Jeff Hall

Only 41 percent of radiologists spotted radiographs that were generated by the large language model GPT-4o, according to newly published research.

The Photonova Spectra photon-counting CT system reportedly offers a variety of features, including 8-bin energy resolution, rapid rotation speed and motion-free imaging.

While smokers with negative baseline CT findings had a higher overall lung cancer risk than never smokers, the authors of a new study of over 30,000 people noted no elevated risk at two years.

Can the radioligand therapy 177Lu-PSMA-617 have an impact in older taxane-naïve men with mCRPC? Could we see the development of more combination therapies for patients with prostate cancer? Will emerging theranostics be effective in other targets beyond PSMA? Daniel George, M.D., addressed these topics and more in a recent interview.

Catch up on the top radiology content of the past week.

New study findings demonstrate that the deep learning software syngo.CT Brain Hemorrhage VB60 reportedly offers a 93.6 percent sensitivity as well as 99.2 percent negative predictive value for intracranial hemorrhage.

The expanded 510(k) clearance, which covers the TrueBeam, TrueBeamSTx, VitalBeam and Edge radiotherapy systems, may facilitate another treatment option for patients with osteoarthritis who have failed to get adequate pain relief from other treatment approaches.

Emerging PSMA PET-based nomograms offered 12 percent and 14 percent higher AUCs than the European Association of Urology risk score for risk stratification of patients with biochemical recurrence of prostate cancer, according to an international, registry-based study involving over 11,000 patients.

Emphasizing patient comfort and features to minimize radiation exposure, the MC2 portable X-ray device is now cleared for imaging of pediatric patients.

In an interview, W. Taylor Kimberly, M.D., discussed recent research showing that multidirectional diffusion-weighted imaging (DWI) on low-field MRI yielded a 95 percent positive predictive value (PPV) for detecting acute ischemic stroke.

More prominent honeycombing, reduced total ground glass opacity (GGO) and less pure GGO on CT scans each demonstrated 75 percent or higher AUCs for differentiating usual interstitial pneumonia (UIP) from non-UIP, according to new research.

Use of an MRI-based deep learning model also led to a greater than 33 percent increase in detecting brain metastases less than or equal to 3 mm, according to newly published research.

In the second part of a two-part interview, David Larson, M.D., M.B.A, and Jason Poff, M.D., shared their insights on the recent study of a structured pre-deployment approach to assessing AI models and potential directions in future radiology research.

Catch up on the top radiology content of the past week.

In the first of a two-part interview, David Larson, M.D., M.B.A., and Jason Poff, M.D., discussed a new study examining the use of a structured pre-deployment method for evaluating the value of a portfolio of AI models for radiology.

In a new podcast episode, Daniel George, M.D., Brian Helfand, M.D., Ph.D., and Ashesh Jani, M.D., discussed recent study findings presented at the ASCO-GU symposium, directions for future research and emerging trends in molecular imaging that may help advance precision medicine for patients with prostate cancer.

Providing 3D ultrasound tomographic views, the Breast Acoustic CT device has an updated configuration that reportedly improves coverage and visualization of posterior breast tissue.

The MRI-based Early Recurrence Individualized Score (MERIS) demonstrated an 18 percent higher C-index than the Barcelona Clinic Liver Cancer (BCLC) staging system for predicting early recurrence of hepatocellular carcinoma (HCC) for patients who had resection of solitary HCC tumors.

Iodine density measurements with photon-counting detector CT offered higher than 93 percent sensitivity and 95 percent specificity for differentiating between bland and neoplastic portal vein thrombosis in newly published research.

Reportedly providing automation of 14 standard and advanced cardiac MRI views in less than 30 seconds, the SmartHeart also facilitates a 75 percent reduction in breath holds for basic MRI views.

Initial breast cancer risk scoring with adjunctive AI assessment of screening mammograms increased by approximately 80 percent in a subsequent screening round for women with breast cancer, according to research findings presented at the European Congress of Radiology.

Utilized within the cloud-native Genesis Radiology Workspace, the newly FDA-cleared View reportedly facilitates anytime, anywhere efficiency with seamless access to enhanced visualization and AI-powered tools.

Catch up on the top radiology content of the past week.

Designed to enhance product stability at higher radioactive concentrations, the PSMA PET imaging agent Pylarify TruVu (piflufolastat F 18) may facilitate enhanced supply flexibility and access for men with suspected recurrence or metastasis of prostate cancer.

While large language models (LLMs) may offer the promise of workflow efficiencies and streamlined documentation in radiology, LLMs may also increase exposure to cybersecurity risks, according to a presentation at the European Congress of Radiology (ECR).

The Harrison.ai software is reportedly the first AI software to gain clearance for acute infarct triage on non-contrast CT scans.

Emerging research reveals higher attrition rates among cardiothoracic radiologists, breast radiologists and nuclear medicine physicians in comparison to other subspecialties in radiology.

The MEDRAD MRXperion MR Injection System is cleared for use with low-field and high-field MRI platforms, including 7T MRI.

Reportedly the first generative AI model in radiology to receive the FDA breakthrough device designation, Cognita Chest X-Ray (CXR) facilitated an 18 percent improvement in CXR interpretation efficiency based on preliminary data.

In addition to significantly reduced interpretation time, researchers found that adjunctive AI with ultrasound resulted in accuracy improvements of greater than 7 percent for nodule echogenicity and 16 percent for nodule margin assessments.