Artificial intelligence in digital pathology: a systematic review and meta-analysis of diagnostic test accuracy.

Ensuring diagnostic performance of artificial intelligence (AI) before introduction into clinical practice is essential. Growing numbers of studies using AI for digital pathology have been reported over recent years. The aim of this work is to examine the diagnostic accuracy of AI in digital pathology images for any disease. This systematic review and meta-analysis included diagnostic accuracy studies using any type of AI applied to whole slide images (WSIs) for any disease. The reference standard was diagnosis by histopathological assessment and/or immunohistochemistry. Searches were conducted in PubMed, EMBASE and CENTRAL in June 2022. Risk of bias and concerns of applicability were assessed using the QUADAS-2 tool. Data extraction was conducted by two investigators and meta-analysis was performed using a bivariate random effects model, with additional subgroup analyses also performed. Of 2976 identified studies, 100 were included in the review and 48 in the meta-analysis. Studies were from a range of countries, including over 152,000 whole slide images (WSIs), representing many diseases. These studies reported a mean sensitivity of 96.3% (CI 94.1-97.7) and mean specificity of 93.3% (CI 90.5-95.4). There was heterogeneity in study design and 99% of studies identified for inclusion had at least one area at high or unclear risk of bias or applicability concerns. Details on selection of cases, division of model development and validation data and raw performance data were frequently ambiguous or missing. AI is reported as having high diagnostic accuracy in the reported areas but requires more rigorous evaluation of its performance.

Quantitative assessment of H&E staining for pathology: development and clinical evaluation of a novel system.

BACKGROUND: Staining tissue samples to visualise cellular detail and tissue structure is at the core of pathology diagnosis, but variations in staining can result in significantly different appearances of the tissue sample. While the human visual system is adept at compensating for stain variation, with the growth of digital imaging in pathology, the impact of this variation can be more profound. Despite the ubiquity of haematoxylin and eosin staining in clinical practice worldwide, objective quantification is not yet available. We propose a method for quantitative haematoxylin and eosin stain assessment to facilitate quality assurance of histopathology staining, enabling truly quantitative quality control and improved standardisation. METHODS: The stain quantification method comprises conventional microscope slides with a stain-responsive biopolymer film affixed to one side, called stain assessment slides. The stain assessment slides were characterised with haematoxylin and eosin, and implemented in one clinical laboratory to quantify variation levels. RESULTS: Stain assessment slide stain uptake increased linearly with duration of haematoxylin and eosin staining (r = 0.99), and demonstrated linearly comparable staining to samples of human liver tissue (r values 0.98-0.99). Laboratory implementation of this technique quantified intra- and inter-instrument variation of staining instruments at one point in time and across a five-day period. CONCLUSION: The proposed method has been shown to reliably quantify stain uptake, providing an effective laboratory quality control method for stain variation. This is especially important for whole slide imaging and the future development of artificial intelligence in digital pathology.

Immune subtyping of melanoma whole slide images using multiple instance learning.

Determining early-stage prognostic markers and stratifying patients for effective treatment are two key challenges for improving outcomes for melanoma patients. Previous studies have used tumour transcriptome data to stratify patients into immune subgroups, which were associated with differential melanoma specific survival and potential predictive biomarkers. However, acquiring transcriptome data is a time-consuming and costly process. Moreover, it is not routinely used in the current clinical workflow. Here, we attempt to overcome this by developing deep learning models to classify gigapixel haematoxylin and eosin (H&E) stained pathology slides, which are well established in clinical workflows, into these immune subgroups. We systematically assess six different multiple instance learning (MIL) frameworks, using five different image resolutions and three different feature extraction methods. We show that pathology-specific self-supervised models using 10x resolution patches generate superior representations for the classification of immune subtypes. In addition, in a primary melanoma dataset, we achieve a mean area under the receiver operating characteristic curve (AUC) of 0.80 for classifying histopathology images into 'high' or 'low immune' subgroups and a mean AUC of 0.82 in an independent TCGA melanoma dataset. Furthermore, we show that these models are able to stratify patients into 'high' and 'low immune' subgroups with significantly different melanoma specific survival outcomes (log rank test, P< 0.005). We anticipate that MIL methods will allow us to find new biomarkers of high importance, act as a tool for clinicians to infer the immune landscape of tumours and stratify patients, without needing to carry out additional expensive genetic tests.

The value of standards for health datasets in artificial intelligence-based applications.

Artificial intelligence as a medical device is increasingly being applied to healthcare for diagnosis, risk stratification and resource allocation. However, a growing body of evidence has highlighted the risk of algorithmic bias, which may perpetuate existing health inequity. This problem arises in part because of systemic inequalities in dataset curation, unequal opportunity to participate in research and inequalities of access. This study aims to explore existing standards, frameworks and best practices for ensuring adequate data diversity in health datasets. Exploring the body of existing literature and expert views is an important step towards the development of consensus-based guidelines. The study comprises two parts: a systematic review of existing standards, frameworks and best practices for healthcare datasets; and a survey and thematic analysis of stakeholder views of bias, health equity and best practices for artificial intelligence as a medical device. We found that the need for dataset diversity was well described in literature, and experts generally favored the development of a robust set of guidelines, but there were mixed views about how these could be implemented practically. The outputs of this study will be used to inform the development of standards for transparency of data diversity in health datasets (the STANDING Together initiative).

Artificial intelligence and medical research databases: ethical review by data access committees.

BACKGROUND: It has been argued that ethics review committees-e.g., Research Ethics Committees, Institutional Review Boards, etc.- have weaknesses in reviewing big data and artificial intelligence research. For instance, they may, due to the novelty of the area, lack the relevant expertise for judging collective risks and benefits of such research, or they may exempt it from review in instances involving de-identified data. MAIN BODY: Focusing on the example of medical research databases we highlight here ethical issues around de-identified data sharing which motivate the need for review where oversight by ethics committees is weak. Though some argue for ethics committee reform to overcome these weaknesses, it is unclear whether or when that will happen. Hence, we argue that ethical review can be done by data access committees, since they have de facto purview of big data and artificial intelligence projects, relevant technical expertise and governance knowledge, and already take on some functions of ethical review. That said, like ethics committees, they may have functional weaknesses in their review capabilities. To strengthen that function, data access committees must think clearly about the kinds of ethical expertise, both professional and lay, that they draw upon to support their work. CONCLUSION: Data access committees can undertake ethical review of medical research databases provided they enhance that review function through professional and lay ethical expertise.

What Works Where and How for Uptake and Impact of Artificial Intelligence in Pathology: Review of Theories for a Realist Evaluation.

BACKGROUND: There is increasing interest in the use of artificial intelligence (AI) in pathology to increase accuracy and efficiency. To date, studies of clinicians' perceptions of AI have found only moderate acceptability, suggesting the need for further research regarding how to integrate it into clinical practice. OBJECTIVE: The aim of the study was to determine contextual factors that may support or constrain the uptake of AI in pathology. METHODS: To go beyond a simple listing of barriers and facilitators, we drew on the approach of realist evaluation and undertook a review of the literature to elicit stakeholders' theories of how, for whom, and in what circumstances AI can provide benefit in pathology. Searches were designed by an information specialist and peer-reviewed by a second information specialist. Searches were run on the arXiv.org repository, MEDLINE, and the Health Management Information Consortium, with additional searches undertaken on a range of websites to identify gray literature. In line with a realist approach, we also made use of relevant theory. Included documents were indexed in NVivo 12, using codes to capture different contexts, mechanisms, and outcomes that could affect the introduction of AI in pathology. Coded data were used to produce narrative summaries of each of the identified contexts, mechanisms, and outcomes, which were then translated into theories in the form of context-mechanism-outcome configurations. RESULTS: A total of 101 relevant documents were identified. Our analysis indicates that the benefits that can be achieved will vary according to the size and nature of the pathology department's workload and the extent to which pathologists work collaboratively; the major perceived benefit for specialist centers is in reducing workload. For uptake of AI, pathologists' trust is essential. Existing theories suggest that if pathologists are able to "make sense" of AI, engage in the adoption process, receive support in adapting their work processes, and can identify potential benefits to its introduction, it is more likely to be accepted. CONCLUSIONS: For uptake of AI in pathology, for all but the most simple quantitative tasks, measures will be required that either increase confidence in the system or provide users with an understanding of the performance of the system. For specialist centers, efforts should focus on reducing workload rather than increasing accuracy. Designers also need to give careful thought to usability and how AI is integrated into pathologists' workflow.

Image analysis of cutaneous melanoma histology: a systematic review and meta-analysis.

The current subjective histopathological assessment of cutaneous melanoma is challenging. The application of image analysis algorithms to histological images may facilitate improvements in workflow and prognostication. To date, several individual algorithms applied to melanoma histological images have been reported with variations in approach and reported accuracies. Histological digital images can be created using a camera mounted on a light microscope, or through whole slide image (WSI) generation using a whole slide scanner. Before any such tool could be integrated into clinical workflow, the accuracy of the technology should be carefully evaluated and summarised. Therefore, the objective of this review was to evaluate the accuracy of existing image analysis algorithms applied to digital histological images of cutaneous melanoma. Database searching of PubMed and Embase from inception to 11th March 2022 was conducted alongside citation checking and examining reports from organisations. All studies reporting accuracy of any image analysis applied to histological images of cutaneous melanoma, were included. The reference standard was any histological assessment of haematoxylin and eosin-stained slides and/or immunohistochemical staining. Citations were independently deduplicated and screened by two review authors and disagreements were resolved through discussion. The data was extracted concerning study demographics; type of image analysis; type of reference standard; conditions included and test statistics to construct 2 × 2 tables. Data was extracted in accordance with our protocol and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses-Diagnostic Test Accuracy (PRISMA-DTA) Statement. A bivariate random-effects meta-analysis was used to estimate summary sensitivities and specificities with 95% confidence intervals (CI). Assessment of methodological quality was conducted using a tailored version of the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. The primary outcome was the pooled sensitivity and specificity of image analysis applied to cutaneous melanoma histological images. Sixteen studies were included in the systematic review, representing 4,888 specimens. Six studies were included in the meta-analysis. The mean sensitivity and specificity of automated image analysis algorithms applied to melanoma histological images was 90% (CI 82%, 95%) and 92% (CI 79%, 97%), respectively. Based on limited and heterogeneous data, image analysis appears to offer high accuracy when applied to histological images of cutaneous melanoma. However, given the early exploratory nature of these studies, further development work is necessary to improve their performance.

Survey of liver pathologists to assess attitudes towards digital pathology and artificial intelligence.

AIMS: A survey of members of the UK Liver Pathology Group (UKLPG) was conducted, comprising consultant histopathologists from across the UK who report liver specimens and participate in the UK National Liver Pathology External Quality Assurance scheme. The aim of this study was to understand attitudes and priorities of liver pathologists towards digital pathology and artificial intelligence (AI). METHODS: The survey was distributed to all full consultant members of the UKLPG via email. This comprised 50 questions, with 48 multiple choice questions and 2 free-text questions at the end, covering a range of topics and concepts pertaining to the use of digital pathology and AI in liver disease. RESULTS: Forty-two consultant histopathologists completed the survey, representing 36% of fully registered members of the UKLPG (42/116). Questions examining digital pathology showed respondents agreed with the utility of digital pathology for primary diagnosis 83% (34/41), second opinions 90% (37/41), research 85% (35/41) and training and education 95% (39/41). Fatty liver diseases were an area of demand for AI tools with 80% in agreement (33/41), followed by neoplastic liver diseases with 59% in agreement (24/41). Participants were concerned about AI development without pathologist involvement 73% (30/41), however, 63% (26/41) disagreed when asked whether AI would replace pathologists. CONCLUSIONS: This study outlines current interest, priorities for research and concerns around digital pathology and AI for liver pathologists. The majority of UK liver pathologists are in favour of the application of digital pathology and AI in clinical practice, research and education.

How, for whom, and in what contexts will artificial intelligence be adopted in pathology? A realist interview study.

OBJECTIVE: There is increasing interest in using artificial intelligence (AI) in pathology to improve accuracy and efficiency. Studies of clinicians' perceptions of AI have found only moderate acceptability, suggesting further research is needed regarding integration into clinical practice. This study aimed to explore stakeholders' theories concerning how and in what contexts AI is likely to become integrated into pathology. MATERIALS AND METHODS: A literature review provided tentative theories that were revised through a realist interview study with 20 pathologists and 5 pathology trainees. Questions sought to elicit whether, and in what ways, the tentative theories fitted with interviewees' perceptions and experiences. Analysis focused on identifying the contextual factors that may support or constrain uptake of AI in pathology. RESULTS: Interviews highlighted the importance of trust in AI, with interviewees emphasizing evaluation and the opportunity for pathologists to become familiar with AI as means for establishing trust. Interviewees expressed a desire to be involved in design and implementation of AI tools, to ensure such tools address pressing needs, but needs vary by subspecialty. Workflow integration is desired but whether AI tools should work automatically will vary according to the task and the context. CONCLUSIONS: It must not be assumed that AI tools that provide benefit in one subspecialty will provide benefit in others. Pathologists should be involved in the decision to introduce AI, with opportunity to assess strengths and weaknesses. Further research is needed concerning the evidence required to satisfy pathologists regarding the benefits of AI.

Faster than light (microscopy): superiority of digital pathology over microscopy for assessment of immunohistochemistry.

AIMS: Digital pathology offers the potential for significant benefits in diagnostic pathology, but currently the efficiency of slide viewing is a barrier to adoption. We hypothesised that presenting digital slides for simultaneous viewing of multiple sections of tissue for comparison, as in those with immunohistochemical panels, would allow pathologists to review cases more quickly. METHODS: Novel software was developed to view synchronised parallel tissue sections on a digital pathology workstation. Sixteen histopathologists reviewed three liver biopsy cases including an immunohistochemical panel using the digital microscope, and three different liver biopsy cases including an immunohistochemical panel using the light microscope. The order of cases and interface was fully counterbalanced. Time to diagnosis was recorded and mean times are presented as data approximated to a normalised distribution. RESULTS: Mean time to diagnosis was 4 min 3 s using the digital microscope and 5 min 24 s using the light microscope, saving 1 min 21 s (95% CI 16 s to 2 min 26 s; p=0.02), using the digital microscope. Overall normalised mean time to diagnosis was 85% on the digital pathology workstation compared with 115% on the microscope, a relative reduction of 26%. CONCLUSIONS: With appropriate interface design, it is quicker to review immunohistochemical slides using a digital microscope than the conventional light microscope, without incurring any major diagnostic errors. As digital pathology becomes more integrated with routine clinical workflow and pathologists increase their experience of the technology, it is anticipated that other tasks will also become more time-efficient.

Generalization of Deep Learning in Digital Pathology: Experience in Breast Cancer Metastasis Detection.

Poor generalizability is a major barrier to clinical implementation of artificial intelligence in digital pathology. The aim of this study was to test the generalizability of a pretrained deep learning model to a new diagnostic setting and to a small change in surgical indication. A deep learning model for breast cancer metastases detection in sentinel lymph nodes, trained on CAMELYON multicenter data, was used as a base model, and achieved an AUC of 0.969 (95% CI 0.926-0.998) and FROC of 0.838 (95% CI 0.757-0.913) on CAMELYON16 test data. On local sentinel node data, the base model performance dropped to AUC 0.929 (95% CI 0.800-0.998) and FROC 0.744 (95% CI 0.566-0.912). On data with a change in surgical indication (axillary dissections) the base model performance indicated an even larger drop with a FROC of 0.503 (95%CI 0.201-0.911). The model was retrained with addition of local data, resulting in about a 4% increase for both AUC and FROC for sentinel nodes, and an increase of 11% in AUC and 49% in FROC for axillary nodes. Pathologist qualitative evaluation of the retrained model´s output showed no missed positive slides. False positives, false negatives and one previously undetected micro-metastasis were observed. The study highlights the generalization challenge even when using a multicenter trained model, and that a small change in indication can considerably impact the model´s performance.

Attention-guided sampling for colorectal cancer analysis with digital pathology.

Improvements to patient care through the development of automated image analysis in pathology are restricted by the small image patch size that can be processed by convolutional neural networks (CNNs), when compared to the whole-slide image (WSI). Tile-by-tile processing across the entire WSI is slow and inefficient. While this may improve with future computing power, the technique remains vulnerable to noise from uninformative image areas. We propose a novel attention-inspired algorithm that selects image patches from informative parts of the WSI, first using a sparse randomised grid pattern, then iteratively re-sampling at higher density in regions where a CNN classifies patches as tumour. Subsequent uniform sampling across the enclosing region of interest (ROI) is used to mitigate sampling bias. Benchmarking tests informed the adoption of VGG19 as the main CNN architecture, with 79% classification accuracy. A further CNN was trained to separate false-positive normal epithelium from tumour epithelium, in a novel adaptation of a two-stage model used in brain imaging. These subsystems were combined in a processing pipeline to generate spatial distributions of classified patches from unseen WSIs. The ROI was predicted with a mean F1 (Dice) score of 86.6% over 100 evaluation WSIs. Several algorithms for evaluating tumour-stroma ratio (TSR) within the ROI were compared, giving a lowest root mean square (RMS) error of 11.3% relative to pathologists' annotations, against 13.5% for an equivalent tile-by-tile pipeline. Our pipeline processed WSIs between 3.3x and 6.3x faster than tile-by-tile processing. We propose our attention-based sampling pipeline as a useful tool for pathology researchers, with the further potential for incorporating additional diagnostic calculations.

Reporting of Artificial Intelligence Diagnostic Accuracy Studies in Pathology Abstracts: Compliance with STARD for Abstracts Guidelines.

Artificial intelligence (AI) research is transforming the range tools and technologies available to pathologists, leading to potentially faster, personalized and more accurate diagnoses for patients. However, to see the use of tools for patient benefit and achieve this safely, the implementation of any algorithm must be underpinned by high quality evidence from research that is understandable, replicable, usable and inclusive of details needed for critical appraisal of potential bias. Evidence suggests that reporting guidelines can improve the completeness of reporting of research, especially with good awareness of guidelines. The quality of evidence provided by abstracts alone is profoundly important, as they influence the decision of a researcher to read a paper, attend a conference presentation or include a study in a systematic review. AI abstracts at two international pathology conferences were assessed to establish completeness of reporting against the STARD for Abstracts criteria. This reporting guideline is for abstracts of diagnostic accuracy studies and includes a checklist of 11 essential items required to accomplish satisfactory reporting of such an investigation. A total of 3488 abstracts were screened from the United States & Canadian Academy of Pathology annual meeting 2019 and the 31st European Congress of Pathology (ESP Congress). Of these, 51 AI diagnostic accuracy abstracts were identified and assessed against the STARD for Abstracts criteria for completeness of reporting. Completeness of reporting was suboptimal for the 11 essential criteria, a mean of 5.8 (SD 1.5) items were detailed per abstract. Inclusion was variable across the different checklist items, with all abstracts including study objectives and no abstracts including a registration number or registry. Greater use and awareness of the STARD for Abstracts criteria could improve completeness of reporting and further consideration is needed for areas where AI studies are vulnerable to bias.

Public governance of medical artificial intelligence research in the UK: an integrated multi-scale model.

There is a growing consensus among scholars, national governments, and intergovernmental organisations of the need to involve the public in decision-making around the use of artificial intelligence (AI) in society. Focusing on the UK, this paper asks how that can be achieved for medical AI research, that is, for research involving the training of AI on data from medical research databases. Public governance of medical AI research in the UK is generally achieved in three ways, namely, via lay representation on data access committees, through patient and public involvement groups, and by means of various deliberative democratic projects such as citizens' juries, citizen panels, citizen assemblies, etc.-what we collectively call "citizen forums". As we will show, each of these public involvement initiatives have complementary strengths and weaknesses for providing oversight of medical AI research. As they are currently utilized, however, they are unable to realize the full potential of their complementarity due to insufficient information transfer across them. In order to synergistically build on their contributions, we offer here a multi-scale model integrating all three. In doing so we provide a unified public governance model for medical AI research, one that, we argue, could improve the trustworthiness of big data and AI related medical research in the future.

How might the public be authentically involved in decisions about medical data sharing for artificial intelligence (AI) research? In this paper, we highlight three ways in which public views are used to improve such decisions, namely, through lay representation on data access committees, through patient and public involvement groups, and through a variety of public engagement events we call "citizen forums." Though each approach has common strengths and weaknesses, we argue that they are unable to support each other due to a lack of proper integration. We therefore propose combining them so that they work in a more coordinated way. The combined model, we argue, could be useful for improving the trustworthiness of big data and AI related medical research in the future.

The ethical challenges of artificial intelligence-driven digital pathology.

Digital pathology - the digitalisation of clinical histopathology services through the scanning and storage of pathology slides - has opened up new possibilities for health care in recent years, particularly in the opportunities it brings for artificial intelligence (AI)-driven research. Recognising, however, that there is little scholarly debate on the ethics of digital pathology when used for AI research, this paper summarises what it sees as four key ethical issues to consider when deploying AI infrastructures in pathology, namely, privacy, choice, equity, and trust. The themes are inspired from the authors' experience grappling with the challenge of deploying an ethical digital pathology infrastructure to support AI research as part of the National Pathology Imaging Cooperative (NPIC), a collaborative of universities, hospital trusts, and industry partners largely located across the North of England. Though focusing on the UK case, internationally, few pathology departments have gone fully digital, and so the themes developed here offer a heuristic for ethical reflection for other departments currently making a similar transition or planning to do so in the future. We conclude by promoting the need for robust public governance mechanisms in AI-driven digital pathology.

Understanding the ethical and legal considerations of Digital Pathology.

Digital Pathology (DP) is a platform which has the potential to develop a truly integrated and global pathology community. The generation of DP data at scale creates novel challenges for the histopathology community in managing, processing, and governing the use of these data. The current understanding of, and confidence in, the legal and ethical aspects of DP by pathologists is unknown. We developed an electronic survey (e-survey), comprising 22 questions, with input from the Royal College of Pathologists (RCPath) Digital Pathology Working Group. The e-survey was circulated via e-mail and social media (Twitter) through the RCPath Digital Pathology Working Group network, RCPath Trainee Committee network, the Pathology image data Lake for Analytics, Knowledge and Education (PathLAKE) digital pathology consortium, National Pathology Imaging Co-operative (NPIC), local contacts, and to the membership of both The Pathological Society of Great Britain and Ireland and the British Division of the International Academy of Pathology (BDIAP). Between 14 July 2020 and 6 September 2020, we collected 198 responses representing a cross section of histopathologists, including individuals with experience of DP research. We ascertained that, in the UK, DP is being used for diagnosis, research, and teaching, and that the platform is enabling data sharing. Our survey demonstrated that there is often a lack of confidence and understanding of the key issues of consent, legislation, and ethical guidelines. Of 198 respondents, 82 (41%) did not know when the use of digital scanned slide images would fall under the relevant legislation and 93 (47%) were 'Not confident at all' in their interpretation of consent for scanned slide images in research. With increasing uptake of DP, a working knowledge of these areas is essential but histopathologists often express a lack of confidence in these topics. The need for specific training in these areas is highlighted by the findings of this study.

Hepatectomy risk assessment with functional magnetic resonance imaging (HEPARIM).

BACKGROUND: Post hepatectomy liver failure (PHLF) remains a significant risk in patients undergoing curative liver resection for cancer, however currently available PHLF risk prediction investigations are not sufficiently accurate. The Hepatectomy risk assessment with functional magnetic resonance imaging trial (HEPARIM) aims to establish if quantitative MRI biomarkers of liver function & perfusion can be used to more accurately predict PHLF risk and FLR function, measured against indocyanine green (ICG) liver function test. METHODS: HEPARIM is an observational cohort study recruiting patients undergoing liver resection of 2 segments or more, prior to surgery patients will have both Dynamic Gadoxetate-enhanced (DGE) liver MRI and ICG testing. Day one post op ICG testing is repeated and R15 compared to the Gadoxetate Clearance (GC) of the future liver remnant (FLR-GC) as measure by preoperative DGE- MRI which is the primary outcome, and preoperative ICG R15 compared to GC of whole liver (WL-GC) as a secondary outcome. Data will be collected from medical records, biochemistry, pathology and radiology reports and used in a multi-variate analysis to the value of functional MRI and derive multivariant prediction models for future validation. DISCUSSION: If successful, this test will potentially provide an efficient means to quantitatively assess FLR function and PHLF risk enabling surgeons to push boundaries of liver surgery further while maintaining safe practice and thereby offering chance of cure to patients who would previously been deemed inoperable. MRI has the added benefit of already being part of the routine diagnostic pathway and as such would have limited additional burden on patients time or cost to health care systems. (Hepatectomy Risk Assessment With Functional Magnetic Resonance Imaging - Full Text View - ClinicalTrials.gov , n.d.) TRIAL REGISTRATION: ClinicalTrials.gov, ClinicalTrials.gov NCT04705194 - Registered 12th January 2021 - Retrospectively registered.