Menu
____________________
vRad is at the forefront of artificial intelligence for radiology—accelerating the speed of care and improving the quality of diagnosis for life-threatening pathologies.
Studies Reviewed Annually
Annual Diagnosis Corrections
Up To
Turnaround Time Improvement
Mistakes happen in healthcare.
To quote the landmark 1999 publication,
“To Err is Human.”
AI provides a
powerful safety net,
catching mistakes quickly and reducing the impact on patient care.
AI as a “safety net” does not guarantee all critical pathologies listed will be detected. AI improves the probability of detection, as indicated by 1,500 corrected diagnoses in 2023.
In 2023, vRad AI identified 1,500 diagnostic errors—involving 8 critical pathologies—leading to a timely clinical correction.
vRad’s AI models for quality assurance are designed to augment our already-robust radiologist QA program. If an error is confirmed by a second radiologist, that radiologist quickly notifies the care team to ensure patient safety.
AI has become a powerful tool to further improve vRad’s overall quality of care.
In 2023, vRad AI identified 1500 diagnostic errors—involving 8 critical pathologies—leading to a timely clinical correction.
vRad’s AI models for quality assurance are designed to augment our already-robust radiologist QA program. If an error is confirmed by a second radiologist, that radiologist quickly notifies the care team to ensure patient safety.
AI has become a powerful tool to further improve vRad’s overall quality of care.
While 1,500 is a significant number of corrections, vRad radiologists read 6.5M studies annually and our well-documented error rate of 1.3 major misses per 1,000 reads is one-third that of published benchmarks such as the Wilson Wong study (JACR, 2005).
While 1,500 is a significant number, vRad radiologists read 6.5M studies annually and our well-documented error rate of 1.3 major misses per 1,000 reads is one-third that of published benchmarks such as the Wilson Wong study (JACR, 2005). AI has become a powerful tool to further improve vRad’s overall quality of care.
All relevant images sent to vRad are automatically reviewed for quality assurance by AI for 8 critical pathologies.
Aortic dissection(Chest, Abd CT)
Epidural lesion(Spine CT)
Intracranial hemorrhage(Head CT)
Pneumoperitoneum(Chest XR)
Pneumothorax(Chest CT)
Pulmonary embolism (Chest CTA)
Splenic laceration(Abdomen CT)
SMA occlusion(Abdomen CT)
AI ensured the timely diagnosis for these patients with life-threatening conditions.
Roughly 85% of vRad studies are from emergency departments so pulmonary embolism and intracranial hemorrhage were some of the first models we developed. In 2023, these were the most corrected misses by volume with 673 and 212 respectively.
In 2023, vRad AI identified 68 missed epidural lesions that were quickly corrected and communicated to the care team. Nobody else utilizes AI for the detection of epidural fluid collections. But for vRad clients it makes sense because of our high volume of emergent studies. A missed epidural lesion can have devastating consequences for the patient and is the most expensive miss in terms of med mal indemnity.
Healthcare providers are struggling under the weight of heavy imaging volume and not enough radiologists. This can prolong the time-to-diagnosis for patient conditions where time is of the essence. AI is helping combat the national radiologist shortage.
Our AI models review all incoming relevant images at vRad for 13 critical findings. If pathology is suspected, the case is immediately distributed to all available radiologists and placed at the top of their worklists. The role of AI in prioritization is invisible to the radiologists.
RESULTS
RESULTS
Inaccurate or missing information on a radiology report can impact patient treatment, reimbursement, and even downstream client applications through data integrations.
Mistakenly reporting left vs. right
Forgetting terminology for reimbursement
Omitting reporting requirements such as Fleischner criteria or LI-RAD score
…and many more.
These examples have been eliminated by the automated radiologist alerts embedded into the vRad Imaging Platform.
In 2004, vRad made the pivotal decision to form an internal software engineering team. Since then, the team has developed thousands of tools that meet the unique needs of our practice of 500 radiologists, reading on a single platform for clients in every state.
We deployed our first AI model in 2015 for the detection and prioritization of Intracranial Hemorrhage. Today, our development team is focused on real-time QA overreads, superior mesenteric artery occlusion verification, and other models for efficiency, quality assurance, and risk reduction.
Published on vRad AI
For additional AI and other radiological research by vRad radiologists please visit the research section of our website.
If you are a radiologist wanting the power of AI at your back, talk to one of our knowledgeable recruiters today.
If you are a healthcare administrator seeking new advantages for your organization and patient population, let’s discuss how you can leverage our AI today.
vRad AI software does not provide any diagnostic or treatment information or recommendations, nor are studies flagged in any way to indicate that an AI model has identified a likelihood of the presence of any pathology. AI functionality is intended for prioritization and quality assurance, and is not intended to diagnose, treat, mitigate, or cure any disease or condition.
3600 Minnesota Drive
Suite 800
Edina, MN 55435
