Key Imaging Features Research Articles

Diagnostic Approach to Intrinsic Abnormality of Spinal Cord Signal Intensity.

Intramedullary cord hyperintensity at T2-weighted MRI is a common imaging feature of disease in the spinal cord, but it is nonspecific. Radiologists play a valuable role in helping narrow the differential diagnosis by integrating patient history and laboratory test results with key imaging characteristics. The authors present an algorithmic approach to evaluating intrinsic abnormality of spinal cord signal intensity (SI), which incorporates clinical evaluation results, time of onset (acute vs nonacute), cord expansion, and pattern of T2 SI abnormality. This diagnostic approach provides a practical framework to aid both trainees and practicing radiologists in workup of myelopathy.©RSNA, 2019.

Molecular image-based convolutional neural network for the prediction of ADMET properties

Convolutional neural network (CNN), is one of the most representative architectures in deep learning and is widely adopted in many fields especially in image classification and object detection. In the last few years, CNN has been aroused more and more attentions in drug discovery domain. In this work, molecular 2-D image-based CNN method was used to establish prediction models of the ADMET properties, including CYP1A2 inhibitory potency, P-glycoprotein (P-gp) inhibitory activity, Blood-Brain Barrier (BBB) penetrating activity, and Ames mutagenicity. The results showed that the predictive power of the established CNN models is comparable to that of the available machine learning models based on manual structural description and feature selection. It can be inferred that CNN can extract efficiently the key image features related to the molecular ADMET properties and offer a useful tool for virtual screening and drug design researches.

Transoral robotic surgery for head and neck malignancies: Imaging features in presurgical workup.

The objective of this article was to review the indications for transoral robotic surgery (TORS) in head and neck malignancies. The role of imaging in patient selection will be specifically reviewed. TORS is a recently developed technique that allows minimally invasive surgeries to be performed in the head and neck. TORS has a role in the de-escalation of oropharyngeal cancers, which allows for lower doses of chemoradiation therapy (this is a technique currently in clinical trials). Additionally, this technique allows for less invasive surgery and decreases associated complications. TORS can also be performed at other subsites. Cross-sectional imaging has a prominent role to help identify suitable candidates for this type of surgery. This article will review important anatomy and staging related to TORS. Additionally, the key imaging features for patient selection (indications and contraindications) will be presented along with case illustrations.

MRI Lesion Features and Distribution for MS Diagnosis

Experts convened in 2018 to discuss key imaging features for the correct interpretation and classification of lesions in multiple sclerosis (MS). They

Non-visualization of the ovaries on pediatric transabdominal ultrasound with a non-distended bladder: Can adnexal torsion be excluded?

The pediatric reproductive organs are optimally imaged with a full bladder. The filling of the bladder, however, often leads to significant delay in diagnosis and can subject the patient to invasive bladder catheterization. As the key imaging feature in ovarian torsion is unilateral ovarian enlargement, we suspected that a torsed ovary is large enough to be visualized even if the bladder is not well distended. The purpose of this study was to retrospectively investigate if clinically suspected adnexal torsion can be excluded based on non-visualization of the ovaries on transabdominal ultrasound (US) with a non-distended bladder in pediatric patients. This retrospective study comprised 349 girls (1-19years old) between Jan. 1, 2013, and July 30, 2018. Three hundred and forty-one of the girls were referred to transabdominal US to assess for adnexal torsion and/or appendicitis, and the ovaries were initially not visualized on US. Their bladders were subsequently filled and rescanned with a distended bladder showing the ovaries. Ovarian volumes and time between US scans were documented. The ratio of the volume of the larger ovary to the smaller one was calculated. Nine girls had surgically proven adnexal torsion and a preoperative transabdominal US with a non-distended bladder. There was an overlap of one girl between the two groups. The negative predictive value (NPV), positive predictive value (PPV), and sensitivity and specificity for exclusion of adnexal torsion based on non-visualization of the ovaries on US with a non-distended bladder were calculated. One of the girls (1/341) who had a US study done with a non-distended bladder in which the ovaries were not visualized had a positive diagnosis of adnexal torsion. In eight of the nine girls who had surgically proven adnexal torsion, the torted ovary was identified with a non-distended bladder. The NPV and PPV for exclusion of adnexal torsion with a non-distended bladder was 1.0 and 0.8, respectively. The specificity and sensitivity were 99.4% and 88.9%, respectively. The mean and median time difference between the initial scan and the scan after bladder filling was 105.1 min (standard deviation [SD] -65.8) and 89.0 min (interquartile range [IQR]- 59.0, 130.5), respectively. Non-visualization of the ovaries with a non-distended bladder on transabdominal US study can help exclude clinically suspected adnexal torsion, alleviating the need for bladder filling and prolonging the wait time in the emergency department. Inclusion of non-visualization of the ovaries as one of the features in a predictive score for adnexal torsion should be considered.

Usefulness of diffusion-weighted MRI in the initial assessment of osseous sarcomas in children and adolescents.

Concern regarding gadolinium deposition in the brain after repeated administration of intravenous gadolinium-based contrast agents has prompted evaluation of imaging alternatives. The study purpose was to determine if magnetic resonance imaging (MRI) using conventional sequences with diffusion-weighted imaging (DWI) instead of gadolinium-based contrast-enhanced MRI is valid for local staging and guiding biopsies in osseous sarcomas. Initial pretreatment MRI with DWI and gadolinium-based contrast-enhanced images in patients ≤ 18years with histopathologically proven osseous sarcomas were included. Two radiologists blinded to collated demographic and clinical data, independently reviewed conventional/DWI and conventional/gadolinium-based contrast-enhanced MRI then conventional sequences alone, recording tumor size, skip lesions, necrosis, neurovascular invasion, enlarged lymph nodes and diffusion restriction. Discrepancies were resolved by a third reader. A single reader measured apparent diffusion coefficient (ADC) values in non-necrotic tumors, then correlated minimum ADC values -- with and without normalization to skeletal muscle -- with relative enhancement. Twenty-one patients (mean age: 11.3±4.2years, 15 [71%] females) had 14 osteosarcomas and 7 Ewing sarcomas, 50% centered in the femur. Conventional/DWI versus conventional/gadolinium-based contrast-enhanced MRI showed agreement for tumor size estimation with significant associations for necrosis (P=0.021), neurovascular involvement (P<0.001) and enlarged lymph nodes (P=0.005). Diagnostic accuracy of conventional/DWI is comparable to conventional/gadolinium-based contrast-enhanced MRI and superior to conventional sequences alone. Comparison between minimum ADC values and relative enhancement showed no correlation (P>0.05). Significant associations of key imaging features in the initial assessment of osseous sarcomas support DWI as an alternative to gadolinium-based contrast-enhanced MRI. The lack of association between ADC values and relative enhancement suggests that they measure independent constructs, DWI dependent upon tumor cellularity and perfusion.

DETERMINING MORPHOMETRIC PROPERTIES OF RADIATA PINE USING LONG WAVE INFRARED SENSING AND BIOLOGICALLY-INSPIRED VISION

Abstract. A miniaturised, high resolution visible and long wave infrared (LWIR) sensor was carried onboard an unmanned aerial vehicle to observe sections of radiata pine forests. The raw irradiance measurements were temporally pre-processed using a biologically-inspired vision (BIV) model to allow information within and across the images to be normalised. This permitted a larger, denser, and more tailored set of key points within the 2D image stacks to be corresponded, thereby improving 3D reconstructions of individual trees derived using structure from motion (SfM). The BIV model comprises multiple layers of processing derived from measured or assumed responses of the photoreceptor cells in the hoverfly. Its pre-processing expands the range of input signal obtained from the LWIR sensor and enhances foreground-background contrast. Morphological image processing techniques were also applied to enhance key image features before structure from motion is applied. The result allows structural properties of individual trees to be characterised in terms of their potential volume and quality; and contrasted with the point clouds obtained from the visible imagery that only depicts the tree canopies.

Deep learning for liver tumor diagnosis part II: convolutional neural network interpretation using radiologic imaging features.

To develop a proof-of-concept "interpretable" deep learning prototype that justifies aspects of its predictions from a pre-trained hepatic lesion classifier. A convolutional neural network (CNN) was engineered and trained to classify six hepatic tumor entities using 494 lesions on multi-phasic MRI, described in Part 1. A subset of each lesion class was labeled with up to four key imaging features per lesion. A post hoc algorithm inferred the presence of these features in a test set of 60 lesions by analyzing activation patterns of the pre-trained CNN model. Feature maps were generated that highlight regions in the original image that correspond to particular features. Additionally, relevance scores were assigned to each identified feature, denoting the relative contribution of a feature to the predicted lesion classification. The interpretable deep learning system achieved 76.5% positive predictive value and 82.9% sensitivity in identifying the correct radiological features present in each test lesion. The model misclassified 12% of lesions. Incorrect features were found more often in misclassified lesions than correctly identified lesions (60.4% vs. 85.6%). Feature maps were consistent with original image voxels contributing to each imaging feature. Feature relevance scores tended to reflect the most prominent imaging criteria for each class. This interpretable deep learning system demonstrates proof of principle for illuminating portions ofa pre-trained deep neural network's decision-making, by analyzing inner layers and automatically describing features contributing to predictions. • An interpretable deep learning system prototypecan explain aspects of its decision-making by identifying relevant imaging features and showing where these features are found on an image, facilitating clinical translation. • By providing feedback on the importance of various radiological features in performing differential diagnosis, interpretable deep learning systems have the potential to interface with standardized reporting systems such as LI-RADS, validating ancillary features and improving clinical practicality. • An interpretable deep learning system could potentially add quantitative data to radiologic reports and serve radiologists with evidence-based decision support.

Ligaments and Lymphatic Pathways in Gastric Adenocarcinoma.

Gastric cancer is a leading cause of cancer-related deaths worldwide and is associated with an overall 5-year survival rate of less than 20%. The most common histologic subtype of gastric cancer is adenocarcinoma. Imaging techniques for evaluating gastric adenocarcinoma include endoscopic US, fluoroscopic upper gastrointestinal imaging, CT, PET/CT, and MRI. Hydrodynamic multiphasic contrast material-enhanced CT is the imaging modality of choice for preoperative clinical staging of regional, nodal, and metastatic involvement. Radiologic manifestations of gastric adenocarcinoma at double-contrast upper gastrointestinal imaging and CT include polyps, ulceration, indistensibility, wall thickening, and abnormal enhancement. There are multiple pathways of disease spread. These pathways include lymphatic dissemination; subperitoneal dissemination along the perigastric ligaments, mesentery, or omentum; direct invasion into adjacent organs; transperitoneal seeding; and hematogenous dissemination. The spread of disease is affected by the location of the tumor in the stomach, and the ligamentous and lymphatic anatomy. Key imaging features that affect clinical staging with use of the TNM classification system for gastric adenocarcinoma, as described in the eighth edition of the AJCC Cancer Staging Manual, are briefly discussed. Accurate radiologic assessment of gastric adenocarcinoma requires identification of perigastric ligament infiltration, regional and metastatic nodal disease, and direct and metastatic organ involvement, all of which directly affect tumor staging, treatment, and prognosis. ©RSNA, 2019.

Non-Contrast-Enhancing Tumor: A New Frontier in Glioblastoma Research.

There is a growing understanding of the prognostic importance of non-contrast-enhancing tumor in glioblastoma, and recent attempts at more aggressive management of this component using neurosurgical resection and radiosurgery have been shown to prolong survival. Optimizing these therapeutic strategies requires an understanding of the features that can distinguish non-contrast-enhancing tumor from other processes, in particular vasogenic edema; however, the limited and heterogeneous manner in which it has been defined in the literature limits clinical translation. This review covers pertinent literature on our growing understanding of non-contrast-enhancing tumor and focuses on key conventional MR imaging features for improving its delineation. Such features include subtle differences in the degree of FLAIR hyperintensity, gray matter involvement, and focal mass effect. Improved delineation of tumor from edema will facilitate more aggressive management of this component and potentially realize associated survival benefits.

Understanding gloss perception through the lens of art: Combining perception, image analysis, and painting recipes of 17th century painted grapes.

To understand the key image features that we use to infer the glossiness of materials, we analyzed the pictorial shortcuts used by 17th century painters to imitate the optical phenomenon of specular reflections when depicting grapes. Gloss perception of painted grapes was determined via a rating experiment. We computed the contrast, blurriness, and coverage of the grapes' highlights in the paintings' images, inspired by Marlow and Anderson (2013). The highlights were manually segmented from the images, and next the features contrast, coverage, and blurriness were semiautomatically quantified using self-defined algorithms. Multiple linear regressions of contrast and blurriness resulted in a predictive model that could explain 69% of the variance in gloss perception. No effect was found for coverage. These findings are in agreement with the instructions to render glossiness of grapes contained in a 17th century painting manual (Beurs, 1692/in press), suggesting that painting practice embeds knowledge about key image features that trigger specific material percepts.

Nontraumatic Orbital Emergencies

Non-traumatic orbital pathology can often have overlapping imaging features and challenging timely diagnosis. Prompt recognition is necessary due to potential risk of permanent visual loss. In this article we present systematic evaluation of non-traumatic orbital pathology with attention to key imaging features, pathogenesis and potential complications.

Magnetic resonance imaging of subchondral insufficiency fractures of the lower limb.

Subchondral insufficiency fracture (SIF) is a non-traumatic condition that has historically been associated with elderly, osteoporotic women and patients with systemic conditions. There has been much work done to determine the pathogenesis of SIF, which has previously been regarded as idiopathic, rapid-progressive osteoarthritis or osteonecrosis of the hip, spontaneous osteonecrosis of the knee (SONK), osteochondral defect (OCD) of the talus and adult-onset Freiberg infraction of the metatarsal head. Early diagnosis and management are crucial to prevent subchondral collapse, secondary osteonecrosis and early-onset osteoarthritis. Magnetic resonance imaging (MRI) plays an important role in the diagnosis of SIF, which is often inconspicuous on initial radiographs. In this article, the authors provide an update on the role of MRI in identifying key imaging features of SIF in various joints of the lower limb to aid in its correct diagnosis.

Demyelinating lesions behaving like aggressive tumours on advanced MRI techniques

Tumefactive demyelinating lesions are a rare disorder in which inflammatory demyelination manifests as solitary or multiple focal brain lesions (greater than 2 cm in size), which can be mistaken for glioma, lymphoma, metastasis and in some cases even brain abscess. The symptomatology of tumefactive demyelinating lesions depends on the white matter area involved and includes quickly progressing neurological deterioration of motor, sensory and visual function, praxis, language and mood impairment, as well as seizures. Recognising the key imaging features in a patient with a prior history of demyelination may expedite appropriate management. Preoperative diagnosis or at least the consideration of a demyelinating process is important to avoid unnecessary surgery. We report three patients with demyelinating lesions who presented with findings suggestive of demyelination on conventional magnetic resonance imaging studies. However, in all patients the lesions showed high perfusion and in two high permeability, which are findings generally seen with high-grade neoplasias. In rare instances, tumefactive demyelinating lesions may show increased perfusion and high permeability, imaging findings more commonly seen in high-grade gliomas. We suggest that if white matter lesions on conventional magnetic resonance imaging are compatible with tumefactive demyelinating lesions, atypical findings of high perfusion/permeability should not dissuade the radiologist from suggesting the presence of tumefactive demyelinating lesions rather than high-grade gliomas.

Connective Tissue Disorders in Childhood: Are They All the Same?

Systemic connective tissue disorders are characterized by the presence of autoantibodies and multiorgan system involvement. Juvenile systemic lupus erythematosus with or without associated antiphospholipid syndrome; juvenile dermatomyositis; sclerodermiform syndromes, including systemic and localized sclerodermas and eosinophilic fasciitis; mixed connective tissue disease; and Sjögren syndrome are the disorders that affect children most frequently. Diagnosis is difficult, because the clinical presentation of patients is diverse, from mild to severe disease. In addition, all organs may be affected. However, a variety of imaging techniques are now available to investigate rheumatic disease in children. These imaging modalities offer the potential for earlier diagnosis and improved assessment of therapeutic response. This article reviews the main connective tissue disorders that affect children, highlighting their key imaging features on images acquired with different diagnostic imaging modalities and correlating these features with clinical and pathologic findings, when available. ©RSNA, 2019.

Lung Masses of Unusual Histologies Mimicking Malignancy: Flurodeoxyglucose Positron Emission Tomography-Computed Tomography Appearance.

18F flurodeoxyglucose positron emission tomography-computed tomography (18F FDG PET-CT) is widely used in the evaluation of patients with lung mass suspicious for malignancy. In addition to malignancy, a variety of benign neoplasms and inflammatory lesions can arise in the lungs, many of which show increased FDG concentration, thereby mimicking malignancy. Awareness of the common mimics of lung cancer and a thorough understanding of their key imaging characteristics on CT as well as FDG PET is helpful in narrowing the differential diagnosis, eventually leading to appropriate therapy. In this article, we enlist these mimics and discuss their metabolic and morphologic characteristics and provide a pathophysiological basis for their FDG uptake.

Role of Cardiac Computed Tomography in the Diagnosis of Left Ventricular Myocardial Diseases.

Multimodality imaging is indicated for the evaluation of left ventricular (LV) myocardial diseases. Cardiac magnetic resonance (CMR) allows morphological and functional assessment of the LV along with soft tissue characterization. Technological advances in cardiac computed tomography (CT) have led to the development of techniques for diagnostic acquisition in LV myocardial disease. Cardiac CT facilitates the characterization of LV myocardial disease based on anatomy, function, and enhancement pattern. LV regional and global functional parameters are evaluated using multi-phasic cine CT images. CT myocardial perfusion facilitates the identification of hemodynamically significant coronary artery stenosis. Cardiac CT with delayed enhancement is used to detect myocardial scarring or fibrosis in myocardial infarction and non-ischemic cardiomyopathy, and for the measurement of extracellular volume fraction in non-ischemic cardiomyopathy. In this review, we review imaging techniques and key imaging features of cardiac CT used for the evaluation of myocardial diseases, along with CMR findings.

Pancreatic Incidentalomas: Review and Current Management Recommendations

AbstractThere has been significant increase in the detection of incidental pancreatic lesions due to widespread use of cross-sectional imaging like computed tomography and magnetic resonance imaging supplemented with improvements in imaging resolution. Hence, accurate diagnosis (benign, borderline, or malignant lesion) and adequate follow-up is advised for these incidentally detected pancreatic lesions. In this article, we would review the various pancreatic parenchymal (cystic or solid) and ductal lesions (congenital or pathological), discuss the algorithmic approach in management of incidental pancreatic lesions, and highlight the key imaging features for accurate diagnosis.

Automated detection of imaging features of disproportionately enlarged subarachnoid space hydrocephalus using machine learning methods.

ObjectiveCreate an automated classifier for imaging characteristics of disproportionately enlarged subarachnoid space hydrocephalus (DESH), a neuroimaging phenotype of idiopathic normal pressure hydrocephalus (iNPH).Methods1597 patients from the Mayo Clinic Study of Aging (MCSA) were reviewed for imaging characteristics of DESH. One core feature of DESH, the presence of tightened sulci in the high-convexities (THC), was used as a surrogate for the presence of DESH as the expert clinician-defined criterion on which the classifier was trained. Anatomical MRI scans were automatically segmented for cerebrospinal fluid (CSF) and overlaid with an atlas of 123 named sulcal regions. The volume of CSF in each sulcal region was summed and normalized to total intracranial volume. Area under the receiver operating characteristic curve (AUROC) values were computed for each region individually, and these values determined feature selection for the machine learning model. Due to class imbalance in the data (72 selected scans out of 1597 total scans) adaptive synthetic sampling (a technique which generates synthetic examples based on the original data points) was used to balance the data. A support vector machine model was then trained on the regions selected.ResultsUsing the automated classification model, we were able to classify scans for tightened sulci in the high convexities, as defined by the expert clinician, with an AUROC of about 0.99 (false negative ≈ 2%, false positive ≈ 5%). Ventricular volumes were among the classifier's most discriminative features but are not specific for DESH. The inclusion of regions outside the ventricles allowed specificity from atrophic neurodegenerative diseases that are also accompanied by ventricular enlargement.ConclusionAutomated detection of tight high convexity, a key imaging feature of DESH, is possible by using support vector machine models with selected sulcal CSF volumes as features.

Imaging Features and Management of Stress, Atypical, and Pathologic Fractures

Traumatic and atraumatic fractures are entities with distinct but often overlapping clinical manifestations, imaging findings, and management protocols. This article is a review of terminology, etiology, and key imaging features that affect management of atraumatic fractures including stress fractures, atypical femoral fractures, and pathologic fractures. The terminology of atraumatic fractures is reviewed, with an emphasis on the distinctions and similarities of stress, atypical, and pathologic fractures. The basic biomechanics of normal bone is described, with an emphasis on the bone remodeling pathway. This framework is used to better convey the shared etiologies, key differences, and important imaging findings of these types of fractures. Next, the characteristic imaging findings of this diverse family of fractures is discussed. For each type of fracture, the most clinically relevant imaging features that guide management by the multidisciplinary treatment team, including orthopedic surgeons, are reviewed. In addition, imaging features are reviewed to help discriminate stress fractures from pathologic fractures in patients with challenging cases. Finally, imaging criteria to risk stratify an impending pathologic fracture at the site of an osseous neoplasm are discussed. Special attention is paid to fractures occurring in the proximal femur because the osseous macrostructure and mix of trabecular and cortical bone of the proximal femur can function as a convenient framework to understanding atraumatic fractures throughout the skeleton. Atraumatic fractures elsewhere in the body also are used to illustrate key imaging features and treatment concepts. ©RSNA, 2018.