Data infrastructures for AI in medical imaging: a report on the experiences of five EU projects.

Artificial intelligence (AI) is transforming the field of medical imaging and has the potential to bring medicine from the era of 'sick-care' to the era of healthcare and prevention. The development of AI requires access to large, complete, and harmonized real-world datasets, representative of the population, and disease diversity. However, to date, efforts are fragmented, based on single-institution, size-limited, and annotation-limited datasets. Available public datasets (e.g., The Cancer Imaging Archive, TCIA, USA) are limited in scope, making model generalizability really difficult. In this direction, five European Union projects are currently working on the development of big data infrastructures that will enable European, ethically and General Data Protection Regulation-compliant, quality-controlled, cancer-related, medical imaging platforms, in which both large-scale data and AI algorithms will coexist. The vision is to create sustainable AI cloud-based platforms for the development, implementation, verification, and validation of trustable, usable, and reliable AI models for addressing specific unmet needs regarding cancer care provision. In this paper, we present an overview of the development efforts highlighting challenges and approaches selected providing valuable feedback to future attempts in the area.Key points• Artificial intelligence models for health imaging require access to large amounts of harmonized imaging data and metadata.• Main infrastructures adopted either collect centrally anonymized data or enable access to pseudonymized distributed data.• Developing a common data model for storing all relevant information is a challenge.• Trust of data providers in data sharing initiatives is essential.• An online European Union meta-tool-repository is a necessity minimizing effort duplication for the various projects in the area.

Position of the AI for Health Imaging (AI4HI) network on metadata models for imaging biobanks.

A huge amount of imaging data is becoming available worldwide and an incredible range of possible improvements can be provided by artificial intelligence algorithms in clinical care for diagnosis and decision support. In this context, it has become essential to properly manage and handle these medical images and to define which metadata have to be considered, in order for the images to provide their full potential. Metadata are additional data associated with the images, which provide a complete description of the image acquisition, curation, analysis, and of the relevant clinical variables associated with the images. Currently, several data models are available to describe one or more subcategories of metadata, but a unique, common, and standard data model capable of fully representing the heterogeneity of medical metadata has not been yet developed. This paper reports the state of the art on metadata models for medical imaging, the current limitations and further developments, and describes the strategy adopted by the Horizon 2020 "AI for Health Imaging" projects, which are all dedicated to the creation of imaging biobanks.

A neural network-based algorithm for simultaneous event positioning and timestamping in monolithic scintillators.

Objective. Monolithic scintillator crystals coupled to silicon photomultiplier (SiPM) arrays are promising detectors for PET applications, offering spatial resolution around 1 mm and depth-of-interaction information. However, their timing resolution has always been inferior to that of pixellated crystals, while the best results on spatial resolution have been obtained with algorithms that cannot operate in real-time in a PET detector. In this study, we explore the capabilities of monolithic crystals with respect to spatial and timing resolution, presenting new algorithms that overcome the mentioned problems.Approach.Our algorithms were tested first using a simulation framework, then on experimentally acquired data. We tested an event timestamping algorithm based on neural networks which was then integrated into a second neural network for simultaneous estimation of the event position and timestamp. Both algorithms are implemented in a low-cost field-programmable gate array that can be integrated in the detector and can process more than 1 million events per second in real-time.Results.Testing the neural network for the simultaneous estimation of the event position and timestamp on experimental data we obtain 0.78 2D FWHM on the (x,y) plane, 1.2 depth-of-interaction FWHM and 156 coincidence time resolution on a25mm×25mm×8mm×LYSO monolith read-out by 643mm×3mmHamamatsu SiPMs.Significance.Our results show that monolithic crystals combined with artificial intelligence can rival pixellated crystals performance for time-of-flight PET applications, while having better spatial resolution and DOI resolution. Thanks to the use of very light neural networks, event characterization can be done on-line directly in the detector, solving the issues of scalability and computational complexity that up to now were preventing the use of monolithic crystals in clinical PET scanners.

Performance evaluation of the LightPath imaging system for intra-operative Cerenkov luminescence imaging.

The performances of an intra-operative optical imaging system for Cerenkov luminescence imaging of resected tumor specimens were evaluated with phantom studies. The spatial resolution, the linearity of the measured signal with the activity concentration and the minimum detectable activity concentration were considered. A high linearity was observed over a broad range of activity concentration (R2⩾0.99 down to ∼40 kBq/ml of 18F-FDG). For 18F-FDG activity distributions 2 mm deep in biological tissue, the measured detection limit was 8 kBq/ml and a spatial resolution of 2.5 mm was obtained. The detection limit of the imaging system is comparable with clinical activity concentrations in tumor specimens, and the spatial resolution is compatible with clinical requirements.

Cerenkov luminescence imaging: physics principles and potential applications in biomedical sciences.

Cerenkov luminescence imaging (CLI) is a novel imaging modality to study charged particles with optical methods by detecting the Cerenkov luminescence produced in tissue. This paper first describes the physical processes that govern the production and transport in tissue of Cerenkov luminescence. The detectors used for CLI and their most relevant specifications to optimize the acquisition of the Cerenkov signal are then presented, and CLI is compared with the other optical imaging modalities sharing the same data acquisition and processing methods. Finally, the scientific work related to CLI and the applications for which CLI has been proposed are reviewed. The paper ends with some considerations about further perspectives for this novel imaging modality.

Cherenkov luminescence measurements with digital silicon photomultipliers: a feasibility study.

BACKGROUND: A feasibility study was done to assess the capability of digital silicon photomultipliers to measure the Cherenkov luminescence emitted by a ß source. Cherenkov luminescence imaging (CLI) is possible with a charge coupled device (CCD) based technology, but a stand-alone technique for quantitative activity measurements based on Cherenkov luminescence has not yet been developed. Silicon photomultipliers (SiPMs) are photon counting devices with a fast impulse response and can potentially be used to quantify ß-emitting radiotracer distributions by CLI. METHODS: In this study, a Philips digital photon counting (PDPC) silicon photomultiplier detector was evaluated for measuring Cherenkov luminescence. The PDPC detector is a matrix of avalanche photodiodes, which were read one at a time in a dark count map (DCM) measurement mode (much like a CCD). This reduces the device active area but allows the information from a single avalanche photodiode to be preserved, which is not possible with analog SiPMs. An algorithm to reject the noisiest photodiodes and to correct the measured count rate for the dark current was developed. RESULTS: The results show that, in DCM mode and at (10-13) °C, the PDPC has a dynamic response to different levels of Cherenkov luminescence emitted by a ß source and transmitted through an opaque medium. This suggests the potential for this approach to provide quantitative activity measurements. Interestingly, the potential use of the PDPC in DCM mode for direct imaging of Cherenkov luminescence, as a opposed to a scalar measurement device, was also apparent. CONCLUSIONS: We showed that a PDPC tile in DCM mode is able to detect and image a ß source through its Cherenkov radiation emission. The detector's dynamic response to different levels of radiation suggests its potential quantitative capabilities, and the DCM mode allows imaging with a better spatial resolution than the conventional event-triggered mode. Finally, the same acquisition procedure and data processing could be employed also for other low light levels applications, such as bioluminescence.