Thread by @rosslevinemd

Part 2 of an awesome week of great collaborative studies coming out; the Beat AML Master Trial! An amazing effort led by Amy Burd @LLSResearch , John Byrd @OSUCCC_James , Brian Druker @OHSUNews and myself @SloanKettering.

nature.com/articles/s4159…

1.This study started in a small hotel conference room at ASH 2015; a group of us got together and wondered if it was time for a genomic-guided clinical trial in AML. The LLS and their CEO @LouisJDeGennaro saw the possibility, and challenged us to make it happen

2.After a year of discussions including Sunday AM calls, many conversations with our amazing colleagues at the FDA, and lots of discussions with academic and industry partners->we decided to move forward and make it happen

3. A critical part was choice of a genomic assay and ensuring we could get it back within 7 days to inform clinical care->Foundation Medicine stepped up and provided this critical feature

4. The trial was designed to see IF such a trial can be done; using genomics to help assign therapy to newly diagnosed AML patients>60 y.o. who did not choose intensive induction rx.

5. The answer to that is yes, we could safely and effectively execute such a trial. That is the primary result from this first report

6. The trial is dynamic, with many arms opening and closing as science and trial results evolved over time.

7. We were able to show that paitents on the trial had similar outcomes to older adults treated off-trial->we cannot conclude that the trial provided better outcomes. It shows that a trial like this can be done safely

8. Lots of important caveats and limitations; small trial esp when considering different arms, venetoclaz/HMA data emerging as new approved first-line therapy in middle of trial, addition of second agent (HMA) means we are testing a treatment approach and not a drug.

9. We believe that this shows that we can test newer (not brand new, need a dose/safety data from phase I) therapies in biologically defined subsets->allowing us to ask specific questions relating to efficacy/mechanism and pave the way to randomized trials.

10. Not every AML drug, or subset, has a biomarker or a targeted drug. This is not the only, or necessarily, the best way to do a trial, but we think it has a role...

11.Thanks to the patients who made this trial happen, the investigators and HCWs at each site, our funders, and the @LLSResearch for their vision and support

12.Thanks to Amy, Brian and John for their leadership and vision. Let’s learn from this and move forward as a field!