Sohrab Shah, PhD, joined Memorial Sloan Kettering Cancer Center earlier this year as the cancer center's first-ever Chief of Computational Oncology in the Department of Epidemiology and Biostatistics. His charge is to marry big data and computational resources with biology to further MSK's mission of providing and improving cancer care.
"We cannot overstate the significance of computational oncology to the future of cancer research or our great fortune in recruiting someone with Dr. Shah's unparalleled expertise and imagination," noted Jose Baselga, MD, PhD, Physician-in-Chief at MSK. "We are thrilled to welcome him to New York City and look forward to a future filled with exciting new developments."
Before this role at MSK, Shah was Associate Professor in the Department of Pathology and Laboratory Medicine and a senior scientist in the Department of Molecular Oncology with British Columbia Cancer at the University of British Columbia (UBC). He was also an associate member in the Department of Computer Science at UBC and at the Genome Sciences Centre at BC Cancer.
Shah told Oncology Times his background in both biology and computational modeling are what uniquely prepare him for this new role at MSK. Here's what he said about his role, the potential in the field of computational oncology, and what's first on his to-do list in the new position.
1 You're in a brand-new role of Chief of Computational Oncology. What does that mean?
"The role is really to lead and develop a new research initiative in computational oncology. And you could define computational oncology as a coming together of computer science and data that is focused entirely on the cancer problem. That can range from research topics where we're trying to measure properties of a cancer at the molecular or genomic level all the way through to patient data that is generated in that context of patient care.
"In the latter example, you could think of imaging data from a cross-sectional PET scan or digital pathology information or genomic info that can inform treatments.
"The goal is really that in the context of patient care these data are being generated to help treat the [individual] patient, but an aggregate of thousands of patients may become an incredible resource using state-of-the-art computational techniques to learn how we can better treat patients from those measurements and make better predictions as to what will happen to a patient after diagnosis and throughout their clinical trajectory.
"In both diagnostics and in research, the dimensionality data or the number of measurements taken per patient or per sample is increasing at an exponential rate. The only way to synthesize the information is computationally. And all machines will generate data with some degree of error-and the nature of error profiles is complicated. [Using] tried and true statistical techniques that employ leading edge and methods, we try to extract the most relevant signal from a potentially noisy measurement.
"This signaling-processing problem exists in many domains-cancer's not unique to that. But what's interesting about cancer in this day and age is that we really have this urgent need now to build to leverage these measurements that are taken in a new way.
"The computing power required for these types of approaches and also the amount of input data required to train those approaches properly-it's really not been possible before. So now is really the time where this type of approach could really transform cancer care and cancer research."
2 So, what's first on your to-do list in this new job?
"The first objective is to build a critical mass of computational scientists to address these problems in a computational research environment. So we have a faculty recruitment under way. And we have a post-doctoral recruitment under way.
"[Cancer] is a monumental societal problem. So my first real goal is to try to attract people who would otherwise go and work with tech companies or elsewhere-and stimulate and inspire them to work on these problems in cancer. Because there's similar computational challenges in cancer [as there are in other fields], but in oncology the impact is much more in health care and helping this societal problem.
"So that's really the first task is to build out our team.
"It's beneficial if people have training in both computational domains and an oncology or biology focus. But that's rare. We already have an incredible environment at MSK whereby clinicians, clinician scientists, and basic scientists can provide all the problem spaces and frame the problems very very well. Where we need to build is in the people who are doing leading, cutting-edge computational work. So ultimately this work becomes a team science. It's a collaborative endeavor where multidisciplinary people are required."
3 What would you say is most important for anyone in cancer care to know about this new role at MSK and how it might change oncology research and practice?
"The main message is that we need to drag cancer diagnostics and prognostics into the future by taking advantage of the datasets that exist and develop in the context of patient care. And the field will likely change and can support clinicians in new and more effective ways using these advanced computational techniques.
"And that's really what our goal is-to support clinicians in a much more effective way so they can treat their patients more effectively.
"I think embracing the leading-edge computational techniques to advance precision cancer care is our objective, and we look forward to working with the clinicians to help accomplish that goal."