Description:
Fast-Track proposals will be accepted.
Direct-to-Phase II proposals will not be accepted. Number of anticipated awards: 3-5 Budget (total costs, per award): Phase I: up to $400,000 for up to 9 months; Phase II: up to $2,000,000 for up to 2 years
PROPOSALS THAT EXCEED THE BUDGET OR PROJECT DURATION LISTED ABOVE MAY NOT BE FUNDED.
Summary
Tremendous progress has been made in cancer imaging in the last decade, and much of this is due to computers that have revolutionized imaging protocols and image analysis. Magnetic resonance imaging (MRI), positron emission tomography (PET), computed tomography (CT), ultrasound, optical imaging, and other modalities have become fundamental tools for cancer research and clinical applications. While these medical imaging modalities are the workhorses of cancer prevention, diagnosis, and monitoring, there is increasing evidence that the accuracy of pattern recognition, predictions, and clinical decision making can be improved by the use of machine learning (artificial intelligence; AI) in image analysis. For example, retrospective review of false negative cases indicates that missed cancer diagnosis is often due to misinterpretation of perceived abnormality. In addition, the amount of imaging data has increased tremendously. Rich data have empowered physicians but also challenged them with computational complexity.
As imaging data are continually growing and readily available, they offer an incredibly abundant resource for scientific and medical discovery, particularly in the application of AI for medical imaging. AI, which is defined as an area of computer science that mimics cognitive functions, such as learning and problem solving, has been progressing rapidly over the past decade. More and more physicians have started to recognize that AI-aided imaging tools (e.g. machine learning, or deep
learning based on learning data) could help them with clinical decision making and improve imaging efficiency that would otherwise not be possible.
Project Goals
The goal of this solicitation is to call for development of AI-aided imaging software for cancer prevention, diagnostics, prognostics, and/or response to therapy. The system that will be developed can be used either as a stand-alone package for clinical applications or a tool for facilitating clinical decision making. The AI-based system may also be used to provide a better mechanistic understanding of tumor development and progress with the idea that this knowledge may lead to better therapeutic targets and improve patient outcome. The data sources for cancer imaging can be from conventional X-ray, MRI, PET, CT, ultrasound, optical imaging, and/or other imaging modalities or imaging devices. Since a single imaging modality may not be sufficient to quantitatively process, reconstruct, and analyze specific cancer imaging, integration of images from multi-imaging modalities or imaging devices that could make the system more robust for their technology development is permitted. The sensitivity and specificity for the cancer prevention, diagnosis, and/or monitoring will depend on the clinical question and unmet need that the tool is attempting to answer. Products addressing cancers of the brain, cervix, colon, head and neck, lung, prostate, and rare cancers as well as childhood cancers are particularly encouraged for this topic. However, proposals may be focused on any single cancer type. Cloud-based AI-aided imaging systems are also encouraged.
To apply for this topic, offerors must outline and indicate the clinical question and unmet clinical need in the areas of cancer prevention, diagnosis, and/or monitoring that their AI-aided imaging system will address. Proposals focused on sharing and archiving imaging information, radiation therapy treatment planning, or mammography will not be considered responsive to this solicitation.
Phase I Activities and Deliverables
• Select one modality, or a set of imaging modalities (e.g., MRI, PET, CT, ultrasound and/or optical imaging, etc.), and data sources that are associated with the modalities for the AI-aided imaging software that will be developed for cancer prevention, diagnosis, and/or monitoring
• Perform a software usability study for the prototype software with at least 25 users
• Demonstrate in a small-scale, proof-of-concept study with animal or human medical image data the feasibility of an algorithm and software package for an AI-aided imaging system for cancer prevention, diagnosis, and/or monitoring. This study should be designed to assess the sensitivity and cancer specificity of the prototype software
• Deliver to NCI the SOPs of the system for cancer prevention, diagnosis, and/or monitoring
• Develop a regulatory strategy/plan and timeline for seeking approval from FDA to market the AI-aided imaging software
Phase II Activities and Deliverables
• Engage with FDA to refine the regulatory strategy
• Refine and modify the software based on usability and feasibility data from Phase I
• Perform a large-scale usability study with at least 100 users
• Perform a large-scale validation study with human medical image data. The study should be designed to show a statistically significant improvement in the performance of the AI-aided image software
• In the first year of the contract, provide the Program and Contract officers with a letter(s) of commercial interest
• In the second year of the contract, provide the Program and Contract officers with a letter(s) of commercial commitment
• By the end of Phase II, submit a regulatory application to FDA to obtain marketing approval for the AI-aided software for cancer prevention, diagnosis, and/or monitoring
