Digitization is opening many new opportunities. Digitization helps pathologists work more efficiently. It removes pain points, facilitates sharing cases for second opinions, and enables the application of AI to support diagnosis. AI provides pathologists greater peace of mind, which allows them to focus on what’s most important in their diagnosis. Furthermore, it’s helping to democratize better patient care in communities and countries around the world without access to the mostly experienced, super-specialized pathologists.
Cancer is a horrible and terrifying disease. There is so much great information out there, but sometimes it is very difficult to filter out the noise. What causes cancer? Can it be prevented? How do you detect it? What are the odds of survival today? What are the different forms of cancer? What are the best treatments? And what is the best way to support someone impacted by cancer?
In this interview series called, “5 Things Everyone Needs To Know About Cancer” we are talking to experts about cancer such as oncologists, researchers, and medical directors to address these questions. As a part of this interview series, I had the pleasure of interviewing David S. Klimstra, M.D.
David Klimstra is Founder and Chief Medical Officer at Paige, the first and only company to receive FDA approval for an AI product in digital pathology. An internationally recognized expert on the pathology of tumors of the digestive system, pancreas, liver, and neuroendocrine system, Dr. Klimstra has published over 550 primary articles, 125 chapters and reviews, and four books. He received his M.D. and completed a residency in anatomic pathology at Yale University and he completed fellowship training in oncologic surgical pathology at Memorial Sloan Kettering Cancer Center, where he practiced for 30 years prior to joining Paige. Dr. Klimstra also served as the Chairman of the Department at Memorial Sloan Kettering Cancer Center from 2012 to 2021.
Thank you so much for joining us in this interview series! Before we dive into the main focus of our interview, our readers would love to “get to know you” a bit better. Can you tell us a bit about your childhood backstory?
Throughout my education, I was always interested in science. My father was a scientist, so I inherited a scientific mindset from him. He was an organic chemist who worked for a pharmaceutical company and contributed to one of the very first oral contraceptives.
Going to medical school seemed like a natural progression of studying biology in college. I had a family friend who was a pathologist who was very supportive of me going to medical school. He worked really hard to motivate me into pathology because pathology is not emphasized as a subspecialty in medical school. So, unless you have some kind of connection that gives you exposure, it’s hard to develop much of an affinity for it. By the time I got into my fourth year, it was clear to me that diagnosing and studying things microscopically was appealing.
What or who inspired you to pursue your career? We’d love to hear the story.
As I mentioned, pathology was not emphasized as a as a subspecialty when I was in medical school, and it’s emphasized even less so now. Bernard Wagner was the family friend who steered me towards pathology. He was he was a very well-known senior pathologist who founded two journals as editor-in-chief. He was always very supportive and introduced me to the concept of analyzing tissues microscopically and drawing conclusions about the nature of disease.
This is not easy work. What is your primary motivation and drive behind the work that you do?
My primary motivation is to help patients. I work in oncology and diagnose cancer, which is one of the most dreaded diagnoses a person can hear. And so, learning to understand how cancers are different, how we can treat some of them very effectively, recognizing what type of cancer it is and what the best treatment is to help the patient is my primary motivation. To do that, you need to learn and understand what ultimately helps the oncologists and surgeons offer better care to patients.
What are some of the most interesting or exciting projects you are working on now? How do you think that might help people?
At Paige, we’re using artificial intelligence to augment the interpretation of pathology, which for 150 years, has predominantly been based on looking through a microscope at tissue samples that are stained and put on glass slides. In the past 15 years or so, we’ve developed technology to create a digital representation of those slides. So now, a pathologist can look at a slide on a computer. This opens up many different opportunities. Once the image is digital, you can use machine learning tools to analyze the digital image, and essentially teach a computer to recognize things that pathologists can recognize — and also things that pathologists cannot yet see. After training an AI model on literally tens of thousands of different cancers, one of the most exciting things to see is the technology develop the capability of recognizing the cancer itself.
When I started, I had a certain amount of skepticism that this would work. But it does work, and it’s very gratifying to see that the patterns we’ve associated with cancer for generations now can be recognized by a computer, which will give us the ability to do so much more. This is exciting because now there’s an opportunity for the computer to interpret things from the pathology that we currently don’t know how to do. One of the most interesting things we’re working on now is predicting likely cancer outcomes, responses to certain therapies, and other important clinical events that can help even more patients properly manage these different cancers and improve their lives.
For the benefit of our readers, can you briefly let us know why you are an authority about the topic of Cancer?
I trained in diagnostic pathology at Yale and was privileged to work as a fellow with a pathologist who is widely regarded as the primary authority on cancer diagnostics. We both went to Memorial Sloan Kettering in 1991 — one of the top cancer centers in the world that has an enormous volume of cancer cases to learn from — where I worked as a pathologist for 30 years. I served as the director of the training program for 13 years, the chief of surgical pathology, which is the largest division of that department, and then the chairman of the department for the last 10 years that I was there. I had the opportunity to work on the front lines, as well as to lead the different divisions of the department and then ultimately the whole department.
While I was the chair, it was a time of incorporation of major new technology into pathology. We developed new techniques for sequencing the genetics of cancer and actually developed the first genetic test of its kind that was approved by the FDA. We also introduced the digital slide scanning that I mentioned earlier. And because of the availability of so many digital images, we were able to create a computational pathology program.
Ok, thank you for all of that. Let’s now shift to the main focus of our interview. Let’s start with some basic definitions so that we are all on the same page. What is exactly cancer?
What causes cancer?
There are so many kinds of cancer. Even though we think of cancer as one disease, it’s really hundreds of diseases, and many of them have very distinct and recognized causes. So, something as simple as smoking is associated with many kinds of cancer. It is obviously associated with lung cancer, but it can also be associated with bladder cancer, pancreas cancer, stomach cancer, and many others since there are direct carcinogenic agents in cigarette smoke. There are also a range of cancers that are caused by viruses that that get into the human DNA and activate genes within the human DNA that cause cancer. We know that melanoma is caused by mutations from UV exposure. There are also hereditary syndromes where you inherit a gene from your family that malfunctions and can ultimately lead to cancer. So, there are many different answers, and each is specific to a different type of cancer.
What is the difference between the different forms of cancer?
Cancer is many different diseases. Even if you just look at lung cancer, there are five major types of lung cancer, and a couple dozen minor types of lung cancer, each of which is distinct, both in terms of how it looks to a pathologist, how it behaves, its symptoms, and how it responds to therapy. We have cured a number of different cancers, but we have hundreds left to go. Part of the problem is, since it’s such a diverse disease, there is no one etiology and no one cure.
I know that the next few questions are huge topics, but we’d love to hear your thoughts regardless. How can cancer be prevented and how can one detect the main forms of cancer?
The biggest challenge with cancer is often that it remains silent for many years before it’s detected. And the successful treatment of cancer is very much dependent on when you catch it. For certain kinds of cancer, there are screening tests that have been developed. If you take Pap smears for instance, you can take a completely asymptomatic person and do a relatively simple procedure that would detect cancer in its earliest stages. Even before there was a vaccine against the virus that causes cervical cancer, the prevalence of that cancer went down because of screening. Mammography is another somewhat more complicated and controversial area, but mammographic screening has been shown to be effective in finding breast cancers at an earlier stage. With colonoscopy, you can also find colon cancers at a stage when they’re just a polyp. Once again, for every different type of cancer, there’s a different potential screening case. The newest thing now is to look in the blood because cancers release some of their DNA, and it circulates in your blood. If that cancer has a very characteristic mutation in the DNA, we can detect those minute quantities of abnormal DNA from a blood test, which offers many possibilities. While it’s not entirely a reality yet, it looks quite promising. However, waiting for symptoms to show up is not a great way to detect cancer, because by the time you start to realize that something’s wrong, it may be later in the stage of the disease.
Cancer used to almost be a death sentence, but it seems that it has changed today. What are the odds of surviving cancer today?
It entirely depends on the kind of cancer. There are some cancers where the odds are almost 100 percent that you will not die of that cancer, and we can sometimes make that prediction based on pathology. Prostate cancer is a good example. People certainly do die from prostate cancer, but many of the cases we’re diagnosing today are not likely to impact the life expectancy of the patient because we can tell that they’re very indolent. Thyroid cancers and some germ cell tumors can often be cured, and a number of lymphomas can be cured. But that still leaves hundreds that don’t fall into those categories. What’s happening with the revolution in precision medicine is that we’re able to identify what are known as “driver mutations,” key genetic changes that really drive cancers. And if you have a drug that can neutralize that specific genetic change, even if you can’t cure the cancer, you can stop the cancer. What is happening now is people who might have had an average survival span of 6 to 8 months are living for years. While the cancer isn’t gone, they’re able to keep it under control with these very targeted agents, which requires recognizing features of the cancer, usually at the genetic level, that a drug will target. The interesting thing is that through Paige.AI, we’re learning new things from digital pathology that we couldn’t recognize before. One of those things is predicting these genetic changes directly from the routine pathology slides, without genetic sequencing, which is expensive and takes a long time. Now, using AI and machine learning, we can — with good likelihood — instantly predict certain important key mutations.
Can you share some of the new cutting-edge treatments for cancer that have recently emerged? What new cancer treatment innovations are you most excited to see come to fruition in the near future?
Another form of treatment is immunotherapy, which is essentially taking advantage of the patient’s own immune system to reject the cancer. We’ve learned how to interfere when cancer prevents the immune system from attacking it. There are a group of drugs that can stimulate the immune system to recognize the cancer and kill it. While using the immune system to fight cancer is not a new idea, it’s only been in the last 10 years or so that effective drugs have been developed. So now there’s many different immuno-oncology agents that work through this mechanism, which can also be combined with other types of therapy for a double whammy that goes after the cancer.
Healing usually takes place between doctor visits. What have you found to be most beneficial to assist a patient to heal?
Healing is a multifactorial process; it’s not entirely related to the medications or the surgery, but has a lot to do with the patient, their support system, their beliefs, and their nutritional status. Helping patients access better diagnosis and treatment options will also ultimately improve the healing process.
From your experience, what are a few of the best ways to support a loved one, friend, or colleague who is impacted by cancer?
Having had multiple family members impacted by cancer, I can speak very personally about how to offer support. If you are a physician, especially a cancer physician, and you’re trying to help a friend or family member impacted by cancer, it’s quite a responsibility. Since there are so many different types of cancer, and because medicine has become so specialized, unless your friend or family member happens to have exactly what you are most knowledgeable about, the medical support you provide is never as important as the encouragement you can provide.
Most important is ensuring patients can get the treatment they need and can get to their appointments. Support also comes through understanding. For example, recognizing that chemotherapy affects how a person thinks, and can cause confusion. So, patience and empathy around what they’re going through, even if they don’t have overt physical signs of illness, is key. The psychological impact of having cancer, especially advanced cancer, is profound.
What are a few of the biggest misconceptions and myths out there about fighting cancer that you would like to dispel?
The number one misconception is that cancer is one disease. Unfortunately, the view of it as a singular disease is harmful, since there are some types that are curable and other types that are not, and the different kinds of treatments needed for each type can range broadly.
Thank you so much for all of that. Here is the main question of our interview. Based on your experiences and knowledge, what are your “5 Things Everyone Needs to Know About Cancer? Please share a story or example for each.
- There are many different kinds of cancers. In fact, there are more than 100 different kinds of cancer, and many don’t have symptoms in their early stages.
- Adhere to routine cancer screening tests. If detected early, cancer tends to be much easier to treat and patients have better outcomes. This is especially true for breast, cervical and colorectal cancers, which can often be cured in the early stages.
- Cancer has a myriad of causes. Many we still don’t know or understand. Things like your diet, tobacco, and alcohol consumption, as well as infections, chronic inflammation, environmental pollutants, and genetics all can play a role.
- Pathology is an interpretive science. When a professional with lots of experience looks at the pattern reflected in tissue samples that we create, and determines that this pattern means its cancer, it’s not like running a blood test. It’s often not an absolute yes or no. It’s an interpretation that requires enormous skill and experience.
- Digitization is opening many new opportunities. Digitization helps pathologists work more efficiently. It removes pain points, facilitates sharing cases for second opinions, and enables the application of AI to support diagnosis. AI provides pathologists greater peace of mind, which allows them to focus on what’s most important in their diagnosis. Furthermore, it’s helping to democratize better patient care in communities and countries around the world without access to the mostly experienced, super-specialized pathologists.
You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. 🙂
I think we’re in an environment now where knowledge sharing for the benefit of patients is a challenge, and it has to be prioritized. It’s a challenge in the academic sector because people are trying to build their careers and get promoted. Having unique knowledge and data can build a credential for an academician, but it doesn’t always necessarily serve the best interests of the patient. In the commercial sector, there are even stronger reasons not to share data, to avoid losing a competitive advantage. A big question for us to answer is how we can democratize access to information, so that whatever discoveries are made can most rapidly move into patient care. It’s a very challenging concept. And so, as a movement, it would require us to completely rethink how we reward academic physicians, and how we commercialize discoveries. If there were a way to change that, so that there was a national effort to put all our resources into one pot, I think we could make much more rapid progress.
How can our readers further follow your work online?
Readers can follow my work through Paige. People can also access my published papers on https://pubmed.ncbi.nlm.nih.gov/?term=klimstra+d&sort=date.
Thank you so much for these insights! This was very inspirational and we wish you continued success in your great work.