From Promise to Practice: The Challenges of Implementing AI in Medical Imaging
By
Eric Topol
Kettled twice. Extra chewy, extra trustworthy.
Summary
The article discusses the evolution and current state of AI in medical imaging, tracing progress from the 2012 AlexNet breakthrough to widespread assessment across virtually all medical image types (X-ray, CT, MRI, ultrasound, pathology, etc.). The author reflects on their own 2017 Nature Medicine review on AI for medical image interpretation and examines the gap between early promise and current implementation challenges, including evidence gaps and missed opportunities in deploying medical AI effectively.
Key quotes
· 3 pulledIn 2012, the era of deep learning AI got legs with the convolutional neural network (AlexNet) that won the ImageNet challenge.
Over 7 years ago, I wrote a review in Nature Medicine entitled High-Performance Medicine that summarized the remarkable progress being made for AI interpretation of medical images.
Now virtually every type of medical images has undergone extensive assessment with AI, including X-ray, CT, MRI, ultrasound, pathology slides, skin abnormalities, electrocardiograms, endoscopy, and retina scans.
You might also wanna read
RCSI report provides framework for safe AI integration in Irish surgical practice
RCSI has launched a major new report titled "Artificial Intelligence and Digital Surgery," developed by a Short Life Working Group appointed
rcsi.com·5d ago
Zocdoc CEO on AI's Role in Transforming Healthcare and Replacing 'Dr. Google'
Zocdoc CEO Oliver Kharraz discusses the future of AI in healthcare, arguing that AI will replace 'Dr. Google' as patients' primary source fo
The Verge·7mo ago
Philips Introduces World's First Digital Pathology Scanner with Native DICOM JPEG and JPEG XL Output
Philips has launched the Pathology Scanner SGi, which is the world's first digital pathology scanner to offer native DICOM JPEG and JPEG XL
philips.com·8mo ago
Medical Journal Warns Against Using ChatGPT for Health Advice After Man Develops Bromism
A US medical journal warns against using ChatGPT for health advice after a 60-year-old man developed bromism, a rare condition caused by bro
theguardian.com·9mo ago
GE HealthCare Unveils AI and Radiopharmaceutical Innovations to Expand Nuclear Medicine Access at SNMMI 2026
GE HealthCare is showcasing new AI-enabled workflows, radiopharmaceuticals, and imaging innovations at the 2026 SNMMI Annual Meeting to help
stocktitan.net·22m ago
Blockchain offers a path to patient-controlled wellness data beyond traditional EHR systems
The article discusses how traditional electronic health record (EHR) systems prioritize billing and institutional workflows over patient con
americaoutloud.news·2h ago