Inactive

POP 28 Weeks from date contract received. NHLBI requires expert support (Computer Programming and Data Processing of a Custom Hybrid CT System) in the detector and hardware geometric calibration of th... POP 28 Weeks from date contract received. NHLBI requires expert support (Computer Programming and Data Processing of a Custom Hybrid CT System) in the detector and hardware geometric calibration of the ultra-high resolution hybrid CT scanner on a per-scan basis, and on-going maintenance and update of the computer programming code for these calibrations; to obtain service in manual segmentation and labeling of pathologic features in the chest scans of lung disease patients as training data for AI algorithms that will be able to automate the process in large cohorts of patients. The performance of the contract requires complete familiarity and expert knowledge of the hybrid CDI-CT scanner developed in LIP/BBC, and ample first-hand experience with the scan and image reconstruction pipeline, including the various calibration steps; familiarity with the data handling pipeline of the CT scanner from the Radiology PACS servers to NHLBI data storage; expertise in writing and modifying the calibration codes; familiarity with the CDI-CT scanner development. The mission of the Laboratory of Imaging Physics (LIP) of BBC, NHLBI is to contribute to the overall mission of NHLBI by developing biomedical imaging methods for better diagnosis and treatment of disease and for enhancing basic science research for the benefit of public health. As part of the effort, we developed a prototype hybrid CT technology for ultra-high resolution chest and body CT scans down to 150 micrometer resolution, to improve the precision in characterizing pathologic changes in the chest of patients with lung disease and in other parts of the body. Preliminary data from 8 patients with the lung disease lymphangioleiomyomatosis Page 2 of 3 proved the concept and pointed to further technical improvement to allow routine use with efficient workflow(1). Simultaneously, the large amount of detailed images from the CT scans need to be analyzed and interpreted in a timely way, which requires automation using AI algorithms. A pre-requisite for training the deep-learning neural network is manual segmentation and labeling of pathologic features in the images of a variety of patients and scan conditions. This is a very laborious task.