Striving for early predictors of immunotherapy response to prevent immune-related adverse events in gastrointestinal cancer: An interview with Dr. Dmitrii Shek
“We do patient-centered research. Old school technologies can explore and elucidate the expression of a single marker or a couple of genes, but for cancer, that’s probably not enough. We need to do something more. We need to explore new dimensions. We need to do multiomics in order to fully understand what's going on with the particular patient.
“Every human and every genome composition is individual. We are so different, but, at the same time, we are so similar. If we treat every patient and if we do multiomic expression analysis for each of the patients, perhaps this is what will give us a new understanding of the tumor. This will give us new information—how we can use therapeutic options which are currently available with the maximum potential benefit for the patient, so the patient can survive another few years. The patient can enjoy their life.” – Dmitrii Shek, MD, PhD, Western Sydney University
An earnest passion for patient care and quality of life, and a burning hope for prolonged survival to beat the odds of some of the world's deadliest cancers, characterizes Dmitrii Shek, MD, PhD of Western Sydney University.
Those qualities were developed through his training and practice as a doctor, and as a grandson. His grandfather’s battle with cholangiocarcinoma, a rare cancer of the bile ducts, left a deep impression on Dr. Shek, personally and intellectually, and drove him to research:
“This was a crucial moment in my career because I decided, ‘Ok, maybe I need to put my medical career on pause,’ and perhaps pursue a research career to better understand the tumor, to better understand cancer, to better understand gastrointestinal cancer, in particular.”
We spoke with Dr. Shek to understand the current state of immunotherapy for gastrointestinal cancers and how high-resolution, high-plex single cell and spatial transcriptomics technologies play a role in supporting his goals to identify predictive biomarkers for immunotherapy-related toxicity and, ultimately, to develop improved, personalized treatment options for the patients he serves.
Can you tell us more about your research focus and how your medical experience motivated the direction you took?
In my PhD project, I was trying to understand why some patients with some cancers are likely to respond and are likely to get the benefit from novel drugs, which are called immune checkpoint inhibitors. These are immunotherapeutic drugs designed to boost the human immune system, in particular, T cells. The T cells begin to proliferate, and they begin to attack the malignant tumor. In melanoma trials, it has been shown that immunotherapy is much more effective than the conventional treatments, such as chemotherapy or other drugs.
But we study why some patients respond and why other patients do not respond. We also study why some patients develop specific immune-related adverse events—such as immunotherapy-related colitis—but some do not.
The scientific and clinical rationality behind this is that patients with colitis can experience diarrhea up to 15 times a day, which is obviously a huge burden for them. And that's why the patients have to cease immunotherapy. They have to receive different immunosuppressive drugs, such as steroids. But, you know, cancer is very smart. Once the immunotherapy stops, there are no barriers for the cancer to grow again. Usually, for the patients who stop immunotherapy, the chances of cancer recurrence are much higher than they were before the immunotherapy started.
What is the significance of these immune-related adverse events? How do they fit into the larger story of progress and improvement in the field of immunotherapy?
Maybe 25, 30 years ago, when patients were diagnosed with advanced metastatic solid cancer, the therapeutic options were really limited. I faced this when my grandfather was diagnosed with cancer. Usually it's a platinum-based chemotherapy. Maybe they can do some radiation therapy. If the cancer is diagnosed at early stages, then surgery.
The results of those therapeutic options, especially if the cancer is diagnosed at advanced or metastatic stages, are not as efficient as they are today with the use of immunotherapy. The first immunotherapeutic drug was approved by the US Food and Drug Administration in 2011. It was successfully used for patients with metastatic melanoma. Since then, seven more immune checkpoint inhibitors have been approved, and they are now used across 40 different malignant tumors. The fact is, they have already shifted the first- and second-line treatment paradigms for many tumors.
We can see every day in the clinical practice that the patient lives a better life—they can survive, they have hope to surpass another year. I know one patient, he's 73 years old, and he unfortunately was diagnosed with malignant pleural mesothelioma, metastatic stage. He was a miner. In Australia, there are a lot of mines, and, unfortunately, there is huge exposure to asbestos, which is one of the major risk factors for the development of this malignant tumor. In the US, mesothelioma is considered a rare cancer, but in Australia, unfortunately, the incidence of this tumor is relatively higher.
After talking with him, the only thing he said was that he wanted to see his grandson turn one. That was his dream. That was his hope. But at the same time, he said he understood that it's very unlikely. It was really an emotional moment for me because I remembered myself, 6 years earlier, when I was in a similar situation—but I was the grandson of a man who was diagnosed with incurable cancer.
Fortunately, this gentleman was treated with immunotherapy, and he is still alive. His grandson is already two years old. He is living his life, and he did not develop any side effects. He had a good response. He's on further follow-up, and we are hoping that he will be able to surpass another few years. For a patient diagnosed with metastatic mesothelioma, that's really a breakthrough.
But this is really crucial to understand: almost one out of two people who are treated with immunotherapy will develop immunotherapy-related adverse events. Cancer is a deadly disease, and that's why everyone is thinking about therapeutic efficacy. And we need to treat the cancer, we need to shrink the tumor. But, at the same time, we forget about the safety profile of those drugs. We forget that every second patient can experience life-threatening, immune-related adverse events.
So this shows just how significant the problem is and that we definitely need to do something. We need to establish some kind of clinical risk algorithm. We need to establish the biomarker score because, if we are able to critique this at the pre-treatment stage, when the therapy starts, perhaps by adjusting the dose or adjusting the regimen, we can prolong the duration of immunotherapy. We can minimize the opportunity for those patients to develop side effects.
This is how we will make immunotherapy a “chronic treatment” and cancer will become a “chronic disease.” The patient will not only talk about 1 year survival, 5 years survival, but 10 years survival, 20 years survival. That's the goal.
Can you introduce us to the clinical trial you helped to start? What's the trial for? What are some of the hoped for outcomes?
The trial is a prospective study. We prospectively recruit eligible cancer patients who are treated with immune checkpoint inhibitors—usually a combination, meaning there are two immune checkpoint inhibitors given at the same time. It's currently considered that this type of treatment is highly effective. But, at the same time, there is a greater likelihood for those patients to develop toxicity. We decided to select this category of people because they are more likely to develop immune-related adverse events.
So we prospectively recruit those patients—we isolate blood and tissue samples from them, we follow up with them, and we collect the clinical and pathological data. By creating this large biobank of information—not just clinical data, but also the biological samples—we want to know if we can establish valid biomarkers of immunotherapy-related toxicity. We implement 10x Genomics technologies on immune cells from those patients at the pre-treatment stage, after cycle 1, after cycle 2, then when the patients develop toxicity; or, on the tissue, which was isolated from either the original tumor or the site of the immunotherapy-related colitis.
Can we establish the population of immune cells? Can we establish gene or protein signatures which will guide us in the pre-treatment stage so clinicians can understand patient outcomes? That can help them say, “Ok, this patient is likely to develop this type of toxicity. We perhaps need to do this and that in order to prevent this in future.”
This is how we will be able to extend the duration of immunotherapy. And the longer the patient receives immunotherapy, the higher the chance that the cancer shrinks to a minimal stage and will not grow again. That's the goal of this study. Today, we have already recruited 250 patients across Western Sydney. To our knowledge, this is a first-of-its-kind study, at least in Western Sydney and in the area where we are currently working. The recruitment is currently ongoing, and we are hoping to reach another 50 patients, maybe by mid-next year.
Once we reach this large cohort of 300 patients, we will continue implementing cutting-edge technologies provided by 10x Genomics in order to prove our preliminary findings and our pilot data. And if it works, then it will probably change the field of immunotherapy, particularly personalized immunotherapy. We will be able to adjust the therapy in the beginning, before the treatment starts, so that the patient will receive the highest likelihood of therapeutic benefit that we can provide.
You mentioned that 10x Genomics products have been valuable in your preliminary work with these clinical samples. What have been some of your most important findings using single cell and/or spatial transcriptomics technologies?
When we were doing the pilot spatial analysis using the Visium technology, we were able to better understand the tumor microenvironment in patients with malignant pleural mesothelioma who were treated with the dual immune checkpoint inhibitors. Some of them responded, some of them did not respond. Some of them developed toxicity, some of them did not develop toxicity. By implementing Visium, we were able to elucidate the full gene expression from the tumor. We were able to understand how the distance between the immune cells and the tumor cells could potentially impact the outcomes of immunotherapy in those patients.
We also established a unique population of macrophages. Macrophages are innate immune cells, and there are two different populations, M1 and M2. The M1 are the cells which play the antigen-presenting role. They benefit our immune system. But, for some reason, there is another population of macrophages—it's like, you know, good cop, bad cop. This M2 population of macrophages is not like the M1—they basically create more fibrotic structures in the tumor. They limit the traffic of T cells into the tumor. They create this kind of protective shield for the tumor so that the immune cells do not reach the tumor cells and do not attack them. We were fortunate to establish that using Visium. Then we also implemented Chromium single cell sequencing. With Chromium, we were able to elucidate the unique features of these macrophages and are currently working on further analysis.
At the Westmead Institute for Medical Research, we finally have a Xenium, which is totally another level. It's the future of basic science for immunology research. We are hoping that, once we finish the clinical trial recruitment, we will [be able to] explore what Xenium can do for our research and how it can help us create more personalized immunotherapy for cancer patients.
What might be the value of combining both single cell and spatial technologies to study these patient samples?
I recently presented at a clinical conference and one of the professors asked me, “What's the rationale of spending so much money for something unknown—something that you do not know whether it will be valid or not, and you do not know whether it will be successfully implemented into the clinical practice or in the diagnostics?”
I explained to him that, by using single cell, we can understand the in-depth gene expression profile within the tissue. By doing spatial, we can basically align this gene expression profile with the tissue. By doing this together, we are able to locate which side of the tumor, which cells, and which genes are responsible for a particular effect and outcome from the immunotherapy. The reason behind implementing both technologies together is to create a better future for the patients.
We do patient-centered research. Old school technologies can explore and elucidate the expression of a single marker or a couple of genes, but, for cancer, that’s probably not enough. We need to do something more. We need to explore new dimensions. We need to do multiomics in order to fully understand what's going on with the particular patient.
Every human and every genome composition is individual. We are so different, but, at the same time, we are so similar. If we treat every patient and if we do multiomic expression analysis for each of the patients, perhaps this is what will give us a new understanding of the tumor. This will give us new information—how we can use therapeutic options which are currently available with the maximum potential benefit for the patient, so the patient can survive another few years, [so] the patient can enjoy their life.
From our experience, it is expensive to treat the patient with immune checkpoint inhibitors. If the patient receives maybe 12 or 14 doses in a year, it will cost roughly $150,000 just for the treatment itself. Some of those patients will experience side effects, so they will have to cease immunotherapy. So then there is another drug option that we have to invest in, or maybe we push the patient into another clinical trial.
But what if, rather than spending a lot of money on treating the patient empirically with the best currently available, we screen the patient from top to bottom in the beginning, at the pre-treatment stage. Then we can understand what's going on with the particular patient, and, after that, we can decide that this drug, or this combination of drugs, is likely to benefit this patient. It will save money for the government. And, more importantly, it will provide the maximum potential benefit for the patient.
That's why I believe these new technologies, the single cell and spatial technologies, the 10x Genomics technologies, are crucial for the future of personalized medicine. Maybe today they are expensive. Maybe today they are considered the new cutting-edge tools to do some fancy research and publish in the big journals. But I see this as the future approach for patient diagnostics, the future approach for creating personalized medicine, and the future approach for increasing the lifespan for patients with cancer.
You mentioned that you plan to incorporate Xenium into your future research. How do you see high-resolution, high-plex spatial tools benefiting your research? How would you want to apply it, what would you be looking for?
From our preliminary data, we established that there are some unique subsets of macrophages and T cells which are likely related to some of the side effects. What we want to do is isolate the colon tissue from the patients who develop immunotherapy-related colitis. We want to use Xenium to understand what's going on within that inflamed tissue.
When the patient develops immunotherapy-related colitis, they unfortunately experience a lot of different clinical symptoms, but they are usually referred to the endoscopy clinic because the immunotherapy-related colitis is still considered a diagnosis of exclusion. Basically, if the patient treated with immunotherapy suddenly develops diarrhea like 12 to 15 times a day, we cannot say with 100% certainty, “Ok, this is immunotherapy-related colitis.” We have to prescribe a lot of different tests. We need to send the patient to the gastroenterologist, to other clinics, because we do not have any predictive biomarkers or valid clinical risk score algorithms. We need to exclude everything else first and then we say, “Ok, it's likely immunotherapy-related colitis.”
Then we send the patient to the endoscopy clinic for a colonoscopy to see what's going on inside the colon. Usually during the colonoscopy, they obtain a few tissue samples. They send those samples to the pathologist, who looks in the microscope and says there are some pathological differences within the colon tissue, within the mucosa layer, etc.
Now we have those tissues, and it will be interesting for us to see what's exactly going on within those tissues, right? Not just doing immunofluorescence staining for a couple of markers, but to see the full plaques, like 5,000 different markers within the tissue, and to understand what's going on there—whether there are T cells, macrophages, or B cells, or maybe some other cells that are relevant to the immunotherapy-related colitis.
And this will give us clues about what we should look at in the blood. Because, eventually, if we want to create a diagnostic method for use in the clinics, it should be blood-based, right? The tissue is so complicated to collect and not all the patients will be happy that, before we start the treatment, we need to do an invasive procedure to obtain a biopsy and do the screening test. Well, it's not very convenient, and everything we do must be patient-centered. We must do what will benefit the patient the most and will be the most convenient for the patient. So it should be blood.
By doing Xenium analysis on the colon samples—on maybe 60 or 100 samples—we will understand if it will give us a clue about which cells are there. And for these cells, we can look into the blood and then potentially establish a reliable diagnostic algorithm on how to predict the likelihood of response or of developing toxicity from the beginning. Then the patient can know exactly what will happen to them within the next few months or years after they receive immunotherapy. They will understand the likelihood of their survival and what type of benefit they will receive.
I think this is the future of medicine. I think that, through implementing cutting-edge research technologies with AI and combining all the data together, we will be able to build a new predictive algorithm for these patients. Although it probably sounds impossible—but, come on, 50 years ago nobody thought that a patient with metastatic mesothelioma or melanoma could survive 5 years. But one person from MD Anderson, who eventually became the Nobel Prize laureate in 2018, thought that it could be possible. So, maybe it all starts with the dream.
Is there anything that you would want to share that we haven't yet addressed or talked about?
I just want to thank the whole 10x Genomics team, the scientific team, the field application scientists, and everyone across the globe working at 10x Genomics for a phenomenal job and for the tremendous support you provide to researchers and clinicians.
We are especially grateful for the staff working in Australia. They are super helpful, and they guided us from the beginning to the end. We were working on FFPE blocks, and there were a lot of questions: what type of staining is better for us? We can do H&E, we can do immunofluorescence staining—what's better for us? Because we were more so doing exploratory analysis, rather than already knowing what we wanted to target or what particular subset of cells or genes we wanted to look into. But we received amazing support. Catherine King helped us through this process. She even came to Western Sydney to help us do Visium for the first time. It was truly amazing, and we were very lucky to receive good results from that. Gerry Ma helped us do single cell sequencing. He was with us in the lab while we were doing library preparation, and we asked him a lot of different questions.
There are multiple technologies, either in single cell or in spatial, currently on the market. We were thinking about which technology we should implement. And, to be honest, we even started with another technology from another company. But, because of their service—there were a lot of technical questions and issues, and when you email someone and do not receive a response for a couple of days or a couple of weeks even, it gets more frustrating, especially when your supervisor is expecting the results from you.
But the 10x Genomics team here in Australia is doing an amazing job. Thanks to all of you guys because it was truly phenomenal to work with you.
I definitely will continue working with 10x Genomics—because 10x has the best service, the best support team, the best field application scientist team, and the best pricing for the value we get.
This interview was edited for length and clarity. We’d like to thank Dr. Shek for his inspiring responses!