High-performance cloud computing (HPCC) is an integration of high-performance computing (HPC) with cloud computing that provides an alternative informatics infrastructure to deliver the supercomputer power needed for developing and reading high quality, multidimensional diagnostic images on desktop or mobile devices. Such infrastructure would advance the use of high-throughput cancer screening like virtual colonoscopy (VC), also known as computed tomography colonography (CTC). The goals of this proposal are to develop and validate the clinical benefits of a mobile HPCC-Virtual Colonoscopy decision support system (HPCC-VC) that integrates novel high-performance electronic cleansing and computer-aided detection schemes. The combination of high performance electronic cleansing (hpEC) with high performance computer aided detection (hpCAD) will allow visualization of the entire mucosal surface of the colon without artifact. Specifically, we hypothesize that the mobile HPCC-VC system will improve the quality of electronic cleansing of non-cathartic CTC (ncCTC) images, will deliver images 100 times faster than conventional approaches, and will improve reader performance in the detection of colonic lesions in the images analyzed on mobile high-resolution display devices. To achieve these goals, an HPCC platform will be established by use of a CPU-cluster-based supercomputing and parallel image processing libraries. Then, an hpEC scheme will be developed that effectively removes the residual fecal materials in ncCTC images. In parallel, a high-resolution hpCAD scheme will be developed to support high-performance detection of colonic lesions in ncCTC. These schemes will be integrated into a mobile HPCC-VC system on the HPCC platform. Necessary preliminary studies in the development of computed EC and CAD schemes show promise;and the development and deployment of the HPCC platform will advance the quality, speed, and utility of high performance, multidimensional imaging for colon cancer screening. A comprehensive reader performance study will be conducted to determine the clinical application and benefit of images analyzed and delivered through an HPCC platform. Successful development of HPCC-VC will demonstrate the clinical benefit of the platform for improved diagnostic imaging and facilitation of accurate, high-throughput colon cancer screening that is highly acceptable to patients. In the longer term, broad adoption and use of the HPCC-VC system will facilitate early and accurate diagnoses, and thus reduce mortality from colon cancer.

Public Health Relevance

Successful development of the HPCC-VC system will demonstrate the clinical benefit of HPCC platform for diagnostic imaging, and will provide a high-throughput colon cancer screening scheme for colorectal lesions that is highly acceptable to patients and highly accurate. Such a system will promote the early diagnosis of colon cancer, and ultimately reduce the mortality due to colon cancer.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Research Project (R01)
Project #
5R01CA166816-02
Application #
8494421
Study Section
Biomedical Computing and Health Informatics Study Section (BCHI)
Program Officer
Henderson, Lori A
Project Start
2012-07-01
Project End
2016-04-30
Budget Start
2013-05-01
Budget End
2014-04-30
Support Year
2
Fiscal Year
2013
Total Cost
$339,387
Indirect Cost
$144,337
Name
Massachusetts General Hospital
Department
Type
DUNS #
073130411
City
Boston
State
MA
Country
United States
Zip Code
02199
Yoshida, Hiroyuki; Wu, Yin; Cai, Wenli (2014) Analysis of scalability of high-performance 3D image processing platform for virtual colonoscopy. Proc SPIE Int Soc Opt Eng 9039:90390U