Background: Acute rejection of renal transplant, i.e., the immunological response of the human body to a foreign kidney, is the main cause of dysfunction in renal transplant patients, surpassing acute tubular necrosis (ATN) and immune drug toxicity. Over 650,000 patients in the US have end-stage renal disease and renal transplant offers the best outcome for these patients. However, 15%-27% of renal transplant patients have ARTR within 5 years, which if not detected and treated promptly, causes renal damage and leads to renal failure of the transplanted kidney. Currently, ARTR is initially evaluated using blood tests and urine sampling, e.g., plasma creatinine and creatinine clearance. However, these indices have low sensitivity, since a significant change in creatinine levels is only detectable after the loss of 60% of renal function. ARTR diagnosis requires renal biopsy (gold standard), which is invasive, takes a week for diagnosis, and costs up to $20,000. Further, needle biopsy is prone to over- or under-estimation of inflammation. The proposed study seeks to develop and validate a new noninvasive computer-aided diagnostic (CAD) software system that can provide highly accurate (over 97% accuracy) diagnosis of ARTR at an early stage by integrating laboratory-based biomarkers (e.g., creatinine clearance (CrCl) and serum plasma creatinine (SPCr)) with imaging-markers that will be extracted from either dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) or diffusion- weighted magnetic resonance imaging (DW-MRI). Objective/Hypotheses: The primary objective of this proposal is to develop and validate a clinically usable system for the early detection of acute renal transplant rejection (ARTR) based on testing the following three hypotheses: (i) The perfusion of the contrast agent in the cortex of the rejected kidney is slower than the perfusion of the contrast agent in the non-rejected kidney; (ii) The estimated apparent diffusion coefficient (ADC) in the cortex of the rejected kidney is smaller than the estimated ADC in the cortex of the non-rejected kidney; and (iii) Fusing laboratory-based biomarkers (e.g., CrCl and SPCr) with imaging-markers that will lead to novel fused indices for distinguishing rejected versus non-rejected kidneys in clinical decision-making.
Specific Aims : (i) Develop a software CAD system for early detection of acute renal transplant rejection based on integrating/fusing the estimated physiological perfusion parameters from DCE-MRI data with laboratory-based biomarkers for both rejected and non-rejected transplanted kidneys; (ii) Develop a software CAD system for early detection of acute renal transplant rejection based on integrating/fusing the estimated physiological diffusion parameters from DW-MRI data with laboratory-based biomarkers for both rejected and non-rejected transplanted kidneys; and (iii) Evaluate the accuracy of the proposed image analysis software system with clinical data that will be collected at University of Louisville (Approved IRB # 16.0041), University of Michigan (Approved IRB # HUM00115031), and the collected MRI data at The Kidney transplant center at Mansoura University with local regulatory approval. Study Design: We will develop the following image analysis techniques to accurately analyze both DCE- MRI and DW-MRI: (i) segmentation of the kidney object from the surrounding abdominal structures for either DCE-MRI or DW-MRI using appearance and prior shape models; (ii) a Laplace-based non-rigid registration approach to handle local deformations in the kidney region caused by physiological effects (e.g., intrinsic pulsatile effects, breathing, or transmitted effects from adjacent structures, such as the bowel) during the data collection; (iii) segmentation of the kidney cortex and generation of either the physiological perfusion parameters (e.g., wash-in slope, wash-out slope, time-to-peak, and peak value) or physiological diffusion parameters (e.g., ADC); (iv) classification of the acute rejection versus non-rejection kidney status and evaluating the method as a diagnostic test; and (v) depiction of parametric maps of the estimated perfusion or diffusion indices to help clinical visual assessment of the transplanted kidney. The MRI and clinical data will be obtained from patients who have undergone kidney transplantation from live donors, and are available to our research team via Dr. Beache and Dr. Dwyer (Radiology Department and The Kidney Transplant Center at the University of Louisville, respectively) and Dr. Samaniego-Picota (Director of the Kidney Transplant Program and the Transplant Ambulatory Care Unit at the University of Michigan). In addition to the data that will be collected at the University of Louisville and the University of Michigan, our group has access to MRI data that had been collected from an additional 100 subjects at Mansoura University Hospital, Egypt. This data is available to our group through Dr. Abou El-Ghar, one of the PI collaborators, and it will be used to test the proposed CAD system on the data collected from another geographical area. Our preliminary diagnostic results, based on analysis of DCE-MRI data from 54 independent in-vivo cases using leave-one-subject-out approach were 98% accurate (sensitivity = 100.00%, specificity = 96%, and Dice similarity coefficient (DSC) = 98%). Moreover, our preliminary results, based on the analysis of DW-MRI data from 75 independent in-vivo cases using leave-one-subject-out were 97% accurate (sensitivity = 98%, specificity = 96%, and DSC = 98%). These results demonstrate the potential of the proposed CAD system to complement current technologies such as nuclear imaging and ultrasonography to determine the type of kidney dysfunction post transplantation.

Agency
National Institute of Health (NIH)
Institute
National Institute of Allergy and Infectious Diseases (NIAID)
Type
Academic Research Enhancement Awards (AREA) (R15)
Project #
1R15AI135924-01A1
Application #
9590370
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Minnicozzi, Michael
Project Start
2018-05-10
Project End
2021-04-30
Budget Start
2018-05-10
Budget End
2021-04-30
Support Year
1
Fiscal Year
2018
Total Cost
Indirect Cost
Name
University of Louisville
Department
Type
University-Wide
DUNS #
057588857
City
Louisville
State
KY
Country
United States
Zip Code
40292