Hazard ratios are so last century: the new measures of success with the IO agents featuring Tom Powles

I’m Joseph Ischia
I’m Nathan Lawrentschuk.

Joseph: And we are Talking Urology. Nathan, it has been a while since we could get together and as we tidily put over 200 days of lockdown behind us, it is genuine pleasure to get back on the horse and pick up on another one of our little areas of interest. As our devoted and learned followers will know, we usually look at landmark papers but today we’re going to talk about something even higher brow and more intellectual. I know, it’s a low bar, but even a one-legged dog has to strive to jump over something. While we jest about legless canines, there is nothing flippant about the enormous change going on in the field of the new immune-oncology agents and their amazing trials which we’re hearing so much about these days. And what I really want to focus on are the new outcomes of success that these trials are using to measure the efficacy of these drugs. Amongst other things, we’re going to be explaining why overall survival (OS) is no longer necessarily the best measure for these newer agents as has been the gold standard for so long in oncology research papers.

Nathan: I didn’t know you had it in you Joseph, the use of a low bar was quite apt. But that’s right folks. We’re looking at the new measures of success of the immunotherapy agents. It’s becoming apparent when you look at the activity observed with different active immunotherapies in clinical trials, that median overall OS, progression-free survival (PFS), disease-free survival (DFS) and response rates etc may not adequately capture the potential outcomes with these therapies and additional metrics are being used such as landmark survival rates, durable responses, or cure (I’m doing air quotes which work well on a podcast), as well as treatment-free survival, or unconventional patterns of response.

Joseph: Nathan, of course we need a guest to help us bring the literature to life so let me begin with a riddle. What do Jackson Pollock and the United Kingdom have in common? One produced Blue Poles, and the other Tom Powles; both works of genius, both generate much excited and sometimes heated debate, and both have their own podcast. While we were not allowed to broadcast from the National Gallery and some of my explanations of the perceived controversy around PSA screening have been likened to an oral representation of Blue Poles, we are privileged to be joined on the show today by Tom Powles. Tom is a Professor of urology cancer at the University of London and the Director of Barts Cancer Centre which is one of the UK’s largest cancer centres with over 4000 new cancer patients per year. Tom has had a major role in the development of biomarkers and new drug strategies in urology cancers: notably, front line immune/targeted therapy combinations in RCC, immune checkpoint inhibitions alone or in combination in bladder cancer and antibody drug conjugates in urothelial cancer.

Nathan: He has authored 10 NEJM or Lancet publications with two first author NEJM publications and 2 first author Nature publications.

Joseph: Gee, Nathan. That’s almost as many as you, the publicator!

Nathan: He authors over 40 peer-reviewed publications per annum. That’s quite the CV. I think it’s fair to say he’s a proper expert on today’s topic.

Joseph: With that CV, we thought we would let him introduce himself.

Tom: I’m Tom Powles. I’m an oncologist from London. It’s a real pleasure to be here with you today.

Joseph: Understated. Australians love him. So, let’s get down to the business of how we measure the success of drugs in trials. In the literature, the concept of a “success” is described using an array of terms that can be subdivided broadly into lack of disease progression, elimination of malignant cells, and survival. Historically, trial endpoints show frequent use of median OS and progression- and response-related endpoints. Because these endpoints are mainly described in the context of chemotherapies that are not generally curative, they may not adequately capture outcomes of new therapeutic modalities with potential for long-term survival ie: immunotherapies. Therefore, different endpoints may logically be more appropriate for these agents.

Nathan: For example, we know that in the older papers looking at cytotoxic drugs, long term survival was typically measured by the median OS. This assessment may not be sufficient for treatments that offer long-term benefit because it does not provide information relating to the small proportion of patients who occupy the tail of the survival curve. So median OS may grossly underestimate the long-term survival of the immune-oncology agents. Likewise, PFS and DFS are appropriate for assessing the activity of agents likely to cause rapid control of tumor growth, but they may be less suitable for therapies where tumour control may develop over a more prolonged time. Endpoints such as DFS or PFS may underestimate the activity of novel therapies if associated with prolonged stable disease or unconventional responses, even though these responses may translate into a prolonged survival benefit. Therefore, for the immune-oncology agents, we now tend to talk about “landmark survival rates”, durable responses, and treatment free survival.

Joseph: Landmark survival refers to identifying the number of patients who are still alive at the pre-specified time point, eg, number of people who are still alive at a clinically relevant 3 years.

Nathan: As you can imagine, the significance of the results can vary according to the landmark time, so the choice of landmark is a critically important consideration and ideally should correspond to a clinically meaningful period of time which might be 12 months in metastatic RCC studies, or 5 years if you were doing a localized RCC study.

Joseph: Let me cut straight to the chase…all of the measures we see in papers – landmark survival, durable response, or cure, treatment-free survival (TFS), or unconventional patterns of response where we’re worried about pseudo progression, so of those measures, what do you think is the most important new measure for these IO agents and why?

Tom: The answer to the question is quite straightforward. Longer term durable response is what we need to drive towards.

Nathan: Succinct but emphatic Tom – I like it! Durable response is that long plateauing of the tail we see in the Kaplan-Meyer curves. Are they cured? Did the investigator’s computer crash and were not able to update the data? So, what are the key one or two things to know about the durable responses? Are they all doomed to die, do you think? Is it just that we haven’t gone long enough or how long is the tale? Are there any cautionary tales that we need to be aware of with these durable responses?

Tom: I started treating patients with immune checkpoint inhibitors. I was really lucky because I work in an institution where we were using them right at the beginning and in fact, I think we have the first urothelial cancer responder whose name is David and I still see him. He’s an extraordinary character because he was told that he only had a few months to live and he had platinum refractory disease and had progressed in lymph nodes, and in some other areas as well. He came to see me and said, what do we do and I said, we’ve got a new drug that may or may not work. And we said, it probably won’t work, and we’re probably wasting your time. But if you want to have a go. Well, I have lots of drugs like that in 1999. But 85% of the time, the drugs in those sorts of stages fail in my experience. David still with us, he has no evidence of disease. And he runs a business and has apparently famously good barbecues on the south coast of England. So, the drugs are transformative for a small group of patients and I always think one day it will come back, I always think that because I am an optimist, but my experience of cancer is in the end, it comes back and that’s not happened with David. And it’s not happened with Andrew. Andrew is another patient we treated six years ago, same sort of story. He plays cricket every weekend, he’s actually 75 years old and his biggest problem right now is he gets pain from his shoulder because he broke it when he was younger. So, they are transformative for some patients. Now, the problem and this is a big problem, is that there is a big group of patients who we treat, they grow straight through and the treatment doesn’t work for them and they die. And I don’t remember some of those patients and there were lots of them because you remember the winners, of course. So, I’m not for a second suggesting that these drugs are transformative for the majority, it is by far the minority of patients.

But you will see on the Kaplan Meier curves of all of the trials that we do in years gone by if you go to some of the EORTC studies and some of GemCis and GemCarbo studies and particularly the second line trials, the Kaplan Meier curves go down to zero or close to zero. Whereas here, we do have this tale now which is present and I think that tale will continue.

Joseph: Another new concept from the immune-oncology agents is the concept of treatment free survival. It is analogous to the treatment free period off ADT for men being treated with intermittent ADT for prostate cancer. It helps capture additional perspective about the patient experience during treatment with immune agents, thus complementing other more commonly used end points. Treatment free survival can measure the time to cessation of immune-oncology agent protocol therapy and time to subsequent systemic anticancer therapy initiation or death (analogous to time to treatment failure and time to second-line therapy, respectively). Treatment free survival can further integrate the possibility of persisting and/or late adverse effects of initial immune-oncology therapy while free of subsequent therapy.

Nathan: So Tom, what can you tell us about Treatment free survival? Is this something that you’re excited about? This opportunity to have people off treatment for a while?

Tom: So, in the Star Wars world and in Star Trek and I would say, Blade Runner, but I think we’ve been and gone with that time, in those sorts of era, the thought of we want to cure patients and make them live for 20 years, and David and Andrew, who I mentioned before, who got into durable remission, are they really going to need to come? They’re both in their 50s, well, actually dangerous, a bit older, but their life expectancy 10 years, are they going to have 10 more years of therapy? At the moment, it looks like neither of them want to stop because someone said to them, they’ll be dead in six months, so they’re not going to stop. We’re going to have to build the system of lifelong systemic therapy for the cancer population in the knowledge that, I don’t know, half of us is going to get cancer, it doesn’t seem, it doesn’t seem like that’s going to be the end chapter of this book. The test is a model the four cycles of therapy, stopping therapy, long term remission, and going on leading your life, sounds more attractive. Now, you might say to me, “Tom, what about diabetes?” You know, we have chronic treatments with that, but remember, insulin is not toxic, and that model for me is probably not going to be the end. I don’t know that.

What I would say to that answer is we are exploring TFS at the moment. The problem with the current experiments that we’ve done is we’ve looked at those patients who have stopped but those patients have stopped because of toxicity. And we know toxicity is associated with response and therefore, to some extent, we are picking the winners, with the current TFS data available to us. Now, my friend Dave McDermott talks about this and I’d love to have a discussion with him and a debate with him about this issue. Ultimately, he’s right about this, we’re going to need to work towards that model. We’re going to need to do the trials. But to do that, we will need to get better drugs

Joseph: There is still so much we have to learn about the future of oncology treatments and how patients will make their treatment choices. Patients may be prepared to sacrifice QoL or even better median OS benefits for chance of cure. How will they decide in a future with both options available? Well, I think we’re going to hear our Oncology colleagues talk a lot more about “precision medicine”. What I mean is that in contrast to a one-size-fits-all approach to cancer treatment, precision medicine is an approach that looks at specific information about a person’s tumor to diagnose and treat the disease. The National Cancer Institute defines precision medicine as “a form of medicine that uses information about a person’s genes, proteins, and environment to prevent, diagnose, and treat disease”.
Nathan: With precision medicine, treatments focus on abnormalities in tumours based on genetic factors, the environment, and lifestyle. It is now believed that between 40 and 50 % of cancers may be treated with precision medicine. The most obvious benefit of precision medicine is that it allows a doctor to tailor cancer treatment based on further information about cancer cells. This both increases the chance that a person will respond to treatment and reduces the chance that a person will have to cope with the side effects of a treatment that does not work.
Joseph: Let me put this concept of side effects and how much will patients put up with another way. With traditional chemo for most cancers, or tyrosine kinase inhibitors in RCC, when you got side effects, you reduced or stopped the treatment. But in an era of 30% durable responses with the IOs would you push on even if you felt terrible hoping to be one of the 30%? In your experience Tom, are patient’s happier to tolerate possibly worse side effects in the hope of that durable response? Have you seen that?

Tom: Yes, they are, of course, but it does depend a bit on how you describe it. If you said, we’re going to give you three drugs instead of two, they probably won’t work, it may or may not make any difference, and we can sequence them. Most patients say, well, give me two and if it goes wrong, we’ll get the one afterwards. If you say we’ll give the three drugs, there’s probably an increased chance of durable remission and that’s going to be our best way of getting in control, yes, there’ll be more side effects. Patients will say yes or give me the three drugs. So, the dialogue in the way we do this is so important.

Joseph: Well, when you need to be reminded by a medical oncologist that chatting to patients is important, I think it is time to read the House of God by Samuel Shem again.

Nathan: So, to summarize in a very small nutshell Joseph, after our tour de force above, as we continue to see novel therapies being introduced with the potential for long-term survival, like immune-oncology drugs, then efforts should be made to use the appropriate endpoints and the related set of value metrics that best describe the clinical and other outcomes of these new treatments.

Joseph: Beautifully summarized, Nathan! I couldn’t have put it better myself! This is a really fascinating topic, and we could talk for ages. Unfortunately, we have run out of time. Thanks so much to Tom for taking time out from his hectic schedule to join us. This episode of Talking Urology was written by Mark Quinlan and Joseph Ischia. Spoken by Nathan Lawrentschuk and Joseph Ischia. Produced by Joseph Ischia and Cara Webb. And of course, none of this would be possible without the generous and non-interventional support of our sponsor, Bristol Myers Squibb. Thanks for listening. We still have some great podcasts coming up so keep tuning in. You can contact us with questions, corrections, or updates at talkingurology@gmail.com.