8/ If this holds up, the pandemic may have given us an unexpected ally in the fight against cancer. This highlights how scientific breakthroughs in one realm often spill over to other areas in unexpected ways-& why investment in mRNA research is still important. jenndowd.substack.com/p/can-mrna-v...

7/ The investigators are planning a randomized clinical trial to directly test this finding. This will be important as the ultimate proof of cause and effect, as well as testing important details such as the optimal timing and dosage of mRNA to make the most of the cancer treatment.

6/ To me this is a much more convincing study than, say, the recent paper claiming COVID infection can “reactivate” dormant cancers, a study that found interesting results IN MICE but overreached the conclusions they could draw from human data: jenndowd.substack.com/p/covid-and-...

5/ This study did a good job of layering multiple sources of evidence, including observational human data, mechanistic mouse models, human immune profiling, and tumor biology that all point in the same direction: mRNA vaccines appear to amplify the tumor-killing effects of checkpoint inhibitors.

4/ The benefits were also specific to mRNA vaccines. Patients who received a pneumonia or influenza vaccine within 100 days of starting therapy saw no survival benefit. If the findings reflected a “healthy vaccinee” effect, we’d expect similar results with other types of vaccines.

3/ For example the survival benefit of mRNA vaccines wasn't observed in patients receiving chemotherapy. This strengthens the case that mRNA vaccines work through an immune mechanism specific to immunotherapy.

2/ The mRNA vaccine appears to activate immune responses that work in synergy with the therapy to attack and kill tumor cells. While the human data was observational, overall I found it pretty compelling given the specificity of the effect.

Thanks for the quick maths :)! I realize it does detract from the dataviz a bit, but especially for trends would appreciate the transparency...alas we can't have everything.

Very encouraging, but why don't they report confidence intervals?!

2/ I will be presenting this at @harvardpopcenter.bsky.social on Weds: popcenter.harvard.edu/blog/calenda...

More details on the paper -- which we started 2+ years ago -- later. We've tried to write it as simply and transparently as possible and walk through every step of the logic.