Objectively Evaluating Innovative Therapies And Making Sense Of Emerging Data
By Kristin Yarema, Ph.D., President and CEO, Poseida Therapeutics
We live in a period of history where innovative therapies for oncology are emerging and changing the dialogue about how we think and talk about treating cancer. It’s an exciting time to be in the industry and to be able to participate in delivering medicines that can change the lives of patients.
Cell therapy is one of those innovative therapies. The excitement in the medical community and industry is understandable. We are observing results in clinical trials that appear extraordinary and may be even resulting in functional cures for some patients in some cancers.
The excitement generated by these results causes everyone to want to jump onboard, move fast and participate in what certainly will be upside for patients, physicians and companies alike. It’s exciting to be sure, but we need to resist the urge to get carried away by the excitement and not objectively evaluate what we are observing in order to understand the real benefit of these various cell therapy approaches for patients, physicians, the health care system and the companies that are pursuing them.
Don’t Simply Follow the Headlines—Do the Analysis
Autologous CAR-T for oncology is one example where the results are extraordinary, but one must take a deeper look to truly understand the outcomes for patients, physicians and the healthcare system. The headlines are impressive, with some companies reporting near 100% overall response rates, or ORR, and reporting that a majority of those patients reach deep responses in some hematologic malignancies. While those headlines represent excellent outcomes for the patients who actually achieve them, we still need to take a deeper look at the bigger picture.
Autologous CAR-T is complex, requiring individualized manufacturing and time to make, test, and release the product and treat the patient. Those complexities can make understanding and comparing the autologous CAR-T data to other modalities a challenge. For example, we heard at ASCO last year that in multiple autologous trials for multiple myeloma and B-cell malignancies, sponsors observed that some meaningful percentage of patients either progress too far in their disease or, sadly, die before their autologous therapy can be manufactured, released and administered. In addition, we know that autologous CAR-T has suffered from manufacturing challenges and that a meaningful percentage of patients’ products can fail manufacturing altogether or not meet specifications. These are not trivial findings and the implications for understanding the data should be explored.
So, how do we factor these real-world dynamics into understanding the data?
“Intent to Treat” or “As Treated” Analysis in CAR-T Clinical Trials for Oncology?
In other areas of drug development, we commonly use Intent-To-Treat, or ITT, analysis to evaluate clinical outcomes. In an ITT analysis the principle is simply that, in assessing the therapy, you should include all patients that you originally intended to treat regardless of what treatment, if any, they received. One other alternative is to look at an “As Treated” analysis, which only evaluates patients who were actually treated as defined in the protocol.
In our view, from a patient and a physician perspective, the best analysis would be an ITT analysis. Simplistically, if you intend to treat a patient, they should be counted in the analysis. If your manufacturing process is too long or you can’t deliver a product that meets specifications – those things matter when comparing to alternatives. As a result, for patients and physicians – ITT is the much better metric to reflect the value and the risk/benefit of a therapy for comparison when lives are literally at stake.
Let’s look a little deeper at an example.
In the case of autologous CAR-T, looking at an ITT analysis would have you include all patients that get enrolled - as you are obviously intending to treat them from the time you decide to subject them to apheresis. As a result, outcomes for patients who progress too far or die while waiting, or who receive a product that does not meet specifications, should be included in the analysis. The implications of looking at the data through an ITT lens are apparent and all patients who died or did not receive treatment should be used in the analysis. In our example, including those patients would, in itself, reduce the objective response rate (ORR) from near 100% to something lower, possibly much lower if we believe what we heard at ASCO.
In addition, one should also include in the analysis all patients who received product that did not meet specifications – whether or not the patient was dosed with such product. While the data on out-of-spec dosing is not readily available from most trials or sponsors, it would be reasonable to assume that out-of-spec product would not perform as well as product that meets specifications. Thus, the ORR and deep response rates would likely be further impacted for at least some autologous CAR-T programs.
What are the real ORR and deep response rates for these autologous products based upon an ITT analysis? It’s a great question.
In the case of allogeneic CAR-T, the ITT analysis falls more in line with what we would expect compared to biologics or small molecules that are readily available. The confounding factors are not present. If the product is available off the shelf where patients do not have to wait to be treated and all product is tested and meets specifications prior to treatment, then it becomes much easier to treat every patient who can benefit - and treat them relatively quickly. Therefore, for an allogeneic CAR-T product that meets those requirements, the ITT analysis of the data tells an accurate story about what patients and their physicians can expect.
Our conclusion is that as we compare cell therapies in oncology, we should be looking at the ITT analysis of the data as that data reflects the patient experience and expectations. Through that lens we believe that the right allogeneic CAR-T will stack up favorably to autologous CAR-T.
Allo is Not Auto and Allo is Awesome
We are fond of saying that “Allo is not Auto” and that “Allo is Awesome.” We still hear some investors and others cite as treated numbers from some autologous CAR-T trials and say the rates are “tough to beat” and “it may be hard for allogeneic CAR-T to be competitive.” However, as outlined above, when one takes a deeper look at the reality of these autologous therapies and what they mean for patients, physicians and the healthcare system - we respectfully disagree.
As we think about allogeneic CAR-T for oncology, and the data we recently presented at ASH in December 2023, we believe more than ever that not only can allogenic CAR-T compare favorably with autologous CAR-T, but in many ways allogeneic has advantages that autologous cannot hope to match. We also believe that we have significant benefits in cost and potential tolerability advantages that further set our allogeneic approach apart. While we acknowledge that our data is early, we demonstrated a number of things that continue to build our confidence that our allogeneic high-Tscm CAR-T platform can lead the way for patients.
