A Roadmap for Foundational Research on Artificial Intelligence in Medical Imaging: From the 2018 NIH/RSNA/ACR/The Academy Workshop.

Imaging research laboratories are rapidly creating machine learning systems that achieve expert human performance using open-source methods and tools. These artificial intelligence systems are being developed to improve medical image reconstruction, noise reduction, quality assurance, triage, segmentation, computer-aided detection, computer-aided classification, and radiogenomics. In August 2018, a meeting was held in Bethesda, Maryland, at the National Institutes of Health to discuss the current state of the art and knowledge gaps and to develop a roadmap for future research initiatives. Key research priorities include: 1, new image reconstruction methods that efficiently produce images suitable for human interpretation from source data; 2, automated image labeling and annotation methods, including information extraction from the imaging report, electronic phenotyping, and prospective structured image reporting; 3, new machine learning methods for clinical imaging data, such as tailored, pretrained model architectures, and federated machine learning methods; 4, machine learning methods that can explain the advice they provide to human users (so-called explainable artificial intelligence); and 5, validated methods for image de-identification and data sharing to facilitate wide availability of clinical imaging data sets. This research roadmap is intended to identify and prioritize these needs for academic research laboratories, funding agencies, professional societies, and industry.

Safety of unprotected carotid artery stent placement in symptomatic and asymptomatic patients: a retrospective analysis of 30-day combined adverse outcomes.

PURPOSE: To retrospectively evaluate the outcome of carotid artery stent placement (CAS) without the use of embolic protection devices (EPDs) in a large cohort of patients. MATERIALS AND METHODS: Institutional review board approval and informed consent from all patients were obtained. Preprocedure color Doppler ultrasonography (US), magnetic resonance (MR) imaging, or computed tomography (CT) were used to evaluate stenosis severity (70% or greater). Clinical findings and combined 30-day complication rates in 400 patients (289 men, 111 women; mean age, 73 years +/- 8 [standard deviation]) who underwent unprotected CAS for asymptomatic (n = 156; 39%) or symptomatic (n = 244, 61%) stenoses were analyzed. Follow-up at 30 days included neurologic evaluation and color Doppler US. RESULTS: Self-expanding stents were successfully deployed in 397 of 400 (99.25%) patients. Among the 397 patients, nine (2.27%) major complications (all in patients with prior symptoms) had occurred at 30 days, including three (0.76%) major (all in patients who had stopped antiplatelet prophylaxis) and six (1.5%) minor strokes--three intraprocedural and three delayed. Minor complications included 16 (4%) transient ischemic attacks, four in asymptomatic and 12 in symptomatic patients. The 30-day combined adverse outcomes (transient ischemic attack, ipsilateral stroke, death) were significantly correlated with prior presence of symptoms (symptomatic, 8.6%; asymptomatic, 2.6%; P < .03). CONCLUSION: Stent placement without EPD was performed with a high technical success rate. For asymptomatic patients, the combined 30-day adverse-outcomes rate was within the limits recommended by the American Heart Association for carotid endarterectomy and compared favorably with results reported for CAS with EPD. When a transient ischemic attack is excluded, the 30-day combined death and stroke rate among patients with prior symptoms also compared favorably with published results.

Human health during space travel: An overview.

This article reviews current challenges to health in space and has a secondary intention to set the tone for this special supplement on 'Extra-Terrestrial Neurosciences'. The effects of microgravity, radiation, isolation, disturbance in circadian rhythms and the hostile environment on the cardiovascular, neurological, immunological and various biological human systems are discussed here. Alterations in physiology, environmental hazards, and mitigative safety measures are briefly discussed along with challenges encountered in providing remote diagnoses and health care during space missions.

A Road Map for Translational Research on Artificial Intelligence in Medical Imaging: From the 2018 National Institutes of Health/RSNA/ACR/The Academy Workshop.

Advances in machine learning in medical imaging are occurring at a rapid pace in research laboratories both at academic institutions and in industry. Important artificial intelligence (AI) tools for diagnostic imaging include algorithms for disease detection and classification, image optimization, radiation reduction, and workflow enhancement. Although advances in foundational research are occurring rapidly, translation to routine clinical practice has been slower. In August 2018, the National Institutes of Health assembled multiple relevant stakeholders at a public meeting to discuss the current state of knowledge, infrastructure gaps, and challenges to wider implementation. The conclusions of that meeting are summarized in two publications that identify and prioritize initiatives to accelerate foundational and translational research in AI for medical imaging. This publication summarizes key priorities for translational research developed at the workshop including: (1) creating structured AI use cases, defining and highlighting clinical challenges potentially solvable by AI; (2) establishing methods to encourage data sharing for training and testing AI algorithms to promote generalizability to widespread clinical practice and mitigate unintended bias; (3) establishing tools for validation and performance monitoring of AI algorithms to facilitate regulatory approval; and (4) developing standards and common data elements for seamless integration of AI tools into existing clinical workflows. An important goal of the resulting road map is to grow an ecosystem, facilitated by professional societies, industry, and government agencies, that will allow robust collaborations between practicing clinicians and AI researchers to advance foundational and translational research relevant to medical imaging.

Adult intestinal intussusception: can abdominal MDCT distinguish an intussusception caused by a lead point?

BACKGROUND: Aim of our study was to assess the ability of computed tomography to distinguish between an intussusception with a lead-point from one without it. METHODS: Approval was granted by the Institutional Review Board. Ninety-three consecutive patients diagnosed with an intussusception on abdominal CT were classified with or without lead-point by surgery, clinical or radiological follow-up. Two radiologists blinded to the classification independently reviewed the CT images for predefined predictive variables. RESULTS: Non-lead-point intussusception was shorter in length (mean 4.9 vs. 11.1 cm for Reader 1 (R1); mean 4.0 vs. 8.9 cm for Reader 2 (R2), respectively, P < 0.001), smaller in axial diameter (mean 3.0 vs. 4.8 cm for R1; mean 2.8 vs. 4.4 cm for R2, P < 0.001, respectively), less likely associated with obstruction (P = 0.002 R1; P = 0.039 R2) and infiltration (P < 0.001 for R1, P = 0.003 R2) than lead-point intussusception. CONCLUSIONS: Abdominal CT is helpful in distinguishing between an intussusception with a lead-point from one without a lead-point. Length, axial diameter, and their product, as well as obstruction and infiltration, all suggest the presence of a lead-point. Analysis of CT findings can reduce unnecessary imaging follow-up or operation.

Giant colonic diverticulum: percutaneous computed tomography-guided treatment.

We present a case of inflamed giant sigmoid diverticulum, which was treated under computed tomographic guidance by dilating the neck and decompressing the inflamed diverticulum. The procedure also involved placement of an internal stent across the neck of the diverticulum, the purpose of which was to decompress the giant diverticulum into the sigmoid colon and prevent acute recurrence. In patients with inflamed giant colonic diverticulum, the percutaneous approach may be useful in treatment of diverticulitis and avoiding a staged Hartmann procedure with end colostomy.

Dropped gallstones and surgical clips after cholecystectomy: CT assessment.

AIM: To describe the spectrum of computed tomographic (CT) findings in patients with dropped gallstones or dropped surgical clips after cholecystectomy. MATERIALS AND METHODS: Seventeen patients diagnosed with dropped gallstones and 26 patients with dropped surgical clips on computed tomography after cholecystectomies were included in this study. The CT scans were evaluated for the number, location, size, and density of dropped gallstones or surgical clips and for the presence of an abscess. The cases were evaluated for the outcome on clinical and/or CT follow-up. RESULTS: The location for dropped gallstones and dropped surgical clips after cholecystectomy was the Morrison's pouch in 17 and 12 patients, respectively. There were 9 abscesses in the study, all located in the Morrison's pouch. Duration from surgery to observation of abscess on computed tomography ranged from 5 days to more than 4.7 years. None of the patients with a dropped surgical clip from cholecystectomy developed an associated abscess. CONCLUSIONS: Subhepatic location was the most common location for dropped gallstones with associated abscess and for dropped surgical clips. Dropped cholecystectomy clips are not associated with increased risk of abscess formation and therefore do not need screening follow-up or operative removal. Abscess formation around dropped gallstone is a more common complication and requires surgical treatment in most when associated with an abscess.

Safety of thrombolytic therapy with urokinase or recombinant tissue plasminogen activator for peripheral arterial occlusion: a comprehensive compilation of published work.

PURPOSE: To report a comprehensive literature review focused on comparing the risk of complications with urokinase versus recombinant tissue plasminogen activator (rtPA) for thrombolytic treatment of peripheral arterial occlusions. METHODS: The English-language literature between 1985 and 2002 was searched for studies that used tissue-derived urokinase or rtPA in the treatment of peripheral arterial occlusions. Forty-eight studies (22 urokinase, 22 rtPA, and 4 that included both treatments) were identified, encompassing 2226 urokinase-treated patients and 1927 rtPA-treated patients. The safety of each thrombolytic agent was assessed based on the incidence of major hemorrhage, intracerebral hemorrhage, major limb amputation, transfusions, and mortality. RESULTS: The review revealed a wide range of study protocols, patient conditions, ages of occlusions, dosages/delivery methods of lytic agents, and criteria for reporting complications. The incidence of major hemorrhage varied widely, but the overall rate was lower among urokinase-treated patients (6.2%) than for patients treated with rtPA (8.4%, p=0.007). The overall incidence of intracerebral hemorrhage was also significantly lower for urokinase (0.4% versus 1.1% for rtPA, p=0.020). The major amputation rate was similar for both treatments (urokinase 7.9%, rtPA 7.2%), but the mortality rate was significantly lower for urokinase (3.0% versus 5.6% for rtPA, p<0.001). The need for transfusions was less frequent with urokinase (11.1% versus 16.1%, p=0.002). CONCLUSIONS: These results from a large body of published literature suggest that urokinase may be associated with a lower incidence of complications than rtPA in the treatment of peripheral arterial occlusions.