Dear AuntMinnie Member, 

Although treatment with Pluvicto (lutetium-177 prostate-specific membrane antigen-617) provides meaningful clinical benefits in men with severe prostate cancer, many patients don’t complete the full course.   

In an effort to predict which patients might not respond to treatment or discontinue the radiopharmaceutical therapy, researchers sought to determine factors that could help stratify risk in a real-world setting. Our report on their work was the most highly viewed story of the week.  

Developments at the U.S. Food and Drug Administration (FDA) also drew attention this week. New updated guidance on clinical decision support (CDS) software could pave the way for wider use of large language models in radiology. 

The agency also recently announced that it has received a petition requesting an alternative, streamlined regulatory pathway for radiology AI. In this proposal, vendors who have already received 510(k) clearances for radiology AI software would then be able to comply with postmarket study requirements for new capabilities instead of submitting new 510(k) applications. If adopted, this approach could significantly expedite commercialization of new software. 

In other popular news, researchers have found that PET imaging with a copper-64 (Cu-64)-labeled nanoparticle appears safe in humans and warrants further development for detecting cancer and sarcoidosis. Also, most women felt positively about learning their breast cancer risk category and screening plan. And an AI algorithm demonstrated high accuracy for predicting lung cancer risk. 

See below for the full list of top stories for the week. 

1. What are Pluvicto discontinuation rates in a real-world setting?

2. LLMs in radiology may find opening in updated FDA guidance

3. Petition to U.S. FDA proposes alternative pathway for radiology AI

4. Nano-PET tracer appears safe in first in-human study

5. Most women positive towards receiving breast cancer risk communication

6. Deep-learning image reconstruction boosts CT venography performance

7. Algorithm demonstrates prediction accuracy for lung cancer risk

8. GPT-4o bests radiologists for protocoling abdominal and pelvic CT scans

9. Lung cancer screening reporting reveals incidental finding risk classes

10. CEM is a reliable tool for evaluating screening recalls

11. 2025’s budget bill could lead to more than 1M missed cancer screenings

12. Theranostic pair ready for testing in lung cancer

13. Clinical protocols help drive appropriate cervical spine CT use

14. LLMs help automate PI-RADS classification from MRI reports

Erik L. Ridley
Editor in Chief
AuntMinnie.com