Read the Full Video Transcript

Pedro Barata: Hi, I'm Pedro Barata. I'm a  GU medical oncologist and associate professor of medicine at Siedman Cancer Center in Cleveland, Ohio. It is my privilege to be joined today by Dr. Manish Kohli. Dr. Kohli is professor and presidential endowed chair at University of Utah in Huntsman Cancer Institute in Salt Lake, Utah. Welcome, Dr. Kohli. Thanks for joining us.

Manish Kohli: Good morning. Thank you for having me.

Pedro Barata: Absolutely. We are going to be chatting today about a very interesting paper that came out in one of the ASCO journals regarding your cost-effectiveness analysis of six IO-based regimens plus Sunitinib in patients with metastatic renal cell carcinoma. I think this is such an important topic and we don't talk as much perhaps as we should about that topic, so kudos to you about that. Maybe to get things started, you ran this cost-effective analysis. A lot of us are less familiar with that. They're not quite sure how do you do it? What is the methodology behind it? Maybe I should start by asking you first, can you summarize for us how did you put this study together? What steps did you take to report this?

Manish Kohli: Yes, indeed. Once again, thank you for having me on UroToday. The first and foremost step was to cross-fertilize across teams and form the right team who can come together on a challenging topic like this. We have pharmacology personnel, we have health economic folks, we have clinical oncologists like myself, and we have data sciences people. What we really do in these kinds of analyses is a cost-effective approach. We utilize a model called a decision analytical model, and this decision analytical model, which is pretty common to use in these kinds of approaches, is called a partitioned survival model. Now, this partition survival model, which is an acronym, PSA, which has got nothing to do with anything to do in prostate cancer. This PSA has got three health states in the model. One health state is when the patient is undergoing these treatments, whatever these IO, IOTKI, or Sunitinib combination treatments are. They are progression-free.

The second health state they may be in is that they're progressing, and the third health state is obviously death. What the partition survival analysis does is it derives the proportion of patients in each of these three health states at each point of their published PFS and overall survival curves from the studies, but it does so on a monthly basis. When it calculates that proportion, then we know how many patients are there in all these different studies and then we utilize another aspect to it, the effectiveness measure. The effectiveness measure is basically QALY, quality adjusted life year, which is based on weighing the time each patient or the proportion of patients, in this case, spend in each of these health states based on the progression-free survival curves and the overall survival curves we were talking about. And then the time they spend in these health states and then they are converted into a utility score which can range from zero, which is bad, to one, which is perfect health. Patient is undergoing that. There is no progression. There is no side effect.

Then these utility scores for each of these treatment strategies which are derived from these published literature are then used to calculate quality affected adjusted life years. That's the effective measure. The cost-effective measure is calculated by the pulling together the cost of acquisition of the drug, the cost of management of grade three and four toxicities, again, from published literature, and over the time that they experienced it. We do this on a 10-year horizon. Then now we have the cost that goes in, the effectiveness that goes in, the amount of the quality adjusted life years, and then we put it all together by noting an incremental cost-effective ratio for each drug. Higher the incremental cost-effective ratio, better is the patient in terms of longevity and quality of life from receiving that treatment. We did this for these published literature trials of about six, 7,000 patients with Sutent, Nivo/IPI, AA, which is  avelumab, axitinib, pembrolenvatinib, pembro/axi, Cabo/Nivo, and so on and so forth. And then we calculated all of that and then we compare these ICERs, or incremental cost-effective ratios, and then we reach a certain conclusion. It's not just survival-based.

Pedro Barata: Very, very important. The methodology is really the meat of this work. It's so important. Let me ask you. You ran that study, of course you get access to public data from available trials. I have a few questions about, that but I'll get to that later. Tell us what the findings are. I will recommend everybody to read the paper, but please summarize for us what are the main take-home points from your analysis?

Manish Kohli: The main take-home points are this, that the incremental cost-effective ratio that we calculated for all these drug combinations with a reference towards Sutent, because Sutent was the control arm in all these trials, the main incremental cost-effective ratio was highest for Nivo/Ipi. This was to say that over a 10-year horizon, if we were to start with Nivo/Ipi and give these patients this treatment the compared to the others, we would buy more quality of life and more longevity of life at a dollar amount of about $290,000 spread over 10 years. If the insurer had a willingness to pay off that amount, this becomes the most cost-effective treatment that is available for metastatic renal cell cancer. If on the other hand, the willingness to pay by the insurer is lower than that and is lowered towards $150,000 over this period of time, then sunitinib would be the most cost-effective drug to give, which buys at a lower cost the amount of incremental cost-effective ratio units. But that is not as good as Nivo/Ipi because there are more quality adjusted life years that are being bought for Nivo/Ipi compared to Sutent.

Pedro Barata: Those are very, very interesting results. I have a couple of followup questions, but maybe the first one, so very important data. Your position, it's almost like for us, there's a cost to access novel therapies and advancements. It's really what is about and we're trying to put a number. How much is it worth for us to offer additional safety and additional efficacy with the novel therapies, which is always a complicated topic. Let me ask you this. You mentioned numbers, you mentioned 150 over 10 years. You mentioned almost 300 over 10 years. My question is do we have an idea, I know UK has done a lot of work on that. Do you have any idea around the world, what is the amount in dollars that the societies might be willing to pay for an additional QALY? Any thoughts about that to put this data into perspective?

Manish Kohli: No, that's the crux of how as societies we are developing and attacking the problem of distributing new drugs into the population for managing this disease burden. That's a societal question. All the figures that I said just now came from two sources in the US. One was the VA Federal Supply Scheme, VAFSS, and the other one was from CMS, or Medicare services. Unfortunately in the US, we do not have a single payer system as opposed to many other countries where there is a single payer system. Within those single payer systems, what is the drug acquisition cost going to be is going to vary from place to place. How are the management of toxicities going to be taken care of, and therefore the cost varies from place to place. What in that particular society is the threshold of the willingness to pay may vary from time to time, not just that. Today, it's 2023.

If, suppose, we were to fast-forward into 2027 and some of these drugs became generic, the cost would be reduced. Because of that, the numerator and the denominator gets affected and the ratios start to change. This is a very dynamic process just like tumor biology, which keeps evolving. This, too, is under the pressure of market forces, which again, changes the complexion of which is a better drug in terms of cost-effectiveness now? It's not a direct answer, but it only is to highlight the issues that are ongoing in deriving these kinds of studies. They're important to do because you place the number onto the table for people to look at and discuss and debate and then hopefully, something will come out of it where we can reach a reasonable willingness to pay where we can serve our patients in a way which is cost-effective as well as effective for the drug biology.

Pedro Barata: Right. Great points and you definitely share the complexity around the topic. Let me go back to your methodology because some of the folks might be hearing this and might be thinking, okay, get it. We tested different IO-based combos with sunitinib, but as we know, those are different patients in all those trials. Some trials have more rate proportion of good risk patients or poor risk patients than others, et cetera. I'm not even going to mention the fact that they were conducted different times, meaning patients were exposed to different systemic therapies. The question I have for you is how do you account for that when we're comparing PFFs, for example, or even OS, where we're talking about different patient populations with different prognosis based on the underlying disease? How do you do that matching, if you will? Is that the accounted in the model?

Manish Kohli: The six trials that we took unfortunately consistently did not report IMDC risk category of patients across the board in all of them consistently. It becomes therefore very difficult to look at the proportion of patients and their survival based on the subset analysis across the board for poor and intermediate versus good. The data that's emerging, not taking into account cost-effectiveness, but otherwise, Nivo/Ipi is not thought to be a good choice for good prognosis in general terms that we know about because there is some data that the survival analysis may be not as good if you compare with other ones. Again, post-hoc secondary analysis. We did not have the luxury of breaking down all the different trials based on risk category, and so that is a potential limitation we look at and are not able to overcome at this point with this particular study at least.

Pedro Barata: Gotcha. No, that's very helpful. In the same token, of course these trials, they read at different time points, different followups. Ipi/Nivo data is out there with more than five years followup, other trials at three, four year followup, et cetera. While the median seemed to be very similar, which sounds to be the case in some cases, those numbers can vary, especially when you talk about the states. One of the states was unfortunately death. Obviously, as you get longer follow up you'll have a better sense, a better estimate, of that number might really be. When you look at these studies, and this is a very recent study that you conducted, so I would argue you got very recent data which is good, but can you comment a little bit about how the followup could potentially or not impact the results?

Manish Kohli: Yeah. One of the assumptions that was made in this cost-effectiveness analysis is that we will calculate the cost of the drug until the patient progresses. That's based on the partition survival analysis model, how long the patient is in a health state of progression free, or until the patient progresses, or until two years of continuous immunotherapy if the patient has not progressed. All of these trials are of course mature enough at least that two years followup is approximately available for all of them, including the one that you mentioned. The two-year cutoff is really coming from what we have seen in metastatic lung cancer trials, that if someone is responding to immunotherapy at two years, probably you'll stop that immunotherapy and wait and watch.

Thereafter, what happens is the cost that we calculate is the cost of the toxicities that we are adding, and so that also comes from the health-related quality of life published literature of these trials as such. I don't think therefore that this analysis is going to change based on the time factor, but I think it will change if the costs change of these drugs and that's very dynamic, and so I wish there could be enough task forces to maintain a model on a six monthly or yearly basis to calculate this, but we don't have that luxury, of course.

Pedro Barata: Gotcha. No, that's very, very helpful. Very enlightening while we are going to read the paper, and these explanations provide additional context to it, which is outstanding. Maybe a final question before I let you go. In other parts of the world, cost-effective analyses are part of healthcare systems and also part of guidelines to some extent. To my understanding, I don't think we have them incorporated as part of ASCO guidelines or NCCN. Is this something that you think we are getting closer to incorporate measures like quality in our decision process when we recommend certain treatments or because the complex is so complex that is it difficult/ but I'm thinking there are pathways out there, there are different levels of recommendation for different therapies. Do you think when you look at public health policies and recommendations and guidance, do you think these kinds of studies are being incorporated in those decision making process? Because to me, this information is very, very helpful. It's not determinant, but it definitely impacts when folks sit around the table and decide, can we pick the best treatments for patients?

Manish Kohli: No, you've hit it on the head. That was our motivation to understand the question to begin with when we first started doing this a year and a half back. Look, we have six different regimens, count Sutent and seven different FDA-approved treatments in first line for metastatic renal cell clear cell type. We have no biomarkers to pick one over the other. Very tough field. I've done biomarker research for over 10 years. It's a very tough field to get to an answer, sadly. Number one. Number two, these trials are such that no big pharma company is going to do a head-to-head trial to know superiority one over the other. This is not like metastatic melanoma where immunotherapy was head-to-head compared with BRAF inhibitors and MEK inhibitor combinations and found that immunotherapy trumped. For some reason, it's not going to happen over here in metastatic renal cell cancer.

How are we going to pick? They're all FDA-approved. My suggestion would be that pharmaceutical and therapeutic committees of different hospitals or different payer systems or institutions need to start looking carefully to do these kinds of analyses. They're all FDA-approved. You are not cutting into the effectiveness of giving somebody metastatic renal cell cancer approved treatments. But P&T committees should find out what is good enough in terms of cost-effectiveness and use that for their own practice in their institutions. There is nothing wrong in that. We are not taking away good treatments from patients because we don't know which of these are better over the other.

We use this single payer system from the VA, which gives us those kinds of numbers. Most times, pharmaceutical therapeutic committees in hospitals don't want to share their cost acquisitions and then it becomes difficult to even do these kinds of studies within their own systems. I think it's important to inculcate this because ultimately, it matters how we are treating our patients and getting denied for one treatment over the other and appealing to their insurance companies again and again to give them the treatment that we think is better than the other, which is mostly decided on survival figures and not decided on taking into account quality of life adjusted years, et cetera. I think there is a role for this. It has to be better defined as we do more of these studies.

Pedro Barata: Right. Now, those are fantastic points. Dr. Kohli, it's been a pleasure talking to you. This has been very, very informative. Not a common topic to be addressed and I'm so happy you were able, first of all, congrats on the work again published at JCO Oncology Practice very recently, and congrats for the outstanding presentation providing us editorial comments on your paper. I welcome everybody to be able to read the paper and dive in the study. It's really worth it. Dr. Kohli, pleasure and I hope we meet again, this time in person and we'll chat more about changes in advancements in kidney cancer. Thank you so much and congratulations.

Manish Kohli: Thank you again on behalf of all our team members. Thanks for having us here.

Pedro Barata: Thank you.