Dr Henry HL Wu1,2,3, Dr YANDONG LANG3, Mr ADNAN AGHA3, Prof EWA M GOLDYS3, Prof CAROL A POLLOCK1,2, A/Prof SONIA SAAD1
1Renal Research, Kolling Institute, Royal North Shore Hospital & The University Of Sydney, Sydney, Australia, 2Department of Renal Medicine, Royal North Shore Hospital, Northern Sydney Local Health District, Sydney, Australia , 3ARC Centre of Excellence for Nanoscale Biophotonics, The University of New South Wales, Sydney, Australia
Biography:
Dr Henry HL Wu is a Registrar at the Department of Renal Medicine, Royal North Shore Hospital; Doctoral Researcher at the Renal Research Laboratory, Kolling Institute, University of Sydney; and Visiting Research Fellow at the ARC Centre of Excellence in Nanoscale BioPhotonics, UNSW. He was previously NIHR Academic Clinical Fellow in Nephrology at the University of Manchester, UK from Aug 2019 to Dec 2021. Henry is Associate Editor for Nutrition & Health, Review Editor for Frontiers in Medicine, and Editorial Board Member in BMC Nephrology and PloS One. He is also a current member of the ISN CardioRenal Toolkit Workgroup.
Aim:
Evaluate the ability of a novel non-invasive methodology (Patent AU2024904153) to identify between post-transplant acute tubular necrosis (ATN), rejection and non-rejection associated interstitial fibrosis and tubular atrophy (IFTA).
Background:
There is no validated method apart from transplant biopsy to accurately identify between the histopathological causes of graft dysfunction. Our group developed a diagnostic methodology based on machine-learning to assess autofluorescence signals emitted from urinary exfoliated proximal tubule cells (PTCs) to reflect renal pathology. Hence, we hypothesized that assessing exfoliated PTCs can discriminate between the causes of kidney transplant complications, which is critical as accurate diagnosis of post-transplant complications would determine the course of appropriate therapeutic intervention.
Methods:
30 patients diagnosed with ATN (10 patients), rejection (10 patients) or IFTA (10 patients) by transplant biopsy were included. Urine collected immediately prior to biopsy were processed and exfoliated PTCs were extracted using a validated immunomagnetic separation method based on anti-CD13 and anti-SGLT2 antibodies. Multispectral autofluorescence images of PTCs were quantitatively analyzed. The optimal sets of autofluorescence feature combinations which significantly discriminate between the groups when comparing ATN versus rejection, ATN versus IFTA, and rejection versus IFTA, were identified. Binary classification was performed by the AutoGluon machine-learning software with 5-fold cross-validation.
Results:
Binary classification using the optimal feature combinations identified for each group comparison resulted in area under the curve values of 0.95±0.05 (ATN versus rejection), 0.92±0.08 (ATN versus IFTA) and 0.91±0.09 (rejection versus IFTA).
Conclusions:
Our results demonstrate that machine-learning assessment of urinary exfoliated PTCs multispectral autofluorescence could accurately differentiate between ATN, graft rejection and IFTA in kidney transplant recipients. Our methodology has tremendous potential going forward as a non-invasive tool to guide post-transplant management.