Clinical Image Interpretation Research Articles

Performance of artificial intelligence chatbots in interpreting clinical images of pressure injuries.

To evaluate the accuracy of AI chatbots in staging pressure injuries according to the National Pressure Injury Advisory Panel (NPIAP) Staging through clinical image interpretation, a cross-sectional design was conducted to assess five leading publicly available AI chatbots. As a result, three chatbots were unable to interpret the clinical images, whereas GPT-4 Turbo achieved a high accuracy rate (83.0%) in staging pressure injuries, notably outperforming BingAI Creative mode (24.0%) with statistical significance (p < 0.001). GPT-4 Turbo accurately identified Stages 1 (p < 0.001), 3 (p = 0.001), and 4 (p < 0.001) pressure injuries, and suspected deep tissue injuries (p < 0.001), while BingAI demonstrated significantly lower accuracy across all stages. The findings highlight the potential of AI chatbots, especially GPT-4 Turbo, in accurately diagnosing images and aiding the subsequent management of pressure injuries.

Visual interpretation of clinical images among dental students using eye-tracking technology.

Tele-consultations are increasingly used for screening and diagnosis. Only a few studies have assessed dental students' visual attention to clinical images. To (i) determine dental students' gaze behavior, visual fixations, and diagnostic competence while viewing clinical images, and (ii) explore potential opportunities to strengthen the teaching-learning approaches. Tobii Pro Nano-device captured the eye-tracking data for 65 dental undergraduate students in this cross-sectional study. The predetermined areas of interest (AOI) for all five clinical photographs were uploaded onto Tobii software. All participants used a think-aloud protocol with no restrictions to view time. A total of 325 clinical pictures were analyzed, and the average view time was 189.25±76.90 s. Most participants started at the center of the image (three frontal photos), spent a significant share of their view time on prominent findings, did not follow a systematic pattern, and exhibited diagnostic incompetence. Also, most participants followed a "Z" viewing pattern (oscillating movement from left to right) for the remaining two pictures. Subjects frequently fixated on the prominent AOI, however, failed to make the correct diagnosis. Their view patterns revealed no sequential viewing. Therefore, emphasizing knowledge about common dental abnormalities and focusing on full coverage of clinical pictures can improve dental students' diagnostic competence and view patterns.

Semi-supervised training using cooperative labeling of weakly annotated data for nodule detection in chest CT.

Machine learning algorithms are best trained with large quantities of accurately annotated samples. While natural scene images can often be labeled relatively cheaply and at large scale, obtaining accurate annotations for medical images is both time consuming and expensive. In this study, we propose a cooperative labeling method that allows us to make use of weakly annotated medical imaging data for the training of a machine learning algorithm. As most clinically produced data are weakly-annotated - produced for use by humans rather than machines and lacking information machine learning depends upon - this approach allows us to incorporate a wider range of clinical data and thereby increase the training set size. Our pseudo-labeling method consists of multiple stages. In the first stage, a previously established network is trained using a limited number of samples with high-quality expert-produced annotations. This network is used to generate annotations for a separate larger dataset that contains only weakly annotated scans. In the second stage, by cross-checking the two types of annotations against each other, we obtain higher-fidelity annotations. In the third stage, we extract training data from the weakly annotated scans, and combine it with the fully annotated data, producing a larger training dataset. We use this larger dataset to develop a computer-aided detection (CADe) system for nodule detection in chestCT. We evaluated the proposed approach by presenting the network with different numbers of expert-annotated scans in training and then testing the CADe using an independent expert-annotated dataset. We demonstrate that when availability of expert annotations is severely limited, the inclusion of weakly-labeled data leads to a 5% improvement in the competitive performance metric (CPM), defined as the average of sensitivities at different false-positiverates. Our proposed approach can effectively merge a weakly-annotated dataset with a small, well-annotated dataset for algorithm training. This approach can help enlarge limited training data by leveraging the large amount of weakly labeled data typically generated in clinical image interpretation.

Should Radiological Assessment be Part of Core General Surgery Competency? A Survey of the General Surgery Residents

Introduction: Radiological assessment is an important skill to develop in general surgery training. Therefore, we aimed to determine general surgery residents’ points of view on receiving formal radiology didactics. Methods: We performed an anonymous survey of general surgery residents throughout the US. The survey queried the residents` postgraduate year, training program type, diagnostic radiology education in their training program, as well as the residents’ comfort level in interpreting various imaging modalities, followed by a series of images to assess the residents’ ability to interpret images showing various surgical disease processes. Results: A total of 365 residents responded to the survey. 76.6% of the respondent states that there is no structured didactic session in their program on radiological studies. However, 66.3% felt that interpretation of radiological images should be used to determine surgical competency and promotion to the next academic year. In terms of accurately reading images - 68.7% of the residents were able to read an x-ray showing cecal volvulus correctly, 51.9% were able to read a cholangiogram correctly, and 95.3% were correctly read an x-ray showing free under the diaphragm. Conclusion: Most residents favored having radiological assessments as part of the competency evaluation. Furthermore, a curriculum and inbuilt training structure that aims to ensure residents develop competent clinical image interpretation abilities may enhance the development and retention of such skills, ultimately influencing patient outcomes.

Ten questions for the cardiologist about cardiac scintigraphy with bone tracers, amyloidosis and the heart

Although traditionally considered a rare and incurable condition, the perspective on cardiac amyloidosis (CA) has changed deeply over the last 10 years following major advances in diagnosis and treatment options. First, the development of non-invasive diagnostic methods, predominantly cardiac scintigraphy with bone tracers, has dramatically changed the epidemiology of the disease, revealing the real prevalence of the transthyretin-related form (ATTR), and has consistently reduced the need for histological amyloid confirmation by endomyocardial biopsy. Second, the availability of new specific therapies, such as tafamidis, has made CA early recognition a clinical need with major impact on patients' management and natural history. Since the demonstration of the essential role of bone tracer scintigraphy to non-invasively differentiate ATTR from immunoglobulin light-chain amyloidosis (AL), this imaging technique has become part of the vocabulary and clinical practice of physicians of many specialties, including cardiologists managing patients with cardiomyopathies.In this paper, we address the main questions regarding cardiac scintigraphy with bone tracers and CA, providing cardiologists with the knowledge to guide clinical indications, evaluation and clinical interpretation of images.

Convolutional neural network assistance significantly improves dermatologists’ diagnosis of cutaneous tumours using clinical images

BackgroundConvolutional neural networks (CNNs) have demonstrated expert-level performance in cutaneous tumour classification using clinical images, but most previous studies have focused on dermatologist-versus-CNN comparisons rather than their combination. The objective of our study was to evaluate the potential impact of CNN assistance on dermatologists for clinical image interpretation. MethodsA multi-class CNN was trained and validated using a dataset of 25,773 clinical images comprising 10 categories of cutaneous tumours. The CNN's performance was tested on an independent dataset of 2107 images. A total of 400 images (40 per category) were randomly selected from the test dataset. A fully crossed, self-control, multi-reader multi-case (MRMC) study was conducted to compare the performance of 18 board-certified dermatologists (experience: 13/18 ≤ 10 years; 5/18＞10 years) in interpreting the 400 clinical images with or without CNN assistance. ResultsThe CNN achieved an overall accuracy of 78.45% and kappa of 0.73 in the classification of 10 types of cutaneous tumours on 2107 images. CNN-assisted dermatologists achieved a higher accuracy (76.60% vs. 62.78%, P < 0.001) and kappa (0.74 vs. 0.59, P < 0.001) than unassisted dermatologists in interpreting the 400 clinical images. Dermatologists with less experience benefited more from CNN assistance. At the binary classification level (malignant or benign), the sensitivity (89.56% vs. 83.21%, P < 0.001) and specificity (87.90% vs. 80.92%, P < 0.001) of dermatologists with CNN assistance were also significantly improved than those without. ConclusionsCNN assistance improved dermatologist accuracy in interpreting cutaneous tumours and could further boost the acceptance of this new technique.

Redesigning Clinical Pathways for Immediate Diabetic Retinopathy Screening Results

Summary Regular diabetic retinopathy (DR) screening and early treatment can prevent DR-associated vision loss. However, some DR screening programs within primary care settings have found low rates of patient adherence to referral recommendations, even following a positive screen for vision-threatening DR. The authors hypothesized that adherence can be increased by providing screening results immediately and improving workflows by engaging patients and, when needed, scheduling a follow-up ophthalmology appointment immediately. A long-term goal of this project is to investigate the potential value of an immediate clinical image interpretation provided by artificial intelligence (AI); however, in this study, optometrists simulated AI by providing immediate reads of the fundus images. Immediate interpretation, which formed the basis for counseling and recommendations while the patient was in the office, resulted in significantly improved adherence among patients who received a recommendation to see a specialist within 1 month, from the historical baseline of 35% to 72% (P < .01 after controlling for cohort characteristics). This suggests that providing results and scheduling follow-up appointments immediately following a DR screening test can substantially improve patient adherence and reduce unnecessary vision loss. The changes were adopted widely within the hospital system and even scaled to include referrals to other specialties.

Nothing about us without us: involving patient collaborators for machine learning applications in rheumatology

Novel machine learning methods open the door to advances in rheumatology through application to complex, high-dimensional data, otherwise difficult to analyse. Results from such efforts could provide better classification of...

Integrating Eye Tracking and Speech Recognition Accurately Annotates MR Brain Images for Deep Learning: Proof of Principle.

To generate and assess an algorithm combining eye tracking and speech recognition to extract brain lesion location labels automatically for deep learning (DL). In this retrospective study, 700 two-dimensional brain tumor MRI scans from the Brain Tumor Segmentation database were clinically interpreted. For each image, a single radiologist dictated a standard phrase describing the lesion into a microphone, simulating clinical interpretation. Eye-tracking data were recorded simultaneously. Using speech recognition, gaze points corresponding to each lesion were obtained. Lesion locations were used to train a keypoint detection convolutional neural network to find new lesions. A network was trained to localize lesions for an independent test set of 85 images. The statistical measure to evaluate our method was percent accuracy. Eye tracking with speech recognition was 92% accurate in labeling lesion locations from the training dataset, thereby demonstrating that fully simulated interpretation can yield reliable tumor location labels. These labels became those that were used to train the DL network. The detection network trained on these labels predicted lesion location of a separate testing set with 85% accuracy. The DL network was able to locate brain tumors on the basis of training data that were labeled automatically from simulated clinical image interpretation.© RSNA, 2020.

PET/MRI Radiomics in Patients With Brain Metastases.

Although a variety of imaging modalities are used or currently being investigated for patients with brain tumors including brain metastases, clinical image interpretation to date uses only a fraction of the underlying complex, high-dimensional digital information from routinely acquired imaging data. The growing availability of high-performance computing allows the extraction of quantitative imaging features from medical images that are usually beyond human perception. Using machine learning techniques and advanced statistical methods, subsets of such imaging features are used to generate mathematical models that represent characteristic signatures related to the underlying tumor biology and might be helpful for the assessment of prognosis or treatment response, or the identification of molecular markers. The identification of appropriate, characteristic image features as well as the generation of predictive or prognostic mathematical models is summarized under the term radiomics. This review summarizes the current status of radiomics in patients with brain metastases.

Interdimensional Travel: Visualisation of 3D-2D Transitions in Anatomy Learning.

Clinical image interpretation is one of the most challenging activities for students when they first arrive at medical school. Interpretation of clinical images concerns the identification of three-dimensional anatomical features in two-dimensional cross-sectional computed tomography (CT) and magnetic resonance imaging (MRI) images in axial, sagittal and coronal planes, and the recognition of structures in ultrasound and plain radiographs. We propose that a cognitive transition occurs when initially attempting to interpret clinical images, which requires reconciling known 3D structures with previously unknown 2D visual information. Additionally, we propose that this 3D-2D transition is required when integrating an understanding of superficial 2D surface landmarks with an appreciation of underlying 3D anatomical structures during clinical examinations.Based on educational theory and research findings, we recommend that 3D and 2D approaches should be simultaneously combined within radiological and surface anatomy education. With a view to this, we have developed and utilised digital and art-based methods to support the 3D-2D transition. We outline our observations and evaluations, and describe our practical implementation of these approaches within medical curricula to serve as a guide for anatomy educators. Furthermore, we define the theoretical underpinnings and evidence supporting the integration of 3D-2D approaches and the value of our specific activities for enhancing the clinical image interpretation and surface anatomy learning of medical students.

Image Quality and Activity Optimization in Oncologic 18F-FDG PET Using the Digital Biograph Vision PET/CT System.

The first Biograph Vision PET/CT system (Siemens Healthineers) was installed at the University Medical Center Groningen. Improved performance of this system could allow for a reduction in activity administration or scan duration. This study evaluated the effects of reduced scan duration in oncologic 18F-FDG PET imaging on quantitative and subjective imaging parameters and its influence on clinical image interpretation. Methods: Patients referred for a clinical PET/CT scan were enrolled in this study, received a weight-based 18F-FDG injected activity, and underwent list-mode PET acquisition at 180 s per bed position (s/bp). Acquired PET data were reconstructed using the vendor-recommended clinical reconstruction protocol (hereafter referred to as "clinical"), using the clinical protocol with additional 2-mm gaussian filtering (hereafter referred to as "clinical+G2"), and-in conformance with European Association of Nuclear Medicine Research Ltd. (EARL) specifications-using different scan durations per bed position (180, 120, 60, 30, and 10 s). Reconstructed images were quantitatively assessed for comparison of SUVs and noise. In addition, clinically reconstructed images were qualitatively evaluated by 3 nuclear medicine physicians. Results: In total, 30 oncologic patients (22 men, 8 women; age: 48-88 y [range], 67 ± 9.6 y [mean ± SD]) received a single weight-based (3 MBq/kg) 18F-FDG injected activity (weight: 45-123 kg [range], 81 ± 15 kg [mean ± SD]; activity: 135-380 MBq [range], 241 ± 47.3 MBq [mean ± SD]). Significant differences in lesion SUVmax were found between the 180-s/bp images and the 30- and 10-s/bp images reconstructed using the clinical protocols, whereas no differences were found in lesion SUVpeak EARL-compliant images did not show differences in lesion SUVmax or SUVpeak between scan durations. Quantitative parameters showed minimal deviation (∼5%) in the 60-s/bp images. Therefore, further subjective image quality assessment was conducted using the 60-s/bp images. Qualitative assessment revealed the influence of personal preference on physicians' willingness to adopt the 60-s/bp images in clinical practice. Although quantitative PET parameters differed minimally, an increase in noise was observed. Conclusion: With the Biograph Vision PET/CT system for oncologic 18F-FDG imaging, scan duration or activity administration could be reduced by a factor of 3 or more with the use of the clinical+G2 or the EARL-compliant reconstruction protocol.

An Ad Hoc Random Initialization Deep Neural Network Architecture for Discriminating Malignant Breast Cancer Lesions in Mammographic Images.

Breast cancer is one of the most common cancers in women, with more than 1,300,000 cases and 450,000 deaths each year worldwide. In this context, recent studies showed that early breast cancer detection, along with suitable treatment, could significantly reduce breast cancer death rates in the long term. X-ray mammography is still the instrument of choice in breast cancer screening. In this context, the false-positive and false-negative rates commonly achieved by radiologists are extremely arduous to estimate and control although some authors have estimated figures of up to 20% of total diagnoses or more. The introduction of novel artificial intelligence (AI) technologies applied to the diagnosis and, possibly, prognosis of breast cancer could revolutionize the current status of the management of the breast cancer patient by assisting the radiologist in clinical image interpretation. Lately, a breakthrough in the AI field has been brought about by the introduction of deep learning techniques in general and of convolutional neural networks in particular. Such techniques require no a priori feature space definition from the operator and are able to achieve classification performances which can even surpass human experts. In this paper, we design and validate an ad hoc CNN architecture specialized in breast lesion classification from imaging data only. We explore a total of 260 model architectures in a train-validation-test split in order to propose a model selection criterion which can pose the emphasis on reducing false negatives while still retaining acceptable accuracy. We achieve an area under the receiver operatic characteristics curve of 0.785 (accuracy 71.19%) on the test set, demonstrating how an ad hoc random initialization architecture can and should be fine tuned to a specific problem, especially in biomedical applications.

PD06-07 AGGRESSIVENESS OF PROSTATE CANCER: CORRELATION BETWEEN PROSPECTIVELY ASSIGNED APPARENT DIFFUSION COEFFICIENT VALUES AND ISUP GRADE GROUP

You have accessJournal of UrologyProstate Cancer: Detection & Screening I1 Apr 2018PD06-07 AGGRESSIVENESS OF PROSTATE CANCER: CORRELATION BETWEEN PROSPECTIVELY ASSIGNED APPARENT DIFFUSION COEFFICIENT VALUES AND ISUP GRADE GROUP Daniel Costa, Yin Xi, Muhammad Aziz, Franto Francis, Claus Roehrborn, Kenneth Goldberg, Niccolo Passoni, Nabeel Shakir, Alberto Diaz de Leon, and Ivan Pedrosa Daniel CostaDaniel Costa More articles by this author , Yin XiYin Xi More articles by this author , Muhammad AzizMuhammad Aziz More articles by this author , Franto FrancisFranto Francis More articles by this author , Claus RoehrbornClaus Roehrborn More articles by this author , Kenneth GoldbergKenneth Goldberg More articles by this author , Niccolo PassoniNiccolo Passoni More articles by this author , Nabeel ShakirNabeel Shakir More articles by this author , Alberto Diaz de LeonAlberto Diaz de Leon More articles by this author , and Ivan PedrosaIvan Pedrosa More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2018.02.427AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES Non-invasive quantitative markers of prostate cancer (PCa) aggressiveness have the potential to improve pre-treatment risk stratification. Previous investigations have shown an inverse correlation between Gleason score and apparent diffusion coefficient (ADC) values derived from diffusion-weighted MRI. Most of these investigations, however, were retrospective. The goal of this study was to determine the ability to use prospectively assigned mean ADC values to discriminate between ISUP grade groups. METHODS Using a prospectively maintained database, we identified all men with abnormal 3T endorectal coil mpMRI examinations (PIRADS v2 score ≥3) between Nov/16-Jul/17, without prior treatment, who subsequently had a positive targeted biopsy or radical prostatectomy. One of 9 radiologists prospectively assigned a mean ADC value for each lesion during clinical image interpretation. The index lesion on MRI - one per patient - was correlated with the site-concordant lesion on biopsy or whole-mount histopathology. Fusion or in-bore targeted biopsy and 3D-printed whole-mount histopathology enabled strict imaging-histology correlation. Logistic regression was used to evaluate the association between mean ADC values in discriminating International Society of Urological Pathology (ISUP) group 1 vs. grade group 2-5. Impact of lesion location (PZ vs. TZ) was also evaluated by stratification. Interquartile ranges (IQRs) were calculated. Diagnostic performances were determined using ROC curve and area under the curve (AUC) with 95% confidence intervals (CIs). Youden′s J statistic was used to determine optimal cutoff values. RESULTS Among the 218 eligible men, 41% underwent surgery and 59% had targeted biopsies. Mean PSA, age, prostate volume and lesion size were 10.2 ng/mL, 65y, 43 cc and 15 mm, respectively. ISUP grade groups 1-5 accounted for, respectively, 21%, 48%, 19%, 7% and 5% of cases. Overall and per-zone mean ADC values were lower for ISUP 2-5 than ISUP 1 (p<0.001,Fig.1). AUCs for the discrimination between ISUP group 1 vs. 2 or higher using mean ADC values were 0.85 (PZ+TZ), 0.91 (PZ) and 0.70 (TZ). Optimal cutoff values were 0.682 for PZ and 0.638 for TZ, resulting in a sensitivity and specificity of 78% and 94% for PZ, 72% and 67% for TZ, respectively. CONCLUSIONS ADC values can be used to predict ISUP grade groups, particularly in PZ lesions, facilitating pre-treatment risk stratification in men undergoing mpMRI and targeted prostate biopsies. © 2018FiguresReferencesRelatedDetails Volume 199Issue 4SApril 2018Page: e153 Advertisement Copyright & Permissions© 2018MetricsAuthor Information Daniel Costa More articles by this author Yin Xi More articles by this author Muhammad Aziz More articles by this author Franto Francis More articles by this author Claus Roehrborn More articles by this author Kenneth Goldberg More articles by this author Niccolo Passoni More articles by this author Nabeel Shakir More articles by this author Alberto Diaz de Leon More articles by this author Ivan Pedrosa More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...

TU‐H‐CAMPUS‐IeP2‐03: Development of 3D Printed Coronary Phantoms for In‐Vitro CT‐FFR Validation Using Data from 320‐ Detector Row Coronary CT Angiography

Purpose:To validate Computed Tomography Fractional Flow Reserve (CT‐FFR) measurements with accurate 3D printed coronary phantoms.Methods:DICOM data from four phases in two patients imaged with a standard 320 × 0.5mm coronary CT acquisition (70–80% cardiac cycle) underwent semi‐automated segmentation using a research workstation. Both patients had a &gt;50% stenosis from the clinical image interpretation. Each volume was saved as a Stereo Lithographic (STL) file with 250 micron resolution. The 3D geometries were qualitatively assessed; the best of the four phases was 3D printed using a Stratasys Eden260V printer in Tango+, a rubber‐like material that roughly emulates mechanical properties of human vasculature. We connected the model to a programmable pump and measured the pressure drop using pressure sensors embedded proximal and distal to the arterial stenosis. Next, the STL files used for the 3D printed models were uploaded in the ANSYS meshing tool (ICEM CFD 16.1). A standard meshing process was applied and the meshed geometry was directly imported in the ANSYS Fluent for Computational Flow Dynamics simulations. The CFD simulations were used to calculate the CT‐FFR and compared to the bench top FFR measured in the 3D printed phantoms.Results:FFR‐CT measurements and phantoms were completed in within an hour after the segmentation. Patient 1 had a 60% stenosis that resulted in a CT‐FFR of 0.68. The second case had a 50% stenosis and a CT‐FFR of 0.75. The average bench top FFR measurements were 0.72 and 0.80, respectively.Conclusion:This pilot investigation demonstrated the use of a bench‐top coronary model for CT‐FFR validation. The measurements and the CFD simulations agreed within 6%. Project supported by Support: Toshiba America Medical Systems Corp.and NIH grant R01‐EB002873.Project supported by Toshiba America Medical Systems Corp.and partial support from NIH grant R01‐EB002873

Ultra‐High Field MRI Clinical Anatomy of the Equine Fetlock to Aid Clinical Image Interpretation

Reasons for performing studyInterpretation of clinical diagnostic imaging of the equine musculoskeletal system requires accurate reference anatomical materials. In the case of MRI, this includes images acquired from low and high field systems. Using ultra‐high field MRI, higher resolution images with small slice thickness and gap suitable for reconstruction in any anatomical plane can be produced.ObjectivesTo describe the normal equine metacarpophalangeal joint using ultra‐high field MRI and compare ultra‐high and high field MR images.Study designLimbs (n = 14) from horses (n = 5) with no history of orthopaedic disease were frozen at ‐20°C until imaging. Fetlock joints were scanned using an ultra‐high field (7T) MRI unit. T1 (280 slices, thickness = 0.356 mm) and T2 (140 slices, thickness = 0.713 mm) weighted sequences were acquired. Limbs were scanned using a 1.5T MRI unit in transverse, sagittal, and dorsal planes using T1 (60 slices, thickness 2.2–3.5 mm) and T2 (60 slices, thickness 2.2–3.5 mm) sequences. Images were analysed using OsiriX software.ResultsUltra‐high field MR images visualised the fine detail of trabecular and subchondral bone, the articular cartilage surface of P1 and McIII, the short, oblique, straight and oblique sesamoidean ligaments including origins and insertions, the fasicular structure of superficial digital flexor tendon, deep digital flexor tendon and ligaments and the regional neurovascular structures. Whilst T1 high field images visualised the majority of tissues, structural detail was reduced compared to ultra‐high field imaging, margins were less well defined in some instances, e.g. articular cartilage, and a number of smaller structures were indistinct, e.g. short and cruciate sesamoidean ligaments.Conclusions7T MRI produced images of the fetlock with excellent anatomical detail. Future publication of open‐source ultra‐high field MRI files could aid the interpretation of 0.3T and 1.5T field MR images, especially with the capacity to reconstruct image slices orientated exactly to match those of interest.Ethical animal research: The study was approved by the Ethics Committee of the School of Veterinary and Life Sciences, University of Glasgow. Explicit owner informed consent for participation in this study was not stated but permission was given for post mortem examination and the use of materials in research. Sources of funding: The Royal Saudi Arabia Cultural Bureau. Competing interests: None.

Physiology's Impact: Applying Mathematics and Advanced Technologies

How do you measure the mechanical forces generated by single cells? What are the cues that guide cell migration? What are the interactions between cells that account for cooperative behaviors among cells? To answer these and many other basic questions about biological function, we must have the

Chronic pancreatitis.

We review selected important clinical observations reported in 2012. Celiac disease is a risk factor for pancreatitis. Patients with recurrent acute pancreatitis likely have chronic pancreatitis, do not benefit from pancreatic sphincterotomy, and may not benefit from biliary sphincterotomy. Analysis of endoscopic ultrasonography (EUS) images with an artificial neural network (ANN) program may improve chronic pancreatitis diagnosis compared with clinical interpretation of images. In a multicenter, randomized controlled trial of chronic pancreatitis patients, 90 000 USP U of pancreatin with meals decreased fat malabsorption compared with placebo. Detection of visceral pain in chronic pancreatitis predicts pain relief from various treatments, but nonvisceral pain due to altered central pain processing may respond to agents such as pregabalin. Predictors of surgical pain relief include onset of symptoms less than 3 years and preoperatively no opioid use and less than five endoscopic procedures. Total pancreatectomy for presumed painful chronic pancreatitis remains controversial. Celiacs are at risk for pancreatitis. The diagnosis of chronic pancreatitis may be enhanced by ANN analysis of EUS imaging. Treatment of fat malabsorption requires 90,000 USP U of lipase with meals. Relief of pain from organ directed treatment of chronic pancreatitis may depend upon timing of interventions and whether pain is visceral or nonvisceral.

Using a High-Speed Movie Camera to Evaluate Slice Dropping in Clinical Image Interpretation with Stack Mode Viewers

The purpose of this study is to verify objectively the rate of slice omission during paging on picture archiving and communication system (PACS) viewers by recording the images shown on the computer displays of these viewers with a high-speed movie camera. This study was approved by the institutional review board. A sequential number from 1 to 250 was superimposed on each slice of a series of clinical Digital Imaging and Communication in Medicine (DICOM) data. The slices were displayed using several DICOM viewers, including in-house developed freeware and clinical PACS viewers. The freeware viewer and one of the clinical PACS viewers included functions to prevent slice dropping. The series was displayed in stack mode and paged in both automatic and manual paging modes. The display was recorded with a high-speed movie camera and played back at a slow speed to check whether slices were dropped. The paging speeds were also measured. With a paging speed faster than half the refresh rate of the display, some viewers dropped up to 52.4% of the slices, while other well-designed viewers did not, if used with the correct settings. Slice dropping during paging was objectively confirmed using a high-speed movie camera. To prevent slice dropping, the viewer must be specially designed for the purpose and must be used with the correct settings, or the paging speed must be slower than half of the display refresh rate.

Identifying Regional Cardiac Abnormalities From Myocardial Strains Using Nontracking-Based Strain Estimation and Spatio-Temporal Tensor Analysis

Myocardial strain is a critical indicator of many cardiac diseases and dysfunctions. The goal of this paper is to extract and use the myocardial strain pattern from tagged magnetic resonance imaging (MRI) to identify and localize regional abnormal cardiac function in human subjects. In order to extract the myocardial strains from the tagged images, we developed a novel nontracking-based strain estimation method for tagged MRI. This method is based on the direct extraction of tag deformation, and therefore avoids some limitations of conventional displacement or tracking-based strain estimators. Based on the extracted spatio-temporal strain patterns, we have also developed a novel tensor-based classification framework that better conserves the spatio-temporal structure of the myocardial strain pattern than conventional vector-based classification algorithms. In addition, the tensor-based projection function keeps more of the information of the original feature space, so that abnormal tensors in the subspace can be back-projected to reveal the regional cardiac abnormality in a more physically meaningful way. We have tested our novel methods on 41 human image sequences, and achieved a classification rate of 87.80%. The regional abnormalities recovered from our algorithm agree well with the patient's pathology and clinical image interpretation, and provide a promising avenue for regional cardiac function analysis.