NATIONAL CANCER INSTITUTE BIOMEDICAL IMAGING PROGRAM New NCI

NATIONAL CANCER INSTITUTE BIOMEDICAL IMAGING PROGRAM

New NCI Initiatives for CAD Laurence P. Clarke Ph. D. Barbara Y. Croft Ph. D. Edward Staab MD. Daniel Sullivan MD.

NCI Division of Cancer Treatment and Diagnosis (DCTD) Cancer Diagnosis Program (CDP) Developmental Therapeutics Program (DTP) Cancer Therapy Evaluation Program (CTEP) Radiation Research Program (RRP) Biomedical Imaging Program(BIP) Diagnostic Imaging Branch Molecular Imaging Branch Image-guided Therapy Branch Imaging Technology Development Branch

NIH Funding for Imaging $Millions

BIP Grant Funding $M

Branch: Imaging Technology • Promoting new areas for imaging technology development. • Development of new funding initiatives for academia and industry, including SBIR/STTR’s. • Development of research resources and centers of excellence. • Promoting Inter-agency programs.

NIH Initiatives: Image Data Bases • NLM/NSF Visible Human Project (1989 -): CT/MRI images of to allow evaluation of image registration and segmentation (normal subjects/cadavers). • NIH/NSF: Human Brain Project (1993 -). Common tools to facilitate research on neuroinformatics (CT/PET/MRI). • NCI Database (2000 -2005): Focused on evaluation of diagnostic software for specific cancer applications and impact of image pre- processing( image registration).

Rationale: Image Database • CAD methods: Rapidly being developed for more efficient analysis of screening studies (breast, lung and colon). • Clinical Model: Spiral CT for lung cancer screening for high risk patients: Opportunity: access to clinical trials. • Physical Model: Excellent for evaluating 3 D CAD, image registration and temporal analysis of serial CT images, multi-modality image registration, image reconstruction.

Objectives: Image Database Resource • Develop consensus for generation of retrospective image databases, histology confirmation and spatial ground truth, metrics and statistical methods for software evaluation. • Provide an objective means to compare CAD, image processing, change analysis, and image reconstruction software methods. • Stimulate development and early identification of promising software methods and reduction of research risk for academia and industry research sites. • Enhance inter-disciplinary research collaboration by academia, federal laboratories and industry.

Subtle and Very Subtle Nodules Imaged on CT scan

CU 0138 -1 Jun 17 1996, (X, Y) resolution: 0. 68359 mm, Slice thickness : 1 mm A. P. Reeves, Cornell University, 1999

CU 0138 -1 3 D Views Volume: 193. 531 (mm 3) Extents (x, y, z): 11, 10, 7 (mm) Length, Width, Height (EOI): 9. 63099, 7. 67993, 4. 99717 (mm) Aspect Ratios (L/H, L/W, W/H): 1. 92729, 1. 25405, 1. 53686 A. P. Reeves, Cornell University, 1999

U 01: Cooperative Agreement • Support 5 academic sites with fundable scores, selected by NCI, with ongoing lung cancer screening programs, to form a Lung Image Data Base Consortium (LIDC). • LIDC: Minimum of 3 sites and two X-ray CT systems. • U 01 “Assistance mechanism” as opposed to a “contract”. • Steering committee: PI and Co-PI’s and NCI Program staff; responsible forming consensus on research tasks. • Concept Support: $7 M for 5 years for 5 sites.

Future Plans: Technology Development • Proposed PA: R 21/R 33 phased innovation for imaging technology development. • RFP for ultrasound research interface • RFA for spiral CT image database • Proposed additional Centers with technology development component

NCI BIP Web. Page • URL = www. nci. nih. gov/dip/ • Email Newsletter : js 95 x@nih. gov • lclarke@mail. nih. gov • croftb@dtpepn. nci. nih. gov