Technology

While ChatGPT has the potential to help streamline responses to imaging-related questions from patients, the authors of a new study found that a third of ChatGPT responses to unprompted questions on medical imaging were not “fully relevant.”

In a multicenter cohort of patients with interstitial lung disease (ILD), a deep learning classification tool demonstrated an 81 percent sensitivity rate and a 77 percent specificity rate for predicting usual interstitial pneumonia on computed tomography (CT) scans.

Artificial intelligence (AI) software assigned high malignancy risk scores to mammography exams completed up to two years prior to breast cancer diagnosis in over 38 percent of screen-detected cancer cases and over 39 percent of interval cancer cases, according to newly published research.

The Allia IGS Pulse reportedly has the first monopolar X-ray tube geared to interventional procedures.

Utilized in conjunction with screening digital mammography or digital breast tomosynthesis (DBT), the artificial intelligence (AI)-powered software cmAngio may help detect and localize breast arterial calcification (BAC), an incidental finding that has been linked to an elevated risk for heart disease and stroke.

In reportedly the eighth FDA 510(k) clearance for Hyperfine’s Swoop portable brain MRI device, clinicians will now have access to enhanced denoising capabilities with diffusion-weighted imaging (DWI) sequences.

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

Geared to clinicians who may lack ultrasound expertise, the artificial intelligence (AI)-powered Caption Guidance technology reportedly provides stepwise instruction for acquiring optimal cardiac ultrasound images.

While the artificial intelligence (AI) triage system significantly reduced wait times for computed tomography pulmonary angiography (CTPA) results that were positive for pulmonary embolism (PE), researchers found no significant differences in the use of adjunctive AI with respect to accuracy rates and specificity rates.

The Rology teleradiology platform may improve access to top tier-imaging interpretation for stand-alone imaging centers, rural facilities and low-income patients.

Integrating clinical data parameters along with chest X-ray findings, the multimodal artificial intelligence (AI) model outperformed the use of clinical parameters only and chest X-rays only for various conditions ranging from congestive heart failure and chronic kidney disease to hypertension and pneumonia.

Leveraging artificial intelligence (AI) algorithms, the qXR-CTR reportedly provides automated cardiothoracic ratios (CTRs) through assessment of plain chest radiographs.

Is the continued rise of artificial intelligence (AI) driven by legitimate “machine learning,” or do the frequently “suspicious” chest X-rays and questionable detection of subtle findings on head CTs reveal a hype-driven train of new products with an ultimately “planned obsolescence”?

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

In a multicenter study examining four commercially available artificial intelligence (AI) software products for chest X-rays in over 2,000 patients, researchers found sensitivity rates ranging between 33 to 61 percent for vague airspace disease and 9 to 94 percent for small pneumothorax and pleural effusion.

The MAGNETOM Viato.Mobile may facilitate improved access to MRI capabilities for patients with serious health conditions who lack geographic proximity to centers with advanced imaging.

The Exo Iris reportedly combines advanced silicon technology with imaging, workflow software and artificial intelligence (AI) to facilitate multidisciplinary use of point-of-care ultrasound.

Catch up on the top AI-related news and research in radiology over the past month.

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

The Pulmonary Embolism (PE) Triage and Notification (PETN) module with RapidAI’s RapidPE platform reportedly demonstrated over 90 percent sensitivity and specificity rates for detection of central PE on computed tomography (CT), according to multicenter study findings presented at the 9th Annual Pulmonary Embolism Symposium.

New research suggests the capability of a mammography-based deep learning model to identify women at high risk of breast cancer led to more than triple the cancer detection rate on breast MRI in comparison to traditional risk assessment tools.

The artificial intelligence (AI)-enabled AIRAscore can reportedly provide quantitative brain volume data within five minutes of assessing brain MRI scans.

The use of automated detection of large vessel occlusion on computed tomography (CT) through artificial intelligence (AI) software reportedly led to an 11.2-minute reduction in triage time from the completion of imaging to initiation of endovascular therapy, according to newly published research.

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

Indicated for the triage and notification of obstructive hydrocephalus on non-contrast brain computed tomography (CT), the artificial intelligence (AI)-enabled software is reportedly the first radiology triage modality to obtain the Food and Drug Administration’s (FDA) Breakthrough Device Designation.

In a prospective study of over 55,000 women who had screening mammography, researchers found that double-reading by a radiologist and artificial intelligence (AI) was non-inferior to double-reading by two radiologists in detecting breast cancer.

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

Through the use of artificial intelligence (AI) and imaging modalities such as ultrasound, CT, and MRI, the newly FDA-cleared VisAble.IO software reportedly enhances planning and real-time assessment for liver tumor ablation procedures.

In separate test sets that included challenging mammography cases, researchers found that artificial intelligence (AI) demonstrated similar sensitivity and specificity for detecting breast cancer in comparison to assessments from over 500 clinicians.

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