An estimated 1.4 million people in the United States suffer from inflammatory bowel disease (IBD), half of whom are believed to have Crohn's disease (CD), one of two primary forms of IBD along with ulcerative colitis. At least 10% are <18. Our response to Innovations in Biomedical and Computational Science and Technology under PAR-07-344, Novel MRI Imaging Tools and Software for Assessing Pediatric Crohn's Disease (pCD), is aimed at developing and refining a new type of parametric imaging- accelerated spatially constrained incoherent motion MRI (aSCIM-MRI)-as a highly accurate quantitative biomarker for cell proliferation, density and size, and tissue perfusion-all indices tat characterize the extent of disease activity (i.e., inflammation) in the tissue micro-structure of te bowel. If successful, this non-invasive, radiation-free technique will constitute a dramatic improvement over current reference standards (i.e., magnetic resonance enterography (MRE), clinical exam, blood tests, and histology), separately, and in combination. Specifically, a-SCIM-MRI is expected to substantially improve our ability to assess inflammatory activity in pCD; monitor response-to-therapy; evaluate the need for surgical intervention; develop treatment plans that are tailored to individual disease profiles; and predict the likelihood of recurrence. T these ambitious ends, we will undertake the following Specific Aims: 1) to determine whether a spatially constrained signal decay model (SCIM-MRI) improves the reliability of fast and slow diffusion quantification from diffusion-weighted MRI (DW-MRI); 2) to accelerate SCIM-MRI acquisition time with a Bayesian model-based reconstruction (aSCIM-MRI); and 3) to assess the efficacy of fast and slow diffusion components at distinguishing active inflammation from fibrosis in pCD as determined by histopathological findings. With the support of the NIH, these anticipated research accomplishments will result in vastly improved management of a most debilitating bowel disease, the diagnosis of which is often elusive, and on diagnosis, challenging to treat; in part because optimal therapies are somewhat limited for children; and in part because the target population is developmentally fragile to begin with, and thus more susceptible to toxicity from strong biologic drugs and to the complications (both short- and long-term) associated with surgery. Given the fact that up to 75% of children with CD must undergo a bowel resection as some point in their lives, and given the fact that these surgeries are rarely curative; the demand to improve assessment capabilities has never been greater. This highly innovative imaging approach, aSCIM-MRI, is therefore expected not only to create a new reference standard by which pCD is evaluated, monitored, and treated; it is also expected to have rapid translational impact once it is introduced into routine clinical imaging. A second, important overall goal of this project is to develop and broadly disseminate open source software will enable the standardized evaluation of other diseases that are presently evaluated with DW-MRI and would benefit from the advanced diagnostic and assessment capabilities of aSCIM-MRI.
Pediatric Crohn's disease (pCD), a chronic inflammatory bowel disease, affects an estimated 80,000 children and adolescents in North America. Traditional radiographic methods of assessment are not sensitive to crucial inflammatory disease processes and expose the patient to the potentially harmful effects of ionizing radiation. Thus, there is an urgent need for an accurate, safe, reliable imaging technique for assessing bowel inflammation in pCD that will enable improved disease management, therapies, and patient outcomes. The primary objective of this project is two-fold: first, to develop and validate a novel imaging acquisition and analysis technique known as aSCIM-MRI that will enable clinicians to more accurately assess bowel inflammation; and second, to make aSCIM-MRI software available in multiple clinical domains that presently evaluate various diseases with MRI and would benefit from the improved diagnostic and assessment capabilities of this computationally-driven technique.
|Lamash, Yechiel; Kurugol, Sila; Warfield, Simon K (2018) Semi-Automated Extraction of Crohns Disease MR Imaging Markers using a 3D Residual CNN with Distance Prior. Deep Learn Med Image Anal Multimodal Learn Clin Decis Support ( 11045:218-226|
|Haghighi, Marzieh; Warfield, Simon K; Kurugol, Sila (2018) AUTOMATIC RENAL SEGMENTATION IN DCE-MRI USING CONVOLUTIONAL NEURAL NETWORKS. Proc IEEE Int Symp Biomed Imaging 2018:1534-1537|
|Sourati, Jamshid; Gholipour, Ali; Dy, Jennifer G et al. (2018) Active Deep Learning with Fisher Information for Patch-wise Semantic Segmentation. Deep Learn Med Image Anal Multimodal Learn Clin Decis Support ( 11045:83-91|
|Kurugol, Sila; Freiman, Moti; Afacan, Onur et al. (2017) Motion-robust parameter estimation in abdominal diffusion-weighted MRI by simultaneous image registration and model estimation. Med Image Anal 39:124-132|
|Kurugol, Sila; Marami, Bahram; Afacan, Onur et al. (2017) Motion-Robust Spatially Constrained Parameter Estimation in Renal Diffusion-Weighted MRI by 3D Motion Tracking and Correction of Sequential Slices. Mol Imaging Reconstr Anal Mov Body Organs Stroke Imaging Treat 10555:75-85|
|Kurugol, Sila; Freiman, Moti; Afacan, Onur et al. (2016) Spatially-constrained probability distribution model of incoherent motion (SPIM) for abdominal diffusion-weighted MRI. Med Image Anal 32:173-83|
|Kurugol, Sila; Freiman, Moti; Afacan, Onur et al. (2015) Motion Compensated Abdominal Diffusion Weighted MRI by Simultaneous Image Registration and Model Estimation (SIR-ME). Med Image Comput Comput Assist Interv 9351:501-9|
|Taimouri, Vahid; Afacan, Onur; Perez-Rossello, Jeannette M et al. (2015) Spatially constrained incoherent motion method improves diffusion-weighted MRI signal decay analysis in the liver and spleen. Med Phys 42:1895-903|
|Stamm, Aymeric; Singh, Jolene; Afacan, Onur et al. (2015) Analytic quantification of bias and variance of coil sensitivity profile estimators for improved image reconstruction in MRI. Med Image Comput Comput Assist Interv 9350:684-691|