Pushing Boundaries, Inciting Innovation
Imagine a floor plan for a scientific research building where you have the offices of a prolific grant funded Department Chair of Diagnostic Radiology and Chair of Cancer Imaging and Metabolism next to a team of creative mathematicians and graduate students. Envision a “collaboratorium” where couches, a blackboard, and an espresso machine invite ad hoc scientific discussion throughout the work day. Conjure up a lab where the way cancer cells grow and metastasize can be modeled on a computer. All this and you have just seen only the first ripple in the vast pond of Integrated Mathematical Oncology (IMO) at Moffitt Cancer Center.
The chief architect of the IMO has been Dr. Robert Gatenby. Dr. Gatenby left the University of Arizona and recruited many of his colleagues, such as Robert Gillies, PhD, to build a new scientific universe at Moffitt that brings these diverse specialists together under one roof. When “the Bobs” came to Tampa, the senior leadership at Moffitt quickly realized that “thinking outside the box” did not apply to a team that did not even see a box to transcend. Indeed, as the Chair of the IMO, Alexander R.A. “Sandy” Anderson, PhD, notes, “Cancer is a dynamic complex multi-scale system that can only truly be understood via the integration of theory and experiments. The goal of the IMO is to use such an integrated approach to better understand, predict, and treat cancer.”
For an NCI-designated comprehensive cancer center whose mission is the prevention and cure of cancer, the strategic investment in the IMO may seem like an unusual choice. Dr. Gatenby agrees, noting that there are few programs of its kind in the world. Yet what may seem non-intuitive at first glance gains a payload of traction as you dig beneath the surface. Consider the “immortality” of cancer cells: what promotes the evolutionary mechanisms underlying these adaptations? What happens at the level of the tumor microenvironment surrounding the neoplastic cells? Can such growth patterns be modeled prospectively by incorporating data from “super” computers? And, perhaps most importantly, are these models true to life in vivo? Moreover, can they be prospectively predictive and potentially modified to maximize treatment effect?
Yes, the “aha” moment now resonates. Moffitt Cancer Center has long been a national leader in the concept of individualized cancer care as espoused for the last decade by former cancer center CEO and current M2Gen CEO William Dalton, MD, PhD. Dr. Dalton and his colleagues recruited “the Bobs”, in fact, to explore new scientific frontiers at Moffitt with bold pursuit of the personalized cancer care dream.
Progress to date has been steady. Thanks to Dr. Anderson’s selective recruitment and the lure of the IMO’s unique environment of open collaboration, the IMO now houses a robust faculty as well as a large research lab where tumors can be modeled, imaged, and treated in animals. In addition, the IMO has a formal relationship with the Mathematical Biology Program at Oxford University dedicated to the training of PhD candidates in mathematical oncology, such as Dr. Jacob Scott.
Those of you who watched the 2012 TedMed talks may recall his standing ovation lecture. You may have seen him come to the stage following prominent healthcare experts such as the national director of the Centers for Disease Control and the Director of the National Institute of Health. You may have wondered how a young astrophysicist and former nuclear sub engineer, now training in radiation oncology at the USF/Moffitt program, became so interested in mathematical oncology that he extended his residency to simultaneously obtain a PhD in the combined Tampa/Oxford Mathematical Biology program and how his work earned him a TedMed invitation. To know Dr. Scott, however, is to see the immediate answer: his passion is to find ways to inject imagination into medicine to cure cancer.
As Dr. Scott notes, “In medicine, we memorize, we recite…. that is not what we need. We need to collaborate, innovate, and develop new pipelines of thought that cross pollinate the best ideas from different disciplines. We have too many dots in biological science that don’t get connected. That is what the IMO is all about. That is what kind of physician I want to be….. A dot connector in the purest sense.”
For example, Dr. Scott became interested in the lack of a coherent understanding of the mechanisms of cancer metastasis despite the detailed knowledge many biologists had of certain aspects of the metastatic process. Frustrated with the lack of overwhelming progress in changing the death sentence of metastasis, Scott joined forces with Alexander R.A. Anderson, PhD. Together with Peter Kuhn, PhD of the Scripps Research Institute Physical Sciences-Oncology Center they developed a unifying theory on the causes of cancer metastasis that was published in a May 24, 2012 Nature Reviews Cancer article.
Their article explores the role of the circulating tumor cells (CTCs) that are in the blood vessels of advanced cancer patients in the development of metastatic disease. They note that although both the site of a metastasis (the soil) and the metastatic cells (the seeds) are needed for cancer dissemination to occur, just how the seeds seek specific soil is not clear.
Dr. Scott thinks that it is indeed just a dilemma like this that mathematical oncology can help clarify. As he noted in the Journal of the National Cancer Institute (JNCI), “we think this process is governed by solvable physical rules that relate to the dynamics of the circulatory flow between different organs and how these organs filter. Although these biological mechanisms are not yet known, we might be able to infer their existence by finding out which measurements do not fit a model that is defined only by physical flow and filtration.”
The IMO has been very successful in grant funding for pioneering research such as this. In 2011, Dr. Anderson and colleagues were awarded a 5 year $3 million dollar grant to model the aggressiveness of prostate cancer. The problem of metastatic prostate cancer is ripe for mathematical modeling: who are those patients that respond to systemic therapy? Some respond transiently and some don’t respond at all. How can models be generated to predict such tumor behavior and guide the clinician to give therapy that maximizes the patient’s time to progression?
Work such as this is ongoing in the IMO and getting closer to clinical reality according to Dr. Gatenby.
“We’re fairly close to developing computational models for every patient’s individual cancer that takes advantage of all their personal data at the genomic, cellular, clinical, demographic and imaging level. The conceptual model is then like a hurricane….. We need models to predict the path of a hurricane since any complicated system cannot be understood intuitively. So, if we view the cancer like a hurricane, we can then determine the various paths that are possible. What if we did not do anything, then based on where the cancer ‘has been’ and where it ‘is now’ scenario X is likely. Alternatively, if we perturb it with various therapies, then the model will predict scenario Y. And if we do perturb it with these therapies, the model would need to be continually updated as we give the therapy so we can get real time information to readjust it’s course.”
Dr. Gatenby explains further, “we are making significant progress in multiple myeloma at Moffitt. In fact, for myeloma, we have developed computational models for individual tumors based on data from bone marrow biopsies, aspirated cells, clinical data, and previous therapies. The goal here is to help guide the clinician with reliable information for the individual patient about the specific drug and dose, as well the timing and duration of administration. The dream is for the patient to come in to Moffitt Cancer Center and a computer model integrating all of their clinical and pathological information can be generated before they go home.”
Drs. Gatenby and Gillies have also infused cutting edge analysis of imaging for response identification. Many patients often find this aspect of care the most frustrating. They visit their physician, have a scan, and wait for the dreaded report. Often, disappointment ensues, with ample uncertainty of what these imaging findings “mean” with respect to their individual prognosis.
This could all change, according to “the Bobs.” Both have been exploring ways of extracting data from the patient’s imaging that can be further characterized and studied to go way beyond the traditional report. Some of this work has been done in collaboration with the MAASTRO center in Europe. The results are yielding new ways of interpreting patient data that can be incorporated into these predictive models to help the individual patient truly gain valuable insight into the course of their disease and prognosis.
Dr. Gatenby reports some early progress in glioblastoma, the lethal brain tumor that Ted Kennedy succumbed to a few years ago.
“We can view cancer not as a self-organized organ like system but as a coalition of habitats each with different blood flow and other adaptive characteristics. Each habitat has a unique pattern of responsiveness to different therapies. In glioblastoma, we can gain a lot of information from the patient’s magnetic resonance imaging (MRI) scan. We have found that there are typically five habitats that we see consistently. We have analyzed this and found that the higher the number of low vascular habitats, the worse the prognosis can be. Thus, by integrating additional features from the patient’s MRI scan, we can gain individual prognostic information that may be helpful for stratifying patients so that we know which patient would benefit from which individual therapy.”
Given that imaging data is like the human brain, with vast areas housing untapped information, how might the radiology program of the future appear? If Dr. Gatenby could wave his magic wand, he would ensure that novel molecular agents exploiting different processes of the cancer cell and its microenvironment be developed so that we could maximize non-invasive prognostic information merely by imaging the patient. For example, there is the potential of nanoparticles to be incorporated as imaging contrast agents. Work has been reported from Moffitt Cancer Center and University of Florida collaborators on the promise of nanotechnology for solving clinical problems in breast cancer. Consider the current difficulty discriminating between benign and malignant lesions within the breast. The current standard of care after careful imaging is typically to proceed with a biopsy if there is radiographic concern for cancer. The future may include using these nanoparticles as part of breast imaging to non-invasively determine whether the lesion is malignant. If cancer is diagnosed, nanotechnology may play a role in sentinel node evaluation with possible elimination of the need for axillary surgery in many patients. Validation of innovative agents such as these would change patient care dramatically.
Ultimately, the ideal radiology reading room would be then be transformed. Imagine that the patient’s scan is up on the viewing station. The radiologist clicks on the image, sends it to a partner vendor who analyzes the data to look at the habitat of the tumor, generate a relevant metric about the status of the disease, and send the information back in real time so that the doctor in the clinic can use the information to guide patient management. The clinician wants to know if the patient is responding, and if not, wants the earliest possible data point to know a “non-responder” so that therapy can be modified immediately.
Dr. Anderson agrees that “Imaging is certainly a crucial part of the toolset that the IMO needs in order to validate and parameterize its models but another important piece of this puzzle is basic science experiments, whether they are in the dish or in the mouse; they can provide critical insight as to the realism of the models and allow them to be tuned and/or modified if they don’t fit with the biological reality. Perhaps the ultimate contribution of the IMO will be to provide a new way to look at an old problem, to give our collaborators a set of mathematical glasses that allow them to see their own research in a completely new light. It’s common that experimentalists develop their own mental map of how different components of a system interact and drive specific outcomes (such as treatment failure), however, this systemic view was derived from many experiments that considered each interaction in isolation. Mathematical models allow us to place the complete system in an integrated framework where we can directly examine the impact that changing one component (or multiple) has on the whole system. This often drives new experiments that ultimately lead to a deeper understanding of the system. This iterative dialogue between theory and experiment is why the I in IMO stands for integration: over biological scales, across disciplines, within a cancer center.
With the IMO, such goals appear within reach. After all, this diverse and interdisciplinary group is already breaching borders and unifying mathematics, computer science, imaging, clinical science, radiological physics and experimental biology. The scientific body of work that is emerging spans multiple specialty fields with a common goal that has been inspiring integration. Moffitt’s NCI designated comprehensive cancer center is modeling the cure to cancer…. Now, as Dr. Scott would say, we just have to connect the dots.
MOFFITT CANCER CENTER 12902 Magnolia Drive, Tampa, FL MOFFITT CANCER CENTER AT
INTERNATIONAL PLAZA 4101 Jim Walter Boulevard, Tampa, FL 1-888-MOFFITT (1-888-663-3488) REFERRING PHYSICIAN LINE (813) 745-3980 or 1-888-860-2778 www.moffitt.org
By Sarah E. Hoffe, MD