Towards Clinical Application of Artificial Intelligence in Ultrasound Imaging

Artificial intelligence real-time automated recognition of the gastric antrum cross-sectional area and motility rhythm via bedside ultrasound: a pilot study

The cross-sectional area (CSA) of the gastric antrum and its motility rhythm reflects the gastrointestinal function of critically ill patients. Monitoring the CSA and motility rhythm is crucial but remains time-consuming and operator dependent. This study aimed to develop an artificial intelligence (AI) system for real-time automated recognition of the gastric antrum CSA and motility rhythm using bedside ultrasound. Gastric antrum ultrasound videos were prospectively collected from West China Hospital to establish training and validation datasets. The AI system's predictions were validated against senior clinicians' annotations to assess accuracy. Additionally, videos were collected to evaluate the performance of the AI system. The antrum motility rhythms of patients and volunteers were preliminarily classified to lay the foundation for the subsequent establishment of gastrointestinal motility rhythm phenotypes in critically ill patients. A total of 907 videos (620 patients and 287 volunteers) were included to develop and validate the AI system from January 2022 to November 2023. 49,240 images were used as training datasets to train the model's ability to locate and segment gastric antrum ultrasound images. The remaining 12,309 images were used as the internal validation dataset, achieving a mean dice coefficient (mDice) of 87.36% and an mean intersection over union (mIOU) of 77.56%. For the external validation dataset, 2334 images were used, resulting in mDice and mIOU values of 86.82% and 76.26%, respectively. Moreover, the AI system demonstrated robust performance in video cut frame analysis, achieving a mDice of 90.23% and a mIOU of 85.16% across 105 videos. The intraclass correlation coefficient (ICC) between human operators and the AI model was good (ICC (2, K): 0.813, 95% CI 0.728-0.871). In terms of antrum motility rhythm phenotypes, we identified several distinct patterns, such as regular movement, minimal movement, and irregular movement, reflecting different statuses, such as fasting, postmeal, postexercise, and postduty. We developed an AI system that is comparable to experienced clinicians in identifying the gastric antrum and measuring its CSA. Furthermore, the system can generate a curve representing the rhythm of antrum movement, reflecting the varying statuses of patients and volunteers. This system may optimize enteral nutrition (EN) protocols by reducing clinicians' workload and minimizing operator dependence.

Overcoming barriers in the use of artificial intelligence in point of care ultrasound

Point-of-care ultrasound is a portable, low-cost imaging technology focused on answering specific clinical questions in real time. Artificial intelligence amplifies its capabilities by aiding clinicians in the acquisition and interpretation of the images; however, there are growing concerns on its effectiveness and trustworthiness. Here, we address key issues such as population bias, explainability and training of artificial intelligence in this field and propose approaches to ensure clinical effectiveness.

Diagnosis Test Accuracy of Artificial Intelligence for Endometrial Cancer: Systematic Review and Meta-Analysis

BACKGROUND

Endometrial cancer is one of the most common gynecological tumors, and early screening and diagnosis are crucial for its treatment. Research on the application of artificial intelligence (AI) in the diagnosis of endometrial cancer is increasing, but there is currently no comprehensive meta-analysis to evaluate the diagnostic accuracy of AI in screening for endometrial cancer.

OBJECTIVE

This paper presents a systematic review of AI-based endometrial cancer screening, which is needed to clarify its diagnostic accuracy and provide evidence for the application of AI technology in screening for endometrial cancer.

METHODS

A search was conducted across PubMed, Embase, Cochrane Library, Web of Science, and Scopus databases to include studies published in English, which evaluated the performance of AI in endometrial cancer screening. A total of 2 independent reviewers screened the titles and abstracts, and the quality of the selected studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool. The certainty of the diagnostic test evidence was evaluated using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system.

RESULTS

A total of 13 studies were included, and the hierarchical summary receiver operating characteristic model used for the meta-analysis showed that the overall sensitivity of AI-based endometrial cancer screening was 86% (95% CI 79%-90%) and specificity was 92% (95% CI 87%-95%). Subgroup analysis revealed similar results across AI type, study region, publication year, and study type, but the overall quality of evidence was low.

CONCLUSIONS

AI-based endometrial cancer screening can effectively detect patients with endometrial cancer, but large-scale population studies are needed in the future to further clarify the diagnostic accuracy of AI in screening for endometrial cancer.

TRIAL REGISTRATION

PROSPERO CRD42024519835; https://www.crd.york.ac.uk/PROSPERO/view/CRD42024519835.

Anxiety about artificial intelligence from patient and doctor-physician

OBJECTIVE

This paper investigates the anxiety surrounding the integration of artificial intelligence (AI) in doctor-patient interactions, analyzing the perspectives of both patients and healthcare providers to identify key concerns and potential solutions.

METHODS

The study employs a comprehensive literature review, examining existing research on AI in healthcare, and synthesizes findings from various surveys and studies that explore the attitudes of patients and doctors towards AI applications in medical settings.

RESULTS

The analysis reveals that patient anxiety encompasses algorithm aversion, robophobia, lack of humanistic care, challenges in human-machine interaction, and concerns about AI's universal applicability. Doctors' anxieties stem from fears of replacement, legal liabilities, emotional impacts of work environment changes, and technological apprehension. The paper highlights the need for patient participation, humanistic care, improved interaction methods, educational training, and policy guidelines to foster public understanding and trust in AI.

CONCLUSION

The paper concludes that addressing AI anxiety in doctor-patient relationships is crucial for successfully integrating AI in healthcare. It emphasizes the importance of respecting patient autonomy, addressing the lack of humanistic care, and improving patient-AI interaction to enhance the patient experience and reduce medical errors.

PRACTICE IMPLICATIONS

The study suggests that future research should focus on understanding the needs and concerns of patients and doctors, strengthening medical humanities education, and establishing policies to guide the ethical use of AI in medicine. It also recommends public education to enhance understanding and trust in AI to improve medical services and ensure professional development and stable work environment for doctors.

Intra- and interobserver agreement of proposed objective transvaginal ultrasound image-quality scoring system for use in artificial intelligence algorithm development

OBJECTIVES

The development of valuable artificial intelligence (AI) tools to assist with ultrasound diagnosis depends on algorithms developed using high-quality data. This study aimed to test the intra- and interobserver agreement of a proposed image-quality scoring system to quantify the quality of gynecological transvaginal ultrasound (TVS) images, which could be used in clinical practice and AI tool development.

METHODS

A proposed scoring system to quantify TVS image quality was created following a review of the literature. This system involved a score of 1-4 (2 = poor, 3 = suboptimal and 4 = optimal image quality) assigned by a rater for individual ultrasound images. If the image was deemed inaccurate, it was assigned a score of 1, corresponding to 'reject'. Six professionals, including two radiologists, two sonographers and two sonologists, reviewed 150 images (50 images of the uterus and 100 images of the ovaries) obtained from 50 women, assigning each image a score of 1-4. The review of all images was repeated a second time by each rater after a period of at least 1 week. Mean scores were calculated for each rater. Overall interobserver agreement was assessed using intraclass correlation coefficient (ICC), and interobserver agreement between paired professionals and intraobserver agreement for all professionals were assessed using weighted Cohen's kappa and ICC.

RESULTS

Poor levels of interobserver agreement were obtained between the six raters for all 150 images (ICC, 0.480 (95% CI, 0.363-0.586)), as well as for assessment of the uterine images only (ICC, 0.359 (95% CI, 0.204-0.523)). Moderate agreement was achieved for the ovarian images (ICC, 0.531 (95% CI, 0.417-0.636)). Agreement between the paired sonographers and sonologists was poor for all images (ICC, 0.336 (95% CI, -0.078 to 0.619) and 0.425 (95% CI, 0.014-0.665), respectively), as well as when images were grouped into uterine images (ICC, 0.253 (95% CI, -0.097 to 0.577) and 0.299 (95% CI, -0.094 to 0.606), respectively) and ovarian images (ICC, 0.400 (95% CI, -0.043 to 0.669) and 0.469 (95% CI, 0.088-0.689), respectively). Agreement between the paired radiologists was moderate for all images (ICC, 0.600 (95% CI, 0.487-0.693)) and for their assessment of uterine images (ICC, 0.538 (95% CI, 0.311-0.707)) and ovarian images (ICC, 0.621 (95% CI, 0.483-0.728)). Weak-to-moderate intraobserver agreement was seen for each of the raters with weighted Cohen's kappa ranging from 0.533 to 0.718 for all images and from 0.467 to 0.751 for ovarian images. Similarly, for all raters, the ICC indicated moderate-to-good intraobserver agreement for all images overall (ICC ranged from 0.636 to 0.825) and for ovarian images (ICC ranged from 0.596 to 0.862). Slightly better intraobserver agreement was seen for uterine images, with weighted Cohen's kappa ranging from 0.568 to 0.808 indicating weak-to-strong agreement, and ICC ranging from 0.546 to 0.893 indicating moderate-to-good agreement. All measures were statistically significant (P < 0.001).

CONCLUSION

The proposed image quality scoring system was shown to have poor-to-moderate interobserver agreement and mostly weak-to-moderate levels of intraobserver agreement. More refinement of the scoring system may be needed to improve agreement, although it remains unclear whether quantification of image quality can be achieved, given the highly subjective nature of ultrasound interpretation. Although some AI systems can tolerate labeling noise, most will favor clean (high-quality) data. As such, innovative data-labeling strategies are needed. © 2025 The Author(s). Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Clinical Application of Artificial Intelligence in Ultrasound Imaging for Oncology

Ultrasound (US) imaging is a widely used tool in oncology because of its noninvasiveness and real-time performance. However, its diagnostic accuracy can be limited by the skills of the examiner when performing manual scanning and by the presence of acoustic shadows that degrade image quality. Artificial intelligence (AI) technologies can support examiners in cancer screening and diagnosis by addressing these limitations. Here, we examine recent advances in AI research and development for US imaging in oncology. Breast cancer has been the most extensively studied cancer, with research predominantly focusing on tumor detection, differentiation between benign and malignant lesions, and prediction of lymph node metastasis. The American College of Radiology developed a medical imaging reporting and data system for various cancers that is often used to evaluate the accuracy of AI models. We will also explore the application of AI in clinical settings for US imaging in oncology. Despite progress, the number of approved AI-equipped software as medical devices for US imaging remains limited in Japan, the United States, and Europe. Practical issues that need to be addressed for clinical application include domain shifts, black boxes, and acoustic shadows. To address these issues, advances in image quality control, AI explainability, and preprocessing of acoustic shadows are essential.

Automated Assessment of the Pulmonary Artery-to-Ascending Aorta Ratio in Fetal Cardiac Ultrasound Screening Using Artificial Intelligence

The three-vessel view (3VV) is a standardized transverse scanning plane used in fetal cardiac ultrasound screening to measure the absolute and relative diameters of the pulmonary artery (PA), ascending aorta (Ao), and superior vena cava, as required. The PA/Ao ratio is used to support the diagnosis of congenital heart disease (CHD). However, vascular diameters are measured manually by examiners, which causes intra- and interobserver variability in clinical practice. In the present study, we aimed to develop an artificial intelligence-based method for the standardized and quantitative evaluation of 3VV. In total, 315 cases and 20 examiners were included in this study. We used the object-detection software YOLOv7 for the automated extraction of 3VV images and compared three segmentation algorithms: DeepLabv3+, UNet3+, and SegFormer. Using the PA/Ao ratios based on vascular segmentation, YOLOv7 plus UNet3+ yielded the most appropriate classification for normal fetuses and those with CHD. Furthermore, YOLOv7 plus UNet3+ achieved an arithmetic mean value of 0.883 for the area under the receiver operating characteristic curve, which was higher than 0.749 for residents and 0.808 for fellows. Our automated method may support unskilled examiners in performing quantitative and objective assessments of 3VV images during fetal cardiac ultrasound screening.

Artificial Intelligence in Human Reproduction

The use of artificial intelligence (AI) in human reproduction is a rapidly evolving field with both exciting possibilities and ethical considerations. This technology has the potential to improve success rates and reduce the emotional and financial burden of infertility. However, it also raises ethical and privacy concerns. This paper presents an overview of the current and potential applications of AI in human reproduction. It explores the use of AI in various aspects of reproductive medicine, including fertility tracking, assisted reproductive technologies, management of pregnancy complications, and laboratory automation. In addition, we discuss the need for robust ethical frameworks and regulations to ensure the responsible and equitable use of AI in reproductive medicine.

Enhancing Lung Ultrasound Diagnostics: A Clinical Study on an Artificial Intelligence Tool for the Detection and Quantification of A-Lines and B-Lines

Background/Objective: A-lines and B-lines are key ultrasound markers that differentiate normal from abnormal lung conditions. A-lines are horizontal lines usually seen in normal aerated lungs, while B-lines are linear vertical artifacts associated with lung abnormalities such as pulmonary edema, infection, and COVID-19, where a higher number of B-lines indicates more severe pathology. This paper aimed to evaluate the effectiveness of a newly released lung ultrasound AI tool (ExoLungAI) in the detection of A-lines and quantification/detection of B-lines to help clinicians in assessing pulmonary conditions. Methods: The algorithm is evaluated on 692 lung ultrasound scans collected from 48 patients (65% males, aged: 55 ± 12.9) following their admission to an Intensive Care Unit (ICU) for COVID-19 symptoms, including respiratory failure, pneumonia, and other complications. Results: ExoLungAI achieved a sensitivity of 91% and specificity of 81% for A-line detection. For B-line detection, it attained a sensitivity of 84% and specificity of 86%. In quantifying B-lines, the algorithm achieved a weighted kappa score of 0.77 (95% CI 0.74 to 0.80) and an ICC of 0.87 (95% CI 0.85 to 0.89), showing substantial agreement between the ground truth and predicted B-line counts. Conclusions: ExoLungAI demonstrates a reliable performance in A-line detection and B-line detection/quantification. This automated tool has greater objectivity, consistency, and efficiency compared to manual methods. Many healthcare professionals including intensivists, radiologists, sonographers, medical trainers, and nurse practitioners can benefit from such a tool, as it assists the diagnostic capabilities of lung ultrasound and delivers rapid responses.

Space Radiology: Emerging Nonsonographic Medical Imaging Techniques and the Potential Applications for Human Spaceflight

Space medicine is a multidisciplinary field that requires the integration of medical imaging techniques and expertise in diagnosing and treating a wide range of acute and chronic conditions to maintain astronaut health. Medical imaging within this domain has been viewed historically through the lens of inflight point-of-care ultrasound and predominantly research uses of cross-sectional imaging before and after flight. However, space radiology, a subfield defined here as the applications of imaging before, during, and after spaceflight, will grow to necessitate the involvement of more advanced imaging techniques and subspecialist expertise as missions increase in length and complexity. While the performance of imaging in spaceflight is limited by equipment mass and volume, power supply, radiation exposure, communication delays, and personnel training, recent developments in nonsonographic modalities have opened the door to their potential for in-mission use. Additionally, improved exam protocols and scanner technology in combination with artificial intelligence algorithms have greatly advanced the utility of possible pre- and postflight studies. This article reviews the past and present of space radiology and discusses possible use cases, knowledge gaps, and future research directions for radiography, fluoroscopy, computed tomography, and magnetic resonance imaging within space medicine, including both the performance of new exam types for new indications and the increased extraction of information from exams already routinely obtained. Through thoughtfully augmenting the use of these tools, medical mission risk may be reduced substantially through preflight screening, inflight diagnosis and management, and inflight and postflight surveillance.

Brain-computer interfaces inspired spiking neural network model for depression stage identification

BACKGROUND

Depression is a global mental disorder, and traditional diagnostic methods mainly rely on scales and subjective evaluations by doctors, which cannot effectively identify symptoms and even carry the risk of misdiagnosis. Brain-Computer Interfaces inspired deep learning-assisted diagnosis based on physiological signals holds promise for improving traditional methods lacking physiological basis and leads next generation neuro-technologies. However, traditional deep learning methods rely on immense computational power and mostly involve end-to-end network learning. These learning methods also lack physiological interpretability, limiting their clinical application in assisted diagnosis.

METHODOLOGY

A brain-like learning model for diagnosing depression using electroencephalogram (EEG) is proposed. The study collects EEG data using 128-channel electrodes, producing a 128×128 brain adjacency matrix. Given the assumption of undirected connectivity, the upper half of the 128×128 matrix is chosen in order to minimise the input parameter size, producing 8,128-dimensional data. After eliminating 28 components derived from irrelevant or reference electrodes, a 90×90 matrix is produced, which can be used as an input for a single-channel brain-computer interface image.

RESULT

At the functional level, a spiking neural network is constructed to classify individuals with depression and healthy individuals, achieving an accuracy exceeding 97.5 %.

COMPARISON WITH EXISTING METHODS

Compared to deep convolutional methods, the spiking method reduces energy consumption.

CONCLUSION

At the structural level, complex networks are utilized to establish spatial topology of brain connections and analyse their graph features, identifying potential abnormal brain functional connections in individuals with depression.

Handheld Ultrasound Bladder Volume Assessment Compared to Standard Technique

Urinary retention is a common complaint encountered in the emergency department (ED). Current tools for the assessment of urinary retention are either bladder volume estimation with a bladder scanner performed by nursing staff or direct visualization and measurement via bedside ultrasound performed by an emergency physician. Newer handheld ultrasound devices such as the Butterfly iQ have been brought to the market to bring ultrasound more conveniently to the bedside. A recently released handheld auto-calculation tool produces a 3D image of the bladder and instant bladder volume measurement in milliliters. However, there is a paucity of data assessing the validity of the new Butterfly iQ at the bedside. This study sought to compare the diagnostic accuracy and rated user convenience of the nursing bladder scanner, the cart-based ultrasound machine, and the Butterfly iQ auto-bladder volume tool. ED patients were prospectively enrolled and underwent bladder measurements in a randomized, pre-determined order with each modality. Measurements were subsequently compared to the gold standard of catheterization. Cart-based ultrasound had the highest agreement to catheterization when compared to the RN scanner and the Butterfly iQ. However, the Butterfly iQ and RN scanner were both considered more convenient measurement modalities than the cart-based ultrasound. The Butterfly iQ serves as a cost-effective alternative to cart-based ultrasound while providing greater general utility compared to bladder scanners.

Deep learning models for interpretation of point of care ultrasound in military working dogs

Introduction

Military working dogs (MWDs) are essential for military operations in a wide range of missions. With this pivotal role, MWDs can become casualties requiring specialized veterinary care that may not always be available far forward on the battlefield. Some injuries such as pneumothorax, hemothorax, or abdominal hemorrhage can be diagnosed using point of care ultrasound (POCUS) such as the Global FAST® exam. This presents a unique opportunity for artificial intelligence (AI) to aid in the interpretation of ultrasound images. In this article, deep learning classification neural networks were developed for POCUS assessment in MWDs.

Methods

Images were collected in five MWDs under general anesthesia or deep sedation for all scan points in the Global FAST® exam. For representative injuries, a cadaver model was used from which positive and negative injury images were captured. A total of 327 ultrasound clips were captured and split across scan points for training three different AI network architectures: MobileNetV2, DarkNet-19, and ShrapML. Gradient class activation mapping (GradCAM) overlays were generated for representative images to better explain AI predictions.

Results

Performance of AI models reached over 82% accuracy for all scan points. The model with the highest performance was trained with the MobileNetV2 network for the cystocolic scan point achieving 99.8% accuracy. Across all trained networks the diaphragmatic hepatorenal scan point had the best overall performance. However, GradCAM overlays showed that the models with highest accuracy, like MobileNetV2, were not always identifying relevant features. Conversely, the GradCAM heatmaps for ShrapML show general agreement with regions most indicative of fluid accumulation.

Discussion

Overall, the AI models developed can automate POCUS predictions in MWDs. Preliminarily, ShrapML had the strongest performance and prediction rate paired with accurately tracking fluid accumulation sites, making it the most suitable option for eventual real-time deployment with ultrasound systems. Further integration of this technology with imaging technologies will expand use of POCUS-based triage of MWDs.

How automated techniques ease functional assessment of the fetal heart: Applicability of two-dimensional speckle-tracking echocardiography for comprehensive analysis of global and segmental cardiac deformation using fetalHQ®

BACKGROUND

Prenatal echocardiographic assessment of fetal cardiac function has become increasingly important. Fetal two-dimensional speckle-tracking echocardiography (2D-STE) allows the determination of global and segmental functional cardiac parameters. Prenatal diagnostics is relying increasingly on artificial intelligence, whose algorithms transform the way clinicians use ultrasound in their daily workflow. The purpose of this study was to demonstrate the feasibility of whether less experienced operators can handle and might benefit from an automated tool of 2D-STE in the clinical routine.

METHODS

A total of 136 unselected, normal, singleton, second- and third-trimester fetuses with normofrequent heart rates were examined by targeted ultrasound. 2D-STE was performed separately by beginner and expert semiautomatically using a GE Voluson E10 (FetalHQ®, GE Healthcare, Chicago, IL). Several fetal cardiac parameters were calculated (end-diastolic diameter [ED], sphericity index [SI], global longitudinal strain [EndoGLS], fractional shortening [FS]) and assigned to gestational age (GA). Bland-Altman plots were used to test agreement between both operators.

RESULTS

The mean maternal age was 33 years, and the mean maternal body mass index prior to pregnancy was 24.78 kg/m2. The GA ranged from 16.4 to 32.0 weeks (average 22.9 weeks). Averaged endoGLS value of the beginner was -18.57% ± 6.59 percentage points (pp) for the right and -19.58% ± 5.63 pp for the left ventricle, that of the expert -14.33% ± 4.88 pp and -16.37% ± 5.42 pp. With increasing GA, right ventricular endoGLS decreased slightly while the left ventricular was almost constant. The statistical analysis for endoGLS showed a Bland-Altman-Bias of -4.24 pp ± 8.06 pp for the right and -3.21 pp ± 7.11 pp for the left ventricle. The Bland-Altman-Bias of the ED in both ventricles in all analyzed segments ranged from -.49 mm ± 1.54 mm to -.10 mm ± 1.28 mm, that for FS from -.33 pp ± 11.82 pp to 3.91 pp ± 15.56 pp and that for SI from -.38 ± .68 to -.15 ± .45.

CONCLUSIONS

Between both operators, our data indicated that 2D-STE analysis showed excellent agreement for cardiac morphometry parameters (ED and SI), and good agreement for cardiac function parameters (EndoGLS and FS). Due to its complexity, the application of fetal 2D-STE remains the domain of scientific-academic perinatal ultrasound and should be placed preferably in the hands of skilled operators. At present, from our perspective, an implementation into clinical practice "on-the-fly" cannot be recommended.

The Promise of Artificial Intelligence-Assisted Point-of-Care Ultrasonography in Perioperative Care

The role of point-of-care ultrasonography in the perioperative setting has expanded rapidly over recent years. Revolutionizing this technology further is integrating artificial intelligence to assist clinicians in optimizing images, identifying anomalies, performing automated measurements and calculations, and facilitating diagnoses. Artificial intelligence can increase point-of-care ultrasonography efficiency and accuracy, making it an even more valuable point-of-care tool. Given this topic's importance and ever-changing landscape, this review discusses the latest trends to serve as an introduction and update in this area.

Evaluation of Deep Learning Model Architectures for Point-of-Care Ultrasound Diagnostics

Point-of-care ultrasound imaging is a critical tool for patient triage during trauma for diagnosing injuries and prioritizing limited medical evacuation resources. Specifically, an eFAST exam evaluates if there are free fluids in the chest or abdomen but this is only possible if ultrasound scans can be accurately interpreted, a challenge in the pre-hospital setting. In this effort, we evaluated the use of artificial intelligent eFAST image interpretation models. Widely used deep learning model architectures were evaluated as well as Bayesian models optimized for six different diagnostic models: pneumothorax (i) B- or (ii) M-mode, hemothorax (iii) B- or (iv) M-mode, (v) pelvic or bladder abdominal hemorrhage and (vi) right upper quadrant abdominal hemorrhage. Models were trained using images captured in 27 swine. Using a leave-one-subject-out training approach, the MobileNetV2 and DarkNet53 models surpassed 85% accuracy for each M-mode scan site. The different B-mode models performed worse with accuracies between 68% and 74% except for the pelvic hemorrhage model, which only reached 62% accuracy for all model architectures. These results highlight which eFAST scan sites can be easily automated with image interpretation models, while other scan sites, such as the bladder hemorrhage model, will require more robust model development or data augmentation to improve performance. With these additional improvements, the skill threshold for ultrasound-based triage can be reduced, thus expanding its utility in the pre-hospital setting.

Implementing a deep learning model for automatic tongue tumour segmentation in ex-vivo 3-dimensional ultrasound volumes

Three-dimensional (3D) ultrasound can assess the margins of resected tongue carcinoma during surgery. Manual segmentation (MS) is time-consuming, labour-intensive, and subject to operator variability. This study aims to investigate use of a 3D deep learning model for fast intraoperative segmentation of tongue carcinoma in 3D ultrasound volumes. Additionally, it investigates the clinical effect of automatic segmentation. A 3D No New U-Net (nnUNet) was trained on 113 manually annotated ultrasound volumes of resected tongue carcinoma. The model was implemented on a mobile workstation and clinically validated on 16 prospectively included tongue carcinoma patients. Different prediction settings were investigated. Automatic segmentations with multiple islands were adjusted by selecting the best-representing island. The final margin status (FMS) based on automatic, semi-automatic, and manual segmentation was computed and compared with the histopathological margin. The standard 3D nnUNet resulted in the best-performing automatic segmentation with a mean (SD) Dice volumetric score of 0.65 (0.30), Dice surface score of 0.73 (0.26), average surface distance of 0.44 (0.61) mm, Hausdorff distance of 6.65 (8.84) mm, and prediction time of 8 seconds. FMS based on automatic segmentation had a low correlation with histopathology (r = 0.12, p = 0.67); MS resulted in a moderate but insignificant correlation with histopathology (r = 0.4, p = 0.12, n = 16). Implementing the 3D nnUNet yielded fast, automatic segmentation of tongue carcinoma in 3D ultrasound volumes. Correlation between FMS and histopathology obtained from these segmentations was lower than the moderate correlation between MS and histopathology.

AI supported fetal echocardiography with quality assessment

This study aimed to develop a deep learning model to assess the quality of fetal echocardiography and to perform prospective clinical validation. The model was trained on data from the 18-22-week anomaly scan conducted in seven hospitals from 2008 to 2018. Prospective validation involved 100 patients from two hospitals. A total of 5363 images from 2551 pregnancies were used for training and validation. The model's segmentation accuracy depended on image quality measured by a quality score (QS). It achieved an overall average accuracy of 0.91 (SD 0.09) across the test set, with images having above-average QS scoring 0.97 (SD 0.03). During prospective validation of 192 images, clinicians rated 44.8% (SD 9.8) of images as equal in quality, 18.69% (SD 5.7) favoring auto-captured images and 36.51% (SD 9.0) preferring manually captured ones. Images with above average QS showed better agreement on segmentations (p < 0.001) and QS (p < 0.001) with fetal medicine experts. Auto-capture saved additional planes beyond protocol requirements, resulting in more comprehensive echocardiographies. Low QS had adverse effect on both model performance and clinician's agreement with model feedback. The findings highlight the importance of developing and evaluating AI models based on 'noisy' real-life data rather than pursuing the highest accuracy possible with retrospective academic-grade data.

Barriers and facilitators of artificial intelligence conception and implementation for breast imaging diagnosis in clinical practice: a scoping review

OBJECTIVE

Although artificial intelligence (AI) has demonstrated promise in enhancing breast cancer diagnosis, the implementation of AI algorithms in clinical practice encounters various barriers. This scoping review aims to identify these barriers and facilitators to highlight key considerations for developing and implementing AI solutions in breast cancer imaging.

METHOD

A literature search was conducted from 2012 to 2022 in six databases (PubMed, Web of Science, CINHAL, Embase, IEEE, and ArXiv). The articles were included if some barriers and/or facilitators in the conception or implementation of AI in breast clinical imaging were described. We excluded research only focusing on performance, or with data not acquired in a clinical radiology setup and not involving real patients.

RESULTS

A total of 107 articles were included. We identified six major barriers related to data (B1), black box and trust (B2), algorithms and conception (B3), evaluation and validation (B4), legal, ethical, and economic issues (B5), and education (B6), and five major facilitators covering data (F1), clinical impact (F2), algorithms and conception (F3), evaluation and validation (F4), and education (F5).

CONCLUSION

This scoping review highlighted the need to carefully design, deploy, and evaluate AI solutions in clinical practice, involving all stakeholders to yield improvement in healthcare.

CLINICAL RELEVANCE STATEMENT

The identification of barriers and facilitators with suggested solutions can guide and inform future research, and stakeholders to improve the design and implementation of AI for breast cancer detection in clinical practice.

KEY POINTS

• Six major identified barriers were related to data; black-box and trust; algorithms and conception; evaluation and validation; legal, ethical, and economic issues; and education. • Five major identified facilitators were related to data, clinical impact, algorithms and conception, evaluation and validation, and education. • Coordinated implication of all stakeholders is required to improve breast cancer diagnosis with AI.

Design and testing of ultrasound probe adapters for a robotic imaging platform

Medical imaging-based triage is a critical tool for emergency medicine in both civilian and military settings. Ultrasound imaging can be used to rapidly identify free fluid in abdominal and thoracic cavities which could necessitate immediate surgical intervention. However, proper ultrasound image capture requires a skilled ultrasonography technician who is likely unavailable at the point of injury where resources are limited. Instead, robotics and computer vision technology can simplify image acquisition. As a first step towards this larger goal, here, we focus on the development of prototypes for ultrasound probe securement using a robotics platform. The ability of four probe adapter technologies to precisely capture images at anatomical locations, repeatedly, and with different ultrasound transducer types were evaluated across more than five scoring criteria. Testing demonstrated two of the adapters outperformed the traditional robot gripper and manual image capture, with a compact, rotating design compatible with wireless imaging technology being most suitable for use at the point of injury. Next steps will integrate the robotic platform with computer vision and deep learning image interpretation models to automate image capture and diagnosis. This will lower the skill threshold needed for medical imaging-based triage, enabling this procedure to be available at or near the point of injury.

The Use of Artificial Intelligence in the Liver Histopathology Field: A Systematic Review

Digital pathology (DP) has begun to play a key role in the evaluation of liver specimens. Recent studies have shown that a workflow that combines DP and artificial intelligence (AI) applied to histopathology has potential value in supporting the diagnosis, treatment evaluation, and prognosis prediction of liver diseases. Here, we provide a systematic review of the use of this workflow in the field of hepatology. Based on the PRISMA 2020 criteria, a search of the PubMed, SCOPUS, and Embase electronic databases was conducted, applying inclusion/exclusion filters. The articles were evaluated by two independent reviewers, who extracted the specifications and objectives of each study, the AI tools used, and the results obtained. From the 266 initial records identified, 25 eligible studies were selected, mainly conducted on human liver tissues. Most of the studies were performed using whole-slide imaging systems for imaging acquisition and applying different machine learning and deep learning methods for image pre-processing, segmentation, feature extractions, and classification. Of note, most of the studies selected demonstrated good performance as classifiers of liver histological images compared to pathologist annotations. Promising results to date bode well for the not-too-distant inclusion of these techniques in clinical practice.

Fetal Ultrasound Segmentation and Measurements Using Appearance and Shape Prior Based Density Regression with Deep CNN and Robust Ellipse Fitting

Accurately segmenting the structure of the fetal head (FH) and performing biometry measurements, including head circumference (HC) estimation, stands as a vital requirement for addressing abnormal fetal growth during pregnancy under the expertise of experienced radiologists using ultrasound (US) images. However, accurate segmentation and measurement is a challenging task due to image artifact, incomplete ellipse fitting, and fluctuations due to FH dimensions over different trimesters. Also, it is highly time-consuming due to the absence of specialized features, which leads to low segmentation accuracy. To address these challenging tasks, we propose an automatic density regression approach to incorporate appearance and shape priors into the deep learning-based network model (DR-ASPnet) with robust ellipse fitting using fetal US images. Initially, we employed multiple pre-processing steps to remove unwanted distortions, variable fluctuations, and a clear view of significant features from the US images. Then some form of augmentation operation is applied to increase the diversity of the dataset. Next, we proposed the hierarchical density regression deep convolutional neural network (HDR-DCNN) model, which involves three network models to determine the complex location of FH for accurate segmentation during the training and testing processes. Then, we used post-processing operations using contrast enhancement filtering with a morphological operation model to smooth the region and remove unnecessary artifacts from the segmentation results. After post-processing, we applied the smoothed segmented result to the robust ellipse fitting-based least square (REFLS) method for HC estimation. Experimental results of the DR-ASPnet model obtain 98.86% dice similarity coefficient (DSC) as segmentation accuracy, and it also obtains 1.67 mm absolute distance (AD) as measurement accuracy compared to other state-of-the-art methods. Finally, we achieved a 0.99 correlation coefficient (CC) in estimating the measured and predicted HC values on the HC18 dataset.

A Quantitative Method for the Evaluation of Deep Vein Thrombosis in a Murine Model Using Three-Dimensional Ultrasound Imaging

Deep vein thrombosis (DVT) is a life-threatening condition that can lead to its sequelae pulmonary embolism (PE) or post-thrombotic syndrome (PTS). Murine models of DVT are frequently used in early-stage disease research and to assess potential therapies. This creates the need for the reliable and easy quantification of blood clots. In this paper, we present a novel high-frequency 3D ultrasound approach for the quantitative evaluation of the volume of DVT in an in vitro model and an in vivo murine model. The proposed method involves the use of a high-resolution ultrasound acquisition system and semiautomatic segmentation of the clot. The measured 3D volume of blood clots was validated to be correlated with in vitro blood clot weights with an R2 of 0.89. Additionally, the method was confirmed with an R2 of 0.91 in the in vivo mouse model with a cylindrical volume from macroscopic measurement. We anticipate that the proposed method will be useful in pharmacological or therapeutic studies in murine models of DVT.

Learning technology for detection and grading of cancer tissue using tumour ultrasound images1

BACKGROUND

Early diagnosis of breast cancer is crucial to perform effective therapy. Many medical imaging modalities including MRI, CT, and ultrasound are used to diagnose cancer.

OBJECTIVE

This study aims to investigate feasibility of applying transfer learning techniques to train convoluted neural networks (CNNs) to automatically diagnose breast cancer via ultrasound images.

METHODS

Transfer learning techniques helped CNNs recognise breast cancer in ultrasound images. Each model's training and validation accuracies were assessed using the ultrasound image dataset. Ultrasound images educated and tested the models.

RESULTS

MobileNet had the greatest accuracy during training and DenseNet121 during validation. Transfer learning algorithms can detect breast cancer in ultrasound images.

CONCLUSIONS

Based on the results, transfer learning models may be useful for automated breast cancer diagnosis in ultrasound images. However, only a trained medical professional should diagnose cancer, and computational approaches should only be used to help make quick decisions.

Automated deep bottleneck residual 82-layered architecture with Bayesian optimization for the classification of brain and common maternal fetal ultrasound planes

Despite a worldwide decline in maternal mortality over the past two decades, a significant gap persists between low- and high-income countries, with 94% of maternal mortality concentrated in low and middle-income nations. Ultrasound serves as a prevalent diagnostic tool in prenatal care for monitoring fetal growth and development. Nevertheless, acquiring standard fetal ultrasound planes with accurate anatomical structures proves challenging and time-intensive, even for skilled sonographers. Therefore, for determining common maternal fetuses from ultrasound images, an automated computer-aided diagnostic (CAD) system is required. A new residual bottleneck mechanism-based deep learning architecture has been proposed that includes 82 layers deep. The proposed architecture has added three residual blocks, each including two highway paths and one skip connection. In addition, a convolutional layer has been added of size 3 × 3 before each residual block. In the training process, several hyper parameters have been initialized using Bayesian optimization (BO) rather than manual initialization. Deep features are extracted from the average pooling layer and performed the classification. In the classification process, an increase occurred in the computational time; therefore, we proposed an improved search-based moth flame optimization algorithm for optimal feature selection. The data is then classified using neural network classifiers based on the selected features. The experimental phase involved the analysis of ultrasound images, specifically focusing on fetal brain and common maternal fetal images. The proposed method achieved 78.5% and 79.4% accuracy for brain fetal planes and common maternal fetal planes. Comparison with several pre-trained neural nets and state-of-the-art (SOTA) optimization algorithms shows improved accuracy.

Mapping the Bibliometrics Landscape of AI in Medicine: Methodological Study

BACKGROUND

Artificial intelligence (AI), conceived in the 1950s, has permeated numerous industries, intensifying in tandem with advancements in computing power. Despite the widespread adoption of AI, its integration into medicine trails other sectors. However, medical AI research has experienced substantial growth, attracting considerable attention from researchers and practitioners.

OBJECTIVE

In the absence of an existing framework, this study aims to outline the current landscape of medical AI research and provide insights into its future developments by examining all AI-related studies within PubMed over the past 2 decades. We also propose potential data acquisition and analysis methods, developed using Python (version 3.11) and to be executed in Spyder IDE (version 5.4.3), for future analogous research.

METHODS

Our dual-pronged approach involved (1) retrieving publication metadata related to AI from PubMed (spanning 2000-2022) via Python, including titles, abstracts, authors, journals, country, and publishing years, followed by keyword frequency analysis and (2) classifying relevant topics using latent Dirichlet allocation, an unsupervised machine learning approach, and defining the research scope of AI in medicine. In the absence of a universal medical AI taxonomy, we used an AI dictionary based on the European Commission Joint Research Centre AI Watch report, which emphasizes 8 domains: reasoning, planning, learning, perception, communication, integration and interaction, service, and AI ethics and philosophy.

RESULTS

From 2000 to 2022, a comprehensive analysis of 307,701 AI-related publications from PubMed highlighted a 36-fold increase. The United States emerged as a clear frontrunner, producing 68,502 of these articles. Despite its substantial contribution in terms of volume, China lagged in terms of citation impact. Diving into specific AI domains, as the Joint Research Centre AI Watch report categorized, the learning domain emerged dominant. Our classification analysis meticulously traced the nuanced research trajectories across each domain, revealing the multifaceted and evolving nature of AI's application in the realm of medicine.

CONCLUSIONS

The research topics have evolved as the volume of AI studies increases annually. Machine learning remains central to medical AI research, with deep learning expected to maintain its fundamental role. Empowered by predictive algorithms, pattern recognition, and imaging analysis capabilities, the future of AI research in medicine is anticipated to concentrate on medical diagnosis, robotic intervention, and disease management. Our topic modeling outcomes provide a clear insight into the focus of AI research in medicine over the past decades and lay the groundwork for predicting future directions. The domains that have attracted considerable research attention, primarily the learning domain, will continue to shape the trajectory of AI in medicine. Given the observed growing interest, the domain of AI ethics and philosophy also stands out as a prospective area of increased focus.

Enhancing medical ultrasound imaging through fractional mathematical modeling of ultrasound bubble dynamics

The classical mathematical modeling of ultrasound acoustic bubble is so far using to improve the medical imaging quality. A clear and visible medical ultrasound image relies on bubble's diameter, wavelength and intensity of the scattered sound. A bubble with diameter much smaller than the sound wavelength is regarded as highly efficient source of sound scattering. The dynamical equation for a medical ultrasound bubble is primarily modeled in classical integer-order differential equation. Then a reduction of order technique is used to convert the modeled dynamic equation for the bubble surface into a system of incommensurate fractional-orders. The incommensurate fractional-order values are calculated directly, by using Riemann stability region. On the basis of stability the convergence and accuracy of the numerical scheme is also discussed in detail. It has been found that the system will remain stable and chaotic for the incommensurate values α1<0.737 and α2<2.80, respectively.

Deep convolutional neural network algorithm for the automatic segmentation of oral potentially malignant disorders and oral cancers

This study aimed to develop an algorithm to automatically segment the oral potentially malignant diseases (OPMDs) and oral cancers (OCs) of all oral subsites with various deep convolutional neural network applications. A total of 510 intraoral images of OPMDs and OCs were collected over 3 years (2006-2009). All images were confirmed both with patient records and histopathological reports. Following the labeling of the lesions the dataset was arbitrarily split, using random sampling in Python as the study dataset, validation dataset, and test dataset. Pixels were classified as the OPMDs and OCs with the OPMD/OC label and the rest as the background. U-Net architecture was used and the model with the best validation loss was chosen for the testing among the trained 500 epochs. Dice similarity coefficient (DSC) score was noted. The intra-observer ICC was found to be 0.994 while the inter-observer reliability was 0.989. The calculated DSC and validation accuracy across all clinical images were 0.697 and 0.805, respectively. Our algorithm did not maintain an excellent DSC due to multiple reasons for the detection of both OC and OPMDs in oral cavity sites. A better standardization for both 2D and 3D imaging (such as patient positioning) and a bigger dataset are required to improve the quality of such studies. This is the first study which aimed to segment OPMDs and OCs in all subsites of oral cavity which is crucial not only for the early diagnosis but also for higher survival rates.

Deep learning algorithm for predicting subacromial motion trajectory: Dynamic shoulder ultrasound analysis

Subacromial motion metrics can be extracted from dynamic shoulder ultrasonography, which is useful for identifying abnormal motion patterns in painful shoulders. However, frame-by-frame manual labeling of anatomical landmarks in ultrasound images is time consuming. The present study aims to investigate the feasibility of a deep learning algorithm for extracting subacromial motion metrics from dynamic ultrasonography. Dynamic ultrasound imaging was retrieved by asking 17 participants to perform cyclic shoulder abduction and adduction along the scapular plane, whereby the trajectory of the humeral greater tubercle (in relation to the lateral acromion) was depicted by the deep learning algorithm. Extraction of the subacromial motion metrics was conducted using a convolutional neural network (CNN) or a self-transfer learning-based (STL)-CNN with or without an autoencoder (AE). The mean absolute error (MAE) compared with the manually-labeled data (ground truth) served as the main outcome variable. Using eight-fold cross-validation, the average MAE was proven to be significantly higher in the group using CNN than in those using STL-CNN or STL-CNN+AE for the relative difference between the greater tubercle and lateral acromion on the horizontal axis. The MAE for the localization of the two aforementioned landmarks on the vertical axis also seemed to be enlarged in those using CNN compared with those using STL-CNN. In the testing dataset, the errors in relation to the ground truth for the minimal vertical acromiohumeral distance were 0.081-0.333 cm using CNN, compared with 0.002-0.007 cm using STL-CNN. We successfully demonstrated the feasibility of a deep learning algorithm for automatic detection of the greater tubercle and lateral acromion during dynamic shoulder ultrasonography. Our framework also demonstrated the capability of capturing the minimal vertical acromiohumeral distance, which is the most important indicator of subacromial motion metrics in daily clinical practice.

Analysis of super-enhancer using machine learning and its application to medical biology

The analysis of super-enhancers (SEs) has recently attracted attention in elucidating the molecular mechanisms of cancer and other diseases. SEs are genomic structures that strongly induce gene expression and have been reported to contribute to the overexpression of oncogenes. Because the analysis of SEs and integrated analysis with other data are performed using large amounts of genome-wide data, artificial intelligence technology, with machine learning at its core, has recently begun to be utilized. In promoting precision medicine, it is important to consider information from SEs in addition to genomic data; therefore, machine learning technology is expected to be introduced appropriately in terms of building a robust analysis platform with a high generalization performance. In this review, we explain the history and principles of SE, and the results of SE analysis using state-of-the-art machine learning and integrated analysis with other data are presented to provide a comprehensive understanding of the current status of SE analysis in the field of medical biology. Additionally, we compared the accuracy between existing machine learning methods on the benchmark dataset and attempted to explore the kind of data preprocessing and integration work needed to make the existing algorithms work on the benchmark dataset. Furthermore, we discuss the issues and future directions of current SE analysis.

Evaluation of Cholesterol Thickness of Blood Vessels Using Photoacoustic Technology

One of the primary indicators of plaque vulnerability is the lipid composition of atherosclerotic plaques. Therefore, the medical industry requires a method to evaluate necrotic nuclei in atherosclerosis imaging with sensitivity. In this regard, photoacoustic imaging is a plaque detection method that provides chemical information on lipids and cholesterol thickness in the arterial walls of the patient. This aspect aims to increase the low-frequency axial resolution by developing a new photoacoustic-based system. A photoacoustic system has been developed to detect the cholesterol thickness of the blood vessels to observe the progression of plaque in the heart's blood vessels. The application of the coherent photoacoustic discontinuous correlation tomography technique, which is based on a novel signal processing, significantly increased the cholesterol oleate's sensitivity to plaque necrosis. By enhancing the quality of thickness detection, the system for measuring the thickness of cholesterol in blood vessels has been reduced to approximately 23 microns. The results show that the phase spectrum peaked at 100 Hz at 58.66 degrees, and at 400 Hz, the phase spectrum was 46.37 degrees. The minimum amplitude is 1.95 at 100 Hz and 17.67 at 400 Hz. In conclusion, it can be stated that photoacoustic imaging as a method based on new technologies is of great importance in medical research, which is based on the use of nonionizing radiation to perform diagnostic processes and measure different types of body tissues.

Annotation Cost Minimization for Ultrasound Image Segmentation Using Cross-Domain Transfer Learning

Deep learning techniques can help minimize inter-physician analysis variability and the medical expert workloads, thereby enabling more accurate diagnoses. However, their implementation requires large-scale annotated dataset whose acquisition incurs heavy time and human-expertise costs. Hence, to significantly minimize the annotation cost, this study presents a novel framework that enables the deployment of deep learning methods in ultrasound (US) image segmentation requiring only very limited manually annotated samples. We propose SegMix, a fast and efficient approach that exploits a segment-paste-blend concept to generate large number of annotated samples based on a few manually acquired labels. Besides, a series of US-specific augmentation strategies built upon image enhancement algorithms are introduced to make maximum use of the available limited number of manually delineated images. The feasibility of the proposed framework is validated on the left ventricle (LV) segmentation and fetal head (FH) segmentation tasks, respectively. Experimental results demonstrate that using only 10 manually annotated images, the proposed framework can achieve a Dice and JI of 82.61% and 83.92%, and 88.42% and 89.27% for LV segmentation and FH segmentation, respectively. Compared with training using the entire training set, there is over 98% of annotation cost reduction while achieving comparable segmentation performance. This indicates that the proposed framework enables satisfactory deep leaning performance when very limited number of annotated samples is available. Therefore, we believe that it can be a reliable solution for annotation cost reduction in medical image analysis.

Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks

Artificial Intelligence (AI) techniques of deep learning have revolutionized the disease diagnosis with their outstanding image classification performance. In spite of the outstanding results, the widespread adoption of these techniques in clinical practice is still taking place at a moderate pace. One of the major hindrance is that a trained Deep Neural Networks (DNN) model provides a prediction, but questions about why and how that prediction was made remain unanswered. This linkage is of utmost importance for the regulated healthcare domain to increase the trust in the automated diagnosis system by the practitioners, patients and other stakeholders. The application of deep learning for medical imaging has to be interpreted with caution due to the health and safety concerns similar to blame attribution in the case of an accident involving autonomous cars. The consequences of both a false positive and false negative cases are far reaching for patients' welfare and cannot be ignored. This is exacerbated by the fact that the state-of-the-art deep learning algorithms comprise of complex interconnected structures, millions of parameters, and a 'black box' nature, offering little understanding of their inner working unlike the traditional machine learning algorithms. Explainable AI (XAI) techniques help to understand model predictions which help develop trust in the system, accelerate the disease diagnosis, and meet adherence to regulatory requirements. This survey provides a comprehensive review of the promising field of XAI for biomedical imaging diagnostics. We also provide a categorization of the XAI techniques, discuss the open challenges, and provide future directions for XAI which would be of interest to clinicians, regulators and model developers.

Artificial intelligence in diagnostic ultrasonography

Artificial intelligence (AI) continues to change paradigms in the field of medicine with new applications that are applicable to daily life. The field of ultrasonography, which has been developing since the 1950s and continues to be one of the most powerful tools in the field of diagnosis, is also the subject of AI studies, despite its unique problems. It is predicted that many operations, such as appropriate diagnostic tool selection, use of the most relevant parameters, improvement of low-quality images, automatic lesion detection and diagnosis from the image, and classification of pathologies, will be performed using AI tools in the near future. Especially with the use of convolutional neural networks, successful results can be obtained for lesion detection, segmentation, and classification from images. In this review, relevant developments are summarized based on the literature, and examples of the tools used in the field are presented.

Advances in image-guided drug delivery for antibacterial therapy

The emergence of antibiotic-resistant bacterial strains is seriously endangering the global healthcare system. There is an urgent need for combining imaging with therapies to realize the real-time monitoring of pathological condition and treatment progress. It also provides guidance on exploring new medicines and enhance treatment strategies to overcome the antibiotic resistance of existing conventional antibiotics. In this review, we provide a thorough overview of the most advanced image-guided approaches for bacterial diagnosis (e.g., computed tomography imaging, magnetic resonance imaging, photoacoustic imaging, ultrasound imaging, fluorescence imaging, positron emission tomography, single photon emission computed tomography imaging, and multiple imaging), and therapies (e.g., photothermal therapy, photodynamic therapy, chemodynamic therapy, sonodynamic therapy, immunotherapy, and multiple therapies). This review focuses on how to design and fabricate photo-responsive materials for improved image-guided bacterial theranostics applications. We present a potential application of different image-guided modalities for both bacterial diagnosis and therapies with representative examples. Finally, we highlighted the current challenges and future perspectives image-guided approaches for future clinical translation of nano-theranostics in bacterial infections therapies. We envision that this review will provide for future development in image-guided systems for bacterial theranostics applications.

Automatic ovarian tumors recognition system based on ensemble convolutional neural network with ultrasound imaging

BACKGROUND

Upon the discovery of ovarian cysts, obstetricians, gynecologists, and ultrasound examiners must address the common clinical challenge of distinguishing between benign and malignant ovarian tumors. Numerous types of ovarian tumors exist, many of which exhibit similar characteristics that increase the ambiguity in clinical diagnosis. Using deep learning technology, we aimed to develop a method that rapidly and accurately assists the different diagnosis of ovarian tumors in ultrasound images.

METHODS

Based on deep learning method, we used ten well-known convolutional neural network models (e.g., Alexnet, GoogleNet, and ResNet) for training of transfer learning. To ensure method stability and robustness, we repeated the random sampling of the training and validation data ten times. The mean of the ten test results was set as the final assessment data. After the training process was completed, the three models with the highest ratio of calculation accuracy to time required for classification were used for ensemble learning pertaining. Finally, the interpretation results of the ensemble classifier were used as the final results. We also applied ensemble gradient-weighted class activation mapping (Grad-CAM) technology to visualize the decision-making results of the models.

RESULTS

The highest mean accuracy, mean sensitivity, and mean specificity of ten single CNN models were 90.51 ± 4.36%, 89.77 ± 4.16%, and 92.00 ± 5.95%, respectively. The mean accuracy, mean sensitivity, and mean specificity of the ensemble classifier method were 92.15 ± 2.84%, 91.37 ± 3.60%, and 92.92 ± 4.00%, respectively. The performance of the ensemble classifier is better than that of a single classifier in three evaluation metrics. Moreover, the standard deviation is also better which means the ensemble classifier is more stable and robust.

CONCLUSION

From the comprehensive perspective of data quantity, data diversity, robustness of validation strategy, and overall accuracy, the proposed method outperformed the methods used in previous studies. In future studies, we will continue to increase the number of authenticated images and apply our proposed method in clinical settings to increase its robustness and reliability.

Introducing AI to the molecular tumor board: one direction toward the establishment of precision medicine using large-scale cancer clinical and biological information

Since U.S. President Barack Obama announced the Precision Medicine Initiative in his New Year's State of the Union address in 2015, the establishment of a precision medicine system has been emphasized worldwide, particularly in the field of oncology. With the advent of next-generation sequencers specifically, genome analysis technology has made remarkable progress, and there are active efforts to apply genome information to diagnosis and treatment. Generally, in the process of feeding back the results of next-generation sequencing analysis to patients, a molecular tumor board (MTB), consisting of experts in clinical oncology, genetic medicine, etc., is established to discuss the results. On the other hand, an MTB currently involves a large amount of work, with humans searching through vast databases and literature, selecting the best drug candidates, and manually confirming the status of available clinical trials. In addition, as personalized medicine advances, the burden on MTB members is expected to increase in the future. Under these circumstances, introducing cutting-edge artificial intelligence (AI) technology and information and communication technology to MTBs while reducing the burden on MTB members and building a platform that enables more accurate and personalized medical care would be of great benefit to patients. In this review, we introduced the latest status of elemental technologies that have potential for AI utilization in MTB, and discussed issues that may arise in the future as we progress with AI implementation.

Digital Technologies: Advancing Individualized Treatments through Gene and Cell Therapies, Pharmacogenetics, and Disease Detection and Diagnostics

Digital technologies are shifting the paradigm of medicine in a way that will transform the healthcare industry. Conventional medical approaches focus on treating symptoms and ailments for large groups of people. These approaches can elicit differences in treatment responses and adverse reactions based on population variations, and are often incapable of treating the inherent pathophysiology of the medical conditions. Advances in genetics and engineering are improving healthcare via individualized treatments that include gene and cell therapies, pharmacogenetics, disease detection, and diagnostics. This paper highlights ways that artificial intelligence can help usher in an age of personalized medicine.

Segmentation-Based Classification Deep Learning Model Embedded with Explainable AI for COVID-19 Detection in Chest X-ray Scans

Background and Motivation: COVID-19 has resulted in a massive loss of life during the last two years. The current imaging-based diagnostic methods for COVID-19 detection in multiclass pneumonia-type chest X-rays are not so successful in clinical practice due to high error rates. Our hypothesis states that if we can have a segmentation-based classification error rate <5%, typically adopted for 510 (K) regulatory purposes, the diagnostic system can be adapted in clinical settings. Method: This study proposes 16 types of segmentation-based classification deep learning-based systems for automatic, rapid, and precise detection of COVID-19. The two deep learning-based segmentation networks, namely UNet and UNet+, along with eight classification models, namely VGG16, VGG19, Xception, InceptionV3, Densenet201, NASNetMobile, Resnet50, and MobileNet, were applied to select the best-suited combination of networks. Using the cross-entropy loss function, the system performance was evaluated by Dice, Jaccard, area-under-the-curve (AUC), and receiver operating characteristics (ROC) and validated using Grad-CAM in explainable AI framework. Results: The best performing segmentation model was UNet, which exhibited the accuracy, loss, Dice, Jaccard, and AUC of 96.35%, 0.15%, 94.88%, 90.38%, and 0.99 (p-value <0.0001), respectively. The best performing segmentation-based classification model was UNet+Xception, which exhibited the accuracy, precision, recall, F1-score, and AUC of 97.45%, 97.46%, 97.45%, 97.43%, and 0.998 (p-value <0.0001), respectively. Our system outperformed existing methods for segmentation-based classification models. The mean improvement of the UNet+Xception system over all the remaining studies was 8.27%. Conclusion: The segmentation-based classification is a viable option as the hypothesis (error rate <5%) holds true and is thus adaptable in clinical practice.

Application of non-negative matrix factorization in oncology: one approach for establishing precision medicine

The increase in the expectations of artificial intelligence (AI) technology has led to machine learning technology being actively used in the medical field. Non-negative matrix factorization (NMF) is a machine learning technique used for image analysis, speech recognition, and language processing; recently, it is being applied to medical research. Precision medicine, wherein important information is extracted from large-scale medical data to provide optimal medical care for every individual, is considered important in medical policies globally, and the application of machine learning techniques to this end is being handled in several ways. NMF is also introduced differently because of the characteristics of its algorithms. In this review, the importance of NMF in the field of medicine, with a focus on the field of oncology, is described by explaining the mathematical science of NMF and the characteristics of the algorithm, providing examples of how NMF can be used to establish precision medicine, and presenting the challenges of NMF. Finally, the direction regarding the effective use of NMF in the field of oncology is also discussed.

Towards deep observation: A systematic survey on artificial intelligence techniques to monitor fetus via Ultrasound Images

Several reviews have been conducted regarding artificial intelligence (AI) techniques to improve pregnancy outcomes. But they are not focusing on ultrasound images. This survey aims to explore how AI can assist with fetal growth monitoring via ultrasound image. We reported our findings using the guidelines for PRISMA. We conducted a comprehensive search of eight bibliographic databases. Out of 1269 studies 107 are included. We found that 2D ultrasound images were more popular (88) than 3D and 4D ultrasound images (19). Classification is the most used method (42), followed by segmentation (31), classification integrated with segmentation (16) and other miscellaneous methods such as object-detection, regression, and reinforcement learning (18). The most common areas that gained traction within the pregnancy domain were the fetus head (43), fetus body (31), fetus heart (13), fetus abdomen (10), and the fetus face (10). This survey will promote the development of improved AI models for fetal clinical applications.

•

Artificial intelligence studies to monitor fetal development via ultrasound images

•

Fetal issues categorized based on four categories — general, head, heart, face, abdomen

•

The most used AI techniques are classification, segmentation, object detection, and RL

•

The research and practical implications are included.

Automated Endocardial Border Detection and Left Ventricular Functional Assessment in Echocardiography Using Deep Learning

Endocardial border detection is a key step in assessing left ventricular systolic function in echocardiography. However, this process is still not sufficiently accurate, and manual retracing is often required, causing time-consuming and intra-/inter-observer variability in clinical practice. To address these clinical issues, more accurate and normalized automatic endocardial border detection would be valuable. Here, we develop a deep learning-based method for automated endocardial border detection and left ventricular functional assessment in two-dimensional echocardiographic videos. First, segmentation of the left ventricular cavity was performed in the six representative projections for a cardiac cycle. We employed four segmentation methods: U-Net, UNet++, UNet3+, and Deep Residual U-Net. UNet++ and UNet3+ showed a sufficiently high performance in the mean value of intersection over union and Dice coefficient. The accuracy of the four segmentation methods was then evaluated by calculating the mean value for the estimation error of the echocardiographic indexes. UNet++ was superior to the other segmentation methods, with the acceptable mean estimation error of the left ventricular ejection fraction of 10.8%, global longitudinal strain of 8.5%, and global circumferential strain of 5.8%, respectively. Our method using UNet++ demonstrated the best performance. This method may potentially support examiners and improve the workflow in echocardiography.

BUSnet: A Deep Learning Model of Breast Tumor Lesion Detection for Ultrasound Images

Ultrasound (US) imaging is a main modality for breast disease screening. Automatically detecting the lesions in US images is essential for developing the artificial-intelligence-based diagnostic support technologies. However, the intrinsic characteristics of ultrasound imaging, like speckle noise and acoustic shadow, always degenerate the detection accuracy. In this study, we developed a deep learning model called BUSnet to detect the breast tumor lesions in US images with high accuracy. We first developed a two-stage method including the unsupervised region proposal and bounding-box regression algorithms. Then, we proposed a post-processing method to enhance the detecting accuracy further. The proposed method was used to a benchmark dataset, which includes 487 benign samples and 210 malignant samples. The results proved the effectiveness and accuracy of the proposed method.

Interdisciplinary Collaboration Opportunities, Challenges and Solutions for Artificial Intelligence in Ultrasound

Ultrasound is one of the most widely utilized imaging tools in clinical practice with the advantages of noninvasive nature and ease of use. However, ultrasound examinations have low reproducibility and considerable heterogeneity due to the variability of operators, scanners, and patients. In recent years, Artificial Intelligence (AI) -assisted ultrasound has matured and moved closer to routine clinical uses. The combination of AI with ultrasound has opened up a world of possibilities for increasing work productivity and precision diagnostics. In this article, we describe AI strategies in ultrasound, from current opportunities, constraints to potential options for AI-assisted ultrasound.

Common and Uncommon Errors in Emergency Ultrasound

Errors in emergency ultrasound (US) have been representing an increasing problem in recent years thanks to several unique features related to both the inherent characteristics of the discipline and to the latest developments, which every medical operator should be aware of. Because of the subjective nature of the interpretation of emergency US findings, it is more prone to errors than other diagnostic imaging modalities. The misinterpretation of US images should therefore be considered as a serious risk in diagnosis. The etiology of error is multi-factorial: it depends on environmental factors, patients and the technical skills of the operator; it is influenced by intrinsic US artifacts, poor clinical correlation, US-setting errors and anatomical variants; and it is conditioned by the lack of a methodologically correct clinical approach and excessive diagnostic confidence too. In this review, we evaluate the common and uncommon sources of diagnostic errors in emergency US during clinical practice, showing how to recognize and avoid them.

Medical Professional Enhancement Using Explainable Artificial Intelligence in Fetal Cardiac Ultrasound Screening

Diagnostic support tools based on artificial intelligence (AI) have exhibited high performance in various medical fields. However, their clinical application remains challenging because of the lack of explanatory power in AI decisions (black box problem), making it difficult to build trust with medical professionals. Nevertheless, visualizing the internal representation of deep neural networks will increase explanatory power and improve the confidence of medical professionals in AI decisions. We propose a novel deep learning-based explainable representation "graph chart diagram" to support fetal cardiac ultrasound screening, which has low detection rates of congenital heart diseases due to the difficulty in mastering the technique. Screening performance improves using this representation from 0.966 to 0.975 for experts, 0.829 to 0.890 for fellows, and 0.616 to 0.748 for residents in the arithmetic mean of area under the curve of a receiver operating characteristic curve. This is the first demonstration wherein examiners used deep learning-based explainable representation to improve the performance of fetal cardiac ultrasound screening, highlighting the potential of explainable AI to augment examiner capabilities.

Brain Tumor Classification Using a Combination of Variational Autoencoders and Generative Adversarial Networks

Brain tumors are a pernicious cancer with one of the lowest five-year survival rates. Neurologists often use magnetic resonance imaging (MRI) to diagnose the type of brain tumor. Automated computer-assisted tools can help them speed up the diagnosis process and reduce the burden on the health care systems. Recent advances in deep learning for medical imaging have shown remarkable results, especially in the automatic and instant diagnosis of various cancers. However, we need a large amount of data (images) to train the deep learning models in order to obtain good results. Large public datasets are rare in medicine. This paper proposes a framework based on unsupervised deep generative neural networks to solve this limitation. We combine two generative models in the proposed framework: variational autoencoders (VAEs) and generative adversarial networks (GANs). We swap the encoder-decoder network after initially training it on the training set of available MR images. The output of this swapped network is a noise vector that has information of the image manifold, and the cascaded generative adversarial network samples the input from this informative noise vector instead of random Gaussian noise. The proposed method helps the GAN to avoid mode collapse and generate realistic-looking brain tumor magnetic resonance images. These artificially generated images could solve the limitation of small medical datasets up to a reasonable extent and help the deep learning models perform acceptably. We used the ResNet50 as a classifier, and the artificially generated brain tumor images are used to augment the real and available images during the classifier training. We compared the classification results with several existing studies and state-of-the-art machine learning models. Our proposed methodology noticeably achieved better results. By using brain tumor images generated artificially by our proposed method, the classification average accuracy improved from 72.63% to 96.25%. For the most severe class of brain tumor, glioma, we achieved 0.769, 0.837, 0.833, and 0.80 values for recall, specificity, precision, and F1-score, respectively. The proposed generative model framework could be used to generate medical images in any domain, including PET (positron emission tomography) and MRI scans of various parts of the body, and the results show that it could be a useful clinical tool for medical experts.

Ultrasound in Telemedicine: A Brief Overview

The delivery of healthcare from a distance, also known as telemedicine, has evolved over the past 50 years, changing the way healthcare is delivered globally. Its integration into numerous domains has permitted high-quality care that transcends the obstacles of geographic distance, lack of access to health care providers, and cost. Ultrasound is an effective diagnostic tool and its application within telemedicine has advanced substantially in recent years, particularly in high-income settings and low-resource areas. The literature in Pubmed from 1960–2020 was assessed with the keywords “ultrasound”, “telemedicine”, “ultrasound remote”, and “tele-ultrasound” to conduct a SWOT analysis (strengths, weaknesses, opportunities, and threats). In addressing strengths and opportunities, we emphasized practical aspects, such as the usefulness of tele-ultrasound and the cost efficiency of it. Furthermore, aspects of medical education in tele-ultrasound were considered. When it came to weaknesses and threats, we focused on issues that may not be solved immediately, and that require careful consideration or further development, such as new software that is not yet available commercially.

Single-Cell Analysis Using Machine Learning Techniques and Its Application to Medical Research

In recent years, the diversity of cancer cells in tumor tissues as a result of intratumor heterogeneity has attracted attention. In particular, the development of single-cell analysis technology has made a significant contribution to the field; technologies that are centered on single-cell RNA sequencing (scRNA-seq) have been reported to analyze cancer constituent cells, identify cell groups responsible for therapeutic resistance, and analyze gene signatures of resistant cell groups. However, although single-cell analysis is a powerful tool, various issues have been reported, including batch effects and transcriptional noise due to gene expression variation and mRNA degradation. To overcome these issues, machine learning techniques are currently being introduced for single-cell analysis, and promising results are being reported. In addition, machine learning has also been used in various ways for single-cell analysis, such as single-cell assay of transposase accessible chromatin sequencing (ATAC-seq), chromatin immunoprecipitation sequencing (ChIP-seq) analysis, and multi-omics analysis; thus, it contributes to a deeper understanding of the characteristics of human diseases, especially cancer, and supports clinical applications. In this review, we present a comprehensive introduction to the implementation of machine learning techniques in medical research for single-cell analysis, and discuss their usefulness and future potential.