Current perspectives in medical image perception

Incorporating Radiologist Knowledge Into MRI Quality Metrics for Machine Learning Using Rank-Based Ratings

BACKGROUND

Deep learning (DL) often requires an image quality metric; however, widely used metrics are not designed for medical images.

PURPOSE

To develop an image quality metric that is specific to MRI using radiologists image rankings and DL models.

STUDY TYPE

Retrospective.

POPULATION

A total of 19,344 rankings on 2916 unique image pairs from the NYU fastMRI Initiative neuro database was used for the neural network-based image quality metrics training with an 80%/20% training/validation split and fivefold cross-validation.

FIELD STRENGTH/SEQUENCE

1.5 T and 3 T T1, T1 postcontrast, T2, and FLuid Attenuated Inversion Recovery (FLAIR).

ASSESSMENT

Synthetically corrupted image pairs were ranked by radiologists (N = 7), with a subset also scoring images using a Likert scale (N = 2). DL models were trained to match rankings using two architectures (EfficientNet and IQ-Net) with and without reference image subtraction and compared to ranking based on mean squared error (MSE) and structural similarity (SSIM). Image quality assessing DL models were evaluated as alternatives to MSE and SSIM as optimization targets for DL denoising and reconstruction.

STATISTICAL TESTS

Radiologists' agreement was assessed by a percentage metric and quadratic weighted Cohen's kappa. Ranking accuracies were compared using repeated measurements analysis of variance. Reconstruction models trained with IQ-Net score, MSE and SSIM were compared by paired t test. P < 0.05 was considered significant.

RESULTS

Compared to direct Likert scoring, ranking produced a higher level of agreement between radiologists (70.4% vs. 25%). Image ranking was subjective with a high level of intraobserver agreement ( 94.9 % ± 2.4 % ) and lower interobserver agreement ( 61.47 % ± 5.51 % ). IQ-Net and EfficientNet accurately predicted rankings with a reference image ( 75.2 % ± 1.3 % and 79.2 % ± 1.7 % ). However, EfficientNet resulted in images with artifacts and high MSE when used in denoising tasks while IQ-Net optimized networks performed well for both denoising and reconstruction tasks.

DATA CONCLUSION

Image quality networks can be trained from image ranking and used to optimize DL tasks.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 1.

From tissue to sound: A new paradigm for medical sonic interaction design

Medical imaging maps tissue characteristics into image intensity values, enhancing human perception. However, comprehending this data, especially in high-stakes scenarios such as surgery, is prone to errors. Additionally, current multimodal methods do not fully leverage this valuable data in their design. We introduce "From Tissue to Sound," a new paradigm for medical sonic interaction design. This paradigm establishes a comprehensive framework for mapping tissue characteristics to auditory displays, providing dynamic and intuitive access to medical images that complement visual data, thereby enhancing multimodal perception. "From Tissue to Sound" provides an advanced and adaptable framework for the interactive sonification of multimodal medical imaging data. This framework employs a physics-based sound model composed of a network of multiple oscillators, whose mechanical properties-such as friction and stiffness-are defined by tissue characteristics extracted from imaging data. This approach enables the representation of anatomical structures and the creation of unique acoustic profiles in response to excitations of the sound model. This method allows users to explore data at a fundamental level, identifying tissue characteristics ranging from rigid to soft, dense to sparse, and structured to scattered. It facilitates intuitive discovery of both general and detailed patterns with minimal preprocessing. Unlike conventional methods that transform low-dimensional data into global sound features through a parametric approach, this method utilizes model-based unsupervised mapping between data and an anatomical sound model, enabling high-dimensional data processing. The versatility of this method is demonstrated through feasibility experiments confirming the generation of perceptually discernible acoustic signals. Furthermore, we present a novel application developed based on this framework for retinal surgery. This new paradigm opens up possibilities for designing multisensory applications for multimodal imaging data. It also facilitates the creation of interactive sonification models with various auditory causality approaches, enhancing both directness and richness.

Interobserver agreement during clinical magnetic resonance imaging of the equine foot

BACKGROUND

Agreement between experienced observers for assessment of pathology and assessment confidence are poorly documented for magnetic resonance imaging (MRI) of the equine foot.

OBJECTIVES

To report interobserver agreement for pathology assessment and observer confidence for key anatomical structures of the equine foot during MRI.

STUDY DESIGN

Exploratory clinical study.

METHODS

Ten experienced observers (diploma or associate level) assessed 15 equine foot MRI studies acquired from clinical databases of 3 MRI systems. Observers graded pathology in seven key anatomical structures (Grade 1: no pathology, Grade 2: mild pathology, Grade 3: moderate pathology, Grade 4: severe pathology) and provided a grade for their confidence for each pathology assessment (Grade 1: high confidence, Grade 2: moderate confidence, Grade 3: limited confidence, Grade 4: no confidence). Interobserver agreement for the presence/absence of pathology and agreement for individual grades of pathology were assessed with Fleiss' kappa (k). Overall interobserver agreement for pathology was determined using Fleiss' kappa and Kendall's coefficient of concordance (KCC). The distribution of grading was also visualised with bubble charts.

RESULTS

Interobserver agreement for the presence/absence of pathology of individual anatomical structures was poor-to-fair, except for the navicular bone which had moderate agreement (k = 0.52). Relative agreement for pathology grading (accounting for the ranking of grades) ranged from KCC = 0.19 for the distal interphalangeal joint to KCC = 0.70 for the navicular bone. Agreement was generally greatest at the extremes of pathology. Observer confidence in pathology assessment was generally moderate to high.

MAIN LIMITATIONS

Distribution of pathology varied between anatomical structures due to random selection of clinical MRI studies. Observers had most experience with low-field MRI.

CONCLUSIONS

Even with experienced observers, there can be notable variation in the perceived severity of foot pathology on MRI for individual cases, which could be important in a clinical context.

Application of radiomics features in differential diagnosis of odontogenic cysts

OBJECTIVES

Cysts in jaws may have similar radiographic features. However, it is important to clarify the diagnosis prior to surgery. The aim of this study was to compare the radiomic features of radicular cysts (RCs), dentigerous cysts (DCs), and odontogenic keratocysts (OKCs) as a non-invasive diagnostic alternative to biopsy.

METHODS

In total, 161 odontogenic cysts diagnosed histopathologically (55 RCs, 53 DCs, and 53 OKCs) were included in the present study. Each cyst was semi-automatically segmented on CBCT images, and radiomic features were extracted by an observer. A second observer repeated 20% of the evaluations and the radiomic features. Those achieving an inter-observer agreement level above 0.850 were included in the study. Consequently, five shape-based and 22 textural features were investigated in the study. Statistical analysis was performed comparing both three cyst features and making pairwise comparisons.

RESULTS

All features included in the study showed statistical differences between cysts, with the exception of one textural feature (NGTDM coarseness) (P < .05). However, only one shape-based feature (shericity) and one textural feature (GLSZM large area emphasis) were statistically different in pairwise comparisons of all three cysts (P < .05).

CONCLUSION

Radiomics features of the RCs, DCs, and OKCs showed significant differences, and may have the potential to be used as a non-invasive method in the differential diagnosis of cysts.

Automatic medical report generation based on deep learning: A state of the art survey

With the increasing popularity of medical imaging and its expanding applications, posing significant challenges for radiologists. Radiologists need to spend substantial time and effort to review images and manually writing reports every day. To address these challenges and speed up the process of patient care, researchers have employed deep learning methods to automatically generate medical reports. In recent years, researchers have been increasingly focusing on this task and a large amount of related work has emerged. Although there have been some review articles summarizing the state of the art in this field, their discussions remain relatively limited. Therefore, this paper provides a comprehensive review of the latest advancements in automatic medical report generation, focusing on four key aspects: (1) describing the problem of automatic medical report generation, (2) introducing datasets of different modalities, (3) thoroughly analyzing existing evaluation metrics, (4) classifying existing studies into five categories: retrieval-based, domain knowledge-based, attention-based, reinforcement learning-based, large language models-based, and merged model. In addition, we point out the problems in this field and discuss the directions of future challenges. We hope that this review provides a thorough understanding of automatic medical report generation and encourages the continued development in this area.

Visual perceptual artefacts in cardiothoracic imaging: Implications for diagnostic accuracy and strategies for mitigation

Visual perceptual artefacts are distortions or illusions in medical image interpretation arising from the human visual system rather than hardware or imaging acquisition processes. These artefacts, emerging at various visual processing stages, such as the retina, visual pathways, visual cortex, and cognitive interpretation stages, impact the interpretation of cardiothoracic images. This review discusses artefacts including Mach bands, Dark Rim, Background Effects, Ambiguous Figures, Subjective Contours, and the Parallax Effect. A thorough literature search was conducted using PubMed and Google Scholar. Search terms included 'visual perceptual artifacts', 'cardiothoracic imaging', 'Mach bands', 'dark rim artifact', 'stress cardiac MRI', and 'radiology visual illusions'. Inclusion criteria encompassed studies focusing on visual perceptual artefacts in cardiothoracic imaging published in peer-reviewed journals. Studies not addressing visual perceptual artefacts or those focusing on imaging technology, hardware, or software-related artefacts were excluded. Duplicate studies were removed, and a final selection of 32 studies was analysed. Several key visual perceptual artefacts relevant to cardiothoracic imaging were identified. Mach bands, caused by lateral inhibition in the retina, create false contrasts at object boundaries, potentially mimicking pathologies such as fractures or pneumothoraces. The dark rim artefact observed on stress cardiac MRI, resulting from Gibbs ringing or truncation artefacts, can simulate perfusion defects and complicate diagnosis. Lateral inhibition can possibly accentuate the appearance of the dark rim artefact. Artefacts also fall under illusions of sensation, perception, and image formation. These illusions present both risks and benefits to diagnostic radiology. Recognizing visual perceptual artefacts is crucial for improving diagnostic accuracy and developing strategies to mitigate their impact. A comprehensive understanding and training on these artefacts are essential for radiologists to enhance interpretive skills, reduce diagnostic errors, and ultimately improve patient care in radiology.

Diagnostic performance of approximal caries in bitewing radiographs from different monitors and room illuminances

OBJECTIVE

To compare the accuracy, duration, and certainty of diagnosing approximal caries in bitewing radiographs displayed in 3 monitors under 2 luminance conditions.

METHODS

A total of 39 teeth without evident caries were selected from 11 patients undergoing partial jaw resection. Before the operation, 13 bitewing radiographs were captured by a digital imaging system. Eight observers evaluated the images under dark (9 lux) and bright (200 lux) conditions, using 2 medical-grade monitors and a commercial monitor. Using histological results as the gold standard, the areas under the receiver operating characteristic curves under different conditions were compared using the Z-test. Multivariate analysis of variance was conducted to assess the impact of various factors on diagnostic duration, while ordinal logistic regression was used to examine factors influencing diagnostic certainty level. It was considered significant when P < .05.

RESULTS

No significant difference was found in the diagnostic accuracy or duration for diagnosis of approximal caries under 2 luminance conditions with the 3 distinct monitors (P > .05). Ambient light, clinical experience, and the pathological grade of approximal caries have influence on the degree of diagnostic confidence (P < .05).

CONCLUSIONS

Different monitors and ambient luminance didn't influence the diagnostic accuracy or evaluation duration. Ambient luminance, clinical experience, and the depth of caries affect the degree of diagnostic confidence.

ADVANCES IN KNOWLEDGE

The study employing bitewing radiographs from real patients indicates that ambient luminance, clinical experience, and the depth of caries affect the degree of diagnostic confidence.

Eye Gaze Guided Cross-Modal Alignment Network for Radiology Report Generation

The potential benefits of automatic radiology report generation, such as reducing misdiagnosis rates and enhancing clinical diagnosis efficiency, are significant. However, existing data-driven methods lack essential medical prior knowledge, which hampers their performance. Moreover, establishing global correspondences between radiology images and related reports, while achieving local alignments between images correlated with prior knowledge and text, remains a challenging task. To address these shortcomings, we introduce a novel Eye Gaze Guided Cross-modal Alignment Network (EGGCA-Net) for generating accurate medical reports. Our approach incorporates prior knowledge from radiologists' Eye Gaze Region (EGR) to refine the fidelity and comprehensibility of report generation. Specifically, we design a Dual Fine-Grained Branch (DFGB) and a Multi-Task Branch (MTB) to collaboratively ensure the alignment of visual and textual semantics across multiple levels. To establish fine-grained alignment between EGR-related images and sentences, we introduce the Sentence Fine-grained Prototype Module (SFPM) within DFGB to capture cross-modal information at different levels. Additionally, to learn the alignment of EGR-related image topics, we introduce the Multi-task Feature Fusion Module (MFFM) within MTB to refine the encoder output information. Finally, a specifically designed label matching mechanism is designed to generate reports that are consistent with the anticipated disease states. The experimental outcomes indicate that the introduced methodology surpasses previous advanced techniques, yielding enhanced performance on two extensively used benchmark datasets: Open-i and MIMIC-CXR.

Combining Image Similarity and Predictive Artificial Intelligence Models to Decrease Subjectivity in Thyroid Nodule Diagnosis and Improve Malignancy Prediction

OBJECTIVES

To evaluate the efficacy of combining predictive artificial intelligence (AI) and image similarity model to risk stratify thyroid nodules, using retrospective external validation study.

METHODS

Two datasets were used to determine efficacy of the AI application. One was Stanford dataset ultrasound images of 192 nodules between April 2017 and May 2018 and the second was private practice consisting of 118 thyroid nodule images between January 2018 and December 2023. The nodules had definitive diagnosis by cytology or surgical pathology. The AI application was used to predict the diagnosis and American College of Radiology Thyroid Imaging and Data System (ACR TI-RADS) score.

RESULTS

In the Stanford dataset, the AI application predicted malignancies with sensitivity of 1.0 and specificity of 0.55. Positive predictive value (PPV) was 0.18 and negative predictive value (NPV) was 1.0. The Area Under the Curve - Receiver Operating Characteristic was 0.78. ACR TI-RADS based clinical recommendation had a polychoric correlation of 0.67. In the private dataset, the AI application predicted malignancies with sensitivity of 0.91 and specificity of 0.95. PPV was 0.8 and NPV was 0.98. The area under the curve - receiver operating characteristic was 0.93 and accuracy was 0.94. ACR TI-RADS based score had a polychoric correlation of 0.94.

CONCLUSION

The AI application showed good performance for sensitivity and NPV between the two datasets and demonstrated potential for 61.5% reduction in the need for fine needle aspiration and strong correlation to ACR TI-RADS. However, PPV was variable between the datasets possibly from variability in image selection and prevalence of malignancy. If implemented widely and consistently among various clinical settings, this could lead to decreased patient burden associated with an invasive procedure and possibly to decreased health care spending.

Virtual and augmented reality systems and three-dimensional printing of the renal model-novel trends to guide preoperative planning for renal cancer

Objective

This study aimed to explore the applications of three-dimensional (3D) technology, including virtual reality, augmented reality (AR), and 3D printing system, in the field of medicine, particularly in renal interventions for cancer treatment.

Methods

A specialized software transforms 2D medical images into precise 3D digital models, facilitating improved anatomical understanding and surgical planning. Patient-specific 3D printed anatomical models are utilized for preoperative planning, intraoperative guidance, and surgical education. AR technology enables the overlay of digital perceptions onto real-world surgical environments.

Results

Patient-specific 3D printed anatomical models have multiple applications, such as preoperative planning, intraoperative guidance, trainee education, and patient counseling. Virtual reality involves substituting the real world with a computer-generated 3D environment, while AR overlays digitally created perceptions onto the existing reality. The advances in 3D modeling technology have sparked considerable interest in their application to partial nephrectomy in the realm of renal cancer. 3D printing, also known as additive manufacturing, constructs 3D objects based on computer-aided design or digital 3D models. Utilizing 3D-printed preoperative renal models provides benefits for surgical planning, offering a more reliable assessment of the tumor's relationship with vital anatomical structures and enabling better preparation for procedures. AR technology allows surgeons to visualize patient-specific renal anatomical structures and their spatial relationships with surrounding organs by projecting CT/MRI images onto a live laparoscopic video. Incorporating patient-specific 3D digital models into healthcare enhances best practice, resulting in improved patient care, increased patient satisfaction, and cost saving for the healthcare system.

Greater benefits of deep learning-based computer-aided detection systems for finding small signals in 3D volumetric medical images

Purpose

Radiologists are tasked with visually scrutinizing large amounts of data produced by 3D volumetric imaging modalities. Small signals can go unnoticed during the 3D search because they are hard to detect in the visual periphery. Recent advances in machine learning and computer vision have led to effective computer-aided detection (CADe) support systems with the potential to mitigate perceptual errors.

Approach

Sixteen nonexpert observers searched through digital breast tomosynthesis (DBT) phantoms and single cross-sectional slices of the DBT phantoms. The 3D/2D searches occurred with and without a convolutional neural network (CNN)-based CADe support system. The model provided observers with bounding boxes superimposed on the image stimuli while they looked for a small microcalcification signal and a large mass signal. Eye gaze positions were recorded and correlated with changes in the area under the ROC curve (AUC).

Results

The CNN-CADe improved the 3D search for the small microcalcification signal ( Δ   AUC = 0.098 , p = 0.0002 ) and the 2D search for the large mass signal ( Δ   AUC = 0.076 , p = 0.002 ). The CNN-CADe benefit in 3D for the small signal was markedly greater than in 2D ( Δ Δ   AUC = 0.066 , p = 0.035 ). Analysis of individual differences suggests that those who explored the least with eye movements benefited the most from the CNN-CADe ( r = - 0.528 , p = 0.036 ). However, for the large signal, the 2D benefit was not significantly greater than the 3D benefit ( Δ Δ   AUC = 0.033 , p = 0.133 ).

Conclusion

The CNN-CADe brings unique performance benefits to the 3D (versus 2D) search of small signals by reducing errors caused by the underexploration of the volumetric data.

Classification images for aerial images capture visual expertise for binocular disparity and a prior for lighting from above

Using a novel approach to classification images (CIs), we investigated the visual expertise of surveyors for luminance and binocular disparity cues simultaneously after screening for stereoacuity. Stereoscopic aerial images of hedges and ditches were classified in 10,000 trials by six trained remote sensing surveyors and six novices. Images were heavily masked with luminance and disparity noise simultaneously. Hedge and ditch images had reversed disparity on around half the trials meaning hedges became ditch-like and vice versa. The hedge and ditch images were also flipped vertically on around half the trials, changing the direction of the light source and completing a 2 × 2 × 2 stimulus design. CIs were generated by accumulating the noise textures associated with "hedge" and "ditch" classifications, respectively, and subtracting one from the other. Typical CIs had a central peak with one or two negative side-lobes. We found clear differences in the amplitudes and shapes of perceptual templates across groups and noise-type, with experts prioritizing binocular disparity and using this more effectively. Contrariwise, novices used luminance cues more than experts meaning that task motivation alone could not explain group differences. Asymmetries in the luminance CIs revealed individual differences for lighting interpretation, with experts less prone to assume lighting from above, consistent with their training on aerial images of UK scenes lit by a southerly sun. Our results show that (i) dual noise in images can be used to produce simultaneous CI pairs, (ii) expertise for disparity cues does not depend on stereoacuity, (iii) CIs reveal the visual strategies developed by experts, (iv) top-down perceptual biases can be overcome with long-term learning effects, and (v) CIs have practical potential for directing visual training.

Training the New Radiologists: Approaches for Education

The field of Radiology is continually changing, requiring corresponding evolution in both medical student and resident training to adequately prepare the next generation of radiologists. With advancements in adult education theory and a deeper understanding of perception in imaging interpretation, expert educators are reshaping the training landscape by introducing innovative teaching methods to align with increased workload demands and emerging technologies. These include the use of peer and interdisciplinary teaching, gamification, case repositories, flipped-classroom models, social media, and drawing and comics. This publication aims to investigate these novel approaches and offer persuasive evidence supporting their incorporation into the updated Radiology curriculum.

Training to Analyze Functional Parameters with Fiberoptic Endoscopic Evaluation of Swallowing: A Scoping Review

Analyzing fiberoptic endoscopic evaluation of swallowing (FEES) is challenging and requires training to ensure the proficiency of health professionals and improve reliability. This scoping review aims to identify and map the available evidence on training health professionals to analyze FEES functional parameters. The method proposed by the Joanna Briggs Institute and the PRISMA-ScR guidelines were followed. The search was performed in MEDLINE, Cochrane Library, Embase, Web of Science, Scopus, CINAHL databases, and in the gray literature. Two blinded independent reviewers screened articles by title and abstract. Then, they read the full text of the included reports, considering the eligibility criteria. Data were extracted using a standardized form. Six studies met the established eligibility criteria, published between 2009 and 2022, with few participants. All these studies addressed training as part of the process to validate a rating scale. No standardized criteria were observed regarding the selection of experts and participants, training structure, and outcome measures to assess participants' competence. The reviewed literature indicates that training must be developed to equip students and health professionals who treat dysphagia, enabling them to analyze the functional parameters of the FEES, considering variables that may influence the participants' performance.

Impact of bias field correction on 0.35 T pelvic MR images: evaluation on generative adversarial network-based OARs' auto-segmentation and visual grading assessment

Purpose

Magnetic resonance imaging (MRI)-guided radiotherapy enables adaptive treatment plans based on daily anatomical changes and accurate organ visualization. However, the bias field artifact can compromise image quality, affecting diagnostic accuracy and quantitative analyses. This study aims to assess the impact of bias field correction on 0.35 T pelvis MRIs by evaluating clinical anatomy visualization and generative adversarial network (GAN) auto-segmentation performance.

Materials and methods

3D simulation MRIs from 60 prostate cancer patients treated on MR-Linac (0.35 T) were collected and preprocessed with the N4ITK algorithm for bias field correction. A 3D GAN architecture was trained, validated, and tested on 40, 10, and 10 patients, respectively, to auto-segment the organs at risk (OARs) rectum and bladder. The GAN was trained and evaluated either with the original or the bias-corrected MRIs. The Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (HD95th) were computed for the segmented volumes of each patient. The Wilcoxon signed-rank test assessed the statistical difference of the metrics within OARs, both with and without bias field correction. Five radiation oncologists blindly scored 22 randomly chosen patients in terms of overall image quality and visibility of boundaries (prostate, rectum, bladder, seminal vesicles) of the original and bias-corrected MRIs. Bennett's S score and Fleiss' kappa were used to assess the pairwise interrater agreement and the interrater agreement among all the observers, respectively.

Results

In the test set, the GAN trained and evaluated on original and bias-corrected MRIs showed DSC/HD95th of 0.92/5.63 mm and 0.92/5.91 mm for the bladder and 0.84/10.61 mm and 0.83/9.71 mm for the rectum. No statistical differences in the distribution of the evaluation metrics were found neither for the bladder (DSC: p = 0.07; HD95th: p = 0.35) nor for the rectum (DSC: p = 0.32; HD95th: p = 0.63). From the clinical visual grading assessment, the bias-corrected MRI resulted mostly in either no change or an improvement of the image quality and visualization of the organs' boundaries compared with the original MRI.

Conclusion

The bias field correction did not improve the anatomy visualization from a clinical point of view and the OARs' auto-segmentation outputs generated by the GAN.

Convenient model of hard tissue simulation for dental radiographic research and instruction

BACKGROUND

The authors describe a technique for building an alternative jawbone phantom using dental gypsum and rice for research and dental radiology instruction.

AIM

To investigate the potential of an alternative phantom to simulate the trabecular bone aspect of the human maxilla in periapical radiographs.

METHODS

Half-maxillary phantoms built from gypsum-ground rice were exposed to X-rays, and the resulting images (experimental group) were compared to standardized radiographic images produced from dry human maxillary bone (control group) (n = 7). The images were blindly assessed according to strict criteria by three examiners for the usual trabecular aspects of the surrounding bone, and significant differences between groups and in assessment reliability were compared using Fisher's exact and kappa tests (α = 0.05).

RESULTS

The differences in the trabecular aspects between groups were not statistically significant. In addition, interobserver agreement among observers was 0.43 and 0.51 for the control and experimental groups, respectively, whereas intraobserver agreement was 0.71 and 0.73, respectively.

CONCLUSION

The tested phantom seemed to demonstrate potential for trabecular bone image simulation on maxillary periapical radiographs.

Machine learning classification of diagnostic accuracy in pathologists interpreting breast biopsies

OBJECTIVE

This study explores the feasibility of using machine learning to predict accurate versus inaccurate diagnoses made by pathologists based on their spatiotemporal viewing behavior when evaluating digital breast biopsy images.

MATERIALS AND METHODS

The study gathered data from 140 pathologists of varying experience levels who each reviewed a set of 14 digital whole slide images of breast biopsy tissue. Pathologists' viewing behavior, including zooming and panning actions, was recorded during image evaluation. A total of 30 features were extracted from the viewing behavior data, and 4 machine learning algorithms were used to build classifiers for predicting diagnostic accuracy.

RESULTS

The Random Forest classifier demonstrated the best overall performance, achieving a test accuracy of 0.81 and area under the receiver-operator characteristic curve of 0.86. Features related to attention distribution and focus on critical regions of interest were found to be important predictors of diagnostic accuracy. Further including case-level and pathologist-level information incrementally improved classifier performance.

DISCUSSION

Results suggest that pathologists' viewing behavior during digital image evaluation can be leveraged to predict diagnostic accuracy, affording automated feedback and decision support systems based on viewing behavior to aid in training and, ultimately, clinical practice. They also carry implications for basic research examining the interplay between perception, thought, and action in diagnostic decision-making.

CONCLUSION

The classifiers developed herein have potential applications in training and clinical settings to provide timely feedback and support to pathologists during diagnostic decision-making. Further research could explore the generalizability of these findings to other medical domains and varied levels of expertise.

Computer-extracted global radiomic features can predict the radiologists' first impression about the abnormality of a screening mammogram

OBJECTIVE

Radiologists can detect the gist of abnormal based on their rapid initial impression on a mammogram (ie, global gist signal [GGS]). This study explores (1) whether global radiomic (ie, computer-extracted) features can predict the GGS; and if so, (ii) what features are the most important drivers of the signals.

METHODS

The GGS of cases in two extreme conditions was considered: when observers detect a very strong gist (high-gist) and when the gist of abnormal was not/poorly perceived (low-gist). Gist signals/scores from 13 observers reading 4191 craniocaudal mammograms were collected. As gist is a noisy signal, the gist scores from all observers were averaged and assigned to each image. The high-gist and low-gist categories contained all images in the fourth and first quartiles, respectively. One hundred thirty handcrafted global radiomic features (GRFs) per mammogram were extracted and utilized to construct eight separate machine learning random forest classifiers (All, Normal, Cancer, Prior-1, Prior-2, Missed, Prior-Visible, and Prior-Invisible) for characterizing high-gist from low-gist images. The models were trained and validated using the 10-fold cross-validation approach. The models' performances were evaluated by the area under receiver operating characteristic curve (AUC). Important features for each model were identified through a scree test.

RESULTS

The Prior-Visible model achieved the highest AUC of 0.84 followed by the Prior-Invisible (0.83), Normal (0.82), Prior-1 (0.81), All (0.79), Prior-2 (0.77), Missed (0.75), and Cancer model (0.69). Cluster shade, standard deviation, skewness, kurtosis, and range were identified to be the most important features.

CONCLUSIONS

Our findings suggest that GRFs can accurately classify high- from low-gist images.

ADVANCES IN KNOWLEDGE

Global mammographic radiomic features can accurately predict high- from low-gist images with five features identified to be valuable in describing high-gist images. These are critical in providing better understanding of the mammographic image characteristics that drive the strength of the GGSs which could be exploited to advance breast cancer (BC) screening and risk prediction, enabling early detection and treatment of BC thereby further reducing BC-related deaths.

A short 18F-FDG imaging window triple injection neuroimaging protocol for parametric mapping in PET

BACKGROUND

In parametric PET, kinetic parameters are extracted from dynamic PET images. It is not commonly used in clinical practice because of long scan times and the requirement for an arterial input function (AIF). To address these limitations, we designed an 18F-fluorodeoxyglucose (18F-FDG) triple injection dynamic PET protocol for brain imaging with a standard field of view PET scanner using a 24-min imaging window and an input function modeled using measurements from a region of interest placed over the left ventricle.

METHODS

To test the protocol in 6 healthy participants, we examined the quality of voxel-based maps of kinetic parameters in the brain generated using the two-tissue compartment model and compared estimated parameter values with previously published values. We also utilized data from a 36-min validation imaging window to compare (1) the modeled AIF against the input function measured in the validation window; and (2) the net influx rate ([Formula: see text]) computed using parameter estimates from the short imaging window against the net influx rate obtained using Patlak analysis in the validation window.

RESULTS

Compared to the AIF measured in the validation window, the input function estimated from the short imaging window achieved a mean area under the curve error of 9%. The voxel-wise Pearson's correlation between [Formula: see text] estimates from the short imaging window and the validation imaging window exceeded 0.95.

CONCLUSION

The proposed 24-min triple injection protocol enables parametric 18F-FDG neuroimaging with noninvasive estimation of the AIF from cardiac images using a standard field of view PET scanner.

Artificial intelligence in fracture detection with different image modalities and data types: A systematic review and meta-analysis

Artificial Intelligence (AI), encompassing Machine Learning and Deep Learning, has increasingly been applied to fracture detection using diverse imaging modalities and data types. This systematic review and meta-analysis aimed to assess the efficacy of AI in detecting fractures through various imaging modalities and data types (image, tabular, or both) and to synthesize the existing evidence related to AI-based fracture detection. Peer-reviewed studies developing and validating AI for fracture detection were identified through searches in multiple electronic databases without time limitations. A hierarchical meta-analysis model was used to calculate pooled sensitivity and specificity. A diagnostic accuracy quality assessment was performed to evaluate bias and applicability. Of the 66 eligible studies, 54 identified fractures using imaging-related data, nine using tabular data, and three using both. Vertebral fractures were the most common outcome (n = 20), followed by hip fractures (n = 18). Hip fractures exhibited the highest pooled sensitivity (92%; 95% CI: 87-96, p< 0.01) and specificity (90%; 95% CI: 85-93, p< 0.01). Pooled sensitivity and specificity using image data (92%; 95% CI: 90-94, p< 0.01; and 91%; 95% CI: 88-93, p < 0.01) were higher than those using tabular data (81%; 95% CI: 77-85, p< 0.01; and 83%; 95% CI: 76-88, p < 0.01), respectively. Radiographs demonstrated the highest pooled sensitivity (94%; 95% CI: 90-96, p < 0.01) and specificity (92%; 95% CI: 89-94, p< 0.01). Patient selection and reference standards were major concerns in assessing diagnostic accuracy for bias and applicability. AI displays high diagnostic accuracy for various fracture outcomes, indicating potential utility in healthcare systems for fracture diagnosis. However, enhanced transparency in reporting and adherence to standardized guidelines are necessary to improve the clinical applicability of AI. Review Registration: PROSPERO (CRD42021240359).

Accuracy of an artificial intelligence-enabled diagnostic assistance device in recognizing normal chest radiographs: a service evaluation

Objectives

Artificial intelligence (AI) enabled devices may be able to optimize radiologists' productivity by identifying normal and abnormal chest X-rays (CXRs) for triaging. In this service evaluation, we investigated the accuracy of one such AI device (qXR).

Methods

A randomly sampled subset of general practice and outpatient-referred frontal CXRs from a National Health Service Trust was collected retrospectively from examinations conducted during November 2022 to January 2023. Ground truth was established by consensus between 2 radiologists. The main objective was to estimate negative predictive value (NPV) of AI.

Results

A total of 522 CXRs (458 [87.74%] normal CXRs) from 522 patients (median age, 64 years [IQR, 49-77]; 305 [58.43%] female) were analysed. AI predicted 348 CXRs as normal, of which 346 were truly normal (NPV: 99.43% [95% CI, 97.94-99.93]). The sensitivity, specificity, positive predictive value, and area under the ROC curve of AI were found to be 96.88% (95% CI, 89.16-99.62), 75.55% (95% CI, 71.34-79.42), 35.63% (95% CI, 28.53-43.23), and 91.92% (95% CI, 89.38-94.45), respectively. A sensitivity analysis was conducted to estimate NPV by varying assumptions of the prevalence of normal CXRs. The NPV ranged from 88.96% to 99.54% as prevalence increased.

Conclusions

The AI device recognized normal CXRs with high NPV and has the potential to increase radiologists' productivity.

Advances in knowledge

There is a need for more evidence on the utility of AI-enabled devices in identifying normal CXRs. This work adds to such limited evidence and enables researchers to plan studies to further evaluate the impact of such devices.

Factors related to the performance of laypersons diagnosing pigmented skin cancer: an explorative study

It is important but challenging for prospective health professionals to learn the visual distinction between potentially harmful and harmless skin lesions, such as malignant melanomas and benign nevi. Knowledge about factors related to diagnostic performance is sparse but a prerequisite for designing and evaluating evidence-based educational interventions. Hence, this study explored how the characteristics of 240 skin lesions, the number of classified lesions and the response times of 137 laypeople were related to performance in diagnosing pigmented skin cancer. Our results showed large differences between the lesions, as some were classified correctly by more than 90% and others by less than 10% of the participants. A t-test showed that for melanomas, the correct diagnosis was provided significantly more often than for nevi. Furthermore, we found a significant Pearson correlation between the number of solved tasks and performance in the first 50 diagnostic tasks. Finally, t-tests for investigating the response times revealed that compared to true decisions, participants spent longer on false-negative but not on false-positive decisions. These results provide novel knowledge about performance-related factors that can be useful when designing diagnostic tests and learning interventions for melanoma detection.

A review on the recent applications of synthetic biopolymers in 3D printing for biomedical applications

3D printing technology is an emerging method that gained extensive attention from researchers worldwide, especially in the health and medical fields. Biopolymers are an emerging class of materials offering excellent properties and flexibility for additive manufacturing. Biopolymers are widely used in biomedical applications in biosensing, immunotherapy, drug delivery, tissue engineering and regeneration, implants, and medical devices. Various biodegradable and non-biodegradable polymeric materials are considered as bio-ink for 3d printing. Here, we offer an extensive literature review on the current applications of synthetic biopolymers in the field of 3D printing. A trend in the publication of biopolymers in the last 10 years are focused on the review by analyzing more than 100 publications. Their application and classification based on biodegradability are discussed. The various studies, along with their practical applications, are elaborated in the subsequent sections for polyethylene, polypropylene, polycaprolactone, polylactide, etc. for biomedical applications. The disadvantages of various biopolymers are discussed, and future perspectives like combating biocompatibility problems using 3D printed biomaterials to build compatible prosthetics are also discussed and the potential application of using resin with the combination of biopolymers to build customized implants, personalized drug delivery systems and organ on a chip technologies are expected to open a new set of chances for the development of healthcare and regenerative medicine in the future.

Eye-Gaze-Guided Vision Transformer for Rectifying Shortcut Learning

Learning harmful shortcuts such as spurious correlations and biases prevents deep neural networks from learning meaningful and useful representations, thus jeopardizing the generalizability and interpretability of the learned representation. The situation becomes even more serious in medical image analysis, where the clinical data are limited and scarce while the reliability, generalizability and transparency of the learned model are highly required. To rectify the harmful shortcuts in medical imaging applications, in this paper, we propose a novel eye-gaze-guided vision transformer (EG-ViT) model which infuses the visual attention from radiologists to proactively guide the vision transformer (ViT) model to focus on regions with potential pathology rather than spurious correlations. To do so, the EG-ViT model takes the masked image patches that are within the radiologists' interest as input while has an additional residual connection to the last encoder layer to maintain the interactions of all patches. The experiments on two medical imaging datasets demonstrate that the proposed EG-ViT model can effectively rectify the harmful shortcut learning and improve the interpretability of the model. Meanwhile, infusing the experts' domain knowledge can also improve the large-scale ViT model's performance over all compared baseline methods with limited samples available. In general, EG-ViT takes the advantages of powerful deep neural networks while rectifies the harmful shortcut learning with human expert's prior knowledge. This work also opens new avenues for advancing current artificial intelligence paradigms by infusing human intelligence.

Machine learning-based diagnosis and risk classification of coronary artery disease using myocardial perfusion imaging SPECT: A radiomics study

This study aimed to investigate the diagnostic performance of machine learning-based radiomics analysis to diagnose coronary artery disease status and risk from rest/stress Myocardial Perfusion Imaging (MPI) single-photon emission computed tomography (SPECT). A total of 395 patients suspicious of coronary artery disease who underwent 2-day stress-rest protocol MPI SPECT were enrolled in this study. The left ventricle myocardium, excluding the cardiac cavity, was manually delineated on rest and stress images to define a volume of interest. Added to clinical features (age, sex, family history, diabetes status, smoking, and ejection fraction), a total of 118 radiomics features, were extracted from rest and stress MPI SPECT images to establish different feature sets, including Rest-, Stress-, Delta-, and Combined-radiomics (all together) feature sets. The data were randomly divided into 80% and 20% subsets for training and testing, respectively. The performance of classifiers built from combinations of three feature selections, and nine machine learning algorithms was evaluated for two different diagnostic tasks, including 1) normal/abnormal (no CAD vs. CAD) classification, and 2) low-risk/high-risk CAD classification. Different metrics, including the area under the ROC curve (AUC), accuracy (ACC), sensitivity (SEN), and specificity (SPE), were reported for models' evaluation. Overall, models built on the Stress feature set (compared to other feature sets), and models to diagnose the second task (compared to task 1 models) revealed better performance. The Stress-mRMR-KNN (feature set-feature selection-classifier) reached the highest performance for task 1 with AUC, ACC, SEN, and SPE equal to 0.61, 0.63, 0.64, and 0.6, respectively. The Stress-Boruta-GB model achieved the highest performance for task 2 with AUC, ACC, SEN, and SPE of 0.79, 0.76, 0.75, and 0.76, respectively. Diabetes status from the clinical feature family, and dependence count non-uniformity normalized, from the NGLDM family, which is representative of non-uniformity in the region of interest were the most frequently selected features from stress feature set for CAD risk classification. This study revealed promising results for CAD risk classification using machine learning models built on MPI SPECT radiomics. The proposed models are helpful to alleviate the labor-intensive MPI SPECT interpretation process regarding CAD status and can potentially expedite the diagnostic process.

Factors that influence development of speech pathology skills required for videofluoroscopic swallowing studies

BACKGROUND

Perceptual, cognitive and previous clinical experience may influence a novice Videofluoroscopic Swallowing Study (VFSS) analyst's trajectory towards competency. Understanding these factors may allow trainees to be better prepared for VFSS training and may allow training to be developed to accommodate differences between trainees.

AIMS

This study explored a range of factors previously suggested in the literature as influencing the development of novice analysts' VFSS skills. We hypothesised that knowledge of swallow anatomy and physiology, visual perceptual skills, self-efficacy and interest, and prior clinical exposure would all influence VFSS novice analysts' skill development.

METHODS & PROCEDURES

Participants were undergraduate speech pathology students recruited from an Australian university, who had completed the required theoretical units in dysphagia. Data assessing the factors of interest were collected-the participants identified anatomical structures on a still radiographic image, completed a physiology questionnaire, completed subsections of the Developmental Test of Visual Processing-Adults, self-reported the number of dysphagia cases they managed on placement, and self-rated their confidence and interest. Data for 64 participants relating to the factors of interest were compared with their ability to accurately identify swallowing impairments following 15 h of VFSS analytical training, using correlation and regression analysis.

OUTCOMES & RESULTS

Success in VFSS analytical training was best predicted by clinical exposure to dysphagia cases and the ability to identify anatomical landmarks on still radiographic images.

CONCLUSIONS & IMPLICATIONS

Novice analysts vary in the acquisition of beginner-level VFSS analytical skill. Our findings suggest that speech pathologists who are new to VFSS may benefit from clinical exposure to dysphagia cases, sound foundational knowledge of anatomy relevant to swallowing and the ability to see the anatomical landmarks on still radiographic images. Further research is required to equip VFSS trainers and trainees for training, to understand differences between learners during skill development.

WHAT THIS PAPER ADDS

What is already known on the subject The existing literature suggests that no vice Video fluoroscopic Swallowing Study (VFSS) analysts training may be influenced by their personal characteristics and experience. What this study adds This study found that student clinicians, clinical exposure to dysphagia cases and their ability to identify anatomical landmarks relevant to swallowing on still radiographic images prior to training best predicted their ability to identify swallowing impairments after training. What are the clinical implications of this work? Given the expense of training health professionals, further research is required into the factors that successfully prepare clinicians for VFSS training, including clinical exposure, foundational knowledge of anatomy relevant to swallowing and the ability to identify the anatomical landmarks on still radiographic images.

Automatic Spine Segmentation and Parameter Measurement for Radiological Analysis of Whole-Spine Lateral Radiographs Using Deep Learning and Computer Vision

Radiographic examination is essential for diagnosing spinal disorders, and the measurement of spino-pelvic parameters provides important information for the diagnosis and treatment planning of spinal sagittal deformities. While manual measurement methods are the golden standard for measuring parameters, they can be time consuming, inefficient, and rater dependent. Previous studies that have used automatic measurement methods to alleviate the downsides of manual measurements showed low accuracy or could not be applied to general films. We propose a pipeline for automated measurement of spinal parameters by combining a Mask R-CNN model for spine segmentation with computer vision algorithms. This pipeline can be incorporated into clinical workflows to provide clinical utility in diagnosis and treatment planning. A total of 1807 lateral radiographs were used for the training (n = 1607) and validation (n = 200) of the spine segmentation model. An additional 200 radiographs, which were also used for validation, were examined by three surgeons to evaluate the performance of the pipeline. Parameters automatically measured by the algorithm in the test set were statistically compared to parameters measured manually by the three surgeons. The Mask R-CNN model achieved an average precision at 50% intersection over union (AP50) of 96.2% and a Dice score of 92.6% for the spine segmentation task in the test set. The mean absolute error values of the spino-pelvic parameters measurement results were within the range of 0.4° (pelvic tilt) to 3.0° (lumbar lordosis, pelvic incidence), and the standard error of estimate was within the range of 0.5° (pelvic tilt) to 4.0° (pelvic incidence). The intraclass correlation coefficient values ranged from 0.86 (sacral slope) to 0.99 (pelvic tilt, sagittal vertical axis).

Providing comparison normal examples alongside pathologic thoracic radiographic cases can improve veterinary students' ability to identify abnormal findings or diagnose disease

Learning by comparison is a frequently employed education strategy used across many disciplines and levels. Interpreting radiographs requires both skills of perception and pattern recognition, which makes comparison techniques particularly useful in this field. In this randomized, prospective, parallel-group study, students enrolled in second and third-year radiology veterinary courses were given a case-based thoracic radiographic interpretation assignment. A cohort of the participants was given cases with side-by-side comparison normal images while the other cohort only had access to the cases. Twelve cases in total were presented to the students, with 10 cases depicting examples of common thoracic pathologies, while 2 cases were examples of normal. Radiographs of both feline and canine species were represented. Correctness of response to multiple choice questions was tracked, as was year and group (group 1: non compare, Control; group 2: compare, Intervention). Students assigned to group 1 had a lower percentage of correct answers than students assigned to group 2 (45% Control vs. 52% Intervention; P = 0.01). This indicates that side-by-side comparison to a normal example is helpful in identifying disease. No statistical significance was noted for the correctness of responses according to the year of training (P = 0.90). The overall poor performance on the assignment, regardless of group or year, shows that students in the early years of undergraduate veterinary radiology training struggle with the interpretation of common pathologies, likely a result of a lack of exposure to a multitude of cases and normal variants.

Recent Advances in Deep Learning and Medical Imaging for Head and Neck Cancer Treatment: MRI, CT, and PET Scans

Deep learning techniques have been developed for analyzing head and neck cancer imaging. This review covers deep learning applications in cancer imaging, emphasizing tumor detection, segmentation, classification, and response prediction. In particular, advanced deep learning techniques, such as convolutional autoencoders, generative adversarial networks (GANs), and transformer models, as well as the limitations of traditional imaging and the complementary roles of deep learning and traditional techniques in cancer management are discussed. Integration of radiomics, radiogenomics, and deep learning enables predictive models that aid in clinical decision-making. Challenges include standardization, algorithm interpretability, and clinical validation. Key gaps and controversies involve model generalizability across different imaging modalities and tumor types and the role of human expertise in the AI era. This review seeks to encourage advancements in deep learning applications for head and neck cancer management, ultimately enhancing patient care and outcomes.

Deep Learning Algorithm for Differentiating Patients with a Healthy Liver from Patients with Liver Lesions Based on MR Images

The problems in diagnosing the state of a vital organ such as the liver are complex and remain unresolved. These problems are underscored by frequently published studies on this issue. At the same time, demand for imaging diagnostics, preferably using a method that can detect the disease at the earliest possible stage, is constantly increasing. In this paper, we present liver diseases in the context of diagnosis, diagnostic problems, and possible elimination. We discuss the dataset and methods and present the stages of the pipeline we developed, leading to multiclass segmentation of the liver in multiparametric MR image into lesions and normal tissue. Finally, based on the processing results, each case is classified as either a healthy liver or a liver with lesions. For the training set, the AUC ROC is 0.925 (standard error 0.013 and a p-value less than 0.001), and for the test set, the AUC ROC is 0.852 (standard error 0.039 and a p-value less than 0.001). Further refinements to the proposed pipeline are also discussed. The proposed approach could be used in the detection of focal lesions in the liver and the description of liver tumors. Practical application of the developed multi-class segmentation method represents a key step toward standardizing the medical evaluation of focal lesions in the liver.

Learning effects in visual grading assessment of model-based reconstruction algorithms in abdominal Computed Tomography

Objectives

Images reconstructed with higher strengths of iterative reconstruction algorithms may impair radiologists' subjective perception and diagnostic performance due to changes in the amplitude of different spatial frequencies of noise. The aim of the present study was to ascertain if radiologists can learn to adapt to the unusual appearance of images produced by higher strengths of Advanced modeled iterative reconstruction algorithm (ADMIRE).

Methods

Two previously published studies evaluated the performance of ADMIRE in non-contrast and contrast-enhanced abdominal CT. Images from 25 (first material) and 50 (second material) patients, were reconstructed with ADMIRE strengths 3, 5 (AD3, AD5) and filtered back projection (FBP). Radiologists assessed the images using image criteria from the European guidelines for quality criteria in CT. To ascertain if there was a learning effect, new analyses of data from the two studies was performed by introducing a time variable in the mixed-effects ordinal logistic regression model.

Results

In both materials, a significant negative attitude to ADMIRE 5 at the beginning of the viewing was strengthened during the progress of the reviews for both liver parenchyma (first material: -0.70, p < 0.01, second material: -0.96, p < 0.001) and overall image quality (first material:-0.59, p < 0.05, second material::-1.26, p < 0.001). For ADMIRE 3, an early positive attitude for the algorithm was noted, with no significant change over time for all criteria except one (overall image quality), where a significant negative trend over time (-1.08, p < 0.001) was seen in the second material.

Conclusions

With progression of reviews in both materials, an increasing dislike for ADMIRE 5 images was apparent for two image criteria. In this time perspective (weeks or months), no learning effect towards accepting the algorithm could be demonstrated.

A comparison of perceived image quality between computer display monitors and augmented reality smart glasses

INTRODUCTION

Augmented-reality (AR) smart glasses provide an alternative to standard computer display monitors (CDM). AR smart glasses may provide an opportunity to improve visualisation during fluoroscopy and interventional radiology (IR) procedures when there can be difficulty in viewing intra-procedural images on a CDM. The aim of this study was to evaluate radiographer perception of image quality (IQ) when comparing CDM and AR smart glasses.

METHODS

38 radiographers attending an international congress evaluated ten fluoroscopic-guided surgery and IR images on both a CDM (1920 × 1200 pixels) and a set of Epson Moverio BT-40 AR smart glasses (1920 × 1080 pixels). Participants provided oral responses to pre-defined IQ questions generated by study researchers. Summative IQ scores for each participant/image were compared between CDM and AR smart glasses.

RESULTS

Of the 38 participants, the mean age was 39 ± 1 years. 23 (60.5%) participants required corrective glasses. In terms of generalisability, participants were from 12 different countries, the majority (n = 9, 23.7%) from the United Kingdom. For eight out of ten images, the AR smart glasses demonstrated a statistically significant increase in perceived IQ (median [IQR] 2.0 [-1.0 to 7.0] points) when compared to the CDM.

CONCLUSION

AR smart glasses appear to show improvements in perceived IQ when compared to a CDM. AR smart glasses could provide an option for improving the experiences of radiographers involved in image-guided procedures and should be subject to further clinical evaluations.

IMPLICATIONS FOR PRACTICE

Opportunities exist to improve perceived IQ for radiographers when reviewing fluoroscopy and IR images. AR smart glasses should be further evaluated as a potential opportunity to improve practice when visual attention is split between positioning equipment and image review.

Anatomical Correlation for Focused Assessment With Sonography in Trauma

When learning the Focused Assessment with Sonography in Trauma (FAST) exam, anatomical orientation can be difficult, especially in the subxiphoid and upper quadrant views. To facilitate understanding in these areas, a novel in-situ cadaver dissection was used to demonstrate anatomy related to the FAST exam. In situ, because the structures remained in normal positions with adjacent organs, layers, and spaces clearly visible from the point of view of the ultrasound probe. These views were then correlated with what was seen on the ultrasound screen. The right upper quadrant and subxiphoid anatomy were viewed in a mirror to match the ultrasound images, and the left upper quadrant was viewed directly from the examiner's position, also matching the view on the ultrasound screen. The in-situ cadaver dissection was developed as a resource to correlate FAST exam ultrasound images in the upper quadrant and subxiphoid regions with related cadaver anatomy.

Automatic determination of pubertal growth spurts based on the cervical vertebral maturation staging using deep convolutional neural networks

BACKGROUND

This study aimed to develop a deep convolutional neural network (CNN) for automatic classification of pubertal growth spurts using cervical vertebral maturation (CVM) staging based on the lateral cephalograms of an Iranian subpopulation.

MATERIAL AND METHODS

Cephalometric radiographs were collected from 1846 eligible patients (aged 5-18 years) referred to the orthodontic department of Hamadan University of Medical Sciences. These images were labeled by two experienced orthodontists. Two scenarios, including two- and three-class (pubertal growth spurts using CVM), were considered as the output for the classification task. The cropped image of the second to fourth cervical vertebrae was used as input to the network. After the preprocessing, the augmentation step, and hyperparameter tuning, the networks were trained with initial random weighting and transfer learning. Finally, the best architecture among the different architectures was determined based on the accuracy and F-score criteria.

RESULTS

The CNN based on the ConvNeXtBase-296 architecture had the highest accuracy for automatically assessing pubertal growth spurts based on CVM staging in both three-class (82% accuracy) and two-class (93% accuracy) scenarios. Given the limited amount of data available for training the target networks for most of the architectures in use, transfer learning improves predictive performance.

CONCLUSIONS

The results of this study confirm the potential of CNNs as an auxiliary diagnostic tool for intelligent assessment of skeletal maturation staging with high accuracy even with a relatively small number of images. Considering the development of orthodontic science toward digitalization, the development of such intelligent decision systems is proposed.

The Role of an Artificial Intelligence Method of Improving the Diagnosis of Neoplasms by Colonoscopy

BACKGROUND

Colorectal cancer (CRC) is the third most common cancer worldwide. Colonoscopy is the gold standard examination that reduces the morbidity and mortality of CRC. Artificial intelligence (AI) could be useful in reducing the errors of the specialist and in drawing attention to the suspicious area.

METHODS

A prospective single-center randomized controlled study was conducted in an outpatient endoscopy unit with the aim of evaluating the usefulness of AI-assisted colonoscopy in PDR and ADR during the day time. It is important to understand how already available CADe systems improve the detection of polyps and adenomas in order to make a decision about their routine use in practice. In the period from October 2021 to February 2022, 400 examinations (patients) were included in the study. One hundred and ninety-four patients were examined using the ENDO-AID CADe artificial intelligence device (study group), and 206 patients were examined without the artificial intelligence (control group).

RESULTS

None of the analyzed indicators (PDR and ADR during morning and afternoon colonoscopies) showed differences between the study and control groups. There was an increase in PDR during afternoon colonoscopies, as well as ADR during morning and afternoon colonoscopies.

CONCLUSIONS

Based on our results, the use of AI systems in colonoscopies is recommended, especially in circumstances of an increase of examinations. Additional studies with larger groups of patients at night are needed to confirm the already available data.

Multiple expressions of "expert" abnormality gist in novices following perceptual learning

With a brief half-second presentation, a medical expert can determine at above chance levels whether a medical scan she sees is abnormal based on a first impression arising from an initial global image process, termed "gist." The nature of gist processing is debated but this debate stems from results in medical experts who have years of perceptual experience. The aim of the present study was to determine if gist processing for medical images occurs in naïve (non-medically trained) participants who received a brief perceptual training and to tease apart the nature of that gist signal. We trained 20 naïve participants on a brief perceptual-adaptive training of histology images. After training, naïve observers were able to obtain abnormality detection and abnormality categorization above chance, from a brief 500 ms masked presentation of a histology image, hence showing "gist." The global signal demonstrated in perceptually trained naïve participants demonstrated multiple dissociable components, with some of these components relating to how rapidly naïve participants learned a normal template during perceptual learning. We suggest that multiple gist signals are present when experts view medical images derived from the tens of thousands of images that they are exposed to throughout their training and careers. We also suggest that a directed learning of a normal template may produce better abnormality detection and identification in radiologists and pathologists.

Visual adaptation to medical images: a comparison of digital mammography and tomosynthesis

Purpose

Radiologists and other image readers spend prolonged periods inspecting medical images. The visual system can rapidly adapt or adjust sensitivity to the images that an observer is currently viewing, and previous studies have demonstrated that this can lead to pronounced changes in the perception of mammogram images. We compared these adaptation effects for images from different imaging modalities to explore both general and modality-specific consequences of adaptation in medical image perception.

Approach

We measured perceptual changes induced by adaptation to images acquired by digital mammography (DM) or digital breast tomosynthesis (DBT), which have both similar and distinct textural properties. Participants (nonradiologists) adapted to images from the same patient acquired from each modality or for different patients with American College of Radiology-Breast Imaging Reporting and Data System (BI-RADS) classification of dense or fatty tissue. The participants then judged the appearance of composite images formed by blending the two adapting images (i.e., DM versus DBT or dense versus fatty in each modality).

Results

Adaptation to either modality produced similar significant shifts in the perception of dense and fatty textures, reducing the salience of the adapted component in the test images. In side-by-side judgments, a modality-specific adaptation effect was not observed. However, when the images were directly fixated during adaptation and testing, so that the textural differences between the modalities were more visible, significantly different changes in the sensitivity to the noise in the images were observed.

Conclusions

These results confirm that observers can readily adapt to the visual properties or spatial textures of medical images in ways that can bias their perception of the images, and that adaptation can also be selective for the distinctive visual features of images acquired by different modalities.

The Oddity Detection in Diverse Scenes (ODDS) database: Validated real-world scenes for studying anomaly detection

Many applied screening tasks (e.g., medical image or baggage screening) involve challenging searches for which standard laboratory search is rarely equivalent. For example, whereas laboratory search frequently requires observers to look for precisely defined targets among isolated, non-overlapping images randomly arrayed on clean backgrounds, medical images present unspecified targets in noisy, yet spatially regular scenes. Those unspecified targets are typically oddities, elements that do not belong. To develop a closer laboratory analogue to this, we created a database of scenes containing subtle, ill-specified "oddity" targets. These scenes have similar perceptual densities and spatial regularities to those found in expert search tasks, and each includes 16 variants of the unedited scene wherein an oddity (a subtle deformation of the scene) is hidden. In Experiment 1, eight volunteers searched thousands of scene variants for an oddity. Regardless of their search accuracy, they were then shown the highlighted anomaly and rated its subtlety. Subtlety ratings reliably predicted search performance (accuracy and response times) and did so better than image statistics. In Experiment 2, we conducted a conceptual replication in which a larger group of naïve searchers scanned subsets of the scene variants. Prior subtlety ratings reliably predicted search outcomes. Whereas medical image targets are difficult for naïve searchers to detect, our database contains thousands of interior and exterior scenes that vary in difficulty, but are nevertheless searchable by novices. In this way, the stimuli will be useful for studying visual search as it typically occurs in expert domains: Ill-specified search for anomalies in noisy displays.

Factors affecting interpretation of dental radiographs

OBJECTIVES

To identify the factors influencing errors in the interpretation of dental radiographs.

METHODS

A protocol was registered on Prospero. All studies published until May 2022 were included in this review. The search of the electronic databases spanned Ovid Medline, PubMed, EMBASE, Web of Science and Scopus. The quality of the studies was assessed using the MMAT tool. Due to the heterogeneity of the included studies, a meta-analysis was not conducted.

RESULTS

The search yielded 858 articles, of which eight papers met the inclusion and exclusion criteria and were included in the systematic review. These studies assessed the factors influencing the accuracy of the interpretation of dental radiographs. Six factors were identified as being significant that affected the occurrence of interpretation errors. These include clinical experience, clinical knowledge, and technical ability, case complexity, time pressure, location and duration of dental education and training and cognitive load.

CONCLUSIONS

The occurrence of interpretation errors has not been widely investigated in dentistry. The factors identified in this review are interlinked. Further studies are needed to better understand the extent of the occurrence of interpretive errors and their impact on the practice of dentistry.

Deep learning-based image reconstruction improves radiologic evaluation of pituitary axis and cavernous sinus invasion in pituitary adenoma

PURPOSE

To compare performance of 1-mm deep learning reconstruction (DLR) with 3-mm routine MRI imaging for the delineation of pituitary axis and identification of cavernous sinus invasion for pituitary macroadenoma.

METHOD

This retrospective study included 104 patients (59.4 ± 13.1 years; 46 women) who underwent an MRI protocol including 1-mm deep learning-reconstructed and 3-mm routine images for evaluating pituitary adenoma between August 2019 and October 2020. Five readers (24, 9, 2 years, and <1 year of experience) assessed the delineation of pituitary axis (gland and stalk) and the presence of cavernous sinus invasion for using a pairwise design. The signal-to-noise ratio (SNR) was measured. Diagnostic performance as well as image preference data were analysed and compared according to the readers' experience using the McNemar test.

RESULTS

For delineation of normal pituitary axis, all readers preferred thin 1-mm DLR MRI over 3-mm MRI (overall superiority, 55.8 %, P <.001), with this preference being greater in the less experienced readers (92.3 % vs. 55.8 % [expert], P <.001). The readers showed higher diagnostic performance for cavernous sinus invasion on 1-mm (AUC, 0.91 and 0.92) than on 3-mm imaging (AUC, 0.87 and 0.88). The SNR of the 1-mm DLR was 1.21-fold higher than that of the routine 3-mm imaging.

CONCLUSION

Deep learning reconstruction-based 1-mm imaging demonstrates improved image quality and better delineation of microstructure in the sellar fossa and is preferred by both radiologists and non-radiologist physicians, especially in less experienced readers.

Idiosyncratic biases in the perception of medical images

Introduction

Radiologists routinely make life-altering decisions. Optimizing these decisions has been an important goal for many years and has prompted a great deal of research on the basic perceptual mechanisms that underlie radiologists' decisions. Previous studies have found that there are substantial individual differences in radiologists' diagnostic performance (e.g., sensitivity) due to experience, training, or search strategies. In addition to variations in sensitivity, however, another possibility is that radiologists might have perceptual biases-systematic misperceptions of visual stimuli. Although a great deal of research has investigated radiologist sensitivity, very little has explored the presence of perceptual biases or the individual differences in these.

Methods

Here, we test whether radiologists' have perceptual biases using controlled artificial and Generative Adversarial Networks-generated realistic medical images. In Experiment 1, observers adjusted the appearance of simulated tumors to match the previously shown targets. In Experiment 2, observers were shown with a mix of real and GAN-generated CT lesion images and they rated the realness of each image.

Results

We show that every tested individual radiologist was characterized by unique and systematic perceptual biases; these perceptual biases cannot be simply explained by attentional differences, and they can be observed in different imaging modalities and task settings, suggesting that idiosyncratic biases in medical image perception may widely exist.

Discussion

Characterizing and understanding these biases could be important for many practical settings such as training, pairing readers, and career selection for radiologists. These results may have consequential implications for many other fields as well, where individual observers are the linchpins for life-altering perceptual decisions.

Early stage of radiological expertise modulates resting-state local coherence in the inferior temporal lobe

Background

The visual system and its inherent functions undergo experience-dependent changes through the lifespan, enabling acquisition of new skills. Previous fMRI studies using tasks reported increased specialization in a number of cortical regions subserving visual expertise. Although ample studies focused on representation of long-term visual expertise in the brain, i.e. in terms of year, monthly-based early-stage representation of visual expertise remains unstudied. Given that spontaneous neuronal oscillations actively encode previous experience, we propose brain representations in the resting state is fundamentally important.

Objective

The current study aimed to investigate how monthly-based early-stage visual expertise are represented in the resting state using the expertise model of radiologists.

Methods

In particular, we investigated the altered local clustering pattern of spontaneous brain activity using regional homogeneity (ReHo). A cohort group of radiology interns (n = 22) after one-month training in X-ray department and matched laypersons (n = 22) were recruited after rigorous behavioral assessment.

Results

The results showed higher ReHo in the right hippocampus (HIP) and the right ventral anterior temporal lobe (vATL) (corrected by Alphasim correction, P < 0.05). Moreover, ReHo in the right HIP correlated with the number of cases reviewed during intern radiologists' training (corrected by Alphasim correction, P < 0.05).

Conclusions

In sum, our results demonstrated that the early stage of visual expertise is more concerned with stabilizing visual feature and domain-specific knowledge into long-term memory. The results provided novel evidence regarding how early-stage visual expertise is represented in the resting brain, which help further elaborate how human visual expertise is acquired. We propose that our current study may provide novel ideas for developing new training protocols in medical schools.

Spotting lesions in thorax X-rays at a glance: holistic processing in radiology

Radiologists often need only a glance to grasp the essence of complex medical images. Here, we use paradigms and manipulations from perceptual learning and expertise fields to elicit mechanisms and limits of holistic processing in radiological expertise. In the first experiment, radiologists were significantly better at categorizing thorax X-rays when they were presented for 200 ms in an upright orientation than when they were presented upside-down. Medical students, in contrast, were guessing in both situations. When the presentation time was increased to 500 ms, allowing for a couple more glances, the radiologists improved their performance on the upright stimuli, but remained at the same level on the inverted presentation. The second experiment circumvented the holistic processing by immediately cueing a tissue within the X-rays, which may or may not contain a nodule. Radiologists were again better than medical students at recognizing whether the cued tissue was a nodule, but this time neither the inverted presentation nor additional time affected their performance. Our study demonstrates that holistic processing is most likely a continuous recurring process which is just as susceptible to the inversion effect as in other expertise domains. More importantly, our study also indicates that holistic-like processing readily occurs in complex stimuli (e.g., whole thorax X-rays) but is more difficult to find in uniform single parts of such stimuli (e.g., nodules).

Automated quality assessment of chest radiographs based on deep learning and linear regression cascade algorithms

OBJECTIVES

Develop and evaluate the performance of deep learning and linear regression cascade algorithms for automated assessment of the image layout and position of chest radiographs.

METHODS

This retrospective study used 10 quantitative indices to capture subjective perceptions of radiologists regarding image layout and position of chest radiographs, including the chest edges, field of view (FOV), clavicles, rotation, scapulae, and symmetry. An automated assessment system was developed using a training dataset consisting of 1025 adult posterior-anterior chest radiographs. The evaluation steps included: (i) use of a CNN framework based on ResNet - 34 to obtain measurement parameters for quantitative indices and (ii) analysis of quantitative indices using a multiple linear regression model to obtain predicted scores for the layout and position of chest radiograph. In the testing dataset (n = 100), the performance of the automated system was evaluated using the intraclass correlation coefficient (ICC), Pearson correlation coefficient (r), mean absolute difference (MAD), and mean absolute percentage error (MAPE).

RESULTS

The stepwise regression showed a statistically significant relationship between the 10 quantitative indices and subjective scores (p < 0.05). The deep learning model showed high accuracy in predicting the quantitative indices (ICC = 0.82 to 0.99, r = 0.69 to 0.99, MAD = 0.01 to 2.75). The automatic system provided assessments similar to the mean opinion scores of radiologists regarding image layout (MAPE = 3.05%) and position (MAPE = 5.72%).

CONCLUSIONS

Ten quantitative indices correlated well with the subjective perceptions of radiologists regarding the image layout and position of chest radiographs. The automated system provided high performance in measuring quantitative indices and assessing image quality.

KEY POINTS

• Objective and reliable assessment for image quality of chest radiographs is important for improving image quality and diagnostic accuracy. • Deep learning can be used for automated measurements of quantitative indices from chest radiographs. • Linear regression can be used for interpretation-based quality assessment of chest radiographs.

Oral presentation assessment and image reading behaviour on brain computed tomography reading in novice clinical learners: an eye-tracking study

BACKGROUND

To study whether oral presentation (OP) assessment could reflect the novice learners' interpretation skills and reading behaviour on brain computed tomography (CT) reading.

METHODS

Eighty fifth-year medical students were recruited, received a 2-hour interactive workshop on how to read brain CT, and were assigned to read two brain CT images before and after instruction. We evaluated their image reading behaviour in terms of overall OP post-test rating, the lesion identification, and competency in systematic image reading after instruction. Students' reading behaviour in searching for the target lesions were recorded by the eye-tracking technique and were used to validate the accuracy of lesion reports. Statistical analyses, including lag sequential analysis (LSA), linear mixed models, and transition entropy (TE) were conducted to reveal temporal relations and spatial complexity of systematic image reading from the eye movement perspective.

RESULTS

The overall OP ratings [pre-test vs. post-test: 0 vs. 1 in case 1, 0 vs. 1 in case 2, p < 0.001] improved after instruction. Both the scores of systematic OP ratings [0 vs.1 in both cases, p < 0.001] and eye-tracking studies (Case 1: 3.42 ± 0.62 and 3.67 ± 0.37 in TE, p = 0.001; Case 2: 3.42 ± 0.76 and 3.75 ± 0.37 in TE, p = 0.002) showed that the image reading behaviour changed before and after instruction. The results of linear mixed models suggested a significant interaction between instruction and area of interests for case 1 (p < 0.001) and case 2 (p = 0.004). Visual attention to the target lesions in the case 1 assessed by dwell time were 506.50 ± 509.06 and 374.38 ± 464.68 milliseconds before and after instruction (p = 0.02). However, the dwell times in the case 2, the fixation counts and the frequencies of accurate lesion diagnoses in both cases did not change after instruction.

CONCLUSION

Our results showed OP performance may change concurrently with the medical students' reading behaviour on brain CT after a structured instruction.

A SWOT analysis of artificial intelligence in diagnostic imaging in the developing world: making a case for a paradigm shift

Diagnostic imaging (DI) refers to techniques and methods of creating images of the body’s internal parts and organs with or without the use of ionizing radiation, for purposes of diagnosing, monitoring and characterizing diseases. By default, DI equipment are technology based and in recent times, there has been widespread automation of DI operations in high-income countries while low and middle-income countries (LMICs) are yet to gain traction in automated DI. Advanced DI techniques employ artificial intelligence (AI) protocols to enable imaging equipment perceive data more accurately than humans do, and yet automatically or under expert evaluation, make clinical decisions such as diagnosis and characterization of diseases. In this narrative review, SWOT analysis is used to examine the strengths, weaknesses, opportunities and threats associated with the deployment of AI-based DI protocols in LMICs. Drawing from this analysis, a case is then made to justify the need for widespread AI applications in DI in resource-poor settings. Among other strengths discussed, AI-based DI systems could enhance accuracies in diagnosis, monitoring, characterization of diseases and offer efficient image acquisition, processing, segmentation and analysis procedures, but may have weaknesses regarding the need for big data, huge initial and maintenance costs, and inadequate technical expertise of professionals. They present opportunities for synthetic modality transfer, increased access to imaging services, and protocol optimization; and threats of input training data biases, lack of regulatory frameworks and perceived fear of job losses among DI professionals. The analysis showed that successful integration of AI in DI procedures could position LMICs towards achievement of universal health coverage by 2030/2035. LMICs will however have to learn from the experiences of advanced settings, train critical staff in relevant areas of AI and proceed to develop in-house AI systems with all relevant stakeholders onboard.

Deep Learning Algorithms for Diagnosis of Lung Cancer: A Systematic Review and Meta-Analysis

We conducted a systematic review and meta-analysis of the diagnostic performance of current deep learning algorithms for the diagnosis of lung cancer. We searched major databases up to June 2022 to include studies that used artificial intelligence to diagnose lung cancer, using the histopathological analysis of true positive cases as a reference. The quality of the included studies was assessed independently by two authors based on the revised Quality Assessment of Diagnostic Accuracy Studies. Six studies were included in the analysis. The pooled sensitivity and specificity were 0.93 (95% CI 0.85−0.98) and 0.68 (95% CI 0.49−0.84), respectively. Despite the significantly high heterogeneity for sensitivity (I2 = 94%, p < 0.01) and specificity (I2 = 99%, p < 0.01), most of it was attributed to the threshold effect. The pooled SROC curve with a bivariate approach yielded an area under the curve (AUC) of 0.90 (95% CI 0.86 to 0.92). The DOR for the studies was 26.7 (95% CI 19.7−36.2) and heterogeneity was 3% (p = 0.40). In this systematic review and meta-analysis, we found that when using the summary point from the SROC, the pooled sensitivity and specificity of DL algorithms for the diagnosis of lung cancer were 93% and 68%, respectively.

Diagnostic captioning: a survey

Diagnostic captioning (DC) concerns the automatic generation of a diagnostic text from a set of medical images of a patient collected during an examination. DC can assist inexperienced physicians, reducing clinical errors. It can also help experienced physicians produce diagnostic reports faster. Following the advances of deep learning, especially in generic image captioning, DC has recently attracted more attention, leading to several systems and datasets. This article is an extensive overview of DC. It presents relevant datasets, evaluation measures, and up-to-date systems. It also highlights shortcomings that hinder DC’s progress and proposes future directions.

Eye Movements as a Window into Decision-Making

For over 100 years, eye movements have been studied and used as indicators of human sensory and cognitive functions. This review evaluates how eye movements contribute to our understanding of the processes that underlie decision-making. Eye movement metrics signify the visual and task contexts in which information is accumulated and weighed. They indicate the efficiency with which we evaluate the instructions for decision tasks, the timing and duration of decision formation, the expected reward associated with a decision, the accuracy of the decision outcome, and our ability to predict and feel confident about a decision. Because of their continuous nature, eye movements provide an exciting opportunity to probe decision processes noninvasively in real time. Expected final online publication date for the Annual Review of Vision Science, Volume 8 is September 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Pitfalls of Echocardiographic Image Perception: How to Overcome Them?

In recent years, the significant development of echocardiography systems has led to a sharp improvement in echocardiographic images' quality. In parallel with this, computerized technologies are also going far forward, which today make it possible to ensure a high level of transmission, storage, and display of echocardiography studies. Despite this, many cardiologists are not familiar with modern computerized technologies' new possibilities and continue to use the old standards. That is why many echocardiography laboratories with the best echocardiography systems work following the old minimalist approach. In this paper, we will look at some of the most common mistakes that result from the improper transmission, storage, and demonstration of echocardiography studies, and describe possible ways to overcome these problems.