Fast-Track proposals will be accepted.
Number of anticipated awards: 2-3
Budget (total costs, per award):
PROPOSALS THAT EXCEED THE BUDGET OR PROJECT DURATION LISTED ABOVE MAY NOT BE FUNDED.
Radiotherapy, both with or without systemic therapy, is administered to over half a million patients annually in the United States alone. The decision regarding what kind of radiotherapy to employ (e.g., Intensity Modulated Radiation Therapy, Proton therapy, Stereotactic Body Radiation Therapy, etc.) and which, if any, drugs to add to radiotherapy for a patient are usually based upon rather crude criteria (e.g., age, TNM stage, histological grade, etc.) that frequently fail to accurately predict the outcome of the treatment administered. Better tools are needed to improve decision-making and thereby decreasing both over and under-treatment.
For more than two decades, almost every patient treated by radiotherapy in the United States has undergone a “treatment planning” CT scan; some patients also undergo MRI or PET scans in addition to CT scans. Recent advances in image analysis, pattern recognition, and data characterization enable high throughput extraction of quantitative imaging features from these images. This emerging field of imaging studies (“Radiomics”) allows us to quantify various tumor phenotypes that can be visualized non-invasively by analyzing numerous imaging features such as tumor shape, boundary features, tumor size, texture, uptake or density distributions, etc. These data can be combined with other patient data and be mined with sophisticated bioinformatics tools to develop models that may improve diagnostic, prognostic, and predictive accuracy.
Radiomic tools thus could be used with treatment planning CT and other scans to extract a treasure trove of information that by using the right tools/algorithms could help greatly improve decision making in radiotherapy. This topic is in line with the Cancer Moonshot Blue Ribbon Panel’s Recommendation to support Development of New Enabling Cancer Technologies.
Radiomic studies for lung cancer using CT and FDG PET images have shown that tumor image features and parameters can describe the nature of disease and predict patient outcomes. Other localized diseases such as brain, breast, kidney lung, liver, and esophageal cancers have also been analyzed with different imaging modalities such as FDG PET, CT, MRI, and ultrasound. Studies have shown that Radiomics has the potential to impact clinical care by contributing to cancer diagnosis, assessing tumor prognosis, assisting in biopsy decision and helping to select the right chemotherapeutic regimen. Computer aided diagnostic tools are being developed for diagnostic radiology but so far the tools for radiotherapy decision making remain sparse. Radiotherapy treatment prescription involves:
1) selecting the type of radiation (e.g., 3-dimensional conformal RT, Intensity modulated RT, Stereotactic body RT, Stereotactic radiosurgery, Proton RT, Carbon ion RT, Low dose rate brachytherapy, High dose rate brachytherapy, etc.);
2) selecting drugs that can enhance the effects of radiation on tumors (e.g., cisplatin, temozolomide, cetuximab, mitomycin, gemcitabine, etc.);
3) selecting drugs that can decrease the effects of radiation on organs-at-risk (e.g., amifostine, memantine, etc.); and,
4) selecting the total dose of radiation, the dose per fraction, the number of fractions of radiation, the sequencing of those fractions with the drugs, etc.
At present the radiotherapy prescription is too often "one size fits all" and based upon relatively rudimentary criteria such as age, TNM stage, and histological grade. Studies that explore the utility of radiomics have indicated that the use of image analysis tools will help refine and personalize cancer decision-making, thereby increasing tumor control and decreasing adverse effects. Furthermore, their use may facilitate more robust "mid-course corrections" (i.e., adaptive therapy), since CT scanning for readjustment of radiotherapy treatment plans is often repeated during a course of radiotherapy on account of weight loss or tumor shrinkage.
The short-term goal of this contract topic is to develop new approaches and refine existing “radiomics” tools for radiotherapy treatment planning images to enable more accurate decision making support for radiation therapy treatment planning. These can include (but are not limited to) the following:
It is expected that the proposed innovation will be driven by clinical practice. Therefore, in addition to standard proposal components; the contract proposal must contain specific discussion of the target patient population and evidence of an existing clinical problem which is addressed by the proposed method. The proposal must also contain an analysis of competitive methods to address the same problem and an explanation of competitive technical advantages of the proposed algorithm. All Phase II or Fast-Track proposals MUST contain a section entitled “Regulatory Plan” that 1) demonstrates an understanding of the regulatory requirements for clearing the software device through the FDA, if appropriate; 2) details the company’s plan to meet the requirements, and 3) explains how the proposed work helps to meet these requirements. If regulatory approval is not expected to be required, the offeror must provide an extensive justification for this. The long-term goal of this program is to eventually commercialize an image analysis software toolkit for decision support in radiation oncology.
Activities not supported by this topic:
Development of algorithms for image acquisition and/or routine image processing tools is not appropriate for this topic and will not be considered for funding. Development of computer aided diagnosis/detection systems not intended for radiation oncology are also not appropriate and will not be considered for funding.
Closing date: October 20, 2017, 5:00 PM Eastern Daylight Time
Apply for this topic on the Contract Proposal Submission (eCPS) website.
Full FY18 Contract Solicitation is available HERE.