| Editor's Choice: Artificial Intelligence | Although large language models produce appeals letters that are generally deemed useful by readers, they likely still require a babysitter, according to a new analysis. Although large language models produce appeals letters that are generally deemed useful by readers, they likely still require a babysitter, according to a new analysis. | | |
| In new Neiman Health Policy Institute research, scientists explored the use of artificial intelligence to aid in diagnosing PE on CTPA, with “AI-informed radiologists" scoring high marks. In new Neiman Health Policy Institute research, scientists explored the use of artificial intelligence to aid in diagnosing PE on CTPA, with “AI-informed radiologists" scoring high marks. | | |
| New findings question the use of artificial intelligence in these settings due to numerous shortcomings, including issues with overdetection and low specificity. New findings question the use of artificial intelligence in these settings due to numerous shortcomings, including issues with overdetection and low specificity. | | |
| RadNet Chaiman and CEO Howard Berger, MD, explains why the company has invested tens of millions into DeepHealth to rapidly build up a new business model. RadNet Chaiman and CEO Howard Berger, MD, explains why the company has invested tens of millions into DeepHealth to rapidly build up a new business model. | | |
| The Scottsdale, Arizona-based radiology practice said Monday it’s seeking to scale AI across all routine imaging, hoping to deliver “preventive insights nationwide.” The Scottsdale, Arizona-based radiology practice said Monday it’s seeking to scale AI across all routine imaging, hoping to deliver “preventive insights nationwide.” | | |
| Leaders voted to OK the new Practice Parameter for Imaging Artificial Intelligence during ACR’s annual meeting, being held in Washington, D.C. Leaders voted to OK the new Practice Parameter for Imaging Artificial Intelligence during ACR’s annual meeting, being held in Washington, D.C. | | |
| New findings support the routine use of deep learning-based risk assessments, as this method can decrease subjectivity, reduce unnecessary imaging and improve diagnostic accuracy. New findings support the routine use of deep learning-based risk assessments, as this method can decrease subjectivity, reduce unnecessary imaging and improve diagnostic accuracy. | | |
| The AI framework was designed to identify radiomic signatures of pre-diagnostic pancreatic ductal adenocarcinoma on CT. The AI framework was designed to identify radiomic signatures of pre-diagnostic pancreatic ductal adenocarcinoma on CT. | | |
| The COlorectal Cancer detection with AI, or COCA, model is a cost-effective, scalable solution that turns routine CT scans into opportunistic exams that can be used to proactively identify CRC. The COlorectal Cancer detection with AI, or COCA, model is a cost-effective, scalable solution that turns routine CT scans into opportunistic exams that can be used to proactively identify CRC. | | |
| Patricia Balthazar, MD, Emory University School of Medicine, says radiology AI needs to be monitored to ensure it is performing as it is supposed to and not wasting time and money. Patricia Balthazar, MD, Emory University School of Medicine, says radiology AI needs to be monitored to ensure it is performing as it is supposed to and not wasting time and money. | | |
| |
|
| ![]() | |
|
