Quantification and Optimization of Standard-of-Care Therapy to Delay the Emergence of Resistant Bone Metastatic Prostate Cancer
Arturo Araujo, Leah M. Cook, Jeremy S. Frieling, Winston Tan, John A. Copland II, Manish Kohli, Shilpa Gupta, Jasreman Dhillon, Julio Pow-Sang, Conor C. Lynch, David Basanta
Read the paper
A few weeks ago, our team got our work entitled "Optimizing the standard of care to delay the emergence of resistance in bone metastatic prostate cancer
," published in the journal Cancers. You can find the article online here
I would love to spend the rest of this (hopefully not too long) post describing all the cool things we did but there will be a different post for that, this one is an ode to team science.
So much about mathematical and computational oncology relies on team science that it is easy to forget how important the interactions between diverse cancer researchers are. A lot of posts in this blog are testament to interdisciplinary teamwork but often we focus on the results rather than the process.
Today’s story began at the Moffitt Cancer Center in 2012 where the Integrated Mathematical Oncology department (thanks Sandy
!) was hosting the 2nd ever IMO workshop
. The idea of these hackathons
is to create teams that include clinical, experimental and mathematical researchers. The ability of the teams to tackle an assigned problem depends on the skills of the different team members but, importantly, in their dedication to the workshop (not always easy for busy clinicians). I am happy to report that our team included plenty of talent: mathematical modelers (of course! Jake Scott and Jill Gallaher), cancer biologists (Leah Cook and Conor Lynch), an epidemiologist (Jong Park), an oncologist (Shilpa Gupta), a pathologist (Jasreman Dhillon) and a surgeon (and chair of GU oncology, Julio PowSang). I am even happier to say that all these people *really* committed to the project in our workshop centered on improving treatment strategies for patients with metastatic castrate resistant prostate cancer.
Any person that has ever attended (or witnessed) an IMO workshop knows two things: that this is an intense 4-5 days and that a lot can be done in a short period of time when minds are focused. The pictures that are shown about this paragraph hopefully reflect this. But as they say (or not) by the time we got 90% of the work done, we only had another 90% of the work to do! And that other half of the work needs to be done when people are not in the same room and completely focused on the project, which is why it took another year to finally have something worth sending to peer review
. This work suggested that it might be possible to use retrospective patient data to inform a mathematical model that assumes a certain degree of intra-tumor heterogeneity in regards key mutations (in the context of metastatic prostate cancer we deemed JAK/STAT, PTEN and AR to be the key ones) to make predictions about the impact of treatments in the standard of care.
During this time Arturo Araujo
joined the project and together with collaborators at the Mayo Clinic (hello Manish, now at Utah as well as Winston Tan and John Copland) did what one does after a workshop (plus a year) where you feel you have done 90% of the work required: start all over again!
With the lessons learned during the workshop and the data gathered thanks to the pilot funding from the IMO department, Arturo decided to rethink our approach to intra tumor heterogeneity. If we are trying to figure out how treatments impact a heterogeneous tumor, does it not make sense to look at this heterogeneity from the perspective of available treatments? With that idea in mind, we created cells lines, using samples from patients from the Mayo Clinic, that are either sensitive, resistant to one of the treatments in the standard of care or resistant to combinations of treatments.
After deriving parameters for the model, we could now begin to see if it could predict how a patient would do under treatment and use retrospective data from patients from the Moffitt Cancer Center (PSA levels and treatment history) to compare predictions with reality. This is important since we are trying to work from first principles: the patient data was not used for data fitting.
These are still early times in the use of mathematical models, parameterized with patient data, as proxies for patients to explore novel. Math modelers, our team included, have been making inroads to bring some basic biology research into clinical use but what I presented here can be better understood as proof of principle. Our next goal is to try this on pre-clinical models (AKA mouse models), only then will we be able to bring this to the hospital. Luckily, and with the help of the NCI PSON
we can now explore that.
These results, being important to us, are not the reason I am writing these lines. What I want is to illustrate that team science’s goal of bringing discoveries from the bench to the bedside (or from the blackboard to the bedside as my colleagues at IMO advocate) is necessarily two-way ... one that involves more than an exchange of ideas and requires people to be in the same (maybe zoom) room.