November 14: Controversy, hierarchy
For Lung Cancer Awareness Month #LCAM I’m going to summarize 30 important lung cancer trials over 30 days. These posts are directed at non-medical professionals, with descriptions of the results and of what makes a good trial. #lcsm 1/17
For Lung Cancer Awareness Month #LCAM I’m going to summarize 30 important lung cancer trials over 30 days. These posts are directed at non-medical professionals, with descriptions of the results and of what makes a good trial. #lcsm 1/17
This is the most recent study we’ve looked at so far, from 2020. You may recall that adjuvant chemotherapy (after surgery) increases the chances of cure for early stage lung cancer (Nov 2), and also that we have good oral medications for lung cancer with EGFR mutation (Nov 7, 12)
This trial enrolled people who had resected lung cancer with EGFR mutation. They were allowed to have adjuvant chemotherapy. They were then randomized to three years of the EGFR TKI osimertinib, or three years of placebo. 3/17
The primary endpoint was disease-free survival (DFS), or the amount of time before the tumour comes back. This has been a controversial choice, as many people feel that the appropriate endpoint for an adjuvant trial should be whether people live longer (overall survival OS). 4/17
The primary endpoint was also planned to be assessed in only a subgroup of enrolled patients, those with stage II-IIIa disease (though people with stage Ib were eligible). There is a hierarchical testing scheme that we will discuss below. 5/17
682 patients had surgery and then adjuvant chemotherapy if they so decided. They were then randomized to osimertinib or placebo. All participants had tumours with either EGFR exon 19 deletion or L858R mutation (the two commonest EGFR mutations) 6/17
It has been noted that several elements of the trial design favour the osimertinib arm
1. DFS rather than OS endpoint
2. Chemo not required
3. MRI brain not required
4. Trial sponsor did not guarantee osimertinib available at crossover for people on placebo arm 7/17
1. DFS rather than OS endpoint
2. Chemo not required
3. MRI brain not required
4. Trial sponsor did not guarantee osimertinib available at crossover for people on placebo arm 7/17
We have seen multiple previous studies (November 7 and 12) where EGFR drugs have large effects on PFS, but less impact on OS. As expected, this study showed a large improvement in DFS in the osimertinib arm (median 18 mo vs >3 years p<0.001). 8/17
There’s a lot of controversy about the application of this result. For years doctors have held that the purpose of adjuvant therapy is to increase the rate of cure, meaning that people live longer overall, rather than just delaying recurrences. 9/17
This argument holds that if people live the same amount of time being treated either adjuvantly or at the time of recurrence, then we might as well just treat at recurrence. This approach would mean that people who are cured by surgery are never exposed to the toxicity of therapy
The counterargument is that time without recurrence is valuable, even if the overall duration of life is not increased. This is also an appealing argument: if you had a certain number of years to live, you would prefer to have as many of them free of cancer as possible. 11/17
Many oncologists have been persuaded by the latter argument, while others say that they are waiting to see the OS results. But OS results may take years to arrive, and as we will see below, may not come at all. 12/17
Yesterday we briefly mentioned co-primary endpoints. We showed how two endpoints could be chosen, each with a p-value <0.05 required for significance, so that the total risk of a false-positive trial (called type 1 error) remained <5%. 13/17
For this trial the plan is to do tests of several endpoints in sequence. If the p-value is low enough with the first test, then the second will be performed. If it’s possible to do a third test while still having the cumulative chance of type 1 error <5% then OS will be compared
I don’t know the right answer as to whether we should be using adjuvant osimertinib. I think it comes down to the value of disease-free time for some, versus the value of avoiding treatment entirely in others who are cured by surgery. 15/17
I feel strongly that OS should be the outcome of adjuvant trials. DFS sets the bar too low, in my opinion. I’m also irked by elements of the trial that tilt the scales in osimertinib’s favour: no amount of statistical analysis can adjust for biases baked into the trial design.
I hope you'll join me again tomorrow. I think you might be surprised at the trial we discuss, but I can genuinely say that it changed my practice.
See you then... 17/17
See you then... 17/17
• • •
Missing some Tweet in this thread? You can try to
force a refresh
