Former Medical Director at GRAIL
Shilpen is a radiation oncologist by training and the Former Medical Director at GRAIL. Shilpen was responsible for running the clinical trials for Grail including the CCGA trial that led to breakthroughs and the upcoming Galleri cancer detection test. Shilpen was deeply involved in testing the different assays and analysing the results from each trial across over 50 cancers. He is now the Medical Director at Genentech where he runs the breast cancer and immuno-oncology division. Read moreView Profile Page
Shilpen, can you provide a short introduction to your background in the industry, please?
I’m a radiation oncologist by trade. I did my training in Baltimore and went on to be on clinical faculty at the University of Washington, and then I was recruited to a small startup called GRAIL. That’s where I worked for a little under two years. Most recently, I was recruited over to Genentech to be a medical director, where I'm focusing on breast cancer and immuno-oncology. At the same time, I continued to have an affiliate appointment as an associate professor at the University of Washington's Department of Global Health and an affiliate appointment in the Division of Public Health at the Fred Hutchinson Cancer Research Center. Because you don’t want to keep it too simple, I also continue to see patients here in the San Francisco Bay Area.
What exactly was your role at GRAIL?
GRAIL is a healthcare company whose mission is to detect cancer early when it could be cured. They were trying to improve the global burden of cancer by developing technology to detect and identify cancer early. The company has been utilizing next-generation sequencing, population-scale, clinical studies, along with some machine learning, to better understand cancer biology and develop a multi-cancer early detection blood test. I was recruited to GRAIL as the medical director in their clinical development. The goal was to provide strategic input, define the strategy and the tactics required to support GRAIL's product pipeline and the eventual commercial launch, as well as post-marketing commitments that would happen after that commercial launch.
I was in charge of helping develop new strategical relationships and maintaining existing relationships with key opinion leaders along the way. At the very basic level, I was doing the medical monitoring of the CCGA study at the time and interacting with our cross-functional team, including clinical operations, biostatistics, and clinical data management. It was working externally but also working internally. The group of people internally was quite large, including biostatistics and regulatory, and all the other stakeholders that are involved. In addition to doing your day to day as a medical monitor for the CCGA study, in my case, we also needed to provide the scientific and medical support for publications and presentations, which we did quite a lot of while I was there. I enjoyed moving the ball forward in that science arena.
Walk me through how a typical clinical trial works. Take the CCGA as an example. How does it work right from drawing the blood.
Just to take a big step back, a clinical trial is a research study conducted, in our case, on human beings. The goal is to answer specific questions. I’ve been a clinical trialist ever since my career started. Sometimes it’s answering specific questions around a new therapy, or as we’ve heard a lot about in the news, vaccines, or in the case of CCGA, diagnostic procedures of diagnostic assays. Or it could be a new way of using the known treatments that currently exist.
The ultimate goal for clinical trials is they're used to determine whether new diagnostics or new drugs or new treatments are primarily safe. That's something underlying for most clinical trials, and of course, effective as well. If you carefully conduct these clinical trials, it's a quick and safe way to determine if whatever you're doing – whether it is a diagnostic or a therapeutic treatment – will help people. Sometimes we’ll test that first in the lab. Sometimes we’ll test that in animal studies. Ultimately the goal is to move into human clinical trials. The CCGA study was a clinical trial looking at a large swath of people.
I'll switch gears to talk about what the study looked at if that makes sense. After a patient provided informed consent – and that's a critical part of it in our CCGA study – the first step was to determine whether there was cancer present or suspected in this individual. If it wasn’t present, we would include them in the non-cancer group, and we would draw their blood sample. Then we would ship that blood sample. It would be received by our sponsor or designee. It would be accessioned and processed and stored until some time when sequencing happened. Ultimately that would be then run in an assay.
The other arm of this study is if we did think that cancer was present, we would draw their blood sample before any kind of therapy, whether that was surgery or chemotherapy or radiation therapy. We would collect their clinical data just like we would for a non-cancer patient. Again, we would take that blood tube and accession it and process it and store it, and then do sequencing, and ultimately run our assays on them.
That’s the general, basic gist of a clinical trial and the CCGA study where we had patients that did not have cancer and patients that did have cancer. I’m happy to highlight anything specific, particularly on informed consent, or give a bit more background on the CCGA study if you think that’s helpful.
I think it’s around 15,000 sample size in patients. How diverse were they?
Diversity is critically important. You are exactly right. 15,000 was the goal of this study in terms of how many we enrolled. Originally, we enrolled 10,000 people and then expanded it because we started accruing so quickly. We wanted to make sure we had the diversity present so that whatever assay would be developed could ultimately be used for any patient. When you look at the demographics of that, it spanned quite a large age range. We had very young patients, particularly on the non-cancer side, and we had very old patients as well. We had a good split in terms of gender, race, and ethnicity. We tried to get some representations across the board. We had patients that were smokers and not smokers. We had a wide range of body mass index as well. The goal for those who did have cancer was obviously to get a variety of clinical stages and a variety in terms of how these patients were diagnosed. Whether they were diagnosed by screening or whether they were diagnosed by clinical presentation.
The reason that this diversity is so important is, again, you want to have a study that is very clinically applicable to as many people as we think it will help.
You get the samples, and then you run them through the sequencers. What is the sequencing process? Do you input the data? What did you then get out of that?
Let me give a little bit of background. We know there have been a number of advances in molecular characterization and genetic sequencing of these malignant tumors, and we’ve seen that happen over the last few decades. That’s laid the basis for this. There is also the Cancer Genome Atlas that represented one of the very major efforts to discover and catalog these genetic mutations that were responsible for cancer by tumor tissue sequencing and analysis, across multiple technology platforms. Some of that work contributed to the development of targeted therapies directed at some genetic variations that confirmed this different sensitivity to drugs based on their molecular mechanism of action.
And then there were a lot of these developments in this next-generation sequencing, this NGS, more specifically, sequencing around these circulating cell-free nucleic acids. Those are both the DNA and the RNA, taken from peripheral blood plasma in cancer patients that ultimately revealed somatic variation arising from a fraction of these nucleic acids, derived from these tumor cells. These circulating cell-free tumor nucleic acids were an important potential source for a non-invasive interrogation of a tumor's presence and its molecular features. This is a huge benefit to patients because taking a blood draw is so much easier than if I stuck a needle in somebody's liver or somebody's body. Or surgical removal.
So this is taking quite the leap by doing this kind of sequencing. To achieve the level of sensitivity and specificity required for cancer screening, you need to have this high-intensity sequencing approach based on a very broad genomic coverage and what I like to call ultra-deep sequencing. That’s partly because there isn’t a high concentration of this floating or cell-free nucleic acid, particularly in cancers that are very early on. You need to have something that's going to pick this up and pick it up accurately, if that makes any sense.
As you can imagine, there is a lot of heterogeneity in different cancers and in terms of the adult population. Everybody can present slightly differently, and cancers all present slightly differently. Getting the appropriate information to kind of file into the sequencing is a challenge. Once you do the sequencing, there is a lot of ways to analyze that data. I’m happy to go into that if it sounds helpful.