential clinical use, primarily owing to methodological defects and the lack of implementation considerations for inference. In this study, all development processes of the deep-learning models are performed based on strict criteria of the "KAIZEN checklist", which is proposed based on previous AI development guidelines to overcome the deficiencies mentioned above. We develop and evaluate two binary-classification deep-learning models to triage COVID-19: a slice model examining a Computed Tomography (CT) slice to find COVID-19 lesions; a series model examining a series of CT images to find an infected patient. We collected 2,400,200 CT slices from twelve emergency centers in Japan. Area Under Curve (AUC) and accuracy were calculated for classification performance. The inference time of the system that includes these two models were measured. For validation data, the slice and series models recognized COVID-19 with AUCs and accuracies of 0.989 and 0.982, 95.9% and 93.0% respectively. For test data, the models' AUCs and accuracies were 0.958 and 0.953, 90.0% and 91.4% respectively. The average inference time per case was 2.83 s. Our deep-learning system realizes accuracy and inference speed high enough for practical use. The systems have already been implemented in four hospitals and eight are under progression. We released an application software and implementation code for free in a highly usable state to allow its use in Japan and globally.

Application of an artificial intelligence-based system in the diagnosis of breast ultrasound images obtained using a smartphone

BACKGROUND

Breast ultrasound (US) is useful for dense breasts, and the introduction of artificial intelligence (AI)-assisted diagnoses of breast US images should be considered. However, the implementation of AI-based technologies in clinical practice is problematic because of the costs of introducing such approaches to hospital information systems (HISs) and the security risk of connecting HIS to the Internet to access AI services. To solve these problems, we developed a system that applies AI to the analysis of breast US images captured using a smartphone.

METHODS

Training data were prepared using 115 images of benign lesions and 201 images of malignant lesions acquired at the Division of Breast Surgery, Gifu University Hospital. YOLOv3 (object detection models) was used to detect lesions on US images. A graphical user interface (GUI) was developed to predict an AI server. A smartphone application was also developed for capturing US images displayed on the HIS monitor with its camera and displaying the prediction results received from the AI server. The sensitivity and specificity of the prediction performed on the AI server and via the smartphone were calculated using 60 images spared from the training.

RESULTS

The established AI showed 100% sensitivity and 75% specificity for malignant lesions and took 0.2 s per prediction with the AI sever. Prediction using a smartphone required 2 s per prediction and showed 100% sensitivity and 97.5% specificity for malignant lesions.

CONCLUSIONS

Good-quality predictions were obtained using the AI server. Moreover, the quality of the prediction via the smartphone was slightly better than that on the AI server, which can be safely and inexpensively introduced into HISs.

Sharing Data Is Essential for the Future of AI in Medical Imaging

If we want artificial intelligence to succeed in radiology, we must share data and learn how to share data.

Digitization of Pathology Labs: A Review of Lessons Learned

Pathology laboratories are increasingly using digital workflows. This has the potential of increasing laboratory efficiency, but the digitization process also involves major challenges. Several reports have been published describing the individual experiences of specific laboratories with the digitization process. However, a comprehensive overview of the lessons learned is still lacking. We provide an overview of the lessons learned for different aspects of the digitization process, including digital case management, digital slide reading, and computer-aided slide reading. We also cover metrics used for monitoring performance and pitfalls and corresponding values observed in practice. The overview is intended to help pathologists, information technology decision makers, and administrators to benefit from the experiences of others and to implement the digitization process in an optimal way to make their own laboratory future-proof.

A data-driven framework for clinical decision support applied to pneumonia management

Despite their long history, it can still be difficult to embed clinical decision support into existing health information systems, particularly if they utilise machine learning and artificial intelligence models. Moreover, when such tools are made available to healthcare workers, it is important that the users can understand and visualise the reasons for the decision support predictions. Plausibility can be hard to achieve for complex pathways and models and perceived "black-box" functionality often leads to a lack of trust. Here, we describe and evaluate a data-driven framework which moderates some of these issues and demonstrate its applicability to the in-hospital management of community acquired pneumonia, an acute respiratory disease which is a leading cause of in-hospital mortality world-wide. We use the framework to develop and test a clinical decision support tool based on local guideline aligned management of the disease and show how it could be used to effectively prioritise patients using retrospective analysis. Furthermore, we show how this tool can be embedded into a prototype clinical system for disease management by integrating metrics and visualisations. This will assist decision makers to examine complex patient journeys, risk scores and predictions from embedded machine learning and artificial intelligence models. Our results show the potential of this approach for developing, testing and evaluating workflow based clinical decision support tools which include complex models and embedding them into clinical systems.

Multilevel comparison of deep learning models for function quantification in cardiovascular magnetic resonance: On the redundancy of architectural variations

Background

Cardiac function quantification in cardiovascular magnetic resonance requires precise contouring of the heart chambers. This time-consuming task is increasingly being addressed by a plethora of ever more complex deep learning methods. However, only a small fraction of these have made their way from academia into clinical practice. In the quality assessment and control of medical artificial intelligence, the opaque reasoning and associated distinctive errors of neural networks meet an extraordinarily low tolerance for failure.

Aim

The aim of this study is a multilevel analysis and comparison of the performance of three popular convolutional neural network (CNN) models for cardiac function quantification.

Methods

U-Net, FCN, and MultiResUNet were trained for the segmentation of the left and right ventricles on short-axis cine images of 119 patients from clinical routine. The training pipeline and hyperparameters were kept constant to isolate the influence of network architecture. CNN performance was evaluated against expert segmentations for 29 test cases on contour level and in terms of quantitative clinical parameters. Multilevel analysis included breakdown of results by slice position, as well as visualization of segmentation deviations and linkage of volume differences to segmentation metrics via correlation plots for qualitative analysis.

Results

All models showed strong correlation to the expert with respect to quantitative clinical parameters (r_{z '} = 0.978, 0.977, 0.978 for U-Net, FCN, MultiResUNet respectively). The MultiResUNet significantly underestimated ventricular volumes and left ventricular myocardial mass. Segmentation difficulties and failures clustered in basal and apical slices for all CNNs, with the largest volume differences in the basal slices (mean absolute error per slice: 4.2 ± 4.5 ml for basal, 0.9 ± 1.3 ml for midventricular, 0.9 ± 0.9 ml for apical slices). Results for the right ventricle had higher variance and more outliers compared to the left ventricle. Intraclass correlation for clinical parameters was excellent (≥0.91) among the CNNs.

Conclusion

Modifications to CNN architecture were not critical to the quality of error for our dataset. Despite good overall agreement with the expert, errors accumulated in basal and apical slices for all models.

Evaluating artificial intelligence algorithms for use in veterinary radiology

Artificial intelligence is increasingly being used for applications in veterinary radiology, including detection of abnormalities and automated measurements. Unlike human radiology, there is no formal regulation or validation of AI algorithms for veterinary medicine and both general practitioner and specialist veterinarians must rely on their own judgment when deciding whether or not to incorporate AI algorithms to aid their clinical decision-making. The benefits and challenges to developing clinically useful and diagnostically accurate AI algorithms are discussed. Considerations for the development of AI research projects are also addressed. A framework is suggested to help veterinarians, in both research and clinical practice contexts, assess AI algorithms for veterinary radiology.

Computational Portable Microscopes for Point-of-Care-Test and Tele-Diagnosis

In bio-medical mobile workstations, e.g., the prevention of epidemic viruses/bacteria, outdoor field medical treatment and bio-chemical pollution monitoring, the conventional bench-top microscopic imaging equipment is limited. The comprehensive multi-mode (bright/dark field imaging, fluorescence excitation imaging, polarized light imaging, and differential interference microscopy imaging, etc.) biomedical microscopy imaging systems are generally large in size and expensive. They also require professional operation, which means high labor-cost, money-cost and time-cost. These characteristics prevent them from being applied in bio-medical mobile workstations. The bio-medical mobile workstations need microscopy systems which are inexpensive and able to handle fast, timely and large-scale deployment. The development of lightweight, low-cost and portable microscopic imaging devices can meet these demands. Presently, for the increasing needs of point-of-care-test and tele-diagnosis, high-performance computational portable microscopes are widely developed. Bluetooth modules, WLAN modules and 3G/4G/5G modules generally feature very small sizes and low prices. And industrial imaging lens, microscopy objective lens, and CMOS/CCD photoelectric image sensors are also available in small sizes and at low prices. Here we review and discuss these typical computational, portable and low-cost microscopes by refined specifications and schematics, from the aspect of optics, electronic, algorithms principle and typical bio-medical applications.

Artificial Intelligence in Clinical Practice: Implementation Considerations and Barriers

The rapid growth of artificial intelligence (AI) in radiology has led to Food and Drug Administration clearance of more than 20 AI algorithms for breast imaging. The steps involved in the clinical implementation of an AI product include identifying all stakeholders, selecting the appropriate product to purchase, evaluating it with a local data set, integrating it into the workflow, and monitoring its performance over time. Despite the potential benefits of improved quality and increased efficiency with AI, several barriers, such as high costs and liability concerns, may limit its widespread implementation. This article lists currently available AI products for breast imaging, describes the key elements of clinical implementation, and discusses barriers to clinical implementation.

Considerations for artificial intelligence clinical impact in oncologic imaging: an AI4HI position paper

To achieve clinical impact in daily oncological practice, emerging AI-based cancer imaging research needs to have clearly defined medical focus, AI methods, and outcomes to be estimated. AI-supported cancer imaging should predict major relevant clinical endpoints, aiming to extract associations and draw inferences in a fair, robust, and trustworthy way. AI-assisted solutions as medical devices, developed using multicenter heterogeneous datasets, should be targeted to have an impact on the clinical care pathway. When designing an AI-based research study in oncologic imaging, ensuring clinical impact in AI solutions requires careful consideration of key aspects, including target population selection, sample size definition, standards, and common data elements utilization, balanced dataset splitting, appropriate validation methodology, adequate ground truth, and careful selection of clinical endpoints. Endpoints may be pathology hallmarks, disease behavior, treatment response, or patient prognosis. Ensuring ethical, safety, and privacy considerations are also mandatory before clinical validation is performed. The Artificial Intelligence for Health Imaging (AI4HI) Clinical Working Group has discussed and present in this paper some indicative Machine Learning (ML) enabled decision-support solutions currently under research in the AI4HI projects, as well as the main considerations and requirements that AI solutions should have from a clinical perspective, which can be adopted into clinical practice. If effectively designed, implemented, and validated, cancer imaging AI-supported tools will have the potential to revolutionize the field of precision medicine in oncology.

Medical-Blocks: A Platform for Exploration, Management, Analysis, and Sharing of Data in Biomedical Research

BACKGROUND

Biomedical research requires healthcare institutions to provide sensitive clinical data to leverage data science and artificial intelligence technologies. However, providing healthcare data to researchers simple and secure, proves to be challenging for healthcare institutions.

OBJECTIVE

We describe and introduce Medical-Blocks, a platform for data exploration, data management, data analysis, and data sharing in biomedical research.

METHODS

The specification requirements for Medical-Blocks included: i) Connection to data sources of healthcare institutions with an interface for data exploration, ii) management of data in an internal file storage system, iii) data analysis through visualization and classification of data, and iv) data sharing via a file hosting service for collaboration. Medical-Blocks should be simple to use via a web-based user interface and extensible with new functionalities by a modular design via microservices ("blocks"). The scalability of the platform should be ensured by containerization. Security and legal regulations were considered during the development.

RESULTS

Medical-Blocks is a web application that runs in the cloud or as a local instance at a healthcare institution. Local instances of Medical-Blocks access data sources such as electronic health records and picture archiving and communications system (PACS) at healthcare institutions. Researchers and clinicians can explore, manage, and analyze the available data through Medical-Blocks. The data analysis involves classification of data for metadata extraction and the formation of cohorts. In collaborations, metadata (e.g., number of patients per cohort) and/or the data itself can be shared through Medical-Blocks locally or via a cloud instance to other researchers and clinicians.

CONCLUSIONS

Medical-Blocks facilitates biomedical research by providing a centralized platform to interact with medical data in collaborative research projects. The access to and management of medical data is simplified. Data can be swiftly analyzed to form cohorts for research and be shared among researchers. The modularity of Medical-Blocks makes the platform feasible for biomedical research where heterogenous medical data is needed.

CLINICALTRIAL

Artificial intelligence in musculoskeletal imaging: a perspective on value propositions, clinical use, and obstacles

Artificial intelligence and deep learning (DL) offer musculoskeletal radiology exciting possibilities in multiple areas, including image reconstruction and transformation, tissue segmentation, workflow support, and disease detection. Novel DL-based image reconstruction algorithms correcting aliasing artifacts, signal loss, and noise amplification with previously unobtainable effectiveness are prime examples of how DL algorithms deliver promised value propositions in musculoskeletal radiology. The speed of DL-based tissue segmentation promises great efficiency gains that may permit the inclusion of tissue compositional-based information routinely into radiology reports. Similarly, DL algorithms give rise to a myriad of opportunities for workflow improvements, including intelligent and adaptive hanging protocols, speech recognition, report generation, scheduling, precertification, and billing. The value propositions of disease-detecting DL algorithms include reduced error rates and increased productivity. However, more studies using authentic clinical workflow settings are necessary to fully understand the value of DL algorithms for disease detection in clinical practice. Successful workflow integration and management of multiple algorithms are critical for translating the value propositions of DL algorithms into clinical practice but represent a major roadblock for which solutions are critically needed. While there is no consensus about the most sustainable business model, radiology departments will need to carefully weigh the benefits and disadvantages of each commercially available DL algorithm. Although more studies are needed to understand the value and impact of DL algorithms on clinical practice, DL technology will likely play an important role in the future of musculoskeletal imaging.

The Applied Data Analytics in Medicine Program: Lessons Learned From Four Years' Experience With Personalizing Health Care in an Academic Teaching Hospital

The University Medical Center (UMC) Utrecht piloted a hospital-wide innovation data analytics program over the past 4 years. The goal was, based on available data and innovative data analytics methodologies, to answer clinical questions to improve patient care. In this viewpoint, we aimed to support and inspire others pursuing similar efforts by sharing the three principles of the program: the data analytics value chain (data, insight, action, value), the innovation funnel (structured innovation approach with phases and gates), and the multidisciplinary team (patients, clinicians, and data scientists). We also discussed our most important lessons learned: the importance of a clinical question, collaboration challenges between health care professionals and different types of data scientists, the win-win result of our collaboration with external partners, the prerequisite of available meaningful data, the (legal) complexity of implementation, organizational power, and the embedding of collaborative efforts in the health care system as a whole.

A Proposed Framework for Machine Learning-Aided Triage in Public Specialty Ophthalmology Clinics in Hong Kong

The public specialty ophthalmic clinics in Hong Kong, under the Hospital Authority, receive tens of thousands of referrals each year. Triaging these referrals incurs a significant workload for practitioners and the other clinical duties. It is well-established that Hong Kong is currently facing a shortage of healthcare workers. Thus a more efficient system in triaging will not only free up resources for better use but also improve the satisfaction of both practitioners and patients. Machine learning (ML) has been shown to improve the efficiency of various medical workflows, including triaging, by both reducing the workload and increasing accuracy in some cases. Despite a myriad of studies on medical artificial intelligence, there is no specific framework for a triaging algorithm in ophthalmology clinics. This study proposes a general framework for developing, deploying and evaluating an ML-based triaging algorithm in a clinical setting. Through literature review, this study identifies good practices in various facets of developing such a network and protocols for maintenance and evaluation of the impact concerning clinical utility and external validity out of the laboratory. We hope this framework, albeit not exhaustive, can act as a foundation to accelerate future pilot studies and deployments.

Radiology Implementation Considerations for Artificial Intelligence (AI) Applied to COVID-19, From the AJR Special Series on AI Applications

Hundreds of imaging-based artificial intelligence (AI) models have been developed in response to the COVID-19 pandemic. AI systems that incorporate imaging have shown promise in primary detection, severity grading, and prognostication of outcomes in COVID-19, and have enabled integration of imaging with a broad range of additional clinical and epidemiologic data. However, systematic reviews of AI models applied to COVID-19 medical imaging have highlighted problems in the field, including methodologic issues and problems in real-world deployment. Clinical use of such models should be informed by both the promise and potential pitfalls of implementation. How does a practicing radiologist make sense of this complex topic, and what factors should be considered in the implementation of AI tools for imaging of COVID-19? This critical review aims to help the radiologist understand the nuances that impact the clinical deployment of AI for imaging of COVID-19. We review imaging use cases for AI models in COVID-19 (e.g., diagnosis, severity assessment, and prognostication) and explore considerations for AI model development and testing, deployment infrastructure, clinical user interfaces, quality control, and institutional review board and regulatory approvals, with a practical focus on what a radiologist should consider when implementing an AI tool for COVID-19.

A DICOM Framework for Machine Learning and Processing Pipelines Against Real-time Radiology Images

Real-time execution of machine learning (ML) pipelines on radiology images is difficult due to limited computing resources in clinical environments, whereas running them in research clusters requires efficient data transfer capabilities. We developed Niffler, an open-source Digital Imaging and Communications in Medicine (DICOM) framework that enables ML and processing pipelines in research clusters by efficiently retrieving images from the hospitals’ PACS and extracting the metadata from the images. We deployed Niffler at our institution (Emory Healthcare, the largest healthcare network in the state of Georgia) and retrieved data from 715 scanners spanning 12 sites, up to 350 GB/day continuously in real-time as a DICOM data stream over the past 2 years. We also used Niffler to retrieve images bulk on-demand based on user-provided filters to facilitate several research projects. This paper presents the architecture and three such use cases of Niffler. First, we executed an IVC filter detection and segmentation pipeline on abdominal radiographs in real-time, which was able to classify 989 test images with an accuracy of 96.0%. Second, we applied the Niffler Metadata Extractor to understand the operational efficiency of individual MRI systems based on calculated metrics. We benchmarked the accuracy of the calculated exam time windows by comparing Niffler against the Clinical Data Warehouse (CDW). Niffler accurately identified the scanners’ examination timeframes and idling times, whereas CDW falsely depicted several exam overlaps due to human errors. Third, with metadata extracted from the images by Niffler, we identified scanners with misconfigured time and reconfigured five scanners. Our evaluations highlight how Niffler enables real-time ML and processing pipelines in a research cluster.

Artificial Intelligence in PET: An Industry Perspective

Artificial intelligence (AI) has significant potential to positively impact and advance medical imaging, including positron emission tomography (PET) imaging applications. AI has the ability to enhance and optimize all aspects of the PET imaging chain from patient scheduling, patient setup, protocoling, data acquisition, detector signal processing, reconstruction, image processing, and interpretation. AI poses industry-specific challenges which will need to be addressed and overcome to maximize the future potentials of AI in PET. This article provides an overview of these industry-specific challenges for the development, standardization, commercialization, and clinical adoption of AI and explores the potential enhancements to PET imaging brought on by AI in the near future. In particular, the combination of on-demand image reconstruction, AI, and custom-designed data-processing workflows may open new possibilities for innovation which would positively impact the industry and ultimately patients.

Machine learning solutions in radiology: does the emperor have no clothes?

KEY POINTS

• Interest in radiomics and machine learning is steadily increasing and is reflected both in research output and number of commercially available solutions.• Currently available commercial products using machine learning are often supported by limited evidence of clinical usefulness and studies are often of low methodological quality.• Ethical and regulatory issues remain open and hinder implementation of machine learning software packages in daily clinical practice.