The complexity and ongoing evolution of the imaging-based data generation methods and post-acquisition analytical schemes in these studies require the operation of the CAT-I to be collaborative, not a service.
At present, all of the research on which CAT-I is based has been technical development that includes mouse, nonhuman primate, and human material or mouse studies in which the 'n' for sample size is quite small and for which the LBS has developed reliable panels of stains. Moving to human work requires development of custom staining panels for the questions relevant to a specific study. CAT-I currently has a panel of validated antibody conjugates [>30 at present and growing], along with additional antibodies chosen by the collaborator once they are also tested, to create 8- to 12-color panels; in some cases, two or more overlapping panels will be needed to fully probe all the markers and cell types of interest. This development of panels requires expendable sample material if available.
Once panel testing is done, CAT-I can begin primary data collection. The images are processed to remove noise, compensated for channel spillover, and background autofluorescence is eliminated. The resulting images can be shared with the collaborator for qualitative assessment of the results, which can often be informative, but the goal is to perform quantitative analysis using the Histo-cytometry method. As with multiplex flow cytometry, there are no widely accepted algorithms for the gating analysis used in Histo-cytometry, and much of this has to be done manually by an expert. CAT-I staff members are working on semi-automating the initial image processing workflow and are in discussions with groups that have developed automated flow gating algorithms to determine whether they can be applied to multiplex imaging datasets. Machine learning algorithms are also under development. However, these preliminary efforts are not at a point where they can be used routinely.
These various considerations make clear that there is substantial, expert-level back-and-forth in each project, both for image production and for intelligent analysis, which is why this would be collaborative science and not core service work for the foreseeable future. Current funding, instrumentation, and staffing should allow three to four major projects (i.e., dozens of samples, new panel development) to be completed in a year, and possibly another two to four smaller studies (i.e., small number of samples, existing staining panels). The first year of operation would be a gradual process of starting small and then, as CAT-I gains experience, taking on more projects when CAT-I’s ability to move projects through the pipeline is better defined.
This affects the way projects will be chosen. In addition to scientific significance and relevance to ICD goals, preference for initial studies will be based on these criteria:
- Capacity to undertake the actual work rapidly; e.g., markers and cells of interest are covered by panels already validated by CAT-I, which would include lymphoid malignancies as an example
- Availability of sufficient clinical material for CAT-I to adjust the staining panels and improve analyses on the fly (e.g., large resections or biopsies and not small cores)
- Likelihood of sufficient sample diversity linked to differential clinical outcome to do useful correlation analysis
- Collaborator(s) with a clear understanding that this is still experimental science
Sample and Data Management
All unstained tissues samples received by CAT-I for collaborative projects will be barcoded and safely stored. These unprocessed samples will be held by the CAT-I or returned to the collaborating investigator upon her/his formal request. Sample identification will conform to guidelines related to handling of human material with potential personally identifiable information. All thin- or thick-cut sections from such samples, stained or unstained, will be barcoded for identification and retained by CAT-I to meet requirements for availability of primary data and for quality assurance checks.
The CAT-I may decide to prepare additional sections and conduct additional analyses that go beyond the original collaborative plan. In such cases, the collaborating investigator will be notified in advance about such possible work, and additional use of the samples will be discussed before the research is conducted. In the case of clinical material, any such additional use of the samples will be in strict accordance with permissions granted in the patient consent for research use of samples.
All raw imaging data will be retained by the CAT-I. Raw data will be provided to collaborators upon request. Copies of all processed image data will be provided to collaborators with appropriate metadata as they are obtained or according to an agreed-upon schedule. Data will be backed up locally and in a redundant manner using central NIAID resources.
All data that are required by regulation to be made available on public repositories will be deposited as necessary. Such deposition will be the responsibility of the CAT-I.
Conflicts of Interest
The various studies to be conducted in collaboration with CAT-I may involve investigators, academic institutions, organizations, or commercial firms with competing interests or overlapping goals. All collaborators will need to agree in writing before a study is approved for execution that they understand it is impossible for the CAT-I staff to isolate themselves intellectually from the observations made during the conduct of potentially competing studies or from having such information affect their interpretations of data emerging from another study. At the same time, the CAT-I staff will do everything possible to avoid directly communicating information learned in the course of one study during discussions of data from another, and imaging data will not be shared without written permission of the collaborating investigator.
The CAT-I is a scientific collaborative enterprise and, as such, will not charge a fee for collaborative work. However, the imaging studies to be undertaken have both a significant direct cost and an amortized cost, mainly in terms of staff and equipment. NIAID and NCI have agreed to cover much of the latter and, at least during start-up operation in the first year, no cost factor will be assigned to projects for staff or equipment.
For commercial entities, a research collaborative agreement (RCA) will be used for proof-of-principle pilot studies at CAT-I expense; further work will require a cooperative research and development agreement (CRADA) that includes an agreed-upon level of support from the company for the project to be executed that accounts for both running costs and infrastructure costs.
For intramural and extramural academic investigators, collaborators will be asked to contribute if possible to the cost of antibody reagents for all but very small projects. If a collaborator cannot cover such costs but the project is judged of high scientific merit, the CAT-I may choose to support the project costs from its own budget.
In addition to execution of projects based on established technology, the CAT-I will constantly examine the landscape for methods or devices developed by academic or commercial entities that could improve sample preparations and analysis. This will require testing published methods or undertaking demonstrations of new equipment. The CAT-I will also seek to improve its present methods and to develop new technologies; this will require staff and instrument time. A balance will be sought between rapid execution of multiple collaborative studies and these investments to ensure that the imaging done by the CAT-I remains state of the art.