Articles tagged with Radiology
Researchers have developed magnet-guided microrobots that can target and treat liver tumors using an MRI device. The robots are guided by a magnetic field and use gravity to navigate to the tumor, preserving healthy cells.
The American College of Radiology has issued a joint statement with four other radiology societies to address the development and use of AI tools in radiology. The statement emphasizes the need for increased monitoring of AI utility and safety, advocating for collaboration among developers, clinicians, purchasers, and regulators.
Researchers developed an AI model using CT images to predict lymph node metastasis in non-functional pancreatic neuroendocrine tumors. The model achieved high accuracy, even for smaller tumors, and can help guide surgical decisions.
A new study found that propranolol, a medication for high blood pressure, can lower anxiety levels in individuals with autism spectrum disorder. The study involved 69 patients and showed significantly reduced anxiety levels compared to a placebo group.
Researchers at the University of Münster found that DBT+SM detects invasive breast cancers more effectively than conventional DM, with a higher rate of early tumor stages. This approach may lead to improved screening effect and reduced breast cancer mortality.
A study found that the No Surprises Act's independent dispute resolution (IDR) process would be financially unfeasible for a large portion of out-of-network claims for hospital-based specialties, particularly radiologists. This could undermine patient access to in-network care due to limited bargaining power.
Experts recommend increasing the use of intravascular ultrasound (IVUS) in lower extremity revascularization procedures to optimize outcomes and reduce complications. The technique provides detailed information about vessel walls, plaque composition, and blood flow characteristics, enabling more accurate diagnosis and treatment planning.
A new AI system developed by the University of Technology Sydney can rapidly detect COVID-19 from chest X-rays with high accuracy. The Custom Convolutional Neural Network (Custom-CNN) model streamlines the detection process, providing a faster and more accurate diagnosis.
A new article in Radiology discusses AI applications for musculoskeletal radiology, highlighting its potential to improve complex tasks like disease prognostication and prediction of clinical outcomes. However, challenges such as data quality and multi-institutional collaboration need to be overcome for successful implementation.
A new framework has been established for standardized imaging of diffuse gliomas using amino acid PET, enabling the evaluation of treatment success and improving therapies. The RANO group has developed criteria that enable reliable imaging of tumor activity and extent.
Researchers at UCLA Health have received a $3 million grant to identify novel biomarkers and develop AI for detecting aggressive prostate cancer. The project aims to improve treatment accuracy and reduce unnecessary interventions.
Researchers analyzed over 24 million assault-related injuries and found that 40% resulted in anoxia, with IPV accounting for 30-40% of neck contusions. The study aims to raise awareness of strangulation among medical providers and advocate for comprehensive screening in IPV patients.
Researchers at Temple University Health System report that the BASHIR Endovascular Catheter reduces blockages in lung arteries and correlates with improved right ventricle function, suggesting better survival outcomes. The catheter also shows significantly lower bleeding rates compared to other devices.
A new study has found that most patients with osteoporosis want to receive information on their fracture risk, but only half of them actually get it. Patients prefer visual representations, such as graphs with a colored traffic-light system, to communicate fracture risk and are more likely to take medication if presented in this way.
Researchers used a stellate ganglion block to restore the sense of smell in patients with long-COVID, showing near 100% resolution of phantosmia in some patients. The treatment was effective in improving symptoms for up to 49% of patients at three months.
A new study published in Radiology shows that patients diagnosed with lung cancer via CT screening have a 20-year survival rate of 81 percent, significantly higher than the average five-year survival rate for all lung cancer patients. Early detection through routine screening can lead to a cure if found early.
Researchers at the University of Cincinnati Cancer Center found that SmartClips, a wireless localization method for breast tumors, is safe and effective. The devices reduce patient discomfort and improve surgery efficiency by allowing surgeons to locate tumors more accurately.
Researchers used dynamic total-body PET scans to visualize immune T cell distribution in recovering patients. The study found increased concentrations of CD8+ T cells in the bone marrow of recovering COVID patients compared to healthy controls.
A simple MRI scan and a new metric for prostatic urethra length may help identify men at risk of chronic urinary side effects after radiation therapy. The study found that longer prostatic urethras are associated with a higher likelihood of these symptoms, which can impact quality of life.
A new method using deep learning can provide as much information from brain images taken with CT as images captured with MRI, enhancing diagnostic support for conditions like dementia and other brain disorders. The software has been trained on 1,117 people and shows promise in diagnosing normal pressure hydrocephalus.
A new study from Linköping University has found that the appearance of the thymus gland in chest CT scans is linked to immune system ageing. The researchers examined over 1,000 Swedish individuals aged 50-64 and found that fatty degeneration of the thymus was more common in men and those with abdominal obesity. Lifestyle factors such a...
Researchers tested the feasibility of using locally run LLMs like Vicuna-13B to label key findings in chest X-ray reports while preserving patient privacy. The results showed moderate to substantial agreement with non-LLM computer programs, suggesting that these models can be a viable option for AI research.
The study developed a highly accurate AI model for fully automated cancer detection, including small and difficult-to-detect tumors. The model could detect visually imperceptible cancer from normal-appearing pancreases substantially early before clinical diagnosis, with a median of 438 days.
A new AI tool developed by Brazilian researchers can detect potentially cancerous lung nodules in CT reports, missing a crucial early diagnosis. The NLP tool achieved an accuracy rate of 97% in identifying suspicious nodules.
A new generative AI model enhances clinical accuracy and textual quality of chest radiograph reports, surpassing traditional methods like teleradiology. Its implementation could enable timely alerts for life-threatening conditions, aiding imaging interpretation and documentation.
A new AI model integrates imaging and non-imaging patient data for improved diagnostic performance on chest X-rays. The multimodal model outperformed other models for diagnosing up to 25 conditions, showing potential as an aid to clinicians in high-pressure diagnoses.
A new size threshold for retropharyngeal lymph nodes is proposed to improve risk stratification and treatment decisions in patients with nasopharyngeal carcinoma. Using a 6-mm threshold, the study found significant differences in overall survival between stage-I and stage-II disease.
A study comparing radiologist and AI performance in interpreting 2,000 chest X-rays found that AI tools achieved moderate sensitivity but high false-positive rates. Radiologists outperformed AI in detecting the absence of disease, especially for complex cases.
A study comparing PI-RADS 2.0 and 2.1 found no significant differences in upgrade (29% vs. 22%) or downgrade (19% vs. 21%) rates from targeted biopsy to radical prostatectomy, suggesting no improvement in prostate cancer grade assessment with the latest PI-RADS update.
An AI system by Lunit identified improperly positioned endotracheal tubes with high sensitivity and specificity, allowing for earlier repositioning and reducing complications. The study included 539 chest radiographs from three institutions and showed promising results in detecting ETT presence and improper position.
The European Society of Cardiology calls for competency-based cardiac imaging delivery to enhance effective and efficient patient care. This approach enables cardiologists to deliver high-quality imaging services, select the most appropriate modality for each clinical scenario, and provide informed decision-making.
Researchers developed a deep-learning model to assess CXR images for probable COVID-19 severity. The model achieved an area under the receiver operating characteristic curve of 0.78 when predicting intensive care need within 24 hours.
A new study by Harvard Medical School researchers found that automated scoring systems for AI-generated radiology reports fail to reliably identify clinical errors. The team designed a new method and composite evaluation tool to better evaluate the performance of AI tools in generating clinically useful and trustworthy reports.
A radiomic-based model using T2-weighted MRI data achieved high accuracy in diagnosing pediatric Crohn disease, outperforming expert radiologists. The model was ensembled with clinical data to further improve performance.
Researchers at Lund University found AI-supported mammography screening to be safe and effective, detecting 20% more cancers than standard double reading without increasing false positives. The study also reduced radiologists' screen-reading workload by 44%, saving approximately five months of time.
A commercial AI tool did not provide additional benefit to mammography with supplementary ultrasound in patients with dense breasts. Mammography with supplementary ultrasound showed higher accuracy, specificity, and lower recall rate compared to AI alone or combined with US.
A new position paper reviews TBS for fracture risk prediction, treatment initiation and monitoring in osteoporosis. The evidence-based guide provides practical guidance for clinicians to integrate TBS into clinical practice.
Lacking Medicare coverage for screening CTC may contribute to greater income-based differences in its use compared to other recommended screening strategies or diagnostic CTC. Medicare coverage of CTC could reduce income-based disparities for individuals avoiding optical colonoscopy due to invasiveness or complication risk.
A synoptic report improved the completeness of pretreatment CT reports in patients with advanced ovarian cancer. Involvement of surgically established unresectable or challenging-to-resect disease sites was mentioned in all disease-specific synoptic reports, compared to 37% of simple structured reports.
Researchers developed a chest imaging protocol using photon-counting CT, allowing for simultaneous evaluation of lung structure, ventilation, vasculature, and perfusion. The protocol showed advantages over standard CT, providing high image quality at lower radiation doses and better spectral resolution.
Using MRI instead of CT with CTA alone in patients presenting to the ED with dizziness was associated with greater critical neuroimaging results and more frequent echocardiography. Patients with dizziness undergoing MRI showed a lower frequency of 90-day ED readmissions compared to those discharged after CT with CTA alone.
Researchers at Paul Scherrer Institute have developed a beamline modification called momentum cooling to increase the transmission of proton beams, allowing for shorter treatment times. This breakthrough could enable faster treatments while maintaining accuracy and sparing healthy tissue, with potential benefits including reduced costs...
A new AI tool utilizing deep-learning algorithms and volumetric thresholds accurately detected splenomegaly in patients undergoing CT scans. The study analyzed a large sample of 8,901 patients, revealing that splenic volume is strongly associated with weight.
A deep learning-based model developed in 7,105 patients predicted 30-day all-cause mortality in patients with CAP with AUCs ranging from 0.77 to 0.80. The model showed higher specificity than the CURB-65 score at the same sensitivity.
A study published in Radiology found that AI algorithms with high diagnostic accuracy improved radiologists' detection of lung cancers on chest X-rays. High-accuracy AI led to more frequent changes in reader determinations, suggesting increased human trust in AI.
The study found that preferential use of pulmonary MRA for diagnosing PE in the general population helped conserve iodinated contrast media. This substitution reduced iodine usage by approximately 27 liters during the specified time period, which is a noteworthy benefit considering the shortage's impact.
Researchers found that ChatGPT 4 can accurately identify appropriate imaging services for breast cancer screening and breast pain. The AI model outperformed its predecessor, ChatGPT 3.5, in providing evidence-backed guidelines for radiologists.
A study found that patients with transient ischemic attack (TIA) who received incomplete neurovascular imaging were at higher odds of experiencing a stroke within 90 days. The research, published in the American Journal of Roentgenology, used Medicare claims data to analyze TIA emergency department encounters.
A recent study published in Radiology found that AI algorithms performed better than the standard Breast Cancer Surveillance Consortium (BCSC) risk model for predicting five-year breast cancer risk. The AI models extracted hundreds of additional mammographic features, leading to improved predictive performance.
A multicenter study found that deep learning-based computer-aided diagnosis significantly improved radiologists' diagnostic performance for breast lesion classification on ultrasound. The use of CAD resulted in a 6.0% upgrade of BI-RADS category 3 assessments to category 4A, with 16.7% being malignant.
Researchers investigate the potential of ChatGPT in clinical radiology, highlighting its positive considerations, such as enhancing patient care and radiologist education, as well as negative concerns, like accuracy and transparency issues. The paper aims to initiate an engaging discussion on the use of AI-powered chatbots in this field.
A study found that expert mammogram radiologists exhibit visual hindsight bias, leading to improved performance on blurry images after prior clear images were viewed. This bias can impact the detection of lesions and has implications for negligence lawsuits and patient outcomes.
A recent study found that COVID-19 infection can lead to increased infarct growth and ongoing tissue consumption even after successful reperfusion through angiography. This suggests a potentially aggressive clinical course for patients with COVID-19 and large-vessel occlusion acute ischemic stroke.
Online radiologists choose studies based on financial attractiveness, leading to delays in high-priority cases. This study found that expedited priority class contained the highest percentage of delayed studies.
Large language models like ChatGPT show incredible potential in radiology, but also struggle with reliability and accuracy. The latest model GPT-4 improves advanced reasoning capabilities and performs better on higher-order thinking questions.
A study of 5302 patients found no significant difference in bleeding events between those temporarily discontinuing and maintaining antithrombotic therapy during image-guided breast biopsies. The findings support the safety of continuing antithrombotic therapy, but counsel patients on bruising risk.
The American College of Radiology (ACR) has introduced new guidelines recommending earlier and more intensive breast cancer screening for high-risk women, including Black women. According to the ACR, risk assessment by age 25 is necessary to determine if screening should start before age 40.
Researchers successfully performed the first in-utero surgery to repair a potentially deadly developmental condition called vein of Galen malformation. The procedure involved treating an aggressive vascular malformation before birth, resulting in improved cardiac output and no signs of negative effects on the brain after birth.
A new onsite deep-learning based FFR-CT algorithm demonstrates excellent diagnostic performance, with high interobserver and intraobserver reproducibility. The algorithm showed AUC of 0.975, sensitivity of 93.5%, and specificity of 97.7% for hemodynamically significant stenosis.
A study published in Radiology found that AI-based decision support systems can impair radiologist accuracy on mammograms, particularly for less experienced radiologists. Even highly experienced radiologists were adversely impacted by the system's judgments, highlighting the need for safeguards to mitigate automation bias.