Automated detection of early signs of irreversible ischemic change on CTA source images in patients with large vessel occlusion.

PURPOSE: To create and validate an automated pipeline for detection of early signs of irreversible ischemic change from admission CTA in patients with large vessel occlusion (LVO) stroke. METHODS: We retrospectively included 368 patients for training and 143 for external validation. All patients had anterior circulation LVO stroke, endovascular therapy with successful reperfusion, and follow-up diffusion-weighted imaging (DWI). We devised a pipeline to automatically segment Alberta Stroke Program Early CT Score (ASPECTS) regions and extracted their relative Hounsfield unit (rHU) values. We determined the optimal rHU cut points for prediction of final infarction in each ASPECT region, performed 10-fold cross-validation in the training set, and measured the performance via external validation in patients from another institute. We compared the model with an expert neuroradiologist for prediction of final infarct volume and poor functional outcome. RESULTS: We achieved a mean area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity of 0.69±0.13, 0.69±0.09, 0.61±0.23, and 0.72±0.11 across all regions and folds in cross-validation. In the external validation cohort, we achieved a median [interquartile] AUC, accuracy, sensitivity, and specificity of 0.71 [0.68-0.72], 0.70 [0.68-0.73], 0.55 [0.50-0.63], and 0.74 [0.73-0.77], respectively. The rHU-based ASPECTS showed significant correlation with DWI-based ASPECTS (rS = 0.39, p<0.001) and final infarct volume (rS = -0.36, p<0.001). The AUC for predicting poor functional outcome was 0.66 (95%CI: 0.57-0.75). The predictive capabilities of rHU-based ASPECTS were not significantly different from the neuroradiologist's visual ASPECTS for either final infarct volume or functional outcome. CONCLUSIONS: Our study demonstrates the feasibility of an automated pipeline and predictive model based on relative HU attenuation of ASPECTS regions on baseline CTA and its non-inferior performance in predicting final infarction on post-stroke DWI compared to an expert human reader.

The unmet promise of trustworthy AI in healthcare: why we fail at clinical translation.

Artificial intelligence (AI) has the potential to revolutionize healthcare, for example via decision support systems, computer vision approaches, or AI-based prevention tools. Initial results from AI applications in healthcare show promise but are rarely translated into clinical practice successfully and ethically. This occurs despite an abundance of "Trustworthy AI" guidelines. How can we explain the translational gaps of AI in healthcare? This paper offers a fresh perspective on this problem, showing that failing translation of healthcare AI markedly arises from a lack of an operational definition of "trust" and "trustworthiness". This leads to (a) unintentional misuse concerning what trust (worthiness) is and (b) the risk of intentional abuse by industry stakeholders engaging in ethics washing. By pointing out these issues, we aim to highlight the obstacles that hinder translation of Trustworthy medical AI to practice and prevent it from fulfilling its unmet promises.

A trustworthy AI reality-check: the lack of transparency of artificial intelligence products in healthcare.

Trustworthy medical AI requires transparency about the development and testing of underlying algorithms to identify biases and communicate potential risks of harm. Abundant guidance exists on how to achieve transparency for medical AI products, but it is unclear whether publicly available information adequately informs about their risks. To assess this, we retrieved public documentation on the 14 available CE-certified AI-based radiology products of the II b risk category in the EU from vendor websites, scientific publications, and the European EUDAMED database. Using a self-designed survey, we reported on their development, validation, ethical considerations, and deployment caveats, according to trustworthy AI guidelines. We scored each question with either 0, 0.5, or 1, to rate if the required information was "unavailable", "partially available," or "fully available." The transparency of each product was calculated relative to all 55 questions. Transparency scores ranged from 6.4% to 60.9%, with a median of 29.1%. Major transparency gaps included missing documentation on training data, ethical considerations, and limitations for deployment. Ethical aspects like consent, safety monitoring, and GDPR-compliance were rarely documented. Furthermore, deployment caveats for different demographics and medical settings were scarce. In conclusion, public documentation of authorized medical AI products in Europe lacks sufficient public transparency to inform about safety and risks. We call on lawmakers and regulators to establish legally mandated requirements for public and substantive transparency to fulfill the promise of trustworthy AI for health.

A deep learning analysis of stroke onset time prediction and comparison to DWI-FLAIR mismatch.

INTRODUCTION: When time since stroke onset is unknown, DWI-FLAIR mismatch rating is an established technique for patient stratification. A visible DWI lesion without corresponding parenchymal hyperintensity on FLAIR suggests time since onset of under 4.5 h and thus a potential benefit from intravenous thrombolysis. To improve accuracy and availability of the mismatch concept, deep learning might be able to augment human rating and support decision-making in these cases. METHODS: We used unprocessed DWI and coregistered FLAIR imaging data to train a deep learning model to predict dichotomized time since ischemic stroke onset. We analyzed the performance of Group Convolutional Neural Networks compared to other deep learning methods. Unlabeled imaging data was used for pre-training. Prediction performance of the best deep learning model was compared to the performance of four independent junior and senior raters. Additionally, in cases deemed indeterminable by human raters, model ratings were used to augment human performance. Post-hoc gradient-based explanations were analyzed to gain insights into model predictions. RESULTS: Our best predictive model performed comparably to human raters. Using model ratings in cases deemed indeterminable by human raters improved rating accuracy and interrater agreement for junior and senior ratings. Post-hoc explainability analyses showed that the model localized stroke lesions to derive predictions. DISCUSSION: Our analysis shows that deep learning based clinical decision support has the potential to improve the accessibility of the DWI-FLAIR mismatch concept by supporting patient stratification.

Personalised simulation of hemodynamics in cerebrovascular disease: lessons learned from a study of diagnostic accuracy.

Intracranial atherosclerotic disease (ICAD) poses a significant risk of subsequent stroke but current prevention strategies are limited. Mechanistic simulations of brain hemodynamics offer an alternative precision medicine approach by utilising individual patient characteristics. For clinical use, however, current simulation frameworks have insufficient validation. In this study, we performed the first quantitative validation of a simulation-based precision medicine framework to assess cerebral hemodynamics in patients with ICAD against clinical standard perfusion imaging. In a retrospective analysis, we used a 0-dimensional simulation model to detect brain areas that are hemodynamically vulnerable to subsequent stroke. The main outcome measures were sensitivity, specificity, and area under the receiver operating characteristics curve (ROC AUC) of the simulation to identify brain areas vulnerable to subsequent stroke as defined by quantitative measurements of relative mean transit time (relMTT) from dynamic susceptibility contrast MRI (DSC-MRI). In 68 subjects with unilateral stenosis >70% of the internal carotid artery (ICA) or middle cerebral artery (MCA), the sensitivity and specificity of the simulation were 0.65 and 0.67, respectively. The ROC AUC was 0.68. The low-to-moderate accuracy of the simulation may be attributed to assumptions of Newtonian blood flow, rigid vessel walls, and the use of time-of-flight MRI for geometric representation of subject vasculature. Future simulation approaches should focus on integrating additional patient data, increasing accessibility of precision medicine tools to clinicians, addressing disease burden disparities amongst different populations, and quantifying patient benefit. Our results underscore the need for further improvement of mechanistic simulations of brain hemodynamics to foster the translation of the technology to clinical practice.

Artificial Intelligence for Clinical Decision Support in Acute Ischemic Stroke: A Systematic Review.

BACKGROUND: Established randomized trial-based parameters for acute ischemic stroke group patients into generic treatment groups, leading to attempts using various artificial intelligence (AI) methods to directly correlate patient characteristics to outcomes and thereby provide decision support to stroke clinicians. We review AI-based clinical decision support systems in the development stage, specifically regarding methodological robustness and constraints for clinical implementation. METHODS: Our systematic review included full-text English language publications proposing a clinical decision support system using AI techniques for direct decision support in acute ischemic stroke cases in adult patients. We (1) describe data and outcomes used in those systems, (2) estimate the systems' benefits compared with traditional stroke diagnosis and treatment, and (3) reported concordance with reporting standards for AI in healthcare. RESULTS: One hundred twenty-one studies met our inclusion criteria. Sixty-five were included for full extraction. In our sample, utilized data sources, methods, and reporting practices were highly heterogeneous. CONCLUSIONS: Our results suggest significant validity threats, dissonance in reporting practices, and challenges to clinical translation. We outline practical recommendations for the successful implementation of AI research in acute ischemic stroke treatment and diagnosis.

Expectations and attitudes towards medical artificial intelligence: A qualitative study in the field of stroke.

INTRODUCTION: Artificial intelligence (AI) has the potential to transform clinical decision-making as we know it. Powered by sophisticated machine learning algorithms, clinical decision support systems (CDSS) can generate unprecedented amounts of predictive information about individuals' health. Yet, despite the potential of these systems to promote proactive decision-making and improve health outcomes, their utility and impact remain poorly understood due to their still rare application in clinical practice. Taking the example of AI-powered CDSS in stroke medicine as a case in point, this paper provides a nuanced account of stroke survivors', family members', and healthcare professionals' expectations and attitudes towards medical AI. METHODS: We followed a qualitative research design informed by the sociology of expectations, which recognizes the generative role of individuals' expectations in shaping scientific and technological change. Semi-structured interviews were conducted with stroke survivors, family members, and healthcare professionals specialized in stroke based in Germany and Switzerland. Data was analyzed using a combination of inductive and deductive thematic analysis. RESULTS: Based on the participants' deliberations, we identified four presumed roles that medical AI could play in stroke medicine, including an administrative, assistive, advisory, and autonomous role AI. While most participants held positive attitudes towards medical AI and its potential to increase accuracy, speed, and efficiency in medical decision making, they also cautioned that it is not a stand-alone solution and may even lead to new problems. Participants particularly emphasized the importance of relational aspects and raised questions regarding the impact of AI on roles and responsibilities and patients' rights to information and decision-making. These findings shed light on the potential impact of medical AI on professional identities, role perceptions, and the doctor-patient relationship. CONCLUSION: Our findings highlight the need for a more differentiated approach to identifying and tackling pertinent ethical and legal issues in the context of medical AI. We advocate for stakeholder and public involvement in the development of AI and AI governance to ensure that medical AI offers solutions to the most pressing challenges patients and clinicians face in clinical care.

Machine Learning-Based Identification of Target Groups for Thrombectomy in Acute Stroke.

Whether endovascular thrombectomy (EVT) improves functional outcome in patients with large-vessel occlusion (LVO) stroke that do not comply with inclusion criteria of randomized controlled trials (RCTs) but that are considered for EVT in clinical practice is uncertain. We aimed to systematically identify patients with LVO stroke underrepresented in RCTs who might benefit from EVT. Following the premises that (i) patients without reperfusion after EVT represent a non-treated control group and (ii) the level of reperfusion affects outcome in patients with benefit from EVT but not in patients without treatment benefit, we systematically assessed the importance of reperfusion level on functional outcome prediction using machine learning in patients with LVO stroke treated with EVT in clinical practice (N = 5235, German-Stroke-Registry) and in patients treated with EVT or best medical management from RCTs (N = 1488, Virtual-International-Stroke-Trials-Archive). The importance of reperfusion level on outcome prediction in an RCT-like real-world cohort equaled the importance of EVT treatment allocation for outcome prediction in RCT data and was higher compared to an unselected real-world population. The importance of reperfusion level was magnified in patient groups underrepresented in RCTs, including patients with lower NIHSS scores (0-10), M2 occlusions, and lower ASPECTS (0-5 and 6-8). Reperfusion level was equally important in patients with vertebrobasilar as with anterior LVO stroke. The importance of reperfusion level for outcome prediction identifies patient target groups who likely benefit from EVT, including vertebrobasilar stroke patients and among patients underrepresented in RCT patients with low NIHSS scores, low ASPECTS, and M2 occlusions.

Generating evidence on privacy outcomes to inform privacy risk management: A way forward?

Effective and efficient privacy risk management (PRM) is a necessary condition to support digitalization in health care and secondary use of patient data in research. To reduce privacy risks, current PRM frameworks are rooted in an approach trying to reduce undesired technical/organizational outcomes such as broken encryption or unintentional data disclosure. Comparing this with risk management in preventive or therapeutic medicine, a key difference becomes apparent: in health-related risk management, medicine focuses on person-specific health outcomes, whereas PRM mostly targets more indirect, technical/organizational outcomes. In this paper, we illustrate and discuss how a PRM approach based on evidence of person-specific privacy outcomes might look using three consecutive steps: i) a specification of undesired person-specific privacy outcomes, ii) empirical assessments of their frequency and severity, and iii) empirical studies on how effectively the available PRM interventions reduce their frequency or severity. After an introduction of these three steps, we cover their status quo and outline open questions and PRM-specific challenges in need of further conceptual clarification and feasibility studies. Specific challenges of an outcome-oriented approach to PRM include the potential delays between concrete threats manifesting and the resulting person/group-specific privacy outcomes. Moreover, new ways of exploiting privacy-sensitive information to harm individuals could be developed in the future. The challenges described are of technical, legal, ethical, financial and resource-oriented nature. In health research, however, there is explicit discussion about how to overcome such challenges to make important outcome-based assessments as feasible as possible. This paper concludes that it might be the time to have this discussion in the PRM field as well.

Assessing Trustworthy AI in Times of COVID-19: Deep Learning for Predicting a Multiregional Score Conveying the Degree of Lung Compromise in COVID-19 Patients.

This article's main contributions are twofold: 1) to demonstrate how to apply the general European Union's High-Level Expert Group's (EU HLEG) guidelines for trustworthy AI in practice for the domain of healthcare and 2) to investigate the research question of what does "trustworthy AI" mean at the time of the COVID-19 pandemic. To this end, we present the results of a post-hoc self-assessment to evaluate the trustworthiness of an AI system for predicting a multiregional score conveying the degree of lung compromise in COVID-19 patients, developed and verified by an interdisciplinary team with members from academia, public hospitals, and industry in time of pandemic. The AI system aims to help radiologists to estimate and communicate the severity of damage in a patient's lung from Chest X-rays. It has been experimentally deployed in the radiology department of the ASST Spedali Civili clinic in Brescia, Italy, since December 2020 during pandemic time. The methodology we have applied for our post-hoc assessment, called Z-Inspection®, uses sociotechnical scenarios to identify ethical, technical, and domain-specific issues in the use of the AI system in the context of the pandemic.

Anatomical labeling of intracranial arteries with deep learning in patients with cerebrovascular disease.

Brain arteries are routinely imaged in the clinical setting by various modalities, e.g., time-of-flight magnetic resonance angiography (TOF-MRA). These imaging techniques have great potential for the diagnosis of cerebrovascular disease, disease progression, and response to treatment. Currently, however, only qualitative assessment is implemented in clinical applications, relying on visual inspection. While manual or semi-automated approaches for quantification exist, such solutions are impractical in the clinical setting as they are time-consuming, involve too many processing steps, and/or neglect image intensity information. In this study, we present a deep learning-based solution for the anatomical labeling of intracranial arteries that utilizes complete information from 3D TOF-MRA images. We adapted and trained a state-of-the-art multi-scale Unet architecture using imaging data of 242 patients with cerebrovascular disease to distinguish 24 arterial segments. The proposed model utilizes vessel-specific information as well as raw image intensity information, and can thus take tissue characteristics into account. Our method yielded a performance of 0.89 macro F1 and 0.90 balanced class accuracy (bAcc) in labeling aggregated segments and 0.80 macro F1 and 0.83 bAcc in labeling detailed arterial segments on average. In particular, a higher F1 score than 0.75 for most arteries of clinical interest for cerebrovascular disease was achieved, with higher than 0.90 F1 scores in the larger, main arteries. Due to minimal pre-processing, simple usability, and fast predictions, our method could be highly applicable in the clinical setting.

Comparing Poor and Favorable Outcome Prediction With Machine Learning After Mechanical Thrombectomy in Acute Ischemic Stroke.

Background and Purpose: Outcome prediction after mechanical thrombectomy (MT) in patients with acute ischemic stroke (AIS) and large vessel occlusion (LVO) is commonly performed by focusing on favorable outcome (modified Rankin Scale, mRS 0-2) after 3 months but poor outcome representing severe disability and mortality (mRS 5 and 6) might be of equal importance for clinical decision-making. Methods: We retrospectively analyzed patients with AIS and LVO undergoing MT from 2009 to 2018. Prognostic variables were grouped in baseline clinical (A), MRI-derived variables including mismatch [apparent diffusion coefficient (ADC) and time-to-maximum (Tmax) lesion volume] (B), and variables reflecting speed and extent of reperfusion (C) [modified treatment in cerebral ischemia (mTICI) score and time from onset to mTICI]. Three different scenarios were analyzed: (1) baseline clinical parameters only, (2) baseline clinical and MRI-derived parameters, and (3) all baseline clinical, imaging-derived, and reperfusion-associated parameters. For each scenario, we assessed prediction for favorable and poor outcome with seven different machine learning algorithms. Results: In 210 patients, prediction of favorable outcome was improved after including speed and extent of recanalization [highest area under the curve (AUC) 0.73] compared to using baseline clinical variables only (highest AUC 0.67). Prediction of poor outcome remained stable by using baseline clinical variables only (highest AUC 0.71) and did not improve further by additional variables. Prediction of favorable and poor outcomes was not improved by adding MR-mismatch variables. Most important baseline clinical variables for both outcomes were age, National Institutes of Health Stroke Scale, and premorbid mRS. Conclusions: Our results suggest that a prediction of poor outcome after AIS and MT could be made based on clinical baseline variables only. Speed and extent of MT did improve prediction for a favorable outcome but is not relevant for poor outcome. An MR mismatch with small ischemic core and larger penumbral tissue showed no predictive importance.

Toward Sharing Brain Images: Differentially Private TOF-MRA Images With Segmentation Labels Using Generative Adversarial Networks.

Sharing labeled data is crucial to acquire large datasets for various Deep Learning applications. In medical imaging, this is often not feasible due to privacy regulations. Whereas anonymization would be a solution, standard techniques have been shown to be partially reversible. Here, synthetic data using a Generative Adversarial Network (GAN) with differential privacy guarantees could be a solution to ensure the patient's privacy while maintaining the predictive properties of the data. In this study, we implemented a Wasserstein GAN (WGAN) with and without differential privacy guarantees to generate privacy-preserving labeled Time-of-Flight Magnetic Resonance Angiography (TOF-MRA) image patches for brain vessel segmentation. The synthesized image-label pairs were used to train a U-net which was evaluated in terms of the segmentation performance on real patient images from two different datasets. Additionally, the Fréchet Inception Distance (FID) was calculated between the generated images and the real images to assess their similarity. During the evaluation using the U-Net and the FID, we explored the effect of different levels of privacy which was represented by the parameter ϵ. With stricter privacy guarantees, the segmentation performance and the similarity to the real patient images in terms of FID decreased. Our best segmentation model, trained on synthetic and private data, achieved a Dice Similarity Coefficient (DSC) of 0.75 for ϵ = 7.4 compared to 0.84 for ϵ = ∞ in a brain vessel segmentation paradigm (DSC of 0.69 and 0.88 on the second test set, respectively). We identified a threshold of ϵ <5 for which the performance (DSC <0.61) became unstable and not usable. Our synthesized labeled TOF-MRA images with strict privacy guarantees retained predictive properties necessary for segmenting the brain vessels. Although further research is warranted regarding generalizability to other imaging modalities and performance improvement, our results mark an encouraging first step for privacy-preserving data sharing in medical imaging.

Generating 3D TOF-MRA volumes and segmentation labels using generative adversarial networks.

Deep learning requires large labeled datasets that are difficult to gather in medical imaging due to data privacy issues and time-consuming manual labeling. Generative Adversarial Networks (GANs) can alleviate these challenges enabling synthesis of shareable data. While 2D GANs have been used to generate 2D images with their corresponding labels, they cannot capture the volumetric information of 3D medical imaging. 3D GANs are more suitable for this and have been used to generate 3D volumes but not their corresponding labels. One reason might be that synthesizing 3D volumes is challenging owing to computational limitations. In this work, we present 3D GANs for the generation of 3D medical image volumes with corresponding labels applying mixed precision to alleviate computational constraints. We generated 3D Time-of-Flight Magnetic Resonance Angiography (TOF-MRA) patches with their corresponding brain blood vessel segmentation labels. We used four variants of 3D Wasserstein GAN (WGAN) with: 1) gradient penalty (GP), 2) GP with spectral normalization (SN), 3) SN with mixed precision (SN-MP), and 4) SN-MP with double filters per layer (c-SN-MP). The generated patches were quantitatively evaluated using the Fréchet Inception Distance (FID) and Precision and Recall of Distributions (PRD). Further, 3D U-Nets were trained with patch-label pairs from different WGAN models and their performance was compared to the performance of a benchmark U-Net trained on real data. The segmentation performance of all U-Net models was assessed using Dice Similarity Coefficient (DSC) and balanced Average Hausdorff Distance (bAVD) for a) all vessels, and b) intracranial vessels only. Our results show that patches generated with WGAN models using mixed precision (SN-MP and c-SN-MP) yielded the lowest FID scores and the best PRD curves. Among the 3D U-Nets trained with synthetic patch-label pairs, c-SN-MP pairs achieved the highest DSC (0.841) and lowest bAVD (0.508) compared to the benchmark U-Net trained on real data (DSC 0.901; bAVD 0.294) for intracranial vessels. In conclusion, our solution generates realistic 3D TOF-MRA patches and labels for brain vessel segmentation. We demonstrate the benefit of using mixed precision for computational efficiency resulting in the best-performing GAN-architecture. Our work paves the way towards sharing of labeled 3D medical data which would increase generalizability of deep learning models for clinical use.

Image-to-image generative adversarial networks for synthesizing perfusion parameter maps from DSC-MR images in cerebrovascular disease.

Stroke is a major cause of death or disability. As imaging-based patient stratification improves acute stroke therapy, dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) is of major interest in image brain perfusion. However, expert-level perfusion maps require a manual or semi-manual post-processing by a medical expert making the procedure time-consuming and less-standardized. Modern machine learning methods such as generative adversarial networks (GANs) have the potential to automate the perfusion map generation on an expert level without manual validation. We propose a modified pix2pix GAN with a temporal component (temp-pix2pix-GAN) that generates perfusion maps in an end-to-end fashion. We train our model on perfusion maps infused with expert knowledge to encode it into the GANs. The performance was trained and evaluated using the structural similarity index measure (SSIM) on two datasets including patients with acute stroke and the steno-occlusive disease. Our temp-pix2pix architecture showed high performance on the acute stroke dataset for all perfusion maps (mean SSIM 0.92-0.99) and good performance on data including patients with the steno-occlusive disease (mean SSIM 0.84-0.99). While clinical validation is still necessary for future studies, our results mark an important step toward automated expert-level perfusion maps and thus fast patient stratification.

Neuropsychological Assessments of Patients With Acquired Brain Injury: A Cluster Analysis Approach to Address Heterogeneity in Web-Based Cognitive Rehabilitation.

We aimed to (1) apply cluster analysis techniques to mixed-type data (numerical and categorical) from baseline neuropsychological standard and widely used assessments of patients with acquired brain injury (ABI) (2) apply state-of-the-art cluster validity indexes (CVI) to assess their internal validity (3) study their external validity considering relevant aspects of ABI rehabilitation such as functional independence measure (FIM) in activities of daily life assessment (4) characterize the identified profiles by using demographic and clinically relevant variables and (5) extend the external validation of the obtained clusters to all cognitive rehabilitation tasks executed by the participants in a web-based cognitive rehabilitation platform (GNPT). We analyzed 1,107 patients with ABI, 58.1% traumatic brain injury (TBI), 21.8% stroke and 20.1% other ABIs (e.g., brain tumors, anoxia, infections) that have undergone inpatient GNPT cognitive rehabilitation from September 2008 to January 2021. We applied the k-prototypes algorithm from the clustMixType R package. We optimized seven CVIs and applied bootstrap resampling to assess clusters stability (fpc R package). Clusters' post hoc comparisons were performed using the Wilcoxon ranked test, paired t-test or Chi-square test when appropriate. We identified a three-clusters optimal solution, with strong stability (>0.85) and structure (e.g., Silhouette > 0.60, Gamma > 0.83), characterized by distinctive level of performance in all neuropsychological tests, demographics, FIM, response to GNPT tasks and tests normative data (e.g., the 3 min cut-off in Trail Making Test-B). Cluster 1 was characterized by severe cognitive impairment (N = 254, 22.9%) the mean age was 47 years, 68.5% patients with TBI and 22% with stroke. Cluster 2 was characterized by mild cognitive impairment (N = 376, 33.9%) mean age 54 years, 53.5% patients with stroke and 27% other ABI. Cluster 3, moderate cognitive impairment (N = 477, 43.2%) mean age 33 years, 83% patients with TBI and 14% other ABI. Post hoc analysis on cognitive FIM supported a significant higher performance of Cluster 2 vs. Cluster 3 (p < 0.001), Cluster 2 vs. Cluster 1 (p < 0.001) and Cluster 3 vs. Cluster 1 (p < 0.001). All patients executed 286,798 GNPT tasks, with performance significantly higher in Cluster 2 and 3 vs. Cluster 1 (p < 0.001).

Hybrid modelling for stroke care: Review and suggestions of new approaches for risk assessment and simulation of scenarios.

Stroke is an example of a complex and multi-factorial disease involving multiple organs, timescales, and disease mechanisms. To deal with this complexity, and to realize Precision Medicine of stroke, mathematical models are needed. Such approaches include: 1) machine learning, 2) bioinformatic network models, and 3) mechanistic models. Since these three approaches have complementary strengths and weaknesses, a hybrid modelling approach combining them would be the most beneficial. However, no concrete approach ready to be implemented for a specific disease has been presented to date. In this paper, we both review the strengths and weaknesses of the three approaches, and propose a roadmap for hybrid modelling in the case of stroke care. We focus on two main tasks needed for the clinical setting: a) For stroke risk calculation, we propose a new two-step approach, where non-linear mixed effects models and bioinformatic network models yield biomarkers which are used as input to a machine learning model and b) For simulation of care scenarios, we propose a new four-step approach, which revolves around iterations between simulations of the mechanistic models and imputations of non-modelled or non-measured variables. We illustrate and discuss the different approaches in the context of Precision Medicine for stroke.

A precision medicine framework for personalized simulation of hemodynamics in cerebrovascular disease.

BACKGROUND: Cerebrovascular disease, in particular stroke, is a major public health challenge. An important biomarker is cerebral hemodynamics. To measure and quantify cerebral hemodynamics, however, only invasive, potentially harmful or time-to-treatment prolonging methods are available. RESULTS: We present a simulation-based approach which allows calculation of cerebral hemodynamics based on the patient-individual vessel configuration derived from structural vessel imaging. For this, we implemented a framework allowing segmentation and annotation of brain vessels from structural imaging followed by 0-dimensional lumped simulation modeling of cerebral hemodynamics. For annotation, a 3D-graphical user interface was implemented. For 0D-simulation, we used a modified nodal analysis, which was adapted for easy implementation by code. The simulation enables identification of areas vulnerable to stroke and simulation of changes due to different systemic blood pressures. Moreover, sensitivity analysis was implemented allowing the live simulation of changes to simulate procedures and disease progression. Beyond presentation of the framework, we demonstrated in an exploratory analysis in 67 patients that the simulation has a high specificity and low-to-moderate sensitivity to detect perfusion changes in classic perfusion imaging. CONCLUSIONS: The presented precision medicine approach using novel biomarkers has the potential to make the application of harmful and complex perfusion methods obsolete.

Synthesizing anonymized and labeled TOF-MRA patches for brain vessel segmentation using generative adversarial networks.

Anonymization and data sharing are crucial for privacy protection and acquisition of large datasets for medical image analysis. This is a big challenge, especially for neuroimaging. Here, the brain's unique structure allows for re-identification and thus requires non-conventional anonymization. Generative adversarial networks (GANs) have the potential to provide anonymous images while preserving predictive properties. Analyzing brain vessel segmentation, we trained 3 GANs on time-of-flight (TOF) magnetic resonance angiography (MRA) patches for image-label generation: 1) Deep convolutional GAN, 2) Wasserstein-GAN with gradient penalty (WGAN-GP) and 3) WGAN-GP with spectral normalization (WGAN-GP-SN). The generated image-labels from each GAN were used to train a U-net for segmentation and tested on real data. Moreover, we applied our synthetic patches using transfer learning on a second dataset. For an increasing number of up to 15 patients we evaluated the model performance on real data with and without pre-training. The performance for all models was assessed by the Dice Similarity Coefficient (DSC) and the 95th percentile of the Hausdorff Distance (95HD). Comparing the 3 GANs, the U-net trained on synthetic data generated by the WGAN-GP-SN showed the highest performance to predict vessels (DSC/95HD 0.85/30.00) benchmarked by the U-net trained on real data (0.89/26.57). The transfer learning approach showed superior performance for the same GAN compared to no pre-training, especially for one patient only (0.91/24.66 vs. 0.84/27.36). In this work, synthetic image-label pairs retained generalizable information and showed good performance for vessel segmentation. Besides, we showed that synthetic patches can be used in a transfer learning approach with independent data. This paves the way to overcome the challenges of scarce data and anonymization in medical imaging.

Explainability for artificial intelligence in healthcare: a multidisciplinary perspective.

BACKGROUND: Explainability is one of the most heavily debated topics when it comes to the application of artificial intelligence (AI) in healthcare. Even though AI-driven systems have been shown to outperform humans in certain analytical tasks, the lack of explainability continues to spark criticism. Yet, explainability is not a purely technological issue, instead it invokes a host of medical, legal, ethical, and societal questions that require thorough exploration. This paper provides a comprehensive assessment of the role of explainability in medical AI and makes an ethical evaluation of what explainability means for the adoption of AI-driven tools into clinical practice. METHODS: Taking AI-based clinical decision support systems as a case in point, we adopted a multidisciplinary approach to analyze the relevance of explainability for medical AI from the technological, legal, medical, and patient perspectives. Drawing on the findings of this conceptual analysis, we then conducted an ethical assessment using the "Principles of Biomedical Ethics" by Beauchamp and Childress (autonomy, beneficence, nonmaleficence, and justice) as an analytical framework to determine the need for explainability in medical AI. RESULTS: Each of the domains highlights a different set of core considerations and values that are relevant for understanding the role of explainability in clinical practice. From the technological point of view, explainability has to be considered both in terms how it can be achieved and what is beneficial from a development perspective. When looking at the legal perspective we identified informed consent, certification and approval as medical devices, and liability as core touchpoints for explainability. Both the medical and patient perspectives emphasize the importance of considering the interplay between human actors and medical AI. We conclude that omitting explainability in clinical decision support systems poses a threat to core ethical values in medicine and may have detrimental consequences for individual and public health. CONCLUSIONS: To ensure that medical AI lives up to its promises, there is a need to sensitize developers, healthcare professionals, and legislators to the challenges and limitations of opaque algorithms in medical AI and to foster multidisciplinary collaboration moving forward.