Early Assessment of Acute Renal Rejection Post-transplantation: A Combined Imaging and Clinical Biomarkers Protocol

Document Type

Conference Proceeding

Source of Publication

2018 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2018

Publication Date

2-14-2019

Abstract

© 2018 IEEE. Non-invasive evaluation of renal transplant function is crucial. Hence, a computer-assisted diagnostic (CAD) system is introduced in this paper to evaluate kidney function post-transplantation. The developed CAD system integrates clinical-based with diffusion weighted (DW) MR image-based biomarkers. The latter are derived from 3D DW-MRIs at multiple strengths and duration of the magnetic field (i.e. b-values). These DW-MRI scans were acquired at multiple geographical areas (Egypt and USA) using different scanner types (GE and Philips). The developed CAD system first segments kidneys using level-sets method and then estimates the DW-MRI image-markers, known as apparent diffusion coefficients (ADCs), from the segmented kidney. Then, the clinical biomarkers (serum creatinine and creatinine clearance) are integrated with the DW-MR image-markers (ADCs) resulted in new integrated markers known as integrated ADCs (IADCs). These IADCs are then used to construct cumulative distribution functions (CDFs) at multiple b-values. Finally, these markers (i.e. CDFs of the IADCs) are used to assess renal transplant status using different classifiers. Our CAD system demonstrates an almost consistent accuracy of 93%, sensitivity of 93%, and specificity of 92% in distinguishing acute rejection (AR) from non-rejection (NR) renal transplants, making the proposed diagnostic platform independent from the geographical area, scanner type, and classifier. These promising preliminary results are of high diagnostic accuracy and suggest that the developed CAD system might be noninvasively able to diagnose renal allograft status.

ISBN

9781538675687

Publisher

Institute of Electrical and Electronics Engineers Inc.

First Page

297

Last Page

302

Disciplines

Computer Sciences

Keywords

CAD, Integrated ADCs, Renal transplants

Scopus ID

85063443936

Indexed in Scopus

yes

Open Access

no

Share

COinS