#Breastradiologist. Deep learning researcher. Assoc. Prof @NYUImaging. I explore the interface of #AI, #LLMs and clinical radiology. Posts = my own and not NYU.
Good radiology cases take time and attention to curate — my board review was put together 7-8 years ago and I go through yearly to see what needs to be refreshed based on feedback from residents. My fellow, CME and conference lectures all took hours of work.
As @francisdeng points out, using AI to generate images but not understanding what it’s generated is a huge pitfall.
It’s disheartening to see poor content proliferate. Try to seek out academic radiologist/social media accounts, or stick to well-curated sources such as RadioGraphics.
I am uncomfortable with the proliferation of AI-generated radiology images in social media education. They're not labeled as AI and undiscerning viewers are routinely fooled into thinking they're real.
@sethmhardy@sethmhardy I was kidding. I use Styku at mine. Lots of guys there that would probably pay for WB-MRI if they thought it had better skeletal muscle estimation.
Radiologists in JAMA this week: "If you are considering buying this [whole body screening MRI], our advice is buyer beware. You may lose more than just your money."
Radiologists in Radiology this week: "Whole-body MRI–derived BC z-scores were used to identify at-risk individuals and predict cardiometabolic outcomes and mortality beyond traditional risk factors."
(the discussion of benefits of WB-MRI is obviously nuanced, but the contradiction of cancer detection vs cardiometabolic screening is timely)
https://t.co/tGTZLiVkYw
Using AI to analyze whole-body MRI scans from more than 66,000 people, researchers showed that skeletal muscle quality and fat distribution are powerful predictors of diabetes, cardiovascular events and mortality. The findings suggest advanced imaging can reveal hidden health risks that traditional measures like BMI often miss.
The study also introduces an open-source AI tool that allows clinicians and researchers to extract powerful insights from scans already being performed—opening the door to earlier detection, better risk stratification and more personalized care.
Read the full story: https://t.co/HE7AOcHqrL
#MSK #MRI
Pathology AI-based prognostic tests continue to roll out. @arteraAI has one of the best known FDA-cleared/CE marked prostate cancer prognostic tests and CPT III coding for it.
They have just received FDA clearance for their new breast product that predicts distant metastasis in early-stage, hormone-positive, HER2- breast cancer based on pathology slide analysis.
1) This makes FDA clearance easier for similar products in the breast space -- there are other good companies working in this space.
2) If they manage a CPTIII code for this one too, there is a valid reimbursement pathway for breast path AI predictive tests.
https://t.co/UUYmjxDtO5
@owl_posting Radiology and pathology on similar tracks, honestly. We digitized a decade earlier, but facing similar issues on complexity, non-human findings, and the same well-named "tailwind for human sign-off."
@francisdeng Good, this is what we need, more single-reader mammo AI implementation data. The only other big study to date was the RadNet one which showed increased CDR.