Love the conclusion especially. It feels like there is a tradition of experts trying to come up with very precise answers to complex questions like this, which are inevitably wrong and a tradition of non-experts throwing their hands up in the air and giving up on being able to get even an approximate understanding of the answers.
The future belongs to the folks who remain unsatisfied.
I don't know if the data are out there, but smoking is just the most extreme form of air pollution. It would be interesting to look at lung cancer rates in countries that still use coal-powered electricity generation like China and India as I remember they both have serious air pollution problems. Also there may a link between illnesses such as COPD and the use of gas hobs for cooking. Keep up the good work.
So fantastic - thanks, especially in this time of rejection of science.
One of the craziest statistics to show how things have changed was pointed out by Steven Pinker: about 100 years ago, being pregnant had the same mortality rate as a breast cancer diagnosis today.
Great write up! Appreciate the big picture perspective as well. Progress in oncology is indeed multi-causal.
I have a small correction to suggest.
Re: Osimerinib and EGFR-mutant lung cancers. The LAURA trial (2024 NEJM) which you referenced was indeed a good example of a breakthrough but the study reported on progression-free survival (39.1 vs 5.6 months). The data on actual overall survival (secondary endpoint) was not finalized at the time it was presented. I just looked it up and as of early 2025 (OS was promising but close - something like 5 month difference in OS with 30ish % data “maturity”).
The FDA typically approves drugs if the PFS difference is compelling enough and clinically meaningful (understandable) as one can infer mortality benefit will follow but this is not always the case.
In cases when the OS data comes out after “maturity” and actually supports the PFS difference then true survival differences can be invoked. To make a long story short, the correct way to interpret the study is that with Osimertinib there was a clinically and statistically significant improvement in PFS (keeping the cancer at bay longer) but the conclusive data on whether this translates to better median survival is still pending. Same thing would apply to the IDH example. I suspect (although not sure) it is still too early to strongly comment or infer benefit in overall survival (despite proven clinical benefit in disease control).
Thanks! I agree, but that's why I described them as extending progression free survival in the post, so I'm not sure which part of the text this refers to.
I was referring to the figure description which uses the term “surviving” multiple times which can be interpreted (interchangeably) as overall survival. Maybe this is just minor or semantics so ignore if others feel the distinction there is very clear!
Here is the excerpt
“This survival curve chart shows the ‘progression-free survival’, meaning the fraction of people who survived without the cancer progressing to the next stage. As time passed, the share still surviving declined. However, those taking osimertinib survived much longer. This is shown for people with stage III EGFR-positive non-small cell lung cancer in a phase three trial comparing those who took osimertinib versus placebo. Source: Shun Lu et al. (2024) Osimertinib after Chemoradiotherapy in Stage III EGFR-Mutated NSCLC.”
Ah, thanks for clarifying! I kept those in shorthand for the figure caption, since it would become quite long otherwise, but also hoped that specifying earlier that it was progression free survival, and briefly explaining what it meant, would be clear enough. I'll keep this in mind in the future!
Love the conclusion especially. It feels like there is a tradition of experts trying to come up with very precise answers to complex questions like this, which are inevitably wrong and a tradition of non-experts throwing their hands up in the air and giving up on being able to get even an approximate understanding of the answers.
The future belongs to the folks who remain unsatisfied.
I don't know if the data are out there, but smoking is just the most extreme form of air pollution. It would be interesting to look at lung cancer rates in countries that still use coal-powered electricity generation like China and India as I remember they both have serious air pollution problems. Also there may a link between illnesses such as COPD and the use of gas hobs for cooking. Keep up the good work.
Really cool! Thanks for the write up.
So fantastic - thanks, especially in this time of rejection of science.
One of the craziest statistics to show how things have changed was pointed out by Steven Pinker: about 100 years ago, being pregnant had the same mortality rate as a breast cancer diagnosis today.
https://bestbookbits.com/steven-pinker-enlightenment-now-book-summary/
Great write up! Appreciate the big picture perspective as well. Progress in oncology is indeed multi-causal.
I have a small correction to suggest.
Re: Osimerinib and EGFR-mutant lung cancers. The LAURA trial (2024 NEJM) which you referenced was indeed a good example of a breakthrough but the study reported on progression-free survival (39.1 vs 5.6 months). The data on actual overall survival (secondary endpoint) was not finalized at the time it was presented. I just looked it up and as of early 2025 (OS was promising but close - something like 5 month difference in OS with 30ish % data “maturity”).
https://www.lungcancerstoday.com/post/laura-updated-os-data-support-new-standard-of-care-for-unresectable-stage-iii-egfr-mutated-nsclc
The FDA typically approves drugs if the PFS difference is compelling enough and clinically meaningful (understandable) as one can infer mortality benefit will follow but this is not always the case.
In cases when the OS data comes out after “maturity” and actually supports the PFS difference then true survival differences can be invoked. To make a long story short, the correct way to interpret the study is that with Osimertinib there was a clinically and statistically significant improvement in PFS (keeping the cancer at bay longer) but the conclusive data on whether this translates to better median survival is still pending. Same thing would apply to the IDH example. I suspect (although not sure) it is still too early to strongly comment or infer benefit in overall survival (despite proven clinical benefit in disease control).
Thanks! I agree, but that's why I described them as extending progression free survival in the post, so I'm not sure which part of the text this refers to.
I was referring to the figure description which uses the term “surviving” multiple times which can be interpreted (interchangeably) as overall survival. Maybe this is just minor or semantics so ignore if others feel the distinction there is very clear!
Here is the excerpt
“This survival curve chart shows the ‘progression-free survival’, meaning the fraction of people who survived without the cancer progressing to the next stage. As time passed, the share still surviving declined. However, those taking osimertinib survived much longer. This is shown for people with stage III EGFR-positive non-small cell lung cancer in a phase three trial comparing those who took osimertinib versus placebo. Source: Shun Lu et al. (2024) Osimertinib after Chemoradiotherapy in Stage III EGFR-Mutated NSCLC.”
Ah, thanks for clarifying! I kept those in shorthand for the figure caption, since it would become quite long otherwise, but also hoped that specifying earlier that it was progression free survival, and briefly explaining what it meant, would be clear enough. I'll keep this in mind in the future!
Another great and interesting post. Thank you.
I keep reading of increased cancer incidence such as lung cancer in non-smokers esp young women.
Any thoughts?