SOWAttachmentNo3-BCR-3-10.pdf

SOW BCR dated March 3, 2010

Statement of Work

Biospecimen Core Resources (BCR) Providing Biospecimen Processing Services

For The Cancer Genome Atlas (TCGA)

INTRODUCTION

Cancer is a complex and heterogeneous disease in which mutations and other genomic and epigenomic abnormalities play a role in both its initiation and progression. Accumulated research data implicate numerous somatic mutations and a more limited number of inherited mutations in carcinogenesis.

Understanding cancer-specific somatic mutations can provide important clues regarding the molecular processes underlying the development and progression of certain tumors. Given cancer’s complexity, it is generally believed that only a fraction of alterations that may be useful as characteristic markers of specific tumor types and/or potential molecular targets have been identified to date. Therefore, to be successful, comprehensive genomic analyses of cancer must overcome a broad range of challenges stemming from the biological complexity and heterogeneity of human tumors and subtypes. An important role in tumor heterogeneity is played by genomic instability, which is inherent to the progression of cancer. The dynamic changes in tumor genomes are influenced by the cellular and biological context, genetic characteristics of individual persons, and environmental factors. Certain similarities exist across tumor types; however any effort to characterize the genomes of tumors in a comprehensive, systematic manner must address the heterogeneity across distinct cancer types and subtypes.

Many research groups are working independently to identify cancer-relevant genomic changes (e.g., alterations in expression profiles, chromosome deletions, amplifications, and/or translocations) in a relatively small number of samples. Cancer-relevant changes have been detected at diverse levels of genomic organization; they include point mutations, chromosomal deletions, amplifications, translocations, and/or changes in the number of individual chromosomes, all of which can, in principle, contribute to transcriptional aberrations. However, in terms of comprehensive cataloging, these individualized, researcher-initiated and uncoordinated approaches are limited. Such studies are inherently fragmentary (e.g., by technology and/or tumor specimens analyzed) and difficult to integrate because of data compatibility issues across studies and tumor types/sub-types. The progress in understanding some cancer-associated molecular alterations and the accompanying advances in technology suggest that it is now possible to obtain comprehensive genomic information from multiple tumor types to catalog most, if not all, of the genomic changes associated with cancer.

The Cancer Genome Atlas (TCGA) Pilot Project.

In 2006, the National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI) initiated a collaboration to pursue a 3-year pilot project to determine the feasibility of more comprehensively cataloging the genomic alterations associated with a small number of different human cancers. The pilot project focused mainly on three tumor types [glioblastoma multiforme (GBM), serous cystadenocarcinoma of the ovary, and squamous carcinoma of the lung]. The goals were to a) assess the technical feasibility of generating and integrating the genomic data and b) verify that these data could be analyzed to yield clinically relevant hypotheses. During the pilot phase of TCGA, the collective genomic and clinical data generated and integrated by all TCGA components provided the initial contributions to a comprehensive Web-based resource describing the genomic profiles of specific cancer types. This resource is known by the same title as the project - The Cancer Genome Atlas.

The pilot project demonstrated that cancer-associated genes and genomic regions can be identified by combining diverse information from genome analyses (as defined in the previous paragraph), with tumor biology and clinical data, and that the sequencing of selected regions can be conducted efficiently and cost-effectively. Achievement of both pilot project goals has set the stage for the next phase of TCGA that promises to rapidly and efficiently generate analogous genomic data for all major cancer types and subtypes. The comprehensiveness and rate of progress of TCGA will depend on both optimization of technical issues and resource availability.

The strength of TCGA is to produce unprecedented multi-dimensional data sets using an appropriate number of samples to provide statistically robust results that sets the stage for a new era in the discovery of new cancer interventions. The integrative analyses leading to the formulation of an unanticipated hypothesis on a potential mechanism of resistance highlights precisely the value and power of such project design, demonstrating how unbiased and systematic cancer genome analyses of large sample cohorts can lead to important discoveries.

Three key “lessons learned” during the pilot project were that, to be able to interpret the results generated by the various characterization platform, the Centers a) had to utilize high quality molecular analytes isolated from well characterized tissue specimens; b) perform experiments utilizing strict standardized protocols; c) and deposit the results in structured and well-described formats. The last lesson strongly impacted on the ability of the various analytical groups to extract meaningful results from the genomic data generated. These lessons learned will be applied to the next phase of TCGA.

The Cancer Genome Atlas Pilot Project as a Coordinated Effort A unique aspect of TCGA Pilot Project was the development and function of an integrated research network. TCGA Pilot Project included multidisciplinary teams of investigators and associated institutions that collectively generated and analyzed biological data. TCGA Pilot Project Research Network was comprised of the following major organizational and functional components:

Biospecimen Source Sites: NCI established relationships with a group of clinical sites and clinical trial networks to provide clinically annotated biospecimens, accrued and distributed to TCGA under appropriate Institute Review Board (IRB) review. The biospecimen network continues to expand under TCGA in order to obtain sufficient samples to support the cancers to be studied and the number of cases to be analyzed per cancer. TCGA’s biospecimen resources include both retrospective collections and prospectively established banking protocols.

Biospecimen Core Resource (BCR): TCGA Pilot Project BCR served as a centralized tissue processing center and provided the molecular biomolecules for the Pilot Project. In addition, the BCR collected and standardized clinical annotation. Standard operating procedures governed clinical data collection, sample collection, pathological examination, biomolecule (e.g., DNA and RNA) extractions, quality control, laboratory data collection, and biomolecule distribution to the Cancer Genome Characterization Centers and the Genome Sequencing Centers. The BCR ensured that samples and data were received by TCGA under appropriate human subjects review and informed consent, and also that Material Transfer Agreements represented the policies of the NIH for this project. (http://cancergenome.nih.gov/policies/guidelines/index.asp).

Cancer Genome Characterization Centers (CGCCs): As a part of TCGA Pilot Project Research Network (see below), the CGCCs conducted high-throughput comprehensive genome-wide analyses using validated technologies (e.g., gene expression profiling, detection of chromosomal segment copy numbers alteration) to reveal the spectrum of genomic changes that exist in human tumors and to identify genomic regions for further characterization by the Network’s Genome Sequencing Centers. Technologies were optimized by each CGCC to increase the rate, sensitivity and specificity of production. In addition, the CGCCs made changes to increase microarray density or content to improve the characterization of biomolecules (for details see http://cancergenome.nih.gov/components/cgcc.asp).

Genome Sequencing Centers (GSCs): High-throughput sequencing analysis of tumor DNA provided by the BCR was performed by the NHGRI large-scale sequencing centers (http://cancergenome.nih.gov/components/gsc.asp). Candidate genomic regions for sequencing were identified on the basis of information from the scientific literature; additional genes and genomic regions were selected based on data generated by the CGCCs (genes sequenced are listed on http://cancergenome.nih.gov/dataportal/data/about/types/sequencing/).

Data Management, Bioinformatics, and Computational Analysis: Data generated by TCGA components were deposited into public databases as soon as they were validated, in general within a few weeks of generation. The information generated by TCGA Pilot Project is centrally managed to develop data standards and controlled vocabularies for each new technology, to establish an informatics infrastructure for data exchange between components of the project and a central repository, to create portals for basic and clinical researchers to easily access the data, and to encourage new computational approaches to analyze the data (http://cancergenome.nih.gov/dataportal/data/about/). Another key function of the bioinformatics component is to provide a secure network and means to protect the integrity and security of research and clinical data (http://cancergenome.nih.gov/dataportal/data/access/).

The Cancer Genome Atlas Continuation and Expansion.

Phase 2 of TCGA will utilize a similar model (summarized in Figure 1, below). The primary change is that there will be two types of NCI supported centers. Genome Characterization Centers (GCC) will receive analytes and characterize them for chromosome segment copy number alterations and translocation, expression changes of all genes, epigenetic changes and mutations. Genome Data Analysis Centers (GDAC) will be data recipient and analysis sites. For the purpose of this SOW, the BCR role remains similar as under the Pilot Project. The BCR will receive clinically annotated biospecimens from a network of clinical sites, process those specimens and data, and transmit the specimens and data to characterizing/sequencing centers and the Data Coordinating Center, respectively. (For a more detailed description of plans for the continuation of TCGA, see the RFA available at http://grants1.nih.gov/grants/guide/rfa-files/RFA-CA-09-010.html) or other information from the NCI Office of Cancer Genomics (OCG) (http://ocg.cancer.gov/). It should be noted that under Phase 2 of TCGA, the BCR is expected to support the distribution of actual tissue samples to characterization centers in order to support more epigenomic studies – this is a change from the Pilot Project in which only molecular analytes were distributed. The term characterization is defined to include both the genomic analyses described above and DNA sequence data generated by NHGRI supported Genome Sequencing Centers (GSC) participating in TCGA.

Figure 1

DCC

GCC

GSC

BCRs

NCI &

NHGRI

Project Team

Research Community

NCBI

GDAC

Genomic data

Tracking info

Mutation & trace2sample

Sequence traces

Genomic & clinical data

Protected & public data

Sequence traces

Analysis results

Analysis pipelines

Clinical, pathology & tracking data

Molecular analytes

Clinical sites

Tissue samples & clinical data

Background to the Biospecimen Core Resource

A major prerequisite of TCGA Pilot Project was the acquisition of high quality biospecimens. To meet this need, NCI established a network of clinical sites providing high quality, clinically annotated biospecimens to a centralized quality control and processing facility. It must be noted that the term “high quality” refers not only to the histological and molecular properties of the tissue, but also to characteristics such as degree and quality of clinical annotation, the existence of appropriate informed consent provisions for the intended use of the biomolecules and data, collection and subsequent distribution to TCGA under an institutional review board (IRB) reviewed protocol, as well as unencumbered access for research use (e.g. intellectual property restrictions).

TCGA project management chose to establish a centralized tissue processing model to ensure that process variables are minimized until their effects on the results of molecular analyses become well understood.

This centralization specifically means that all operations to process tissue and data for any single cancer studied by TCGA occur at one or two BCRs, utilizing the same protocols. This minimization of variance refers to the processes of biospecimen receipt, logistical and physical management, processing into analytes (the molecular extracts from tissue such as DNA, RNA), the subdivision of tissue, and finally distribution of tissue or analytes to the research sites with rigorous QC of all intermediate and final products along the workflow. In order to minimize such variability in biospecimen processing, TCGA Pilot Project established one centralized Biospecimen Core Resource (BCR) to perform the biospecimen processing according to a uniform set of Standard Operating Protocols (SOP). The Pilot Project BCR operated a two stage mechanism, first to qualify and quality control biospecimens for histopathological characteristics, followed by processing to isolate biomolecules (DNA and RNA), which were also quality controlled before being shipped to characterization centers. For Phase 2 of TCGA, a second BCR will be established.

Summary of main biospecimen requirements The main requirements for the biospecimens, as obtained from Tissue Source Sites (TSS), included:

Biospecimens were primary tumors, untreated, with a case-matched source of germline DNA (blood or uninvolved tissue, preferably from the same organ site as the tumor).

Tumor tissue, or uninvolved normal adjacent tissue if applicable, was snap frozen quickly after devitalization to preserve biologically-relevant biochemical state.

Tissue specimens and blood needed to be of sufficient quantity to support co-isolation of DNA and RNA from the same sample for distribution to all characterization centers. Minimally, the yield requirement is approximately 20ug each for DNA and RNA, but that number was subject to change with analytical technology platform improvements.

Clinical data, including vital status, were collected (or collectable) and available, both technically and administratively, for distribution to TCGA.

Administrative polices of the TSS, including human subjects requirements, intellectual property restrictions, and material transfer agreements, were in line with NIH policies for TCGA.

Lessons Learned from the Pilot Project phase.

Per objectives of any pilot project, many lessons were learned about how a BCR should operate to support TCGA, and are included in the Statement of Work for the BCR supporting TCGA’s continuation.

A summary of those lessons included:

The BCR was the primary day-to-day interface with tissue source sites, and as such was in the best position to understand the number of potential cases and samples available, the degree of difficulty the TSS would have to access those cases, and, as sample receipt began, the rate at which samples from those cases passed the quality control metrics. Therefore, the BCR was the major source of this data to TCGA project management and investigators performing genomic characterization and data analysis. TCGA Pilot Project BCR systems (including informatics) required enhancement to capture and report such data.

The number of TSS required to meet biospecimen needs was larger than expected and the administrative requirements for TSS participating in TCGA are not typical. The Pilot Project BCR expended more effort and resources than anticipated to establish the documentation prerequisite to receiving tissue and clinical data. Such documentation included execution of MTAs in which are embodied project policies on intellectual property, redistribution of biospecimen derivatives, and ability to receive and redistribute clinical data that may be regulated under HIPAA. Additionally, the BCR receipt of specimens and data was dependent upon the independent establishment of funding for the TSS, so the BCR had to be familiar with this need.

This factor also required the BCR to support more resources for logistics (e.g. shipping) and for logistics training at TSS.

SOPs required frequent development, review and improvement. SOPs were required to cover all processes and included results capture and other metrics to enable process improvement and to discover where processes failed in that event. The Pilot Project BCR required process management improvements to cover all aspects of workflow.

Workflow informatics (e.g. LIMS) to guide processes, and capture status and results, were required. In a pilot project context, workflow informatics requires constant technical development and improvement to support capture and reporting of developed or modified experimental protocols. The Pilot Project BCR required capital investment in commercial LIMS and subsequent information technology (IT) support to meet the experimental protocol support, data capture and reporting requirements of TCGA Pilot Project. Workflow informatics was also required to be linked to production equipment with automated upload of data to the LIMS. Bar coding of samples was essential.

Reporting needs for management and operations during TCGA Pilot Project entailed collecting more operational and processing metrics than initially specified. The Pilot Project BCR was required to modify pathology and laboratory processes (and supporting informatics) in order to collect these data.

New cancers being studied by TCGA necessitated development of processing customization to capture histopathological Quality Control metrics, and to most efficiently yield DNA and RNA.

Further, variable composition and cellularity of tissues and variable rates of processing success meant that the mass of tissue per case required from a TSS could change as experience was gained. Additionally, changes over time in clinical practice have varied the size of tissue specimens being available for research use. The Pilot Project BCR applied resources to several pilot projects and development of cancer-specific tissue processing protocols and their optimization. Such projects included, for example, evaluation of TCGA utility of DNA isolated from immortalized cell lines and evaluation of the effects of necrotic tissue composition of RNA expression.

Characterization technologies were changing over the course of TCGA Pilot Project, so BCR had to be able in adopt new analyte extraction and handling protocols, for example, to isolate different types (e.g. whole RNA for micro-RNA expression analysis) of nucleic acids for new types of characterization (e.g. sequencing for expression). The quantity of nucleic acid required also changed; the Pilot Project BCR was able to “rescue” samples originally too small to yield sufficient DNA once Whole Genome Amplification (WGA) was validated by copy number analysis and found to be adequate.

Clinical informatics requires constant technical support, including development of new data models and data elements for each new cancer, and modifications during the study course of any specific cancer. During the Pilot Project, experience showed that investigators using the data would ask for new information to be obtained from TSS or from different analyses of specimens.

The Pilot Project BCR required greater than expected clinical informatics support.

Macro-dissection and portioning of samples were valuable tools that increased the tumor cellularity of samples resulting in a higher pass rate through pathology QC. These techniques also slowed down production and increased costs.

TSS were expected to histopathologically prescreen samples to meet a minimum level of tumor cellularity before shipping those samples to the BCR for full review. Several sites did not perform this prescreen, which created a bottleneck in the high-throughput production process.

The solution required a significant increase in effort to screen through a large number of samples to find the few samples that would pass full pathology QC metrics.

Related NCI projects The NCI has established several strategic initiatives that interface with and serve supporting roles for TCGA pilot project. These include major institutional efforts focused on biomedical informatics and biospecimen research and biobanking practices. Two such projects warrant specific mention in relation to BCRs providing services to TCGA.

NCI Office of Biorepositories and Biospecimen Research

In 2005, in response to growing needs and requirements for high quality clinical materials, associated clinical data, and human subjects protections, NCI established the Office of Biorepositories and Biospecimen Research (OBBR) (http://biospecimens.cancer.gov). The OBBR will coordinate and manage the implementation of biospecimen best practices and biospecimen research on behalf of the NCI.

In 2007 OBBR published the NCI Best Practices for Biospecimen Resources, which provided NCI biorepositories with a set of technical, as well as ethical and legal standards that were generally well-established within the biorepository community. NCI Best Practices will be informed over the coming years with more detailed evidence-based standards that will be supported by a comprehensive Biospecimen Research Network (BRN). Data yielded by the BRN will be used to develop data-driven SOPs for biorepository operations. OBBR is also working to establish the cancer Human Biobank (caHUB), an NCI-launched entity that will be a broad source of knowledge and also operationally active in support of NCI projects dependent on biospecimens. As a result of the Best Practices, the research network, caHUB, and the overall coordination of all NCI biorepository activities, the OBBR’s initiatives will lead to improved accessibility to high-quality biospecimens for the broader cancer research community.

The BCR shall interact with OBBR in the development and deployment of protocols for biospecimen procurement and quality control.

Cancer Biomedical Informatics Grid (caBIG)

The NCI’s cancer Biomedical Informatics Grid, or caBIG®, is an interoperable network of data systems built to connect individuals and institutions across the cancer research community (https://cabig.nci.nih.gov/overview). NCI’s goal is to enable the sharing of cancer research information and analytical tools to speed innovative approaches for the prevention and treatment of cancer. caBIG® is being developed under the leadership of the National Cancer Institute’s Center for Biomedical Informatics and Information Technology (CBIIT), and encompasses the development of significant infrastructure, application software, and data standards in many domains. The caBIG® project is divided into several workspaces, including two of specific relevance to this SOW: Tissue Banks and Pathology Tools (TBPT) and Vocabulary and Common Data Elements (VCDE). TBPT encompasses applications, data models and data standards supporting tissue banks and pathology. VCDE develops terminology and common data elements, the requirement for which specifically applies to data generated and transmitted by the BCR since the BCR’s informatics systems will be required to interoperate with other cancer research community systems. All data generated from TCGA Pilot Project have been made publicly available via tools and infrastructure that are compatible with the cancer Biomedical Informatics Grid (caBIG®) (https://cabig.nci.nih.gov/).

The BCR shall interact with CBIIT in the development of data standards and the transmission of TCGA data to TCGA’s Data Coordinating Center (DCC).

Biospecimen Core Resource (BCR) Statement of Work

Independently, and not as an agent of the Government, the Contractor shall perform the services described below. The Contractor shall provide qualified personnel, material, equipment and facilities not otherwise provided by the Government during performance of this contract.

The Contractor shall function as a fundamental resource and play a key role (see Figure 1 above,) in TCGA. Key goals shall include ensuring that standards are developed, implemented and maintained for all aspects of cancer-related biospecimen management and processing to support genomic characterization. Processing variables must be minimized until their effects on the results of molecular analyses become well understood, especially at the “front-end” of the workflow for TCGA. The “front end” refers to annotated biospecimen receipt, logistical and physical management, processing into analytes (the molecular extracts from tissue such as DNA, RNA), and finally distribution of those analytes and portions of original tissue to the research sites with rigorous QA/QC of all functions. High throughput, production ready SOPs shall govern operations under which biospecimens are processed into biomolecules and their aliquots that will be shared across TCGA characterization centers. The Contractor shall coordinate the distribution of aliquots of molecular analytes. The primary reason for this centralized approach is to ensure consistent quality of the biomaterials and associated data distributed to the characterization centers and provide the foundation for data comparability across characterization platforms and laboratories. All data from transmitted from the BCR (but not necessarily the software applications performing these functions) for work performed under this contract shall be ® silver-grade compliant. For definition of caBIG® compatibility guidelines please see https://cabig.nci.nih.gov/guidelines_documentation.

Because this project involves use of biospecimens and information about them that potentially relates to clinical care, clinical analysis of tissue samples shall be at the level of standard of care. Therefore, for analyses involving histopathologic examination (i.e., interpretation of clinical findings of tissue samples) the Contractor, or its subcontractors, shall require Clinical Laboratory Improvement Amendments (CLIA) laboratory accreditation. Since the Contractor shall transmit biospecimens and data between a clinical setting to a research setting, the Contractor shall also develop and implement ethical, legal and policy SOPs and/or BCR policies that ensure the protection of all aspects of patient privacy. Such SOPs and policies shall be aligned with NIH policies. The Contractor shall be a primary repository for clinical information accompanying the specimens it collects and manages; consequently, the bioinformatics management systems of the Contractor shall in certain instances have the ability to link patient’s demographic information and clinical data to other data such as molecular profiles (e.g. DNA sequence data) generated from their tissue specimens. Therefore, in addition to requirements for SOPs and policies to support the biospecimen collection and management aspects of TCGA, the Contractor shall implement failsafe systems to ensure patient protection and adhere to the highest current and future standards for ethical conduct of research on human subjects and for maintaining the privacy and confidentiality of protected health information.

TCGA will operate with a centralized and standardized protocol-driven tissue processing and distribution model to ensure that process variables are minimized. Contractor protocols shall include rigorous QC of all intermediate and final products along the workflow. The Contractor shall minimize variation in the following three areas:

1. The Contractor shall operate under logistics and informatics standards governing transfer and receipt of biospecimens from various retrospective and prospective clinical collection protocols.

This will maximize the quality and comparability of the “input.”

2. High throughput, production ready standard operating procedures shall govern internal operations under which biospecimens are reviewed and processed into biomolecules. Protocols shall be shared across research laboratories.

3. The Contractor shall strictly control and document the distribution of aliquots of molecular analytes and other biospecimens. The primary reason for this approach is to ensure consistent quality and comparability of the biomaterials and associated data to the analysis centers and provide the foundation for data comparison across platforms and laboratories.

BCR Capabilities required to support this Statement of Work

There will be two contracts awarded for the BCR function. When two BCRs are operating, NCI will coordinate their efforts to assign different tissue source sites, and possibly cancers, to each Contractor and to share knowledge and systems. Specific requirements for Contractor’s that address coordination are below.

During performance of the contract, the Contractor shall maintain the below required capabilities including but are not limited to expertise, personnel, protocols, systems, and technology to:

Procure biospecimens under appropriate ethical, legal and contractual guidelines adherent to NIH policy for this project.

Manage a biorepository, including biospecimen storage management, tracking and logistics, shipping/receiving, analyte distribution, request processing, and recipient feedback capture.

Management includes the ability of live electronic tracking and reporting of inventory and inventory item status.

Perform biospecimen pathology verification, including the capture of structured standardized data and digital images to confirm accuracy of original diagnosis and to qualify specimens for research use.

Perform specimen processing, including co-isolation of DNA and RNA from a tissue sample, and biomolecule QC.

Deploy BCR operating and management infrastructure, including governance and oversight interfaces, advisory board(s), informatics, reporting systems, and necessary laboratory certifications, and personnel credentials.

Manage interactions with the NCI Contracting Officer’s Technical Representative (COTR), Center for Biomedical Informatics and Information Technology (CBIIT), Office of Biorepositories and Biospecimen Research (OBBR), and TCGA management structure.

Develop Standard Operating Procedures, where none yet exist, in coordination with the TCGA Project Manager for all activities

Interact with staff at other TCGA BCR to harmonize all SOP and data standards between the organizations.

Implement a QA program against which BCR can be evaluated, and that can be integrated in the broader TCGA Quality Management System.

The Contractor shall provide all necessary personnel, expertise and other resources as approved by NCI to provide a centralized facility to uniformly process tissues and associated data for any one cancer being studied and distribute standardized sets of biomolecule aliquots to multiple TCGA sites performing genomic characterization. Tissues shall be received by the Contractor from contributing Tissue Source Sites (TSSs) that will be vetted by NCI to meet a set of preliminary evaluation criteria The Contractor shall receive tissues and clinical data from those TSS according to direction from the NCI Contracting Officer Technical Representative (COTR). Tissue specimens shall be subject to QC, processed into molecular analytes (also subject to QC), and distributed. The Contractor shall coordinate receipt, internal processing and distribution work flow to provide consistently high quality biomolecules to the multiple cancer genome characterization and genome sequencing centers. All operations shall be governed by

SOP.

A Contractor shall be able to perform, as a single entity, all the biospecimen processing capabilities described in this SOW needed to support any one cancer studied by TCGA. Key activities of the Contractor may include developing SOP to govern receipt, processing, management and distribution of biospecimens at a level of quality that supports genomic characterization.

Definitions Biospecimen: includes all biological specimens, including tissue in any format (e.g. solid frozen formalin fixed paraffin embedded, cell pellets, whole blood, etc.) and molecular analytes in any format.

Case: an individual from whom one primary untreated tumor specimen plus one other untreated specimen providing a source of germline DNA are obtained.

Case-matched: all specimens and/or data being referenced come from the same patient case.

Characterization: the combination of both genomic analyses, such as RNA expression, epigenetics, and copy number alteration performed by NCI-supported Genome Characterization Centers (GCC), and sequence data generated by NHGRI-supported Genome Sequencing Centers (GSC).

Tissue Source Site (TSS): a TSS is an administratively and functionally unique source of biospecimens and data as determined from the point of view of the BCR’s work to receive those materials from that site.

Current Assumptions of the BCR Work The Government is providing the below information that is currently available concerning the volume and magnitude of the overall BCR work. The main requirement for a particular cancer studied by TCGA is based upon the statistical power goals set by the project, defined by the successful characterization (as defined above) of 500 cases, each case on all platforms, and resulting in data from each deposited at the DCC. For some cancer types it is possible that fewer than, or more than, 500 cases will be required to result in statistically significant data. For these tumors types, processing samples will stop when statistical significance has been achieved as determined by TCGA working group tasked with determining such end points. In such instances, the Government will notify the Contractor in writing noting the change in the number of cases. A fraction of cases will additionally include a case- matched additional “normal” control solid tissue specimen derived from the same organ site as the tumor. These samples may be used to study “normal” expression and epigenetic characteristics. Another fraction of cases will also include one or more case-matched additional tumor biospecimens that are post-treatment recurrences or metastasis of the primary tumor. Allowance for these variables is made in the requirements below.

Many variables affect the ultimate resources required of the Contractor to support TCGA. Many such variables are unknown at the start of a specific cancer’s analysis, and can only be derived heuristically.

Key examples of unknown variables that affect resource levels include:

The number of TSS with which the Contractor shall need to establish administrative (e.g. MTA), logistics (e.g. shipping), and technical (e.g. clinical data transmission) relationships per cancer is unknown. Some TSS have significant retrospective collections, most have proven to be in the range of 40 to 100 cases. For those TSS establishing prospective collections, the annual rate of case enrollment and sample accrual is not yet known. The NCI has established minimum numbers for case accrual rates and minimum shipment sizes.

The rate at which tissue specimens of a particular cancer, or as provided by a particular TSS, shall pass histopathologic criteria as set forth by TCGA project management. Current specification is that tissue must comprise 80% tumor cells; however, this criterion is subject to review and modification as new technologies become validated.

Tissue specific yields of DNA and RNA per mg. of tissue are highly dependent upon the type of tissue.

The number of biospecimens arriving at the Contractor’s site per month is variable, and can cause fluctuation in resource levels needed to receive and process materials. Nevertheless, the Contractor shall establish a minimum resource level, so that baseline operating capacity exists at all times.

The clinical data collection requirements (i.e. what are the cancer specific data elements) for any particular cancer are not known until TCGA project management convenes the cancer-specific working groups to define those data elements. This may occur after the first shipment of samples from a TSS to the Contractor.

In order to describe the magnitude and variances inherent in performance of this SOW, the Government provides the following information.

For one process described below, pathology verification, two scenarios are provided because the results of this process have proven highly variable. The below information is the basis of the more detailed maximum numerical storage, processing throughput and sample number requirements in the following sections of the SOW.

The following applies to the entirety of the SOW:

The Contractor shall assume a baseline processing level of 200 cases worth of analytes distributed per month 9 months after contract execution.

The Contractor shall assume that each BCR contractor shall have the responsibility to process six

(6) cancers alone and another eight (8) cancers will be shared between the BCR contractors over the term of the contract. For the six (6) cancers processed alone, the Contractor shall be responsible for all 500 cases per cancer output. For the shared cancers, assume 250 cases per cancer output.

The following assumptions apply to the processing of one (1) full cancer (i.e. 500 cases output):

The maximum period for completely processing all cancer types is five (5) years, provided that the contract does not state a shorter period. Individual cancers are to be completed within two (2) years of the start of that cancer type.

The Contractor shall establish administrative and operational relationships with 4 - 16 TSS per cancer. After contract award, the Contractor’s shall submit budget documents provided to NCI under this SOW shall include financial details sufficient for NCI to understand the per TSS costs/savings as the number varies. For example, the BCR may choose to provide budget for ten

(10) TSS and provide variance for +/- six (6) TSS. The provision of these budgets is required during the performance of work under this contract as new cancers and new TSS are engaged.

For the majority of cases, the Contractor shall receive tumor biospecimens on which one surface has been sectioned and histopathologically reviewed to contain tumor cell composition within 10% of the minimum requirement (this value is initially set at 80% for any cancer to be studied by TCGA, but subject to change by TCGA project management). This “pre-screen” will usually have been performed by the TSS. The Contractor, however will propose to perform up to 10,000 pre-screens, and shall submit a budget, with separate line items, that permits the government to understand Contractor prices for this specific activity

For the purpose of increasing success rates through the process of pathology Quality Control, assume that 10% of the frozen tumor tissue specimens shall require some level of manual handling, such as macro-dissection, to enrich for tumor material, prior to being subject to the full pathology quality control regimen.

For the process of pathology Quality Control, the Contractor shall develop plans against two pathology success rate scenarios: assume that 60% (Scenario A) or 40% (Scenario B) of tissue specimens shall meet TCGA pathology quality control metrics. To meet the per cancer output requirements, these numbers are rounded to 1250 and 1875 cases, respectively, for number of cases worth of solid tissue samples that need to enter the BCR pipeline.

For the process of molecular analyte extraction and quality control, assume that 67% of the cases entering molecular processing from pathology QC will pass TCGA molecular analyte metrics.

This number is rounded to 750 cases worth of specimens required to enter molecular processing.

The Contractor should assume that 100% of the cases of molecular analyte aliquots distributed from the Contractor to characterization centers shall be successfully analyzed and yield data sets across all platforms. I.e. the Contractor shall have met its goals upon successful distribution of 500 cases worth of analytes to all characterization centers. However, the Contractor must note that this does NOT mean that occasionally additional aliquots of analyte shall need to be shipped to a particular center due to local failure – this eventuality is addressed in the overage amount under the requirements regarding number of aliquots to be shipped (below).

For the germline DNA source for each cancer, the Contractor shall receive a mix of blood, blood components, DNA and / or normal untreated solid tissue. The Contractor shall assume that 100% of these samples successfully pass through pathology (if relevant), and 67% successfully pass through molecular extraction. The Contractor shall assume that the BCR will receive a minimum of 100 normal untreated solid frozen tissue biospecimens. Of these samples, 100 shall be tissue from the same organ as the tumor, and must be processed through both pathology and molecular extraction for co-isolation of DNA and RNA, for the purpose of providing control analytes for the expression and epigenetic characterization of normal tissue.

Additionally, for each cancer, the Contractor shall assume that the BCR shall receive 100 treated recurrent frozen solid tumor biospecimens. These samples shall be processed through both pathology and molecular extraction. Assume 50% of these samples successfully pass pathology, and 67% of those successfully pass through molecular extraction.

An undetermined number of small production related projects may need to be performed for such purposes as to develop or optimize cancer-specific, tissue-specific, or analyte-specific SOPs.

Sufficient resources required to process samples for these projects shall come from the resources required for production and are accounted for by the broad range (10%-20%) of overages specified in this SOW.

As stated above, the numerical and scale requirements in following detailed SOW sections are predicated on the per cancer assumptions stated above.

Multiple BCRs TCGA expects to analyze up to 20 cancers over the five (5) year course of the project. NCI will determine when a particular cancer is studied. Due to the processing throughput requirements of this number, TCGA will require at least two (2) BCRs. BCR Contractors shall coordinate with each other to harmonize their SOP, and to additionally coordinate with the Data Coordinating Center (DCC) contractor to harmonize data submission formats.

Specific requirements in support of multiple BCRs

Each TSS shall distribute samples and data to only a single BCR contractor that will be assigned by the NCI COTR.

Cancers processed by a particular BCR shall be assigned by the NCI COTR.

All laboratory operations undertaken by BCRs in support of TCGA shall occur according SOP that are harmonized between the BCR contractors. The BCR contractors shall coordinate with each other to develop, improve and deploy SOP that do not differ between BCRs. Inter-BCR communications are a requirement of the Project Management Plan (described below).

All data provided by the BCR contractors to the DCC and TCGA project management and other TCGA components shall be comparable to each other and integrate-able. The BCR contractors shall coordinate with each other and with the DCC on all data standards for data that are distributed in support of the work under this contract. Coordination shall include the standards, format and the mechanism under which data are provided to the recipient.

1. Biospecimen and Associated Data Procurement The contractor shall receive samples and associated data from Tissue Source Sites (TSS) selected by NCI.

(Biospecimen qualification criteria for TCGA Pilot Project are located at:

http://cancergenome.nih.gov/components/hcbcr_process.asp. These criteria are updated as data improve.)

1.1. Specific requirements for biospecimen procurement:

The Contractor shall have sufficient expertise and be of sufficient scale to procure biospecimens from the number (within the range of 4 – 16) of TSSs to meet the biospecimen needs for each cancer. TSS interactions shall include the following activities as necessary to ensure that samples are handled, shipped and received with full maintenance of quality:

o Develop and deploy a TSS communication protocol that governs the Contractor’s communications with the TSS, NCI and potentially other organizations as needed to perform the work herein.

o Develop and deploy a TSS support system to rapidly analyze, understand and correct tissue sample quality issues.

o Conduct biospecimen and data transfer training of TSS repository operators or other staff who shall physically handle tissues and abstract clinical data, respectively. Training shall be conducted using the NCI supplied video for this purpose, but shall be supplemented by direct communication as necessary to ensure proper transfer.

o Travel to TSS for onsite training may be conducted after approval by NCI Contracting Officer. The Contractor should assume that 50% of TSS shall require onsite training by one person.

o Engage in regular contact (at least every 2 weeks) with active TSS (i.e. TSS that have not yet depleted their inventory) to determine TSS administrative status, sample and data review and retrieval status, remaining inventory, and plans for future shipments to BCR.

The Contractor shall execute Material Transfer Agreements (MTA) with each TSS for transfer of identified biospecimens. The NCI shall provide draft MTA documents including a Cover Memo specifying the minimum requirements and NIH policy stipulations for the MTA. The MTA may be modified during negotiations providing that the enumerated key requirements and policies in the Cover Memo are maintained as part of the terms. The contractor shall seek guidance and authorization from the NCI COTR prior to agreeing to any significant changes in MTA terms.

The currently (as of January 2009) used Cover Memo and MTA template are in Appendix 2. The MTA and terms therein, while required under this contract, is a standalone legal agreement between the Contractor and the TSS.

The Contractor shall obtain from contributing TSS necessary human subjects documentation to verify that TSS IRB approval has been obtained to participate in TCGA. Such approval may, but is not limited to, include:

o Materials from deceased donors:

Were originally obtained under IRB approved protocol. The Contractor shall obtain one or more illustrative copies of the original collection protocol’s informed consent.

Are being delivered to TCGA / Contractor either under an IRB approved protocol, an IRB waiver, or documentation that the IRB does not consider the use of these materials from deceased individuals to be human subjects research.

o Materials from living donors:

Have been obtained under an IRB approved protocol, Are being delivered to TCGA / Contractor either under an IRB approved protocol or an IRB waiver.

The Contractor shall obtain one or more illustrative copies of the informed consent under which still-living donors were enrolled.

The Contractor shall retain up-to-date copies of the aforementioned human subjects documentation, and deliver copies of such documentation to NCI and to TCGA network entities as designated by the COTR.

The Contractor shall be of sufficient scale to support the, per cancer, receipt of:

o Up to 1750 (pathology Scenario A) or 2500 (pathology Scenario B) cases worth of biological material, including:

up to 1750 (2500) OCT embedded or cryovial frozen solid tumor tissues up to 1750 (2500) normal frozen solid tissues and / or blood samples (whole blood or fraction and/or buccal cell pellets in DMSO and/or DNA isolated from blood up to 6000 slide mounted sections and corresponding whole slide scanned images from TSS diagnostic blocks, (assumes 8 slides/images per successful case);

up to 1750 (2500) cases worth of diagnostic surgical pathology reports in paper or electronic format o Note: the numbers above include a 20% overage contingency.

The Contractor shall coordinate with designated TSS to ensure that biospecimens are packed appropriately to protect them from loss, damage, and temperature variations during transit. The Contractor shall use vapor phase liquid nitrogen containers to receive frozen biospecimens. The shipping and handling of biospecimens shall meet relevant government regulations for packaging, labeling and shipping of such samples. Air shipments shall conform to the International Air Transport Association (IATA) standards, and ground shipments in the United States shall conform to the Department of Transportation (DOT) standards.

Upon receipt, samples shall be immediately visually inspected to ensure integrity after shipping.

Irregularities in sample quality or quantity shall be immediately communicated to the TSS and the NCI COTR via email per the TSS communication plan required above. This written confirmation should be received within 24 hours of arrival of shipment at a BCR, excluding weekends and holidays. For example, if a shipment is received at 5pm on a Friday afternoon, the TSS will be notified no later than Monday at 5pm unless Monday is a holiday (in which case the site would be notified no later than Tuesday at 5pm).

Biospecimens shall be uniquely identified and tagged with machine readable mechanisms (e.g., barcodes, RFID) upon receipt and registration into an electronic inventory tracking system. If biospecimens are received with such machine-readable labeling already in place, and the technical format of that labeling is supported by the Contractor BCR workflow systems, samples do not need to be relabeled.

All non-transient samples and derivatives, and subdivisions and aliquots thereof, during all steps from receipt, storage, through processing, and shipment of aliquots to TCGA GCC and GSC shall be uniquely identified, tagged and tracked in that electronic inventory…