Accuracy of virtual crossmatch (VXM) prediction of physical crossmatch (PXM) results of donor specific antibody (DSA) in routine pretransplant settings-a single-center experience.
Complement-dependent cytotoxicity crossmatch
Crossmatch prediction
Flow cytometry crossmatch
HLA antibody
Renal transplantation
Journal
Transplant immunology
ISSN: 1878-5492
Titre abrégé: Transpl Immunol
Pays: Netherlands
ID NLM: 9309923
Informations de publication
Date de publication:
06 2022
06 2022
Historique:
received:
26
10
2021
revised:
15
03
2022
accepted:
15
03
2022
pubmed:
23
3
2022
medline:
25
5
2022
entrez:
22
3
2022
Statut:
ppublish
Résumé
Virtual crossmatch (VXM) is a new powerful tool in pre-transplant risk assessment. However, the ability of VXM to predict physical crossmatch (PXM) results remains controversial. Our work evaluated the predictive potential of VXM results, measured by SAB (single antigen bead assay), for CDCXM (complement-dependent cytotoxicity crossmatch) and FLXM (flow cytometry crossmatch) results of DSA (donor specific antibody) in sensitized patients. In total, 261 CDCXM and FLXM measurements were performed for 180 potential kidney transplant candidates, each with a single HLA-A, B, or -DR DSA against a potential deceased donor. Analysis was conducted with two SAB datasets of four-month distant and collected prior to and after PXM results. Optimal MFI (mean fluorescence intensity) thresholds and likelihood ratios were assigned based on low (<2000 MFI), medium (2001-5000 MFI) and high risk (>5000 MFI). The impact of VXM predictability was determined by the ROC curves comparison. In addition, inter-assay changes of MFI were evaluated. The accuracy of VXM to predict CDCXM was inferior to that of FLXM with the AUC (area under ROC curve) of 0.644 vs. 0.849. In contrast, the initial ROC analysis showed that the VXM prediction was good for both T-FLXM with ROC value of 0.849 and by B-FLXM with ROC value of 0.706 for a single antigen of HLA-A, B, or -DR DSA. In fact, the best VXM prediction was for FLXM with good sensitivity for B-FLXM against HLA-DR-specific DSA (0.851). Similar results of VXM predictability were observed for pre- and post-crossmatch ROC curves. VXM predictability is better for positive/negative FLXM than for positive/negative CDCXM results to evaluate a single HLA-A, B, -DR DSA disparity. This may be related to the fact that VXM and FLXM rely on binding of antibodies to beads or cells, respectively. In contrast, VXM is less predictive for CDCXM because the latter measures complement-dependent cytotoxic function. We intend expand VXM analysis to correlate their results with FLXM results to select low/medium risk patients for kidney transplantation in Poland.
Sections du résumé
BACKGROUND
Virtual crossmatch (VXM) is a new powerful tool in pre-transplant risk assessment. However, the ability of VXM to predict physical crossmatch (PXM) results remains controversial. Our work evaluated the predictive potential of VXM results, measured by SAB (single antigen bead assay), for CDCXM (complement-dependent cytotoxicity crossmatch) and FLXM (flow cytometry crossmatch) results of DSA (donor specific antibody) in sensitized patients.
METHODS
In total, 261 CDCXM and FLXM measurements were performed for 180 potential kidney transplant candidates, each with a single HLA-A, B, or -DR DSA against a potential deceased donor. Analysis was conducted with two SAB datasets of four-month distant and collected prior to and after PXM results. Optimal MFI (mean fluorescence intensity) thresholds and likelihood ratios were assigned based on low (<2000 MFI), medium (2001-5000 MFI) and high risk (>5000 MFI). The impact of VXM predictability was determined by the ROC curves comparison. In addition, inter-assay changes of MFI were evaluated.
RESULTS
The accuracy of VXM to predict CDCXM was inferior to that of FLXM with the AUC (area under ROC curve) of 0.644 vs. 0.849. In contrast, the initial ROC analysis showed that the VXM prediction was good for both T-FLXM with ROC value of 0.849 and by B-FLXM with ROC value of 0.706 for a single antigen of HLA-A, B, or -DR DSA. In fact, the best VXM prediction was for FLXM with good sensitivity for B-FLXM against HLA-DR-specific DSA (0.851). Similar results of VXM predictability were observed for pre- and post-crossmatch ROC curves.
CONCLUSION
VXM predictability is better for positive/negative FLXM than for positive/negative CDCXM results to evaluate a single HLA-A, B, -DR DSA disparity. This may be related to the fact that VXM and FLXM rely on binding of antibodies to beads or cells, respectively. In contrast, VXM is less predictive for CDCXM because the latter measures complement-dependent cytotoxic function. We intend expand VXM analysis to correlate their results with FLXM results to select low/medium risk patients for kidney transplantation in Poland.
Identifiants
pubmed: 35314352
pii: S0966-3274(22)00057-0
doi: 10.1016/j.trim.2022.101583
pii:
doi:
Substances chimiques
Antibodies
0
HLA Antigens
0
HLA-A Antigens
0
Isoantibodies
0
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
101583Informations de copyright
Copyright © 2022. Published by Elsevier B.V.