Cancer in Parallel
Center for High Performance Multimodal Data Analysis in Cancer Research
Project Description
The proposed project will focus on the design, development and evaluation of high-throughput multimodal data analysis, automated image interpretation and clinical decision support algorithms and software to advance and foster research and discovery in oncology. This work will be carried out through a close collaboration between investigators from the NCI-designated Comprehensive Cancer Institute of New Jersey, Rutgers University, Robert Wood Johnson University Hospital and IBM using IBM server technology.
According to the World Health Organization, cancer causes 7 million deaths each year, or 12.5% of deaths worldwide. More than 11 million people are diagnosed with cancer every year, and it is estimated that there will be 16 million new cases every year by 2020.
Therapies and treatment regimens which are currently used in the battle against cancer are often based upon classification strategies which are limited in terms of their capacity to distinguish among stages and subtypes of diseases which may present with similar clinical profiles. Recent advances in high-throughput genetics and proteomics present new opportunities for revolutionizing the methods and reliability for characterizing subclasses of cancer and studying the underlying mechanisms of this disease. Drs. Foran and Bhanot have undertaken an ambitious project to design, develop and evaluate a multi-modality decision support approach for assessing and managing cancer that employs an automated, evidence-based strategy for systematically evaluating clinical, genomic, proteomic and image-based data.
Tumors are currently being routinely analyzed at the genomic, RNA, and protein expression level using the latest technologies to confirm clinico-pathologic correlations which have been established with whole tissue sections. This ongoing project takes an interdisciplinary approach to leverage advances in molecular biology and early detection through state-of-the-art imaging and pattern recognition techniques. Through the proposed collaboration with investigators at IBM w the cross-disciplinary team will be in an excellent position to translate the existing and emerging clinical and experimental results from our efforts into a functional clinical decision support system.
As the amount of information being archived digitally, for this project and around the globe continues to grow it has become increasingly difficult to index, analyze and annotate information to facilitate it's being relocated and retrieved at a future time. This is an especially challenging problem throughout the cancer community due to the fact that imaged specimens and radiological data sets and other graphical information is often indexed using a single text-based description or alphanumeric label to refer to its vast spatial and morphological informational content.
As part of our prior and proposed work, innovative approaches are being developed to expedite the process of navigating and filtering through the visual contents of biomedical images contained within distributed databases, automatically, without the requirement for interactively opening and reviewing them. While the mechanisms for content-based access to alphanumeric data have been extensively studied and are now considered relatively well understood, content-based image/ video annotation and indexing still remains an extremely active area of research with potential applications in a wide range of fields including remote sensing, cinematography, medicine, and engineering.
Through the collaboration with IBM our team will advance this work to the next level of progression which is to actually demonstrate the impact of these new technologies on the workflow in the clinical environment leading to improved therapy planning, patient care and discovery.
Key Elements and Focus
A major concentration of research for the Foran laboratory thus far has been the development of a family of advanced data analysis and mining technologies for early detection and characterization of cancers which are sometimes confused with one another during routine microscopic screening. Through the NIH-funded PathMiner project we have already developed, a web-based set of deployable tools for interactive telemedicine, intelligent archiving, and automated decision support in pathology. Hematopathology was chosen as the test domain for these feasibility studies since molecular diagnostics are routinely performed on these specimens, which facilitated the establishment of a "ground-truth" database for which there was independent confirmation of the diagnosis. Another focus has been the development of a robotic system for automated analysis of protein expression in cancer tissue microarrays which features distributed imaging and data management capabilities through a collaborative environment which supports impromptu telemedicine applications in pathology and oncology.
Tissue microarray is an extremely new technology which can provide insight regarding the underlying mechanisms of disease progression and patient response to therapy and holds promise for advancing cancer biology, clinical oncology, and drug discovery. Future progress in several key areas of research will rely upon the capacity of investigators to evaluate expression patterns in cancer tissue arrays. One of the central objectives of the proposed project is to design, develop, and evaluate, TMA-Miner, a Grid-enabled content-based image retrieval system for performing quick, reliable characterization and comparative analysis of cancer tissue microarrays The proposed system leverages a library of more than 100, 000 protein expression signatures which has been generated as part of our “Help Defeat Cancer” project which was launched on IBM’s World Community Grid in July, 2006.
In the proposed project with IBM we plan to establish the requisite computational power to optimize the performance of multimodal data indexing, automated image interpretation and clinical decision support algorithms thus allowing us to unlock the rich informational content extracted from massive data sets of imaged cells, tissues, microarrays and the full range of radiological imaging studies. This will be accomplished through the development of a reliable meta-classifier which will be devised to optimally combine the salient genomic, proteomic and image-based signatures from each of these modalities simultaneously making it possible to perform high-throughput screening and analysis which can hasten the rate of progress in investigative cancer research, drug discovery and diagnostics. Our team has the capacity and requisite knowledge to build such a system, but this undertaking will require much greater computational resources than we currently have at our disposal.
Impact on IBM
The proposed project will impact and benefit both IBM’s business and research sectors in following ways: (1) it will help maintain IBM leadership position in healthcare technologies with an improved capacity to deploy IBM technologies and products in the clinical environment: The Cancer Institute of New Jersey (CINJ) is a high-profile internationally recognized institution in fact it is one of only 39 Comprehensive Cancer Centers throughout the country which has received special designation by the National Cancer Institute of the National Institutes of Health. As such this project will serve to raise the visibility of IBM throughout the cancer and clinical communities with the potential to increase sales of products and services while addressing I/T and computational needs throughout Rutgers University, Robert Wood Johnson Medical School and University Hospitals.
The success of this project will help indoctrinate the next generation of researchers and physicians while simultaneously opening up major opportunities for stronger presence and market share of IBM technologies and products in the healthcare environment. (2) Access to clinical collaboration and multimodal clinical information ranging from molecular, cellular, tissue-level as well as complex datasets arising from the latest diagnostic radiology imaging studies. The Cancer Institute of New Jersey has already established shared resources for Analytical Cytometry/Image Analysis, Tissue Analytic Services (including Tissue Microarray), DNA Synthesis and Sequencing, Affymetrix- and cDNA Microarray, Research Pharmacy, Immunohistochemistry, Cell Culture and a Cell Line Bank, and Radiation Oncology.
The information rich multimodal data and clinical resources at CINJ provide a truly unique and ideal environment for investigators from both IBM and CINJ to develop and implement innovative scalable solutions for research and discovery, which will also help significantly strengthen IBM’sleadership position in the healthcare market andattract the brightest and most talented students to IBM. Many of whom will have received cross-disciplinary training in computational biology, engineering, informatics and medicine. The proposed project will also serve to forge productive collaborations between the clinical and research faculty at The Cancer Institute of New Jersey, University Hospitals, Robert Wood Johnson Medical School, Rutgers University and scientists throughout IBM. (4) After the first stage of the proposed project, The Cancer Institute of New Jersey will need to expand and upgrade its computional power and resources at which time they plan to purchase additional computing equipment and technologies from IBM. This will be accomplished through competitive extramural funding leverage through the success of the proposed project as well as through discretionary and capital improvement funding sources.
Collaboration with IBM
Dr. Leiguang Gong from IBM Watson Research Center is an expert in medical imaging and informatics. He and his colleagues will work closely with Dr. Foran’s team to design, develop and evaluate state-of-the art image processing, machine learning and pattern recognition methods by exploiting the rich informational content of the multi-modal data (image-based, genomic, proteomic) at The Cancer Institute of New Jersey, Rutgers University, Robert Wood Johnson Medical School and University Hospitals , by leveraging the computational power of IBM s latest technologies and platforms and talents and skills of their respective investigative teams.
Expected results
(1) This project will significantly strengthen the collaborative relationship between The Cancer Institute of New Jersey and IBM Research. Specifically, it will provide IBM researchers access to CINJ clinical research resources (e.g. physicians including medical oncologists, pathologists and radiologists as well as correlated clinical cases, data, lab facilities, etc. (2) It will provide an invaluable testbed for demonstrating IBM system server technology in clinical research, education, and practice. Since a huge amount of data originating from a full range of modalities will be used in the project, the performance measurement results from the project will be used as an important evaluation tool for developing blade-based high performance multimodal data indexing and analysis solutions.(3) The proposed project will help IBM attract most talented students throughout CINJ diverse educational programs (medical, computational biology, engineering, oncology, informatics), and will serve to provide the students and the rest of the clinical and research communities with the latest in terms of IBM’s technologies and to work closely with investigators throughout IBM. (4) The high-profile of the proposed projects and the fact that they are being conducted in collaboration with one of only 39 NCI-designated Comprehensive Cancer Centers in the country will also serve to increase profitable sales of products and services while addressing university I/T needs. The success of this project will help indoctrinate the next generation of researchers and physicians while simultaneously opening up major opportunities for stronger presence and market share of IBM technologies and products in healthcare.
Milestones
- 10/2007 - CINJ will allocate a designated office space and computing resources for IBM researchers to work in CINJ lab.
- 06/2008 - Joint publications to report initial research results of the project.
- 01/2008 - CINJ will recommend its most talented students to IBM Research for summer internships. At the same time, IBM researcher will have the opportunity to become immersed in the clinical environment and to interact with pathologist, medical oncologists, and radiologists and residents in training.
- 01/2009 - Joint research grant (e.g NIH) applications by CINJ and IBM researchers.
- 11/2009 - Joint demonstration of multimodal data analysis and clinical decision support using algorithms and computational tools developed using IBM technology.
University Relationship
(1) Cancer Institute of New Jersey (CINJ) has already been an active player IBM’s Help Defeat Cancer Initiative, using the World Community Grid. In 2006 IBM collaborated with investiagators at CINJ to launch the “Help Defeat Cancer” project which has already generated a mixed set of nearly 100, 000 tissue discs originating from immunostained microarrays from several NIH designated Cancer Centers across the country. The research carried in this project over the past 9 months has already helped to leverage nearly $4 million dollars in new funding for the Foran laboratory as a result of competitive funding from the National Institutes of Health which in turn will be used to help boot-strap the proposed project. (2) The CINJ was established in 1991 with a mission to put an end to the devastating effects of cancer. It is one of only 39 National Cancer Institute (NCI)-designated Comprehensive Cancer Centers in the nation - the only one in New Jersey. As part of this elite group of centers nationwide, CINJ is a leader in laboratory, clinical, prevention, and population research. CINJ delivers advanced comprehensive care to adults and children and conducts cutting-edge cancer research. Top-notch scientists are at the forefront of developing methods to treat and prevent cancer. (3) Dr. David Foran of CINJ is an internationally recognized leader in biomedical informatics and imaging research. He is a charter member of an NIH Study Section at the newly established National Institute of Biomedical Imaging & Bioengeering at NIH. His work has been supported through competitive extramural funding from the Whitaker Foundation, the NJ Commission on Science & Technology, the federal Defense Advanced Research Projects Agency (DARPA), the Radiological Society of North America, the National Institutes of Health (NIH) and the private sector to conduct research in these areas. (4) Dr. Gyan Bhanot is an internationally recognized computational biologist related to cancer whose research utilizes a combination of microarray, mass spec and SNP polymorphism data to identify, quantify and explain cancer initiation, progression, and metastasis. He also a leading expert in Evolutionary Genetics including human migration, phylogeny, disease association studies and patterns of mutations correlated with longevity and complex disease phenotypes. This research involves collaborations with medical oncologists at CINJ and researchers throughout CINJ, UMDNJ, CABM and BioMaPS as well as collaborators from the Institute for Advanced Study, Princeton, Broad Institute, IBM Research, Tokyo Gerontological Institute and Boston University. (5) Dr. Leiguang Gong of IBM Watson Research Center is an expert in medical imaging and informatics. He worked with Dr. Foran on several successful medical imaging projects in the mid 90's at Rutgers. He will collaborate closely with Foran’s team to design, develop and evaluate state-of-the art image processing, machine learning and clinical decision support algorithms by exploiting the rich informational content of the multimodal data (image-based, genomic, proteomic) at CINJ, by leveraging the computational power of IBM ‘s latest technologies and platforms.
University Investment
The Cancer Institute of New Jersey (CINJ) is a matrix style cancer center under the auspices of the University of Medicine & Dentistry of New Jersey-Robert Wood Johnson Medical School. The Director serves as an Associate Dean for Oncology Programs and is responsible for integrating research at the medical school with the Robert Wood Johnson University Hospital, the Saint Peter’s Hospital, and several schools and departments of Rutgers University. CINJ was awarded its first P30 Cancer Center Support Grant (CA72720) on March 1, 1997. CINJ has 124 members, whose peer-reviewed, funded research projects, as defined by the CCSG guidelines, total more than $26.5 million in annual direct costs, over $6 million of which is from NCI. The Cancer Institute of New Jersey has established shared resources for Analytical Cytometry/Image Analysis, Tissue Analytic Services (including Tissue Microarray), DNA Synthesis and Sequencing, Affymetrix- and cDNA Microarray, Research Pharmacy, Immunohistochemistry, Cell Culture and a Cell Line Bank. Shared resources for Transgenic Animals and Proteomics are being developed.
Reiss’ laboratory occupies 950 net sq. ft. within The Cancer Institute of New Jersey (CINJ) which is a 3 story, 75,000 sq. ft. state-of-the-art research facility that includes all of the shared equipment and core resources that are essential for modern molecular- and cellular cancer research. Dr. Reiss’ laboratory has all the small equipment required for this project. The laboratory is fully equipped to conduct the studies proposed. The laboratory includes a chemical hood, two thermal cyclers, a Novex apparatus for cold SSCP, sequencing gel apparatus, inverted microscope for microdissection, Forma (-80oC) freezer, Forma (-20oC) freezer. Warm rooms, cold rooms and tissue culture facilities, including Zeiss upright and Nikon fluorescent inverted microscopes with digital camera, Coulter counter, liquid nitrogen storage for cells are located in the immediate vicinity of Dr. Reiss’ laboratory. In addition, shared equipment in the Cancer Institute of New Jersey includes Beckman X80 and XL80 ultracentrifuges, a Beckman GS-6 centrifuge, a Sorvall RC-5C superspeed refrigerated centrifuges, Beckman liquid scintillation spectrophotometers LS6000SC and LS6500SC, Dynatech MR5000/7000 microplate reader, floor shaker-incubators, hybridization ovens, Turner designs luminometer, speed-vac apparatus, Kodak film processors, graphics workstation, Molecular Imager FX and a gel documentation system.
The Center for Biomedical Imaging & Informatics has a core facility which occupies 725 sq. ft. of dedicated space, within the Department of Pathology & Laboratory Medicine at UMDNJ-Robert Wood Johnson Medical School and a Tissue Imaging Core laboratory which occupies 1, 200 sq. ft. at The Cancer Institute of New Jersey. There are also several high performance imaging and computer resources located throughout the Departments of Pathology and Radiology at Robert Wood Johnson University Hospital. (2) The image analysis hardware installed in the core laboratories includes: 3 high-performance image processing workstations, one of these workstations includes Crystal eyes for 3-D viewing of molecular graphics models, six (1GHz-2GHz) Pentium 4 based workstations. (3) A DataCube Image Array processor with a high resolution color monitor and high resolution Occulus TCX image acquisition boards interfaced to Sony and Optronics high resolution cameras; a Nikon Scantouch flat bed scanner; a Pinnacle CD authoring device and software; a Sony programmable VCR, and a Focus Graphics Image recorder, a network-based Oracle Database system, a robotic, Olympus microscope, and a high-resolution, high-throughput Virtual Microscope (whole slide scanner). A Nuance multispectral imaging system which allows for accurate spectral classification and unmixing. (4) The laboratory also has access to a research grade electron microscope. These resources are interconnected and linked via ethernet with Internet access, thus linking this facility to the resources of Rutgers University, the Molecular Biology computing laboratory in the Center for Advancement in Biology and Medicine (CABM), the Library of Science and Medicine, the Academic Computing Center, and computers in the Surgical Pathology Suite at Robert Wood Johnson University Hospitals. The laboratory is also equipped with a Zeiss standard epiflourescence light microscope coupled via a Varo Electronoptics microchannel plate image intensifier to a Dage model 70 newvicon video camera. (5) Additional image acquisition devices include a Paultek high resolution CCD camera with image intensifier, integration and thermoelectric cooling options. This unit can be used for low light level image acquisition on a photo-microscope. It is coupled with one of the G3 image analysis workstations. The hematopathology lab is equipped with a multi-headed microscope and a Coulter flow cytometer, and a portable telemedicine system.
NSF Center for Autonomic Computing (CAC), Rutgers, University. The Center for Autonomic Computing (CAC), an National Science Foundation (NSF) research center established between University of Florida, University of Arizona and Rutgers combines resources from these universities, private companies, and federal government to make all kinds of computer systems and applications – from humble desktop computers to complex air traffic control systems and scientific and engineering applications – more reliable, more secure, and more efficient.
An autonomic computing system/application is any system/application that is designed to function with minimal management even as conditions, users and usage patterns change. Autonomic computing has a wide range of applications, and it can greatly reduce the growing costs of administrating computer systems and applications. Autonomics can protect against loss of service in systems performing critical functions, including those managing power grids, stock markets, and hospital networks. Autonomics can also greatly improve the speed, efficiency and robustness of complex systems that utilize a large number of hardware and software components, and enable them to self-optimize to use resources more effectively to improve productivity and conserve energy.
CAC is organized under the auspices of NSF's successful Industry/University Cooperative Research Centers (I/UCRC) program. The growing list of its industry members includes leading industry organizations and government agencies. The center is funded by membership fees from industry partners, university matching funds, and by the National Science Foundation's I/UCRC Program. The center's Web site is www.nsfcac.org .
Visibility and Publicity
- One important goal of the project is to extend the CINJ study to all major public databases covering all stages of breast cancer disease (initiation, progression and metastases). The equipment in this SUR proposal will be used to store the microarray data, generate web based analysis protocols and clinical decision support tools to apply to the data as well as to create a clinically useful profile for risk analysis. This profile will be tested in the clinic at CINJ with the participation of medical oncologists and pathologists at CINJ. Two years after the award date, we will release the web site to the general community as a public resource for research and clinical use.
- Joint publications and presentations at international conferences to report initial research results of the project.
- Joint research grant (e.g NIH) applications by CINJ and IBM researchers.
- Joint demonstration of multimodal data analysis and clinical decision support using jointly developed algorithms and computational tools using IBM hardware platform.
- A designated website will be setup for this SUR project.
- The public relations department at The Cancer Institute of NJ has already pledged its commitment to maximize newspaper, radio and television coverage of the proposed project.
