Mark Shapiro, vice president of clinical development at contract research organization Clinipace, considers a recent JAMA study on cancer drug pricing in a guest post for Focus. A rebuttal to the post by one of the authors of the JAMA study, Dr. Vinay Prasad, can be found here.
The last five years have been exceptionally busy in terms of oncology drug development. Two interesting articles that reflect back on this period of time were recently published. The first, which I consider here, is “Five Years of Cancer Drug Approvals: Innovation, Efficacy, and Costs” by Drs. Sham Mailankody and Vinay Prasad (JAMA Oncol. 2015; 1(4):539-540). This paper considers whether or not all of the recently approved cancer drugs merit their prices.
According to Mailankody and Prasad, from 2009 to 2013 the FDA approved 51 new oncology medications for a total of 63 indications. They found that the price of new oncology medications is not well correlated with novelty or percentage improvement in the endpoint used to justify approval. The authors go on to note that, “drugs approved based on RR (response rate) were priced highest, with median costs per year of treatment of $137,952. This was greater than the price of drugs approved on the basis of OS [overall survival] (median cost, $112,370).” Their concern is that drugs approved on the basis of a less rigorous endpoint are being priced even higher than drugs that improve survival. I find this interesting but unsurprising, as I explain below.
Effect of Relative Improvement on Cancer Drug Pricing
Mailankody and Prasad made two observations: The first observation regarding relative improvement, and the second regarding novelty. The researchers attempt to relate both to the annualized price of these new drugs. As I noted in a previous blog post, Making Sense of Rising Oncology Drug Costs, patients often do not take these drugs for years, so annualized cost isn’t an appropriate measure of the price of a medication. Typically patients take oncology drugs until they progress, so progression-free survival (PFS) is probably a better measure of how long a patient might be expected to take a particular treatment. For drugs used in later lines of therapy, PFS may mean only a few months of treatment, so the annualized cost dramatically overstates the actual cost. In fact, for drugs used in patients for whom treatment will likely be short, one would expect higher prices, or else the drug wouldn’t be commercially viable.
At the other end of the spectrum, drugs that dramatically improve the duration of PFS might also command higher prices. However, both cases are limited to what the market will bear. It would be more meaningful to see their regression analysis redone using expected cost and cost relative to the comparison treatment used to show extended PFS or OS. Perhaps those comparisons would reveal a statistically significant correlation and also provide some insight into the nuanced logic of those negotiating these prices.
For some drugs, both factors might apply. One example is trametinib, the ninth drug on the researchers’ list. The primary trial on which approval was granted showed a difference in PFS of just over three months. In that case, that was a dramatic improvement versus chemotherapy (4.5 versus 1.3 months, a 246% improvement in median PFS), which shows just how aggressive the V600E and V600K BRAF mutant forms of metastatic melanoma can be. At a price of about $125,280 per year, as reported by Mailankody and Prasad, this drug falls roughly in the middle of the 51 drugs in their study. Of course, if the median patient takes it for 4.5 months, the expected cost is around $45,000, which is still pricey, but this was a groundbreaking medication that substantially improved patient care. Trametinib received AA based on PFS, but later showed an improvement in OS despite the crossover of control patients (Flaherty et al. N Engl J Med 2012; 367:107-114). This point about crossover is worth noting.
In fact, this question about total costs of new, targeted therapies versus older chemotherapies for metastatic melanoma was examined in detail using claims data by Chang et al. (“Comparative healthcare costs in patients with metastatic melanoma in the USA.” Melanoma Res. August 2015. Vol 25(4):312-320.) Using patient claims data from July 2009 to September 2012, they evaluated the fully loaded costs of the available treatments, vemurafenib and ipilimumab (new, targeted therapies which were introduced in 2011), and older chemotherapies: dacarbazine, paclitaxel and temozolomide.
What they found was that total monthly cost for vemurafenib, an orally administered targeted therapy, was $17,129 per month versus $18,272 for dacarbazine, $17,412 for temozolomide and $17,613 for paclitaxel. These differences were not statistically significant. The older chemotherapies had been shown to extend PFS, but not OS. Vemurafenib, on the other hand, extended OS by more than four months. Melanoma patients were suddenly getting a treatment that extended lifespan at the same fully loaded cost as previously available treatments. On the other hand, ipilimumab, a groundbreaking biologic CTLA-4 inhibitor that also extended OS on average by four months, costs almost four times as much. However, in the case of ipilimumab, 22 percent of patients lived longer than three years, meaning that it provides a durable response for a subset of patients.
Novelty in Drug Development: First-in-Class vs. Best-in-Class
The issue of novelty in drug development is a tricky one, especially when it comes to oncology. Trametinib is a novel drug. While it was the first targeted MEK inhibitor, there were approved drugs that targeted other steps in the RAF/MEK/ERK pathway. Although trametinib isn’t the only MEK inhibitor either, it was the first approved. It should be noted that the vagaries of drug development are such that the first drug approved in a particular class often was not the first to enter development. In fact, when many companies start their five-to-seven (or more) yearlong race towards approval at about the same time, no one can predict whose drug will be approved first or in what order the next-in-class drugs will be approved. Portfolio management decisions can alter the pace of development, especially at large companies that make sophisticated tradeoffs among investments in multiple assets. Smaller companies may have fewer options and pursue development in a more single-minded and expeditious manner, but may be subject to external financing pressures. In both cases, “go/no-go” decisions are made at many points along the development path.
Of course, being first-in-class offers some advantages, even while clearing or clarifying a path for those drug developers moving more slowly. Slower developers may benefit from class-specific information about safety, pharmacokinetics and the interplay of both with clinical efficacy. On the other hand, the pool of potential clinical trial subjects can shrink as patients begin receiving the approved drug in the course of normal care. In addition, slower developers may end up having to pay wholesale or increasingly commercial prices for a competitor’s approved product to use as the comparator agent in their later-stage studies. As a result, study enrollment for next-in-class products can be more difficult and the studies themselves potentially more costly. In addition, if the first-in-class agent significantly extends PFS or OS, that longer duration becomes a new benchmark, meaning next-in-class drugs will require longer and therefore more costly studies. Finally, the FDA review and approval process for next-in-class oncology drugs is also typically longer since AA is reserved for those drugs that are especially promising and offer benefits over existing therapy.
Drug developers who are behind their peers have to take some big risks and truly believe that their product is better than existing therapies at the very least for some targeted subset of patients. In this way, companies continue looking for “best-in-class” treatments long after the “first-in-class” drug is approved. When this happens, we frequently see new experimental agents being pitted against approved drugs from the same class. This ultimately works to the benefit of cancer patients, although it can make treatment algorithms complex and confusing for oncologists.
Improving Outcomes Based on Multiple Drugs Within Classes
Experience has shown that even similar drugs are not identical in the clinic, where each patient is different and each tumor is not only unique to a patient, but evolves over time. Furthermore, a mutation that makes a tumor immune to one agent may leave that tumor susceptible to another agent in the same class. Ultimately, the order of marketing approval has no bearing on which drug will work best for a specific patient or when it might be effective during their course of therapy. Therefore, we shouldn’t downplay the value of having multiple drugs in a class. I have often used variants of the following chart to underscore the value having multiple drugs within a class brings for patients.
Figure 1: Schematic representation of the proportion of patients with Disease X helped by Drugs A-F.
In this example, even with six drugs of the same class, patients for whom this class of drugs does not work remain. Note also that five of the six drugs, Drugs A-E, treat at least a small proportion of patients not treated by any other drug in the class. Only Drug A is wholly redundant. However, if one considers the order of approval as A to F, then the innovator drug only became wholly redundant after a sixth drug in the class, Drug F, was approved. I think this is a better representation of what is actually occurring in drug development.
One example that illustrates this concept well is the progress made in treating advanced or metastatic renal cell carcinoma (mRCC). Several drugs on Mailankody and Prasad’s list are VEGF-TKIs used for mRCC. In fact, this space is so crowded that Drs. Rosalie Fisher and James Larkin dedicated a review (“Individualising treatment choices in a crowded treatment algorithm” [EJC SUPPLEMENTS 11(2013)160–168]) that goes into great depth about what all the studies tell us and how to interpret them in terms of best treating patients. They note:
“Immunotherapy, generally interferon-alpha (IFNa), was standard treatment until 2005, when it was replaced by the first inhibitor of the vascular endothelial growth factor receptors (VEGFRs), sunitinib. Since then, another six agents which target either the VEGF or the mammalian target of rapamycin (mTOR) pathways have been developed and approved for use in advanced RCC.”
In addition to sunitinib and sorafenib, the first two agents in this class, next-in-class agents pazopanib and axitinib are also approved for various uses in mRCC. Fisher and Larkin describe the complex treatment algorithm and underlying studies to support it as follows:
“Accepted second-line treatments for mRCC are the VEGFR–TKIs sorafenib, sunitinib, pazopanib and axitinib, and the oral mTOR inhibitor everolimus. Frequently, the decision is influenced by which first-line treatment the patient has received; for example, there is evidence that sorafenib, sunitinib, pazopanib and axitinib have clinical activity after prior cytokine therapy … The main controversy exists in the decision between everolimus and axitinib, when patients have been previously treated with a VEGFR–TKI.”
Interestingly, in 10 years we have seen six new agents from two classes largely replace the previous therapies, which were immunotherapies, typically interferon alpha (IFN) or IL-2. Furthermore, the complex treatment algorithm described above is set for major changes in the very near future as several new therapies based on immune checkpoint inhibition of PD-1, PD-L1 and CTLA-4 are in late stage testing or nearing FDA approval. Like ipilimumab, the CTLA-4 inhibitor discussed above, these immunotherapies are promising to provide longer and more robust responses to subsets of patients than was possible with earlier generation therapies.
Today, sunitinib and axitinib remain first-line treatments, but PD-1 inhibitors are becoming the treatment of choice in second-line, and may be moving into first-line based on the early stoppage of the CheckMate-25 study after the DSMB found that the nivolumab arm showed superior overall survival (European Society for Medical Oncology). Interestingly, this sets the stage for a number of possible third-line treatments and more first- and second-line combination regiments, but today these patients remain healthy enough to receive additional rounds of therapy because all of these newer agents are less toxic and more tolerable than treatments available before 2005, when sunitinib was approved. At that time:
“Sunitinib was compared with IFN in a randomized phase III trial and resulted in a statistically superior response rate (RR 47% versus 12%) and progression-free survival (PFS) time (median 11 months versus 5 months). The median overall survival time was 26.4 months for sunitinib compared to 21.8 months for IFN-treated patients, which only became significant when those patients who crossed over from IFN to sunitinib were excluded from the analysis (emphasis added).”
Again, the issue of why PFS does not always correlate with OS is raised. This was one of the key concerns raised by Drs. Chul Kim and Vinay Prasad in their follow-up article, “Cancer Drugs Approved on the Basis of a Surrogate End Point and Subsequent Overall Survival: An Analysis of 5 Years of US Food and Drug Administration Approvals” (JAMA Intern Med. Published online October 19, 2015).