Skills Lab: Understanding Misconceptions About PFS in Cancer Care

Medscape

10 hours ago

This transcript has been edited for clarity.
Hello, everyone. This is Dr Bishal Gyawali, from Queens University, Kingston, Canada. I'm back again in this Skills Lab Series on Medscape.
Thank you for joining us in the past few lectures. Today is going to be the last lecture regarding surrogate endpoints. We will discuss how physicians and patients understand surrogate endpoints. Do they clearly understand, and do they make the right clinical decisions based on their understanding of surrogate endpoints?
First, let's talk about patient understanding of progression-free survival (PFS), the most commonly used endpoint in advanced cancer drug trials. In fact, colleagues from the FDA conducted a fantastic study, in which they asked participants about their understanding of surrogate endpoints, and how they did this was very clever.
They showed a direct-to-consumer television advertisement about a cancer drug, and the advertisement mentioned that the drug improved progression-free survival. It did not mention anything about overall survival (OS) or anything else. It just said that the drug improved progression-free survival.
There were almost 1000 participants in this study and, after watching that advertisement that mentioned improved progression-free survival, more than 90% of the participants thought that the drug improved survival. There was no mention of overall survival, but just by hearing that it improved progression-free survival, they assumed that the drug improved overall survival, which is quite telling.
The researchers went one step further. They did a fantastic intervention. Now the participants saw the same advertisement again, but this time the advertisement specifically mentioned that the survival data were unknown. It improved progression-free survival, but we don't know whether it improves [overall] survival.
Even after that disclosure, 40% of the participants continued to believe that the drug improved survival, and I think this is because we call it progression-free survival. When we call it that, they hear the last word, "survival," and they inherently assume that if a drug improves progression-free survival, it improves their survival. Even after disclosing that survival effects are unknown, 40% of the participants continued to believe that the drug improves survival.
One more very interesting study was conducted in 100 patients with advanced cancer in Canada. These were actual patients who were receiving treatment for cancer. PFS and OS were explained to these patients, and they were given a hypothetical scenario about a drug that improved PFS and had some toxicities but did not improve OS.
In this experiment, 17% of the patients said they would prefer to receive the drug even if there was no PFS benefit. This is similar to what we see. Some of our patients just want something. They just want a drug. Even if there is no PFS or OS benefit, they just do not want to give up.
Of the patients, 26% said they would want the drug for some PFS benefit, within the range of 3-9 months. If the drug improved PFS between 3 and 9 months, even if it did not improve OS, patients would still want to get the drug. However, the majority of them, 51%, would decline the drug, regardless of PFS benefit.
This is important. What this is telling us is that if the drug does not offer any OS advantage, more than half of our patients would decline a drug with toxicities, irrespective of how big the magnitude of the PFS benefit is.
Therefore, we wrote a paper in The Lancet Oncology back in 2022, in which we called on the community to change the name of progression-free survival because it is confusing, and patients assume it is a survival advantage when it's actually not. If a drug does not improve survival, then they would probably not want a drug with PFS benefit alone.
We said that it's time for a new name, and we proposed that instead of PFS, progression-free survival, we should call it PFI — progression-free interval — because it gets rid of the word "survival" and patients won't be confused.
Now, let's talk about physicians' understanding of surrogate endpoints. Do physicians understand this well? This is a very recent paper that we published in ESMO Open . This was a mixed-methods study of oncologists' perceptions on endpoints, benefit, price, and value of cancer drugs.
We conducted this study among oncologists in India, so there might be some limitations about how applicable it is elsewhere, but I don't think the results we found are very different from what we would find in any other country.
First, we found that 20% of oncologists rarely use cancer drugs that improve only response rate but not survival. Almost an equal percentage said that they would often use drugs that improved response rate but not survival. In regard to PFS, 20% of oncologists rarely used cancer drugs that did not improve survival and only improved PFS, but 48% of them often used such a drug.
We asked them the reasons for this. First, let's talk about response rate. Among the oncologists who said they would not use a drug that improves only response rate, the most common reason was lack of any effect on survival or quality of life because the drug just shrinks the tumor. It does not help patients live longer or better, so what's the point? The second was toxicity, and the drug will definitely have many side effects. That was the reason why they did not want to use a drug with response alone.
Among oncologists who said they would use the drug based on response rate alone, when we asked why, the most common answer was for symptom relief or improved quality of life. This is quite interesting because there are no data to say that response rate leads to improved quality of life or symptom relief. If it is in a tricky location, then it can, but there is no evidence that, in general, shrinking a tumor leads to improved quality of life. This indicates that oncologists' understanding of surrogate endpoints is also not optimal.
The second most common reason was to downstage disease, which is understandable. The third one is also very interesting because they said "something is better than nothing." That means they know that response rate does not lead to improved clinical outcomes, but they can't just do nothing for their patients. They have to do something and the patient wants something, so something is better than nothing.
The fourth one is very interesting. They talk about the impact of drug approval and having it included in guidelines. Even if they know that it's based on response rate alone and they don't want to use the drug, if it's approved and if it's in the guidelines, then they feel like they're forced to use it.
Some of the oncologists also said that it's on a case-by-case basis. For example, if they have a drug that is cheaper or a biosimilar for a drug that improves response rate alone, they would use it. For rare cancers or where there is no other option, as we discussed, they would use such a drug.
The most common reason for not using a drug based on PFS alone was the lack of effect on survival because it was not known whether the drug improved survival or not, which is very appropriate. Reasons in support of PFS were very interesting. The number-one reason here was cost. Oncologists said that a drug that improves PFS costs less than a drug that improves OS. The patients cannot afford the expensive drug, so they're using a cheaper drug.
Again, this is very interesting because that's not true, especially in the context of the United States, Canada, and other high-income countries where we have done these studies. There is no correlation between whether the drug is approved based on PFS or OS and the price of the drug. A drug approved based on PFS can cost even more than a drug that is approved based on OS.
The second reason was the impact of medical literature. If, in the editorials, commentaries, and CMEs, people are promoting a drug that is based on PFS alone, then that leads to their use. Then, on a case-by-case basis, people also talk about cost-effectiveness and for rare cancers or lack of alternatives.
Overall, I'm trying to show that when oncologists understand that it's only response rate or PFS, they still use these drugs based on some misguided assumptions that delaying PFS or improving response rate leads to improved quality of life, which is objectively untrue, and that drugs approved based on surrogate endpoints cost less than drugs approved based on OS, which is also untrue.
It's also interesting that we have this culture of always trying to do something rather than having that difficult discussion about end of life. Doing something is always easier than doing nothing, so we end up prescribing these drugs, although we know that it does not benefit patients. I think these findings are quite interesting, and I want our entire oncology community to be aware of these biases within ourselves.
To close the loop, we also asked these oncologists about whether magnitude and price matter. They said that, for a drug that impacts only PFS, they would want to see at least a 4.5-month improvement in PFS, with the range of answers from 1.5 to 12 months, so quite a big range there. The price is interesting. This is not applicable to high-income countries. This is applicable only in the Indian context. They said that 120 USD per month was appropriate for such a drug.
For OS, again, they wanted an improvement of 4.5 months, with a range of 2-24 months. They said 360 USD per month was justified.
It's not about the exact number, but what they are telling us here is that a drug that improves OS can cost up to three times more than a drug that improves PFS alone. If you think about it the other way around, a drug that only improves PFS but not OS should be costing one third of the drugs that improve OS. That's not the case in reality.
We talked about all the surrogate endpoints in these lectures and today is the end of this topic. To clarify, just because a trial's endpoint is overall survival does not necessarily mean that the benefits are meaningful.
There can be several other issues, which we'll discuss in our subsequent lectures. The next few lectures will talk about statistical analysis, sample sizes, power calculations, and so on. Stay tuned.
Thank you.