Q&A with Ira Mellman, PhD, Vice President, Cancer Immunology, Genentech – Speaker at PMWC 2018 Silicon Valley

Q: Immunotherapy is considered a huge game-changer and holds a lot of promises. Is the hype around immunotherapy justified? What are some of the exciting developments in the recent years?

A: I do believe in the promise of immunotherapy to significantly help people living with cancer.

But it’s important to find the right people to give it to, as it’s certainly not a one-size-fits-all approach. The immune system is incredibly complex, and each person’s immune system may respond differently. Some of the most exciting recent developments are related to improving our ability to identify both responders and non-responders to immunotherapy – and for the non-responders, determining why.

As my colleague Dan Chen noted last year, one of the most common biomarkers for immunotherapy is measuring the levels of a protein called PD-L1, but there are some issues with consistency and reliability with this approach. The identification of new, quantitative, independent biomarkers will likely be key to realizing the promise of immunotherapy. One such genomic biomarker, tumor mutational burden (TMB), has been shown to help predict response immunotherapy across a range of tumor types, including lung, bladder and melanoma, and the list is growing.

Technological advances like liquid biopsies are also helping. The ability to measure these genomic biomarkers from pieces of tumor DNA from the blood can help expand the promise of immunotherapy to patients for whom tissue biopsies are not feasible, such as the estimated 1 in 3 patients with advanced lung cancer who cannot undergo a tissue biopsy. This is why we have a new Phase 3 trial that investigates TMB measured from the blood as a non-invasive biomarker of response to first-line immunotherapy in advanced lung cancer patients.

Another exciting development is the incorporation of “real world” clinical and genomic data to our trial design. At Genentech we’re investing in technologies and companies that are experts in the organization of vast amounts of healthcare data, including tumor genetics, treatments, lifestyle, clinical endpoints and patient-reported outcomes for thousands or even hundreds of thousands of people. This “big data” approach can help us better understand patterns of who responds or doesn’t respond to immunotherapy. Importantly, for those that don’t respond, we have a wealth of information to use to generate new testable hypotheses about why, and we can take those insights from the clinic straight back to the lab and develop new approaches that match the clinical and genomic profiles of those individuals.

Q: Will immunotherapy over time become the standard-of-care for most cancers? What are some of the main challenges we need to overcome to make immunotherapy applicable to more cancers? What are some innovative approaches we need to take to move the needle?

A: Immunotherapy certainly has the potential to become the standard of care for many cancers. But it may be in combination with other types of therapies that involve manipulating the immune system in different ways to recognize or attack a specific type of cancer. Navigating the complexity of each person’s immune system is one of our main challenges, but fortunately the science is moving quickly.

One of the first major roadmaps for teasing apart the biology was the cancer-immunity cycle. The cycle described the diverse ways immune cells can detect, locate, infiltrate and destroy a cancer. Each of these steps in the cycle could be targeted to enhance an immune response, and this paved the way for designing “rational combinations” that fine-tune this response based on the type of cancer is present. Depending on the cancer, a given immunotherapy may be most effective with a chemotherapy, targeted therapy or another immunotherapy, and at Genentech we have dozens of ongoing clinical trials testing each of these possibilities guided by unique biomarkers.

We’ve also learned that the cycle alone can’t explain why some people respond but others don’t. There are many factors beyond the cycle that are involved. Biomarkers that characterize not only the tumor, but also the patient’s immune system can help to tailor cancer immunotherapies in a personalized way.

Understanding the state of a person’s immune system, or their cancer-immune set point, could therefore inform what approach is best for each individual. For example, some tumors don’t express any mutated proteins that can be recognized by immune cells, while in other cases immune cells may be blocked from entering the tumor by physical barriers like blood vessels. Combining this knowledge with tumor biology could bridge scientific and clinical insights and inform the most effective rational combination strategy.

But it will certainly take a village to achieve this vision, and one innovate approach we are taking is to “crowdsource” the collective wisdom of the scientific research community to inform and evolve this framework. We are now sponsoring an online, interactive version of the set point framework that allows scientists to add, edit or remove factors based on their own validated findings. Tapping into the knowledge of the entire community, in real-time, could help rapidly advance the field of personalized immunotherapy.

Q: Cancer vaccines is a growing, exciting area. What new and exciting developments make this a promising solution?

A: The challenge with cancer vaccines is that cancer cells very closely resemble healthy cells, and historically it’s been very hard to get the immune system to recognize the cancer as “non-self”. Finding cancer-specific mutations called neoantigens has been like finding a needle in a haystack.

Fortunately, there have been rapid improvements in next-generation sequencing technology, and today we can sequence the whole genome of a tumor in just days. This also allows us to identify neoantigens in individual tumor samples, and develop cancer vaccines tailor-made for each patient. It’s the ultimate form of personalized medicine.

What’s particularly exciting is how quickly we can now create these personalized cancer vaccines. We can go from sequencing to neoantigen identification to treatment administration all in a matter of weeks – crucial efficiency in what can often be a race against time.

This approach is very promising, but it is also likely that these cancer vaccines will be most effective when combined with other cancer immunotherapies, like checkpoint inhibitors. Rational combinations guided by the cancer-immunity cycle become important here. You can imagine the “one-two punch” when combined with a personalized cancer vaccine: the vaccine “primes” the immune system to recognize the disease, and the checkpoint inhibitor “removes the brakes” to enable the immune system to attack the cancer.

Q: Immunotherapy has potential for dangerous side effects (e.g. autoimmune diseases): how advanced is our knowledge in understanding to fine tune the potency of these new medicines in order to identify and stay within a safe therapeutic window?

A: I think it’s important to note that as a monotherapy, most responses to immunotherapy have generally been durable. As they become more widely available we’ll be able to understand more about their long-term safety and efficacy profile from real world data.

But given that combinations may be the future of cancer immunotherapy, here is where we’ll have to pay special attention to safety. Combining checkpoint inhibitors (e.g., anti-CTLA4 and anti-PD1) has indeed proven to be quite toxic and should only be used for certain patients.

This is a relatively new field, and we only have a few major classes of checkpoint inhibitor (anti-CTLA4, PD1, PD-L1) and cell based immunotherapies (CAR-T). In the short-term, our approach will be based on adjusting the dosing of specific combinations to make different regimens as tolerable as possible. In the long-term, we will hopefully have a larger repertoire of validated targets to combine that will be tailored to the characteristics of each individual and their cancer.

Q: Cancer immunotherapy is a relatively young field: what do we know already on the long-term cure rates, relapses, efficacy profile in relapse patients, etc.?

A: Due to cancer immunotherapy’s relative youth we don’t yet have concrete or reliable long-term data for most checkpoint inhibitors, particularly around cure rates and relapse rates. But what we do know is that for these types of immunotherapies the overall response rate across all tumors studied so far is around 10-30%.

We now have the opportunity to significantly increase those percentages through the approaches I’ve described above. Using a multiple biomarker strategy combined with a deep understanding of the state of each person’s immune system can help us determine why a given tumor might be “escaping” an immune response, and provide us with the tools to prevent that escape.

For example, one hypothesis is that immunotherapies only work in people whose immune system has previously recognized cancer cells. Biomarkers that reflect pre-existing immunity can help us find those people.

In clinical trials we often study large populations, but as individuals we are all unique. With a robust approach to personalized cancer immunotherapy, we can convert that 10-30% average response rate into a 100% response rate for a specific individual.

To learn more about Ira Mellman and Dan Chen’s work and their crowdsourcing initiative, visit A Community Map of Cancer Immunity.