Multimodal Deep Learning Improves Recurrence Risk Prediction in Pediatric Low-Grade Gliomas.

BACKGROUND: Postoperative recurrence risk for pediatric low-grade gliomas (pLGGs) is challenging to predict by conventional clinical, radiographic, and genomic factors. We investigated if deep learning of MRI tumor features could improve postoperative pLGG risk stratification. METHODS: We used pre-trained deep learning (DL) tool designed for pLGG segmentation to extract pLGG imaging features from preoperative T2-weighted MRI from patients who underwent surgery (DL-MRI features). Patients were pooled from two institutions: Dana Farber/Boston Children's Hospital (DF/BCH) and the Children's Brain Tumor Network (CBTN). We trained three DL logistic hazard models to predict postoperative event-free survival (EFS) probabilities with 1) clinical features, 2) DL-MRI features, and 3) multimodal (clinical and DL-MRI features). We evaluated the models with a time-dependent Concordance Index (Ctd) and risk group stratification with Kaplan Meier plots and log-rank tests. We developed an automated pipeline integrating pLGG segmentation and EFS prediction with the best model. RESULTS: Of the 396 patients analyzed (median follow-up: 85 months, range: 1.5-329 months), 214 (54%) underwent gross total resection and 110 (28%) recurred. The multimodal model improved EFS prediction compared to the DL-MRI and clinical models (Ctd: 0.85 (95% CI: 0.81-0.93), 0.79 (95% CI: 0.70-0.88), and 0.72 (95% CI: 0.57-0.77), respectively). The multimodal model improved risk-group stratification (3-year EFS for predicted high-risk: 31% versus low-risk: 92%, p<0.0001). CONCLUSIONS: DL extracts imaging features that can inform postoperative recurrence prediction for pLGG. Multimodal DL improves postoperative risk stratification for pLGG and may guide postoperative decision-making. Larger, multicenter training data may be needed to improve model generalizability.

Automated algorithm for medical data structuring, and segmentation using artificial intelligence within secured environment for dataset creation.

Objective: Routinely collected electronic health records using artificial intelligence (AI)-based systems bring out enormous benefits for patients, healthcare centers, and its industries. Artificial intelligence models can be used to structure a wide variety of unstructured data. Methods: We present a semi-automatic workflow for medical dataset management, including data structuring, research extraction, AI-ground truth creation, and updates. The algorithm creates directories based on keywords in new file names. Results: Our work focuses on organizing computed tomography (CT), magnetic resonance (MR) images, patient clinical data, and segmented annotations. In addition, an AI model is used to generate different initial labels that can be edited manually to create ground truth labels. The manually verified ground truth labels are later included in the structured dataset using an automated algorithm for future research. Conclusion: This is a workflow with an AI model trained on local hospital medical data with output based/adapted to the users and their preferences. The automated algorithms and AI model could be implemented inside a secondary secure environment in the hospital to produce inferences.

Intercellular pathways of cancer treatment-related cardiotoxicity and their therapeutic implications: the paradigm of radiotherapy.

Advances in cancer therapeutics have improved patient survival rates. However, cancer survivors may suffer from adverse events either at the time of therapy or later in life. Cardiovascular diseases (CVD) represent a clinically important, but mechanistically understudied complication, which interfere with the continuation of best-possible care, induce life-threatening risks, and/or lead to long-term morbidity. These concerns are exacerbated by the fact that targeted therapies and immunotherapies are frequently combined with radiotherapy, which induces durable inflammatory and immunogenic responses, thereby providing a fertile ground for the development of CVDs. Stressed and dying irradiated cells produce 'danger' signals including, but not limited to, major histocompatibility complexes, cell-adhesion molecules, proinflammatory cytokines, and damage-associated molecular patterns. These factors activate intercellular signaling pathways which have potentially detrimental effects on the heart tissue homeostasis. Herein, we present the clinical crosstalk between cancer and heart diseases, describe how it is potentiated by cancer therapies, and highlight the multifactorial nature of the underlying mechanisms. We particularly focus on radiotherapy, as a case known to often induce cardiovascular complications even decades after treatment. We provide evidence that the secretome of irradiated tumors entails factors that exert systemic, remote effects on the cardiac tissue, potentially predisposing it to CVDs. We suggest how diverse disciplines can utilize pertinent state-of-the-art methods in feasible experimental workflows, to shed light on the molecular mechanisms of radiotherapy-related cardiotoxicity at the organismal level and untangle the desirable immunogenic properties of cancer therapies from their detrimental effects on heart tissue. Results of such highly collaborative efforts hold promise to be translated to next-generation regimens that maximize tumor control, minimize cardiovascular complications, and support quality of life in cancer survivors.

Deep learning-assisted survival prognosis in renal cancer: A CT scan-based personalized approach.

This paper presents a deep learning (DL) approach for predicting survival probabilities of renal cancer patients based solely on preoperative CT imaging. The proposed approach consists of two networks: a classifier- and a survival- network. The classifier attempts to extract features from 3D CT scans to predict the ISUP grade of Renal cell carcinoma (RCC) tumors, as defined by the International Society of Urological Pathology (ISUP). Our classifier is a 3D convolutional neural network to avoid losing crucial information on the interconnection of slides in 3D images. We employ multiple procedures, including image augmentation, preprocessing, and concatenation, to improve the performance of the classifier. Given the strong correlation between ISUP grading and renal cancer prognosis in the clinical context, we use the ISUP grading features extracted by the classifier as the input to the survival network. By leveraging this clinical association and the classifier network, we are able to model our survival analysis using a simple DL-based network. We adopt a discrete LogisticHazard-based loss to extract intrinsic survival characteristics of RCC tumors from CT images. This allows us to build a completely parametric survival model that varies with patients' tumor characteristics and predicts non-proportional survival probability curves for different patients. Our results demonstrated that the proposed method could predict the future course of renal cancer with reasonable accuracy from the CT scans. The proposed method obtained an average concordance index of 0.72, an integrated Brier score of 0.15, and an area under the curve value of 0.71 on the test cohorts.

Multiphysics simulation of magnetoelectric micro core-shells for wireless cellular stimulation therapy via magnetic temporal interference.

This paper presents an innovative approach to wireless cellular stimulation therapy through the design of a magnetoelectric (ME) microdevice. Traditional electrophysiological stimulation techniques for neural and deep brain stimulation face limitations due to their reliance on electronics, electrode arrays, or the complexity of magnetic induction. In contrast, the proposed ME microdevice offers a self-contained, controllable, battery-free, and electronics-free alternative, holding promise for targeted precise stimulation of biological cells and tissues. The designed microdevice integrates core shell ME materials with remote coils which applies magnetic temporal interference (MTI) signals, leading to the generation of a bipolar local electric stimulation current operating at low frequencies which is suitable for precise stimulation. The nonlinear property of the magnetostrictive core enables the demodulation of remotely applied high-frequency electromagnetic fields, resulting in a localized, tunable, and manipulatable electric potential on the piezoelectric shell surface. This potential, triggers electrical spikes in neural cells, facilitating stimulation. Rigorous computational simulations support this concept, highlighting a significantly high ME coupling factor generation of 550 V/m·Oe. The high ME coupling is primarily attributed to the operation of the device in its mechanical resonance modes. This achievement is the result of a carefully designed core shell structure operating at the MTI resonance frequencies, coupled with an optimal magnetic bias, and predetermined piezo shell thickness. These findings underscore the potential of the engineered ME core shell as a candidate for wireless and minimally invasive cellular stimulation therapy, characterized by high resolution and precision. These results open new avenues for injectable material structures capable of delivering effective cellular stimulation therapy, carrying implications across neuroscience medical devices, and regenerative medicine.

Modeling Extracellular Vesicles-Mediated Interactions of Cells in the Tumor Microenvironment.

Interactions of cells via extracellular vesicles (EVs) manipulate various actions, including cancer initiation and progression, inflammation, anti-tumor signaling and cell migration, proliferation and apoptosis in the tumor microenvironment. EVs as the external stimulus can activate or inhibit some receptor pathways in a way that amplify or attenuate a kind of particle release at target cells. This can also be carried out in a biological feedback-loop where the transmitter is affected by the induced release initiated by the target cell due to the EVs received from the donor cell, to create a bilateral process. In this paper, at first we derive the frequency response of internalization function in the framework of a unilateral communication link. This solution is adapted to a closed-loop system to find the frequency response of a bilateral system. The overall releases of the cells, given by the combination of the natural release and the induced release, are reported at the end of this paper and the results are compared in terms of distance between the cells and reaction rates of EVs at the cell membranes.

Optimization of Extracellular Vesicle Release for Targeted Drug Delivery.

Targeted drug delivery is a promising approach for many serious diseases, such as glioblastoma multiforme, one of the most common and devastating brain tumor. In this context, this work addresses the optimization of the controlled release of drugs which are carried by extracellular vesicles. Towards this goal, we derive and numerically verify an analytical solution for the end-to-end system model. We then apply the analytical solution either to reduce the disease treatment time or to reduce the amount of required drugs. The latter is formulated as a bilevel optimization problem, whose quasiconvex/quasiconcave property is proved here. For solving the optimization problem, we propose and utilize a combination of bisection method and golden-section search. The numerical results demonstrate that the optimization can significantly reduce the treatment time and/or the required drugs carried by extracellular vesicles for a therapy compared to the steady state solution.

Multimodal deep learning for personalized renal cell carcinoma prognosis: Integrating CT imaging and clinical data.

BACKGROUND AND OBJECTIVE: Renal cell carcinoma represents a significant global health challenge with a low survival rate. The aim of this research was to devise a comprehensive deep-learning model capable of predicting survival probabilities in patients with renal cell carcinoma by integrating CT imaging and clinical data and addressing the limitations observed in prior studies. The aim is to facilitate the identification of patients requiring urgent treatment. METHODS: The proposed framework comprises three modules: a 3D image feature extractor, clinical variable selection, and survival prediction. Based on the 3D CNN architecture, the feature extractor module predicts the ISUP grade of renal cell carcinoma tumors linked to mortality rates from CT images. Clinical variables are systematically selected using the Spearman score and random forest importance score as criteria. A deep learning-based network, trained with discrete LogisticHazard-based loss, performs the survival prediction. Nine distinct experiments are performed, with varying numbers of clinical variables determined by different thresholds of the Spearman and importance scores. RESULTS: Our findings demonstrate that the proposed strategy surpasses the current literature on renal cancer prognosis based on CT scans and clinical factors. The best-performing experiment yielded a concordance index of 0.84 and an area under the curve value of 0.8 on the test cohort, which suggests strong predictive power. CONCLUSIONS: The multimodal deep-learning approach developed in this study shows promising results in estimating survival probabilities for renal cell carcinoma patients using CT imaging and clinical data. This may have potential implications in identifying patients who require urgent treatment, potentially improving patient outcomes. The code created for this project is available for the public on: GitHub.

Regional Left Ventricular Fiber Stress Analysis for Cardiac Resynchronization Therapy Response.

Cardiac resynchronization therapy (CRT) is an effective treatment for a subgroup of heart failure (HF) patients, but more than 30% of those selected do not improve after CRT implantation. Imperfect pre-procedural criteria for patient selection and optimization are the main causes of the high non-response rate. In this study, we evaluated a novel measure for assessing CRT response. We used a computational modeling framework to calculate the regional stress of the left ventricular wall of seven CRT patients and seven healthy controls. The standard deviation of regional wall stress at the time of mitral valve closure (SD_MVC) was used to quantify dyssynchrony and compared between patients and controls and among the patients. The results show that SD_MVC is significantly lower in controls than patients and correlates with long-term response in patients, based on end-diastolic volume reduction. In contrast to our initial hypothesis, patients with lower SD_MVC respond better to therapy. The patient with the highest SD_MVC was the only non-responder in the patient cohort. The distribution of fiber stress at the beginning of the isovolumetric phase seems to correlate with the degree of response and the use of this measurement could potentially improve selection criteria for CRT implantation. Further studies with a larger cohort of patients are needed to validate these results.

The End-to-End Molecular Communication Model of Extracellular Vesicle-Based Drug Delivery.

A closer look at nature has recently brought more interest in exploring and utilizing intra-body communication networks composed of cells as intrinsic, perfectly biocompatible infrastructures to deliver therapeutics. Naturally occurring cell-to-cell communication systems are being manipulated to release, navigate, and take-up soluble cell-derived messengers that are either therapeutic by nature or carry therapeutic molecular cargo. One example of such structures is extracellular vesicles (EVs) which have been recently proven to have pharmacokinetic properties, opening new avenues for developing the next generation biotherapeutics. In this paper, we study theoretical aspects of the EV transfer within heart tissue as a case study by utilizing an information and communication technology-like approach in analyzing molecular communication systems. Our modeling implies the abstraction of the EV releasing cells as transmitters, the extracellular matrix as the channel, and the EV receiving cells as receivers. Our results, derived from the developed analytical models, indicate that the release can be modulated using external forces such as electrical signals, and the transfer and reception can be affected by the extracellular matrix and plasma membrane properties, respectively.The presented modeling provides initial results for the EV biodistributions and contribute to avoiding unplanned administration, often resulting in side- and adverse effects.

Implants with Sensing Capabilities.

Because of the aging human population and increased numbers of surgical procedures being performed, there is a growing number of biomedical devices being implanted each year. Although the benefits of implants are significant, there are risks to having foreign materials in the body that may lead to complications that may remain undetectable until a time at which the damage done becomes irreversible. To address this challenge, advances in implantable sensors may enable early detection of even minor changes in the implants or the surrounding tissues and provide early cues for intervention. Therefore, integrating sensors with implants will enable real-time monitoring and lead to improvements in implant function. Sensor integration has been mostly applied to cardiovascular, neural, and orthopedic implants, and advances in combined implant-sensor devices have been significant, yet there are needs still to be addressed. Sensor-integrating implants are still in their infancy; however, some have already made it to the clinic. With an interdisciplinary approach, these sensor-integrating devices will become more efficient, providing clear paths to clinical translation in the future.

Brain-on-a-chip: Recent advances in design and techniques for microfluidic models of the brain in health and disease.

Recent advances in biomaterials, microfabrication, microfluidics, and cell biology have led to the development of organ-on-a-chip devices that can reproduce key functions of various organs. Such platforms promise to provide novel insights into various physiological events, including mechanisms of disease, and evaluate the effects of external interventions, such as drug administration. The neuroscience field is expected to benefit greatly from these innovative tools. Conventional ex vivo studies of the nervous system have been limited by the inability of cell culture to adequately mimic in vivo physiology. While animal models can be used, their relevance to human physiology is uncertain and their use is laborious and associated with ethical issues. To date, organ-on-a-chip systems have been developed to model different tissue components of the brain, including brain regions with specific functions and the blood brain barrier, both in normal and pathophysiological conditions. While the field is still in its infancy, it is expected to have major impact on studies of neurophysiology, pathology and neuropharmacology in future. Here, we review advances made and limitations faced in an effort to stimulate development of the next generation of brain-on-a-chip devices.

Robot-assisted magnetic capsule endoscopy; navigating colorectal inclinations.

Purpose: To investigate the interaction of a robot assisted magnetically driven wireless capsule endoscope (WCE) with colonic tissue, as it traverses the colorectal bends in the dorsal and ventral directions, relying only on the feedback from a 3D accelerometer. We also investigate the impact of shell geometry and water insufflation on WCE locomotion.Methods: A 3D printed incline phantom, lined with porcine colon, was used as the experimental platform, for controlled and repeatable results. The tilt angle of WCE was controlled to observe its influence on WCE locomotion. The phantom was placed underwater to observe the effects of water insufflation. The experiments were repeated using the two capsule shell geometries to observe the effect of shell geometry on WCE locomotion.Results: Friction between WCE and intestinal tissue increased when the tilt angle of the WCE was lower than the angle of the incline of the phantom. Increasing the WCE tilt angle to match the angle of the incline reduced this friction. Water insufflation and elliptical capsule shell geometry reduced the friction further.Conclusion: Tilting of the WCE equal to, or more than the angle of the incline improved the WCE locomotion. WCE locomotion was also improved by using elliptical capsule shell geometry and water insufflation.Abbreviations: CRC: colorectal cancer; GI: gastrointestinal; MRI: magnetic resonance imaging; WCE: wireless capsule endoscope.

Targeted Drug Delivery for Cardiovascular Disease: Modeling of Modulated Extracellular Vesicle Release Rates.

Invasive and medical therapy has led to major improvements in cardiovascular disease management, but important challenges remain open. The discovery of a nano-sized system of extracellular vesicles (EVs) is opening new possibilities for reprogramming malfunctioning cells and indicates that EVs can be employed in therapeutic biomedical applications as engineered drug vehicles. Molecular communication (MC) has applications for treating cells with directed drug delivery, employing special targeting transmembrane proteins. In this paper, we propose a novel drug delivery system for cardiovascular diseases using an EV-mediated MC platform and exemplify the potential use in hypertrophic cardiomyopathy. We utilize intracellular calcium signaling as a natural mediator of EVs released from synthetic cells and model the release rate. We propose to use the cells as a therapeutic release system with a control signal input which modulates the EVs release rate as the output signal. We also study the frequency domain of the proposed model and estimate the transfer function of the therapeutic release system model numerically where the root-mean-square error for two separate estimated output signals are 0.0353 and 0.0124. The proposed EV-mediated targeted drug delivery system can make breakthroughs in future healthcare, in cardiovascular and other diseases where targeting is required.

Visualizing Extracellular Vesicles and Their Function in 3D Tumor Microenvironment Models.

Extracellular vesicles (EVs) are cell-derived nanostructures that mediate intercellular communication by delivering complex signals in normal tissues and cancer. The cellular coordination required for tumor development and maintenance is mediated, in part, through EV transport of molecular cargo to resident and distant cells. Most studies on EV-mediated signaling have been performed in two-dimensional (2D) monolayer cell cultures, largely because of their simplicity and high-throughput screening capacity. Three-dimensional (3D) cell cultures can be used to study cell-to-cell and cell-to-matrix interactions, enabling the study of EV-mediated cellular communication. 3D cultures may best model the role of EVs in formation of the tumor microenvironment (TME) and cancer cell-stromal interactions that sustain tumor growth. In this review, we discuss EV biology in 3D culture correlates of the TME. This includes EV communication between cell types of the TME, differences in EV biogenesis and signaling associated with differing scaffold choices and in scaffold-free 3D cultures and cultivation of the premetastatic niche. An understanding of EV biogenesis and signaling within a 3D TME will improve culture correlates of oncogenesis, enable molecular control of the TME and aid development of drug delivery tools based on EV-mediated signaling.

Molecular Communications in Viral Infections Research: Modeling, Experimental Data, and Future Directions.

Hundreds of millions of people worldwide are affected by viral infections each year, and yet, several of them neither have vaccines nor effective treatment during and post-infection. This challenge has been highlighted by the COVID-19 pandemic, showing how viruses can quickly spread and impact society as a whole. Novel interdisciplinary techniques must emerge to provide forward-looking strategies to combat viral infections, as well as possible future pandemics. In the past decade, an interdisciplinary area involving bioengineering, nanotechnology and information and communication technology (ICT) has been developed, known as Molecular Communications. This new emerging area uses elements of classical communication systems to molecular signalling and communication found inside and outside biological systems, characterizing the signalling processes between cells and viruses. In this paper, we provide an extensive and detailed discussion on how molecular communications can be integrated into the viral infectious diseases research, and how possible treatment and vaccines can be developed considering molecules as information carriers. We provide a literature review on molecular communications models for viral infection (intra-body and extra-body), a deep analysis on their effects on immune response, how experimental can be used by the molecular communications community, as well as open issues and future directions.

On Mathematical Analysis of Active Drug Transport Coupled With Flow-Induced Diffusion in Blood Vessels.

Blood vessels are flow-induced diffusive molecular channels equipped with transport mechanisms across their walls for conveying substances between the organs in the body. Mathematical modeling of the blood vessel as a molecular transport channel can be used for the characterization of the underlying processes and higher-level functions in the circulatory system. Besides, the mathematical model can be utilized for designing and realizing nano-scale molecular communication systems for healthcare applications including drug delivery systems. In this paper, a continuous-time Markov chain framework is proposed to simply model active transport mechanisms e.g. transcytosis, across the single-layered endothelial cells building the inner vessel wall. Correspondingly, a general homogeneous boundary condition over the vessel wall is introduced. Coupled with the derived boundary condition, the flow-induced diffusion problem in an ideal vessel structure with a cylindrical shape is accurately formulated which takes into account variation in all three dimensions. The corresponding concentration Green's function is analytically derived in terms of a convergent infinite series. Particle-based simulation results confirm the proposed analysis. Also, the effects of system parameters on the concentration Green's function are examined.

Optimal pacing sites in cardiac resynchronization by left ventricular activation front analysis.

Cardiac resynchronization therapy (CRT) can substantially improve dyssynchronous heart failure and reduce mortality. However, about one-third of patients who are implanted, derive no measurable benefit from CRT. Non-response may partly be due to suboptimal activation of the left ventricle (LV) caused by electrophysiological heterogeneities. The goal of this study is to investigate the performance of a newly developed method used to analyze electrical wavefront propagation in a heart model including myocardial scar and compare this to clinical benchmark studies. We used computational models to measure the maximum activation front (MAF) in the LV during different pacing scenarios. Different heart geometries and scars were created based on cardiac MR images of three patients. The right ventricle (RV) was paced from the apex and the LV was paced from 12 different sites, single site, dual-site and triple site. Our results showed that for single LV site pacing, the pacing site with the largest MAF corresponded with the latest activated regions of the LV demonstrated during RV pacing, which also agrees with previous markers used for predicting optimal single-site pacing location. We then demonstrated the utility of MAF in predicting optimal electrode placements in more complex scenarios including scar and multi-site LV pacing. This study demonstrates the potential value of computational simulations in understanding and planning CRT.

Toward real-time polyp detection using fully CNNs for 2D Gaussian shapes prediction.

To decrease colon polyp miss-rate during colonoscopy, a real-time detection system with high accuracy is needed. Recently, there have been many efforts to develop models for real-time polyp detection, but work is still required to develop real-time detection algorithms with reliable results. We use single-shot feed-forward fully convolutional neural networks (F-CNN) to develop an accurate real-time polyp detection system. F-CNNs are usually trained on binary masks for object segmentation. We propose the use of 2D Gaussian masks instead of binary masks to enable these models to detect different types of polyps more effectively and efficiently and reduce the number of false positives. The experimental results showed that the proposed 2D Gaussian masks are efficient for detection of flat and small polyps with unclear boundaries between background and polyp parts. The masks make a better training effect to discriminate polyps from the polyp-like false positives. The proposed method achieved state-of-the-art results on two polyp datasets. On the ETIS-LARIB dataset we achieved 86.54% recall, 86.12% precision, and 86.33% F1-score, and on the CVC-ColonDB we achieved 91% recall, 88.35% precision, and F1-score 89.65%.

Theoretical Aspects of Resting-State Cardiomyocyte Communication for Multi-Nodal Nano-Actuator Pacemakers.

The heart consists of billions of cardiac muscle cells-cardiomyocytes-that work in a coordinated fashion to supply oxygen and nutrients to the body. Inter-connected specialized cardiomyocytes form signaling channels through which the electrical signals are propagated throughout the heart, controlling the heart's beat to beat function of the other cardiac cells. In this paper, we study to what extent it is possible to use ordinary cardiomyocytes as communication channels between components of a recently proposed multi-nodal leadless pacemaker, to transmit data encoded in subthreshold membrane potentials. We analyze signal propagation in the cardiac infrastructure considering noise in the communication channel by performing numerical simulations based on the Luo-Rudy computational model. The Luo-Rudy model is an action potential model but describes the potential changes with time including membrane potential and action potential stages, separated by the thresholding mechanism. Demonstrating system performance, we show that cardiomyocytes can be used to establish an artificial communication system where data are reliably transmitted between 10 s of cells. The proposed subthreshold cardiac communication lays the foundation for a new intra-cardiac communication technique.