Collaborative Cancer Cloud demonstrates feasibility of data-sharing model, transitioning proof of concept to projects aimed at technology deployment

Groundwork is laid for further development; tools released open source and will be further developed by OHSU through national consortium

OHSU, Intel and its partners have successfully validated the Collaborative Cancer Cloud platform architecture, finalizing their work on the project.

The Collaborative Cancer Cloud is an advanced, federated analytics platform that enables clinicians and researchers to securely study large data sets across multiple institutions for potentially lifesaving discoveries and treatments. The project launched in 2015 and expanded in 2016 with the establishment of additional partners Dana Farber Cancer Institute and Ontario Institute for Cancer Research.

By completing the proof of concept, the organizations demonstrated the viability of securely sharing genomic information on a massive scale using a distributed architecture that protects privacy and local, institutional control of data. The model system made it possible for disparate research and medical institutions to connect over a secure network to conduct analyses of pooled data without having to download it.

Work done on the CCC lays the groundwork for other organizations to develop it further as part of an integrated infrastructure of tools and systems required to support scalable, widespread use in precision medicine.

“This has been an extremely fruitful project with Intel,” said Mary Stenzel-Poore, Ph.D., who as senior associate dean for research in the OHSU School of Medicine was instrumental in cementing the partnership with Intel. “Now that we have successfully proven the concept of a secure, federated platform to aggregate and analyze large data sets, the stage is set for the precision medicine ecosystem to take forward the concepts demonstrated by the Collaborative Cancer Cloud. By publicly releasing the software developed for the project we have opened up the platform to the entire community.”

Tools and expertise developed in the collaboration are being put to use at OHSU. A high-profile example is the $15 million Kids First Data Resource Center, a National Institutes of Health-funded effort to build the world’s largest database on pediatric cancer and birth defects. Adam Margolin, Ph.D., professor of biomedical engineering, and director of computational biology in the OHSU School of Medicine and the OHSU Knight Cancer Institute, is co-principal investigator. Dr. Margolin and collaborators from the Ontario Institute for Cancer Research – one of the CCC partners – decided to form a consortium to compete for the pediatric project, based on their work together on CCC.

“Our work with Intel has helped us become a major player in a national consortium for pediatric cancer,” said Dr. Margolin. “It’s also been a crucial building block for the OHSU School of Medicine’s Computational Biology Program. Through our work together, we’ve gained valuable experience and developed impactful software tools enabling large-scale data integration and data management. This experience is already being applied to other research initiatives underway on campus.”

Much of the software developed for the project has been released on GitHub, the popular web repository for hosting open-source software projects.

In the months ahead, researchers who worked on the project expect to report their scientific findings in peer-reviewed journals. One study, for example, used the data-sharing platform to search for mutation hotspots among thousands of breast cancer genomes available from public data sources as well as from OHSU, Dana-Farber Cancer Institute and the Ontario Institute for Cancer Research, constituting the largest breast cancer genomics study performed to date.

Additionally, a number of shared tools have been developed as a result of the CCC collaboration.

  • Genomics kernel library (GKL): a collection of kernels to speedup compression and decompression of genomics data, open sourced directly by Intel and used in The Broad Institute’s GATK and Illumina’s Isaac.
  • GenomicsDB: a scalable database capable of supporting hundreds of thousands of genomes and open sourced directly by Intel.
  • Graphical user interface (GUI): the interactive graphical mode for users to submit and monitor workflows will be open sourced by Intel and is used in the Broad-Intel Genomics Stack.
  • Task execution system (TES): an API permitting deployment of computational tasks across multiple cloud environments, adopted by the Global Alliance for Genomics and Health (GA4GH), The Broad Institute, Google, and Seven Bridges Genomics, among others.

The CCC also established the technical capability required for Dana-Farber, Ontario Institute and OHSU to continue sharing aggregate data among themselves in order to conduct analyses as part of any research projects that are jointly defined and undertaken together in the future. And Intel and OHSU continue to collaborate on other projects as part of a shared goal to accelerate scientific progress in understanding complex diseases.

Beyond the CCC collaboration, the strong relationship with Intel has resulted in additional benefits to OHSU’s research community. The ExaCloud, for instance, at OHSU’s West Campus Data Center is a resource developed in close collaboration with Intel to support very large scale computation. The primary ExaCloud cluster includes over 6,600 of Intel’s Xeon microprocessors. And Intel experts are getting ready to work with OHSU faculty to advance a new MRI technology developed at OHSU.

“The CCC is a strong, successful beginning,” said Dr. Stenzel-Poore, who is now chief of research operations in the OHSU Knight Cancer Institute. “OHSU will continue to pursue partnerships and opportunities in the data sharing and data management space as part of our ongoing efforts in precision medicine. By joining forces with health care and industry partners, we will move the vision forward of ‘All in One Day’ cancer care.”

Related Reading 


NIH selects OHSU Knight Cancer Institute to co-lead largest national database for pediatric cancer, rare disease research

The "Kids First” grant caps a momentous time for the institute, which recently garnered three prestigious grant awards totaling more than $17 million in funding to OHSU

Adam Margolin, Ph.D.
Adam Margolin, Ph.D., professor of biomedical engineering, and director of computational biology in the OHSU School of Medicine and the OHSU Knight Cancer Institute. (OHSU/Kristyna Wentz-Graff)

The National Institutes of Health announced today that it is establishing a pediatric cancer and rare disease data program that will allow clinicians and researchers to access multiple datasets in one location -- a first for the pediatric research community.

The NIH Health Common Fund's Gabriella Miller Kids First Pediatric Research Program is awarding a $15 million grant to establish the Kids First Data Resource Center. The five-year award, contingent on available funding, will allow scientists to harness genomic data from thousands of pediatric patients to collaboratively engage in shared research on behalf of children diagnosed with cancer and other diseases.

"There is an unprecedented amount of genomic data available, and these data hold the promise of yielding breakthroughs for debilitating pediatric diseases. The idea behind the DRC is to unlock this potential by allowing researchers and clinicians across the world to easily apply advanced analytics to as much pediatric cancer and rare disease data as possible,” said Adam Margolin, Ph.D., co-principal investigator on the project, professor of biomedical engineering, and director of computational biology in the OHSU School of Medicine and the OHSU Knight Cancer Institute.

OHSU will collaborate with investigators from the Center for Data Driven Discovery in Biomedicine (D3b) at the Children's Hospital of Philadelphia, biomedical data analysis firm Seven Bridges, the Ontario Institute for Cancer Research, the University of Chicago, and Children's National Health System.

The Data Resource Center will provide new analytic tools and access to the largest cross-disease collection of annotated clinical and genomic sequence data from children with cancer and structural birth defect and their families. At the inception of the project, the team will immediately work to make available genomic sequencing data, and associated tools, from 6,000 pediatric patient samples.

By enabling researchers to find and perform large-scale analyses across pediatric cohorts, the DRC will act as a centralized hub from which researchers and clinicians can access phenotypic and genetic sequence data for novel research. Other shared data resources exist in pockets across the country, but the DRC is the largest culling of that information in one resource, according to Margolin.

"NICHD is committed to supporting research on birth defects as part of its overall focus on improving the health of children,” said Diana Bianchi, M.D., NICHD director. "The Kids First initiative provides a unique opportunity to use DNA sequencing information to gain a better understanding of the underlying causes of birth defects and childhood cancers.”

Momentum for OHSU's computational biology program
Being tapped to co-lead an endeavor of this magnitude highlights a momentous few months for the OHSU computational biology program, and continues to fulfill a promise made to donors during the $1 billion fundraising challenge.

"One initiative we promised was the strengthening of computational biology,” said Brian Druker, M.D., director of the Knight Cancer Institute. "Adam has recruited seven faculty, a total of 50 people, and has brought in more than $20 million in grants funding to build a computational program that is bringing nationwide recognition to OHSU. We recognized that this resource would be a necessary part of our overall effort to end cancer as we know it, and we are extremely pleased to see this remarkable return on our investment.”

Over the last three years, the computational biology program has been awarded more than 30 grants, with computational biology personnel as co-principal investigators, totaling more than $22 million to computational biology, $41 million to OHSU, and $90 million for multi-institutional grants.

In June and July, OHSU was selected to join three prestigious national consortia for computational cancer biology research. Margolin will serve as the computational lead and co-principal investigator on the Kids First Award as well as:

  • $9.2 million National Cancer Institute grant to study triple-negative breast cancer as a Research Center in the NCI's Cancer Systems Biology Consortium. Co-investigators include: Joe Gray, Ph.D., Rosalie Sears, Ph.D., and Claire Tomlin, Ph.D.
  • A $5.8 million NCI grant to study drug response and resistance in acute myeloid and chronic lymphocytic leukemias as part of the NCI's Cancer Target Discovery and Development Network. Co-investigators include Druker and Jeff Tyner, Ph.D.

This year, OHSU also was awarded a $2.1 million grant to serve as the center for RNA sequencing and pathway-based analysis within the NCI's Genome Data Analysis Network. Investigators are Margolin and Paul Spellman, Ph.D.

In addition, the OHSU Knight Cancer Institute recently was awarded the National Cancer Institute's highest distinction as a Comprehensive Cancer Center. By joining four prominent national computational consortia, the Knight Cancer Institute should be well positioned to include leadership in computational biology within its slate of comprehensive clinical and research activities to benefit cancer patients.


Three questions for Young Chang

Young Hwan Chang, Ph.D., is assistant professor of biomedical engineering, Computational Biology Program, OHSU School of Medicine.

What’s been the most interesting development in your area in the last two years? 

Our lab is interested in computational and interdisciplinary approaches to deliver scientific advances and clinical innovation for studying cancer. One of our main focuses is to develop quantitative image analysis tools and build spatial systems biology models in order to understand how cancer cells adapt to their microenvironment (ME) and elucidate how tumor cell-ME interactions influence tumor cell physiology and response to therapy. Over the past two years, machine learning (ML) and deep learning (DL) have literally invaded our daily lives as well as many research areas such as computer vision, autonomous vehicles, robotics and computational biology. We expect these methods to enable further characterization of the complex datasets and advance in our understanding of the spatial relationships between immune cells, stromal cells, cancer cells and their microenvironment.

What projects are you currently working on and are there opportunities for fellow faculty to participate? 

I am currently developing a general framework for analyzing multiplexed imaging, histopathological imaging or multiscale imaging to provide quantitative measures and architectural analysis of proteins, cells and tissues. For example, we can interpret multi-nodal aspects of immune complexity and functionality in tumor biopsy material elucidated by highly multiplexed imaging techniques, and provide quantitative analysis of complex dissociated populations with geographical and spatial associations by using image analysis and bioinformatics tools.

I am eagerly looking to explore collaborations and work with other faculty who would be interested in quantitative image analysis for their research interests.

What is the most important aspect of support that OHSU provides to you currently and how would you like this or other support to grow in the future? 

The most important aspect of support that OHSU provides is the collaborative and team science environment. Since arriving about two years ago, I have been involved in very exciting projects, mostly collaborative efforts between biologists, basic scientists, pathologists and engineers like me. I am grateful for the opportunity to be a part of a great academic team and to work closely with many great colleagues. I think if we can keep this spirit of a great team in cancer research, we can prove that something can be done that many people said was impossible.


The first pilot funds for cancer early detection have been awarded

cancer collage

Five research teams at OHSU have been awarded Cancer Early Detection And Research seed grants. CEDAR, a new center in the Knight Cancer Institute, is working to understand the biology of the progression from normal tissue to pre-malignancy to cancer. The center is seeking to develop new diagnostic technologies, find ways discriminate lethal from non-lethal disease, and ultimately identify therapies to treat early tumors.

The awards, up to $120,000 each, aim to cultivate outstanding translational research, assist in the generation of preliminary data, and lead to national funding. Awardees include young scientists and senior investigators who assembled multi-disciplinary teams to explore new areas and ideas with high potential for impact in patient care:

Risk Stratification of Early Melanomas by Deep Histopathological Analysis

Noah Hornick, Erik Burlingame, Young H. Chang, Guillaume Thibault, Eric Smith, Tracy Pawlitschek, Kevin White, Sancy Leachman

Melanoma’s propensity to metastasize is what makes it the deadliest form of skin cancer. Determining the likelihood that a cancer will metastasize at the time of diagnosis is therefore central to the design of a treatment plan; more aggressive cancer needs more aggressive treatment. The research herein proposed will apply sophisticated machine learning tools to the analysis of standard biopsy slides used to diagnose melanoma, in an effort to find features that can be used to separate cancers that are likely to metastasize from those that are not. If successful, this technology could be easily and inexpensively applied to existing biopsy samples at diagnosis, improving treatment with minimal additional cost.

Target and biomarker identification through mapping of signaling networks that control the evolution of premalignant epithelial lesions

John L. Muschler, Emek Demir

We hypothesize that advanced mapping of the diverse regulatory pathways controlling the evolution of pre-malignant lesions will identify vulnerabilities in these lesions that can be targeted for early intervention, and identify biomarkers that can be used to assess risks in premalignant lesions. Preliminary data generated using powerful animal models of pre-malignancy, coupled to targeted gene modifications, have already identified several potent and novel regulatory pathways controlling the evolution of pre-cancerous lesions. We propose here to develop workflows for the efficient and robust mapping of the critical signaling networks controlling pre-malignant disease evolution, exploiting our preliminary results and unique animal models. This will be achieved by bringing together experts in cancer cell biology, animal modeling of cancers, and computer-based bioinformatic analyses of signaling networks built on gene activity and protein composition data, and computational network mapping.

Harnessing Circulating Hybrid Cell Biology and Ultrasensitive Single Cell Imaging Technology for Early Detection of Pancreatic Cancer

Young H. Chang, Summer Gibbs, Brett Sheppard, Tania Vu, Melissa Wong

Non-invasive assays that can identify cancer at the earliest possible stage represent the holy grail of the early detection quest. We have identified a novel circulating tumor cell – the circulating hybrid cell (CHC) – in the peripheral blood of cancer patients. CHCs hold high potential as a non-invasive biomarker to detect high risk early disease states and to monitor the progression of early stage to late stage pancreatic cancer. We will use a new molecular-sensitive single cell profiling methodology, along with protein functional tissue array studies, to establish a panel of CHC-based biomarkers that distinguishes early pancreatic disease from late stage cancer. Successful identification of CHC-based biomarkers will enable next step design of a CHC-based platform technology to monitor early to late stages of disease along the cancer continuum.


OHSU Knight Cancer Institute selected to join prestigious national consortium, receive $9.2 million

OHSU investigators will focus on improving therapeutic strategies for treatment-resistant triple negative breast cancer

Joe Gray, Ph.D.
Joe Gray, Ph.D.,(left) director of the OHSU Center for Spatial Systems Biomedicine (OCSSB) and the OHSU Knight Cancer Institute associate director for biophysical oncology, September 26, 2014. (OHSU/Chris Hornbecker)

The National Cancer Institute has awarded a research team at the OHSU Knight Cancer Institute $9.2 million over five years to serve as a Research Center in the NCI’s Cancer Systems Biology Consortium, or CSBC.

The OHSU Knight Cancer Institute is one of nine research institutions nationwide tapped to join the consortium, including Stanford University, Yale University, Massachusetts Institute of Technology, Memorial Sloan Kettering Cancer Center, among others.

Joe Gray, Ph.D.
Joe Gray, Ph.D. (OHSU/Chris Hornbecker)

The OHSU Knight Cancer Institute’s project aims to develop strategies for improving treatment-resistant triple negative breast cancer, an aggressive form of breast cancer that lacks key receptors known to fuel most breast cancers: estrogen receptors, progesterone receptors and human epidermal growth factor receptor 2 (HER2).

Using advanced microscopy, the team will leverage tools for quantitative analysis and visualization of images generated, together with computational approaches for integrating diverse molecular data types. Through analysis of core cell lines, patient-derived cultures and primary tumors, the team aims to uncover molecular networks that underlie disease progression and therapeutic response.

Joe Gray, Ph.D., director of the OHSU Center for Spatial Systems Biomedicine (OCSSB) and the OHSU Knight Cancer Institute associate director for biophysical oncology will lead the investigative team as a principal investigator.

“Triple negative breast cancer is a particularly difficult form of the disease to treat,” said Gray. “Our goals in the CSBC Research Center are to identify the mechanisms by which these cancers evolve and adapt to become resistant to treatment, and to develop new strategies to counter these mechanisms. Our multidisciplinary approach treats these cancers as adaptive systems that can be controlled using multiple drug combinations.”

Co-principal investigators on the project include: Rosalie Sears, Ph.D., professor of molecular and medical genetics in the OHSU School of Medicine and a senior member of the Knight Cancer Institute; Claire Tomlin, Ph.D., the Charles A. Desoer Professor of Engineering in the Department of Electrical Engineering and Computer Sciences at the University of California, Berkeley; Adam Margolin, Ph.D., associate professor of biomedical engineering and director of computational biology in the OHSU School of Medicine and the Knight Cancer Institute.

Overall research themes of the consortium’s Research Centers address important questions in basic cancer research, including the emergence of drug resistance, the mechanisms underlying cancer metastasis, and the role of the immune system in cancer progression and treatment. The interdisciplinary investigators of the CSBC will integrate experimental biology with mathematical and computational modeling to gain insight into processes relevant to cancer initiation, progression and treatment options.

The consortium brings together clinical and basic science cancer researchers with physician-scientists, engineers, mathematicians and computer scientists to tackle key questions in cancer biology from a novel point of view.

“Cancer is a complex disease and it challenges our traditional approaches, making it hard to predict tumor growth and drug response,” said Daniel Gallahan, Ph.D., deputy director of NCI’s Division of Cancer Biology. “Cancer systems biologists embrace that complexity and use many different types of data to build mathematical models that allow us to make predictions about whether a tumor will metastasize or what drug combinations will be effective.”

In addition to applying systems biology approaches to gain important insight into cancer, each consortium Research Center supports an outreach program to promote training in interdisciplinary science, disseminate important research findings to the community, and to engage the public in cancer systems biology research. Sage Bionetworks in Seattle serves as the consortium’s Coordinating Center, facilitating data and resource sharing and collaborative scientific activities across the nine Research Centers as well as two new Research Projects.

More information can be found on the project website.


Physician studying gene data in cancer immunotherapy lands $240,000 grant

ReidThompson_IMG_1958 (1)Reid Thompson, M.D., Ph.D.

Reid Thompson, M.D., Ph.D., a staff physician with the VA Portland Health Care System and assistant professor in the OHSU Department of Radiation Medicine, will explore genomic predictors of response to cancer immunotherapy with new funding from the Sunlin and Priscilla Chou Foundation.

The $240,000 award will be distributed over three years.

Thompson joined the VA staff and OHSU faculty in August 2016. His research focus is exploring genomic datasets such as the Million Veteran Program, which aims to collect DNA and health information from one million volunteers to help define how genes affect health.

The Chou Foundation award will support research by Thompson that is using publicly available databases to seek genomic correlates that predict responses to immunotherapy in melanoma patients. This work will be conducted under the mentorship of Adam Margolin, Ph.D., director of OHSU’s Computational Biology Program.

Thompson is a graduate of the Albert Einstein College of Medicine and he completed his residency at the University of Pennsylvania.


Dr. Abhinav Nellore recruited through Collaborative Recruitment Pool Program

Another faculty recruitment has been finalized through the school’s Collaborative Recruitment Pool program.

Abhinav Nellore, Ph.D., joined OHSU Nov. 1 as an assistant professor of biomedical engineering working in the Computational Biology Program, with a joint appointment in the Department of Surgery.

Dr. Nellore received his Ph.D. in theoretical physics from Princeton University in 2010. After a brief stint in industry as a consultant and a software developer, he became a postdoctoral fellow in the lab of Dr. Jun Song at UCSF in 2012 working on computational genomics. As a postdoc at Johns Hopkins University between 2013 and 2016, he worked with Drs. Jeff Leek and Ben Langmead on developing Rail-RNA, scalable software for analyzing many thousands of RNA sequencing samples.

New sequencing technologies are producing ever larger amounts of data, which present a tremendous opportunity, explains Dr. Nellore. However, these massive sequencing datasets cannot be analyzed without scalable software.

“My research focuses on developing such scalable software and running it on existing sequencing data,” said Dr. Nellore. “My software is designed especially for distributed-computing settings, including clusters on commercial cloud computing services such as Amazon Web Services. It is distinguished by its use of the MapReduce programming model, which facilitates borrowing strength across many sequencing samples to augment their analysis.”

“Dr. Nellore brings computational biology – a much-needed new dimension – to the scientific program in the Department of Surgery,” said John Hunter, M.D., F.A.C.S., former chair of surgery and interim dean of the OHSU School of Medicine. “Cancer biology is a focus of many of our surgical oncologists. Sorting through large volumes of genomic and proteomic data is critical to predicting tumor behavior and likely response to conventional and novel therapeutic agents. In addition, there are many less traditional methods for applying computational skills to our understanding of other surgical conditions from inflammatory disease to transplantation and response to traumatic injury as well as obesity and diabetes.”

This recruitment represents a joint effort between the Department of Surgery and the Department of Biomedical Engineering, in collaboration with the school’s Computational Biology Program.

“Dr. Nellore is truly pushing the boundary of big data genomics analysis,” said Adam Margolin, Ph.D., computational biology director. “He has demonstrated the ability to meaningfully analyze and manage the vast majority of RNA-seq data generated by the international community. We are thrilled for Dr. Nellore to join the Computational Biology program as we build a team that will break new ground in advancing biological discovery through interpretation of maximal amounts of data relevant to a given research question. His co-recruitment with the departments of Surgery and Biomedical Engineering offers the opportunity to drive Dr. Nellore’s research program through close, synergistic interactions with clinicians and biologists addressing critical translational research questions.”

Boosting collaboration across disciplines

Dr. Nellore joins other faculty members recruited through the Collaborative Recruitment Pool mechanism:

  • Naoki Oshimori, Ph.D., assistant professor of cell, developmental and cancer biology, has joint appointments in the Department of Dermatology and the Department of Otolaryngology, Head and Neck Surgery.
  • Jeffrey Nolz, Ph.D., assistant professor of molecular microbiology and immunology, has joint appointments in the Departments of Radiation Medicine and Cell, Developmental and Cancer Biology.
  • Damien Fair, PA-C, Ph.D., assistant professor of behavioral neuroscience, has a joint appointment in the Department of Psychiatry.
  • Hiroyuki Nakai M.D., Ph.D., associate professor of molecular and medical genetics, has a joint appointment in the Department of Pediatrics and membership in the Pape Family Pediatric Research Institute.

A recruitment by the Department of Physiology and Pharmacology, Department of Medicine, Division of Cardiology, and the OHSU Knight Cardiovascular Institute is in progress.

About the Collaborative Recruitment Pool Program

The Collaborative Recruitment Pool program, which began in FY 2010, seeks to facilitate research partnerships between the school’s clinical and basic science departments in ways that foster the rapid translation of discoveries into clinical therapies and treatments.

The program recruits new scientists in the basic sciences who will support growth in research areas identified for strategic development at OHSU, and who have the potential to contribute to translational research, including interdisciplinary and interdepartmental scientific programs.


Three questions for Reid Thompson

Reid Thompson, M.D., Ph.D.

Reid Thompson, M.D., Ph.D., is assistant professor of radiation medicine, with a joint appointment in computational biology, OHSU School of Medicine, and a staff physician, VA Portland Health Care System.

What’s been the most interesting development in your area in the last two years?

The whole notion of precision medicine has been gaining significant traction in recent years.  While we have increasingly incorporated evidence-based algorithms into clinical care, the wealth and proliferation of molecularly-targeted agents, particularly in the field of oncology, has been astounding. At the same time, our capacity to obtain detailed genomic and other large-scale molecular data has grown by exponential proportion as sequencing technologies have undergone a generational leap. Moreover, we have entered an increasingly data-centric era with electronic health information feeding into databases from all directions and devices.

It is at the nexus of these transformative currents that President Obama announced the launch of the Precision Medicine Initiative in his 2015 State of the Union address. The initiative itself prescribes concrete federal-level steps to facilitate research efforts; however, its biggest significance perhaps is in emphasizing the exciting and profound potential to revolutionize how we improve health and treat disease. Corresponding efforts such as the Veterans Administration’s “Million Veteran Program” (MVP) to sequence 1 million genomes are well underway. In fact, the VA just announced over 500,000 patients accrued to the MVP, already making it the largest such biobank in existence and a potential treasure trove for future exploration.

What projects are you currently working on and are there opportunities for fellow faculty to participate?

The prevailing success and emphasis of precision medicine approaches in oncology to-date has been targeting actionable cancer-specific mutations (a recent example of this approach is the NCI-MATCH trial). However, it is increasingly clear that tolerability of and resistance to cancer therapy may also be a function of host physiology, environment and genetics. In brief, cancer cannot and should not be considered in isolation from the patients it affects. My over-arching research emphasis therefore is to investigate biomarkers of treatment-related toxicity and host-specific modifiers of efficacy.

At present, I am investigating potential genomic predictors of immunotherapy responsiveness in patients with melanoma and other cancers. I am also interested in exploring genetic modifiers of the side effects of other cancer therapies (e.g. radiotherapy), and will be working under the auspices of the computational biology program on multiple omics-related projects.

I am always eager to explore opportunities for collaboration, and in particular am looking to work with other faculty who would leverage the MVP over the coming years.

A hypothetical: if you could have one tool that would solve a seemingly impenetrable problem in your work, what would it do? You have unlimited resources to design this tool, so think big.

Imagine an electronic health record system that not only facilitates patient care, but could provide you with the latest evidence tailored to the patient sitting right in front of you. Imagine that the patient’s own genome could inform whether you would prescribe drug A or drug B, whether you might avoid a treatment entirely due to excess individualized risk, whether you might tell someone to get that scan every three months instead of annually. This truly is the heart and the dream of precision medicine, but it is and will remain unachievable without us being able to integrate our point-of-care observations and data with robust and deep learning.

I see the current structure of our electronic health record systems as perhaps the biggest obstacle to research and to the implementation of a rapid learning health system that could integrate point-of-care clinical decision support. Thinking “big” here, I would completely replace existent platforms with a new paradigm to capture structured patient information and facilitate templated observations. Data would flow seamlessly behind the scenes and constantly add to our aggregate knowledge. This new ecosystem would enable patients and their providers to interact with personalized health data (e.g. imaging, genomics, mobile health streams, biomarker levels) and the latest scientific evidence on an unprecedented level, all with the click of a button or two.

This dream probably sounds crazy, but it may surprise you to learn that we have the technology and the know-how to make this happen – today! What we lack is the audacity and the resources to actually pave a new way.


2016 Race for the Cure: OHSU scientist on cutting edge of breast cancer research

Lynne Terry | The Oregonian/OregonLiveBy Lynne Terry | The Oregonian/OregonLive 
Email the author | Follow on Twitter 
on September 18, 2016 at 8:00 AM, updated September 23, 2016 at 12:07 PM

Joe Gray hopes to unlock the secrets of breast cancer.

His work could pave the way for the development of powerful new therapies that target individual tumors, vastly improving and saving lives.

It's an ambitious undertaking, but that's what Gray does. In the 30 years that he's focused on breast cancer, he's helped develop technologies that have thrust the field forward and cut mortality rates.

His latest project, still in the planning stages, will enlist breast and other cancer patients in the Portland area in a clinical trial.

Gray, a scientist at Oregon Health & Science University, is devising a trial unlike most others. His team, including the trial's leader, Dr. Raymond Bergan, will give participants cancer drugs and then biopsy their tumors as the trial progresses to check the results.

If the drugs don't work, the scientists will try something else. The idea is to come up with an optimum treatment for each patient.

"This is absolutely going to teach us as researchers and clinicians collectively how to do a better job of treating the individual patient," Gray said. "This is the essence of precision medicine."

Gray started his career as a nuclear physicist at Lawrence Livermore National Laboratory in California in the early 1970s. After nearly two decades, he moved to the University of California at San Francisco, and became directly involved in breast cancer research. His interest in working on cancer stems from his family. His father died of lung cancer at 64 when Gray was 28.

Five years ago, he moved to OHSU, where he's continued his research. The university considers Gray quite a catch. He's one of the top researchers in his field.

"It's staggering the impact he's had on cancer," said Colin Collins, a former collaborator who is director of the Laboratory for Advanced Genome Analysis at the Vancouver Prostate Centre in British Columbia.

Tall and trim, Gray looks like a scientist with his short-cropped gray hair, wire-rimmed glasses and pin-stripped shirt. He peppers his speech with words like "whilst" but has a friendly, inviting grin. Although 70, he looks a decade younger and still has lofty goals to accomplish.

Gray's inventions have helped scientists sharpen the diagnosis of cancer, for the breast in particular, and to analyze the range of abnormalities in tumors. His goal now is to find ways to distinguish between lethal and non-lethal tumors and to be able to understand how cancer cells evolve so doctors can attack them with the appropriate drug.

He's already achieved that for a type of breast cancer that affects about 15 percent of patients. Using his technology, scientists can determine who will benefit from the drug Herceptin. The mortality rate of those patients has dropped by 60 percent thanks in part to Gray's work, said Paul Spellman, professor of molecular and medical genetics at OHSU.

Gray and Bergan plan to start the clinical trial at OHSU in January. They aim to personalize cancer treatment and understand why some drugs don't work.

Hundreds of drugs are being developed for cancer, including about 100 for breast cancer alone, Gray said. Before they're approved and marketed, they have to be tested in a clinical trial.

Most trials test one drug and pose one question, Gray said: Does it work?

"Ninety-five percent of the time they don't work as expected," Gray said.

He wants to learn from those failures as they happen to pave the way for effective therapies.

OHSU will recruit about 50 people for the trial. All must have a late stage form of either acute myeloid lymphoma, a blood cancer or breast, prostate or pancreas cancer that has spread to other parts of the body.

They also must have gone through the standard of care in terms of treatment, essentially reaching the end of the line.

"This is the program I would want to be on if I had metastatic cancer," Gray said.

The trial will be one of the few to take biopsies of patients while treating them.

Another trial, at the University of California at San Francisco where Gray used to work, involves scientists taking biopsies of breast cancer patients during treatment. The scientists haven't been able to analyze the tissue quickly enough to use that information to select new treatment.

Gray, however, has a secret weapon: a laboratory of custom-built, high-precision microscopes. They sit in the basement of OHSU's Collaborative Life Sciences Building on the South Waterfront, surrounded by an underground moat. That eliminates vibrations from the building, which would make it impossible to snap pictures unveiling the inner machinery of a tumor.

"Joe is working with the highest imaging technologies that are available in the world today," Spellman said.

The microscopes enable his team to create hand-sized mockups of tumor cells. Each cancer is made up of proteins, formed into cells, formed into sub-tumors, formed into a mass. Understanding how the tissue works requires understanding how it is organized at various levels.

The study will involve a multidisciplinary group, one of Gray's hallmarks.

"If we want to capitalize on the incredible new information that we are able to generate, we need to have teams of people who will work together," said Dr. Gordon Mills, chair of the Systems Biology department at the University of Texas MD Anderson Cancer Center in Houston. "He's the ultimate collaborator and colleague to work with."

Gray plans to start taking biopsies of cancer patients this month to test his lab's ability to analyze the tissue quickly. The project has to be approved by the university's ethics board before it can start.

It is expected to last two years. If all goes well, it could be expanded. The information that emerges will likely add more pieces to Gray's overarching goal to put together a Google Earth-style holistic view of cancer.

-- Lynne Terry