Generating automated kidney transplant biopsy reports combining molecular measurements with ensembles of machine learning classifiers

Precision diagnostics in transplanted organs using microarray-assessed gene expression: concepts and technical methods of the Molecular Microscope® Diagnostic System (MMDx)

There is a major unmet need for improved accuracy and precision in the assessment of transplant rejection and tissue injury. Diagnoses relying on histologic and visual assessments demonstrate significant variation between expert observers (as represented by low kappa values) and have limited ability to assess many biological processes that produce little histologic changes, for example, acute injury. Consensus rules and guidelines for histologic diagnosis are useful but may have errors. Risks of over- or under-treatment can be serious: many therapies for transplant rejection or primary diseases are expensive and carry risk for significant adverse effects. Improved diagnostic methods could alleviate healthcare costs by reducing treatment errors, increase treatment efficacy, and serve as useful endpoints for clinical trials of new agents that can improve outcomes. Molecular diagnostic assessments using microarrays combined with machine learning algorithms for interpretation have shown promise for increasing diagnostic precision via probabilistic assessments, recalibrating standard of care diagnostic methods, clarifying ambiguous cases, and identifying potentially missed cases of rejection. This review describes the development and application of the Molecular Microscope® Diagnostic System (MMDx), and discusses the history and reasoning behind many common methods, statistical practices, and computational decisions employed to ensure that MMDx scores are as accurate and precise as possible. MMDx provides insights on disease processes and highly reproducible results from a comparatively small amount of tissue and constitutes a general approach that is useful in many areas of medicine, including kidney, heart, lung, and liver transplants, with the possibility of extrapolating lessons for understanding native organ disease states.

Cherry on Top or Real Need? A Review of Explainable Machine Learning in Kidney Transplantation

Research on solid organ transplantation has taken advantage of the substantial acquisition of medical data and the use of artificial intelligence (AI) and machine learning (ML) to answer diagnostic, prognostic, and therapeutic questions for many years. Nevertheless, despite the question of whether AI models add value to traditional modeling approaches, such as regression models, their "black box" nature is one of the factors that have hindered the translation from research to clinical practice. Several techniques that make such models understandable to humans were developed with the promise of increasing transparency in the support of medical decision-making. These techniques should help AI to close the gap between theory and practice by yielding trust in the model by doctors and patients, allowing model auditing, and facilitating compliance with emergent AI regulations. But is this also happening in the field of kidney transplantation? This review reports the use and explanation of "black box" models to diagnose and predict kidney allograft rejection, delayed graft function, graft failure, and other related outcomes after kidney transplantation. In particular, we emphasize the discussion on the need (or not) to explain ML models for biological discovery and clinical implementation in kidney transplantation. We also discuss promising future research paths for these computational tools.

Basics and Art of Immunosuppression in Liver Transplantation

Liver transplantation is one of the most challenging areas in the medical field. Despite that, it has already been established as a standard treatment option, especially in decompensated cirrhosis and selected cases of hepatocellular carcinoma and acute liver failure. Complications due to graft rejection, including mortality and morbidity, have greatly improved over time due to better immunosuppressive agents and management protocols. Currently, immunosuppression in liver transplant patients makes use of the best possible combinations of effective agents to achieve optimal immunosuppression for long-term graft survival. Induction agents are no longer used routinely, and the aim is to provide minimal immunosuppression in the maintenance phase. Currently available immunosuppressive agents are mainly classified as biological and pharmacological agents. Though the protocols may vary among the centers and over time, the basics of effective use usually remain similar. Most protocols use the combination of multiple agents with different mechanisms of action to reduce the dose and minimize the side effects. Along with the improvement in operative and perioperative techniques, this art of immunosuppression has contributed to the recent progress made in the outcomes of liver transplants. In this review, we will discuss the various types of immunosuppressive agents currently in use, the different protocols of immunosuppression used, and the art of optimal use for achieving maximum immunosuppression without increasing toxicity. We will also discuss the practical aspects of various immunosuppression regimens, including drug monitoring, and briefly discuss the concepts of immunosuppression minimization and withdrawal.

Distinct Molecular Processes Mediate Donor-derived Cell-free DNA Release From Kidney Transplants in Different Disease States

BACKGROUND

Among all biopsies in the Trifecta-Kidney Study ( ClinicalTrials.gov NCT04239703), elevated plasma donor-derived cell-free DNA (dd-cfDNA) correlated most strongly with molecular antibody-mediated rejection (AMR) but was also elevated in other states: T cell-mediated rejection (TCMR), acute kidney injury (AKI), and some apparently normal biopsies. The present study aimed to define the molecular correlates of plasma dd-cfDNA within specific states.

METHODS

Dd-cfDNA was measured by the Prospera test. Molecular rejection and injury states were defined using the Molecular Microscope system. We studied the correlation between dd-cfDNA and the expression of genes, transcript sets, and classifier scores within specific disease states, and compared AMR, TCMR, and AKI to biopsies classified as normal and no injury (NRNI).

RESULTS

In all 604 biopsies, dd-cfDNA was elevated in AMR, TCMR, and AKI. Within AMR biopsies, dd-cfDNA correlated with AMR activity and stage. Within AKI, the correlations reflected acute parenchymal injury, including cell cycling. Within biopsies classified as MMDx Normal and archetypal No injury (NRNI), dd-cfDNA still correlated significantly with rejection- and injury-related genes. TCMR activity (eg, the TCMR Prob classifier) correlated with dd-cfDNA, but within TCMR biopsies, top gene correlations were complex and not the top TCMR-selective genes.

CONCLUSIONS

In kidney transplants, elevated plasma dd-cfDNA is associated with 3 distinct molecular states in the donor tissue: AMR, recent parenchymal injury (including cell cycling), and TCMR, potentially complicated by parenchymal disruption. Moreover, subtle rejection- and injury-related changes in the donor tissue can contribute to dd-cfDNA elevations in transplants considered to have no rejection or injury.

Comparing Plasma Donor-derived Cell-free DNA to Gene Expression in Endomyocardial Biopsies in the Trifecta-Heart Study

BACKGROUND

Plasma donor-derived cell-free DNA (dd-cfDNA) is used to screen for rejection in heart transplants. We launched the Trifecta-Heart study (ClinicalTrials.gov No. NCT04707872), an investigator-initiated, prospective trial, to examine the correlations between genome-wide molecular changes in endomyocardial biopsies (EMBs) and plasma dd-cfDNA. The present report analyzes the correlation of plasma dd-cfDNA with gene expression in EMBs from 4 vanguard centers and compared these correlations with those in 604 kidney transplant biopsies in the Trifecta-Kidney study (ClinicalTrials.gov No. NCT04239703).

METHODS

We analyzed 137 consecutive dd-cfDNA-EMB pairs from 70 patients. Plasma %dd-cfDNA was measured by the Prospera test (Natera Inc), and gene expression in EMBs was assessed by Molecular Microscope Diagnostic System using machine-learning algorithms to interpret rejection and injury states.

RESULTS

Top transcripts correlating with dd-cfDNA were related to genes increased in rejection such as interferon gamma-inducible genes (eg, HLA-DMA ) but also with genes induced by injury and expressed in macrophages (eg, SERPINA1 and HMOX1 ). In gene enrichment analysis, the top dd-cfDNA-correlated genes reflected inflammation and rejection pathways. Dd-cfDNA correlations with rejection genes in EMB were similar to those seen in kidney transplant biopsies, with somewhat stronger correlations for TCMR genes in hearts and ABMR genes in kidneys. However, the correlations with parenchymal injury-induced genes and macrophage genes were much stronger in hearts.

CONCLUSIONS

In this first analysis of Trifecta-Heart study, dd-cfDNA correlates significantly with molecular rejection but also with injury and macrophage infiltration, reflecting the proinflammatory properties of injured cardiomyocytes. The relationship supports the utility of dd-cfDNA in clinical management of heart transplant recipients.

Molecular diagnosis of antibody-mediated rejection: Evaluating biopsy-based transcript diagnostics in the presence of donor-specific antibodies but without microvascular inflammation, a single-center descriptive analysis

Biopsy-based transcript diagnostics may identify molecular antibody-mediated rejection (AMR) when microvascular inflammation (MVI) is absent. In this single-center cohort, biopsy-based transcript diagnostics were validated in 326 kidney allograft biopsies. A total of 71 histological AMR and 35 T cell-mediated rejection (TCMR) cases were identified as molecular AMR and TCMR in 55% and 63%, respectively. Among 121 cases without MVI (glomerulitis + peritubular capillaritis = 0), 45 (37%) donor-specific antibody (DSA)-positive and 76 (63%) DSA-negative cases were analyzed. Twenty-one out of the 121 (17%) cases showed borderline changes, or TCMR, while BK nephropathy was excluded. None of the 45 DSA-positive patients showed molecular AMR. Among 76 DSA-negative patients, 2 had mixed molecular AMR/TCMR. All-AMR phenotype scores (sum of R4-R6) exhibited median values of 0.13 and 0.12 for DSA-positive and DSA-negative patients, respectively (P = .84). A total of 13% (6/45) DSA-positive and 11% (8/76) DSA-negative patients showed an all-AMR phenotype score > 0.30 (P = .77). Patients with a higher all-AMR phenotype score showed 33% more histologic TCMR (P = .005). The median all-AMR phenotype scores of glomerular basement membrane double contours = 0 and glomerular basement membrane double contours > 0 biopsies were 0.12 and 0.10, respectively (P = .35). Biopsy-based transcript diagnostics did not identify molecular AMR in cases without MVI. Follow-up biopsies and outcome data should evaluate the clinical relevance of subthreshold molecular alterations.

Molecular states associated with dysfunction and graft loss in heart transplants

BACKGROUND

We explored the changes in gene expression correlating with dysfunction and graft failure in endomyocardial biopsies.

METHODS

Genome-wide microarrays (19,462 genes) were used to define mRNA changes correlating with dysfunction (left ventricular ejection fraction [LVEF] ≤ 55) and risk of graft loss within 3 years postbiopsy. LVEF data was available for 1,013 biopsies and survival data for 779 patients (74 losses). Molecular classifiers were built for predicting dysfunction (LVEF ≤ 55) and postbiopsy 3-year survival.

RESULTS

Dysfunction is correlated with dedifferentiation-decreased expression of normal heart transcripts, for example, solute carriers, along with increased expression of inflammation genes. Many genes with reduced expression in dysfunction were matrix genes such as fibulin 1 and decorin. Gene ontology (GO) categories suggested matrix remodeling and inflammation, not rejection. Genes associated with the risk of failure postbiopsy overlapped dysfunction genes but also included genes affecting microcirculation, for example, arginase 2, which reduces NO production, and endothelin 1. GO terms also reflected increased glycolysis and response to hypoxia, but decreased VEGF and angiogenesis pathways. T cell-mediated rejection was associated with reduced survival and antibody-mediated rejection with relatively good survival, but the main determinants of survival were features of parenchymal injury. Both dysfunction and graft loss were correlated with increased biopsy expression of BNP (gene NPPB). Survival probability classifiers divided hearts into risk quintiles, with actuarial 3-year postbiopsy survival >95% for the highest versus 50% for the lowest.

CONCLUSIONS

Dysfunction in transplanted hearts reflects dedifferentiation, decreased matrix genes, injury, and inflammation. The risk of short-term loss includes these changes but is also associated with microcirculation abnormalities, glycolysis, and response to hypoxia.

Development and Validation of a Multiclass Model Defining Molecular Archetypes of Kidney Transplant Rejection: A Large Cohort Study of the Banff Human Organ Transplant Gene Expression Panel

Gene expression profiling from formalin-fixed paraffin-embedded (FFPE) renal allograft biopsies is a promising approach for feasibly providing a molecular diagnosis of rejection. However, large-scale studies evaluating the performance of models using NanoString platform data to define molecular archetypes of rejection are lacking. We tested a diverse retrospective cohort of over 1400 FFPE biopsy specimens, rescored according to Banff 2019 criteria and representing 10 of 11 United Network of Organ Sharing regions, using the Banff Human Organ Transplant panel from NanoString and developed a multiclass model from the gene expression data to assign relative probabilities of 4 molecular archetypes: No Rejection, Antibody-Mediated Rejection, T Cell-Mediated Rejection, and Mixed Rejection. Using Least Absolute Shrinkage and Selection Operator regularized regression with 10-fold cross-validation fitted to 1050 biopsies in the discovery cohort and technically validated on an additional 345 biopsies, our model achieved overall accuracy of 85% in the discovery cohort and 80% in the validation cohort, with ≥75% positive predictive value for each class, except for the Mixed Rejection class in the validation cohort (positive predictive value, 53%). This study represents the technical validation of the first model built from a large and diverse sample of diagnostic FFPE biopsy specimens to define and classify molecular archetypes of histologically defined diagnoses as derived from Banff Human Organ Transplant panel gene expression profiling data.

Artificial Intelligence in Kidney Disease: A Comprehensive Study and Directions for Future Research

Artificial intelligence (AI) has emerged as a promising tool in the field of healthcare, with an increasing number of research articles evaluating its applications in the domain of kidney disease. To comprehend the evolving landscape of AI research in kidney disease, a bibliometric analysis is essential. The purposes of this study are to systematically analyze and quantify the scientific output, research trends, and collaborative networks in the application of AI to kidney disease. This study collected AI-related articles published between 2012 and 20 November 2023 from the Web of Science. Descriptive analyses of research trends in the application of AI in kidney disease were used to determine the growth rate of publications by authors, journals, institutions, and countries. Visualization network maps of country collaborations and author-provided keyword co-occurrences were generated to show the hotspots and research trends in AI research on kidney disease. The initial search yielded 673 articles, of which 631 were included in the analyses. Our findings reveal a noteworthy exponential growth trend in the annual publications of AI applications in kidney disease. Nephrology Dialysis Transplantation emerged as the leading publisher, accounting for 4.12% (26 out of 631 papers), followed by the American Journal of Transplantation at 3.01% (19/631) and Scientific Reports at 2.69% (17/631). The primary contributors were predominantly from the United States (n = 164, 25.99%), followed by China (n = 156, 24.72%) and India (n = 62, 9.83%). In terms of institutions, Mayo Clinic led with 27 contributions (4.27%), while Harvard University (n = 19, 3.01%) and Sun Yat-Sen University (n = 16, 2.53%) secured the second and third positions, respectively. This study summarized AI research trends in the field of kidney disease through statistical analysis and network visualization. The findings show that the field of AI in kidney disease is dynamic and rapidly progressing and provides valuable information for recognizing emerging patterns, technological shifts, and interdisciplinary collaborations that contribute to the advancement of knowledge in this critical domain.

Using Regression Equations to Enhance Interpretation of Histology Lesions of Kidney Transplant Rejection

BACKGROUND

The Banff system for histologic diagnosis of rejection in kidney transplant biopsies uses guidelines to assess designated features-lesions, donor-specific antibody (DSA), and C4d staining. We explored whether using regression equations to interpret the features as well as current guidelines could establish the relative importance of each feature and improve histologic interpretation.

METHODS

We developed logistic regression equations using the designated features to predict antibody-mediated rejection (AMR/mixed) and T-cell-mediated rejection (TCMR/mixed) in 1679 indication biopsies from the INTERCOMEX study ( ClinicalTrials.gov NCT01299168). Equations were trained on molecular diagnoses independent of the designated features.

RESULTS

In regression and random forests, the important features predicting molecular rejection were as follows: for AMR, ptc and g, followed by cg; for TCMR, t > i. V-lesions were relatively unimportant. C4d and DSA were also relatively unimportant for predicting AMR: by AUC, the model excluding them (0.853) was nearly as good as the model including them (0.860). Including time posttransplant slightly but significantly improved all models. By AUC, regression predicted molecular AMR and TCMR better than Banff histologic diagnoses. More importantly, in biopsies called "no rejection" by Banff guidelines, regression equations based on histology features identified histologic and molecular rejection-related changes in some biopsies and improved survival predictions. Thus, regression can screen for missed rejection.

CONCLUSIONS

Using lesion-based regression equations in addition to Banff histology guidelines defines the relative important of histology features for identifying rejection, allows screening for potential missed diagnoses, and permits early estimates of AMR when C4d and DSA are not available.

Evolution of human kidney allograft pathology diagnostics through 30 years of the Banff classification process

The second half of the previous century witnessed a tremendous rise in the number of clinical kidney transplants worldwide. This activity was, however, accompanied by many issues and challenges. An accurate diagnosis and appropriate management of causes of graft dysfunction were and still are, a big challenge. Kidney allograft biopsy played a vital role in addressing the above challenge. However, its interpretation was not standardized for many years until, in 1991, the Banff process was started to fill this void. Thereafter, regular Banff meetings took place every 2 years for the past 30 years. Marked changes have taken place in the interpretation of kidney allograft biopsies, diagnosis, and classification of rejection and other non-rejection pathologies from the original Banff 93 classification. This review attempts to summarize those changes for increasing the awareness and understanding of kidney allograft pathology through the eyes of the Banff process. It will interest the transplant surgeons, physicians, pathologists, and allied professionals associated with the care of kidney transplant patients.

Kidney Allograft Monitoring by Combining Donor-Derived Cell-Free DNA and Molecular Gene Expression: A Clinical Management Perspective

Donor-derived cell-free DNA (dd-cfDNA) may safely assess kidney allograft rejection. Molecular Microscope (MMDx®) gene expression may offer increased precision to histology. This single-center retrospective study monitored kidney transplant recipients for rejection at specified time intervals by utilizing creatinine (SCr), proteinuria, donor-specific antibodies (DSAs), and dd-cfDNA. A clinically indicated biopsy sample was sent for histopathology and MMDx®. Patients were categorized into rejection (Rej) and non-rejection (NRej) groups, and further grouped according to antibody-mediated rejection (ABMR) subtypes. Rej and NRej groups included 52 and 37 biopsies, respectively. Median follow-up duration was 506 days. DSAs were positive in 53% and 22% of patients in both groups, respectively (p = 0.01). Among these groups, pre- and post-intervention median SCr, proteinuria, and dd-cfDNA at 1 month, 2 months, and at the last follow-up revealed significant difference for dd-cfDNA (all p = 0.01), however, no difference was found for SCr and proteinuria (p > 0.05). The AUC was 0.80 (95% CI: 0.69-0.91), with an optimal dd-cfDNA criterion of 2.2%. Compared to histology, MMDx® was more likely to diagnose ABMR (79% vs. 100%) with either C4d positivity or negativity and/or DSA positivity or negativity. Hence, a pre- and post-intervention allograft monitoring protocol in combination with dd-cfDNA, MMDx®, and histology has aided in early diagnosis and timely individualized intervention.

Relating Molecular T Cell-mediated Rejection Activity in Kidney Transplant Biopsies to Time and to Histologic Tubulitis and Atrophy-fibrosis

BACKGROUND

We studied the variation in molecular T cell-mediated rejection (TCMR) activity in kidney transplant indication biopsies and its relationship with histologic lesions (particularly tubulitis and atrophy-fibrosis) and time posttransplant.

METHODS

We examined 175 kidney transplant biopsies with molecular TCMR as defined by archetypal analysis in the INTERCOMEX study ( ClinicalTrials.gov #NCT01299168). TCMR activity was defined by a molecular classifier.

RESULTS

Archetypal analysis identified 2 TCMR classes, TCMR1 and TCMR2: TCMR1 had higher TCMR activity and more antibody-mediated rejection ("mixed") activity and arteritis but little hyalinosis, whereas TCMR2 had less TCMR activity but more atrophy-fibrosis. TCMR1 and TCMR2 had similar levels of molecular injury and tubulitis. Both TCMR1 and TCMR2 biopsies were uncommon after 2 y posttransplant and were rare after 10 y, particularly TCMR1. Within late TCMR biopsies, TCMR classifier activity and activity molecules such as IFNG fell progressively with time, but tubulitis and molecular injury were sustained. Atrophy-fibrosis was increased in TCMR biopsies, even in the first year posttransplant, and rose with time posttransplant. TCMR1 and TCMR2 both reduced graft survival, but in random forests, the strongest determinant of survival after biopsies with TCMR was molecular injury, not TCMR activity.

CONCLUSIONS

TCMR varies in intensity but is always strongly related to molecular injury and atrophy-fibrosis, which ultimately explains its effect on survival. We hypothesize, based on the reciprocal relationship with hyalinosis, that the TCMR1-TCMR2 gradient reflects calcineurin inhibitor drug underexposure, whereas the time-dependent decline in TCMR activity and frequency after the first year reflects T-cell exhaustion.

Utility of Banff Human Organ Transplant Gene Panel in Human Kidney Transplant Biopsies

BACKGROUND

Microarray transcript analysis of human renal transplantation biopsies has successfully identified the many patterns of graft rejection. To evaluate an alternative, this report tests whether gene expression from the Banff Human Organ Transplant (B-HOT) probe set panel, derived from validated microarrays, can identify the relevant allograft diagnoses directly from archival human renal transplant formalin-fixed paraffin-embedded biopsies. To test this hypothesis, principal components (PCs) of gene expressions were used to identify allograft diagnoses, to classify diagnoses, and to determine whether the PC data were rich enough to identify diagnostic subtypes by clustering, which are all needed if the B-HOT panel can substitute for microarrays.

METHODS

RNA was isolated from routine, archival formalin-fixed paraffin-embedded tissue renal biopsy cores with both rejection and nonrejection diagnoses. The B-HOT panel expression of 770 genes was analyzed by PCs, which were then tested to determine their ability to identify diagnoses.

RESULTS

PCs of microarray gene sets identified the Banff categories of renal allograft diagnoses, modeled well the aggregate diagnoses, showing a similar correspondence with the pathologic diagnoses as microarrays. Clustering of the PCs identified diagnostic subtypes including non-chronic antibody-mediated rejection with high endothelial expression. PCs of cell types and pathways identified new mechanistic patterns including differential expression of B and plasma cells.

CONCLUSIONS

Using PCs of gene expression from the B-Hot panel confirms the utility of the B-HOT panel to identify allograft diagnoses and is similar to microarrays. The B-HOT panel will accelerate and expand transcript analysis and will be useful for longitudinal and outcome studies.

Multi-omic molecular profiling and network biology for precision anaesthesiology: a narrative review

Technological advancement, data democratisation, and decreasing costs have led to a revolution in molecular biology in which the entire set of DNA, RNA, proteins, and various other molecules - the 'multi-omic' profile - can be measured in humans. Sequencing 1 million bases of human DNA now costs US$0.01, and emerging technologies soon promise to reduce the cost of sequencing the whole genome to US$100. These trends have made it feasible to sample the multi-omic profile of millions of people, much of which is publicly available for medical research. Can anaesthesiologists use these data to improve patient care? This narrative review brings together a rapidly growing literature in multi-omic profiling across numerous fields that points to the future of precision anaesthesiology. Here, we discuss how DNA, RNA, proteins, and other molecules interact in molecular networks that can be used for preoperative risk stratification, intraoperative optimisation, and postoperative monitoring. This literature provides evidence for four fundamental insights: (1) Clinically similar patients have different molecular profiles and, as a consequence, different outcomes. (2) Vast, publicly available, and rapidly growing molecular datasets have been generated in chronic disease patients and can be repurposed to estimate perioperative risk. (3) Multi-omic networks are altered in the perioperative period and influence postoperative outcomes. (4) Multi-omic networks can serve as empirical, molecular measurements of a successful postoperative course. With this burgeoning universe of molecular data, the anaesthesiologist-of-the-future will tailor their clinical management to an individual's multi-omic profile to optimise postoperative outcomes and long-term health.

Allograft tissue under the microscope: only the beginning

PURPOSE OF REVIEW

To review novel modalities for interrogating a kidney allograft biopsy to complement the current Banff schema.

RECENT FINDINGS

Newer approaches of Artificial Intelligence (AI), Machine Learning (ML), digital pathology including Ex Vivo Microscopy, evaluation of the biopsy gene expression using bulk, single cell, and spatial transcriptomics and spatial proteomics are now available for tissue interrogation.

SUMMARY

Banff Schema of classification of allograft histology has standardized reporting of tissue pathology internationally greatly impacting clinical care and research. Inherent sampling error of biopsies, and lack of automated morphometric analysis with ordinal outputs limit its performance in prognostication of allograft health. Over the last decade, there has been an explosion of newer methods of evaluation of allograft tissue under the microscope. Digital pathology along with the application of AI and ML algorithms could revolutionize histopathological analyses. Novel molecular diagnostics such as spatially resolved single cell transcriptomics are identifying newer mechanisms underlying the pathologic diagnosis to delineate pathways of immunological activation, tissue injury, repair, and regeneration in allograft tissues. While these techniques are the future of tissue analysis, costs and complex logistics currently limit their clinical use.

Antibody-mediated Rejection Without Detectable Donor-specific Antibody Releases Donor-derived Cell-free DNA: Results From the Trifecta Study

BACKGROUND

Trifecta (ClinicalTrials.gov #NCT04239703) is a prospective trial defining relationships between donor-derived cell-free DNA (dd-cfDNA), donor-specific antibody (DSA), and molecular findings in kidney transplant biopsies. Previous analyses of double results showed dd-cfDNA was strongly associated with rejection-associated molecules in the biopsy. The present study analyzed the triple results in 280 biopsies, focusing on the question of dd-cfDNA levels in DSA-negative antibody-mediated rejection (AMR).

METHODS

Molecular Microscope Diagnostic System biopsy testing was performed at Alberta Transplant Applied Genomics Centre, dd-cfDNA testing at Natera, Inc, and central HLA antibody testing at One Lambda Inc. Local DSA and histologic diagnoses were assigned per center standard-of-care.

RESULTS

DSA was frequently negative in both molecular (56%) and histologic (51%) AMR. DSA-negative AMR had slightly less molecular AMR activity and histologic peritubular capillaritis than DSA-positive AMR. However, all AMRs-DSA-positive or -negative-showed elevated %dd-cfDNA. There was no association between dd-cfDNA and DSA in biopsies without rejection. In AMR, %dd-cfDNA ≥1.0 was more frequent (75%) than DSA positivity (44%). In logistic regression, dd-cfDNA percent (area under the curve [AUC] 0.85) or quantity (AUC 0.86) predicted molecular AMR better than DSA (AUC 0.66). However, the best predictions incorporated both dd-cfDNA and DSA, plus time posttransplant (AUC 0.88).

CONCLUSIONS

DSA-negative AMR has moderately decreased mean molecular and histologic AMR-associated features compared with DSA-positive AMR, though similarly elevated dd-cfDNA levels. In predicting AMR at the time of indication biopsies in this population, dd-cfDNA is superior to DSA, reflecting the prevalence of DSA-negative AMR, but the optimal predictions incorporated both dd-cfDNA and DSA.

The Molecular Diagnosis Might Be Clinically Useful in Discrepant Kidney Allograft Biopsy Findings: An Analysis of Clinical Outcomes

BACKGROUND

The Molecular Microscope Diagnostic System (MMDx) may overcome histology shortcomings. Previous studies have simply examined discrepant findings but have not attempted to determine clinical endpoints. To measure performance, clinical outcomes are strongly required.

METHODS

This single-center cohort study described discrepancies between MMDx and histology from 51 kidney transplant recipients (KTRs) and analyzed 72 indication biopsies, including 21 follow-up biopsies. Clinical performance was assessed by a combined endpoint of graft failure, rejection on follow-up biopsy, de novo donor-specific antibody, and improvement of kidney allograft function upon antirejection treatment.

RESULTS

MMDx agreed in 33 (65%) and differed in 18 (35%) of 51 KTRs. Most discrepancies occurred in biopsies called no rejection by MMDx and rejection by histology (15/24, 63%). In contrast, in biopsies called rejection by MMDx, 3 were classified as no rejection by histology (3/27, 11%). Discrepant findings between MMDx and histology occurred following delayed graft function and MMDx from biopsies with a low percentage of cortex. Among 15 biopsies classified as no rejection by MMDx but rejection by histology, the clinical course suggested no rejection in 9 cases. Six KTRs reached the endpoint, showing predominant t ≥ 2 lesions.

CONCLUSIONS

The most often occurring discrepancy is rejection by histology but no rejection by MMDx. As more KTRs do not meet the combined endpoint for rejection, MMDx might be clinically useful in these discrepant cases. Although strong histological findings have priority in indicating the treatment, clinical implementation of MMDx could strengthen treatment strategies.

The Molecular Microscope Diagnostic System: Assessment of Rejection and Injury in Heart Transplant Biopsies

This review describes the development of the Molecular Microscope Diagnostic System (MMDx) for heart transplant endomyocardial biopsies (EMBs). MMDx-Heart uses microarrays to measure biopsy-based gene expression and ensembles of machine learning algorithms to interpret the results and compare each new biopsy to a large reference set of earlier biopsies. MMDx assesses T cell-mediated rejection (TCMR), antibody-mediated rejection (AMR), recent parenchymal injury, and atrophy-fibrosis, continually "learning" from new biopsies. Rejection-associated transcripts mapped in kidney transplants and experimental systems were used to identify TCMR, AMR, and recent injury-induced inflammation. Rejection and injury emerged as gradients of intensity, rather than binary classes. AMR was one-third donor-specific antibody (DSA)-negative, and many EMBs first considered to have no rejection displayed minor AMR-like changes, with increased probability of DSA positivity and subtle inflammation. Rejection-associated transcript-based algorithms now classify EMBs as "Normal," "Minor AMR changes," "AMR," "possible AMR," "TCMR," "possible TCMR," and "recent injury." Additionally, MMDx uses injury-associated transcript sets to assess the degree of parenchymal injury and atrophy-fibrosis in every biopsy and study the effect of rejection on the parenchyma. TCMR directly injures the parenchyma whereas AMR usually induces microcirculation stress but relatively little initial parenchymal damage, although slowly inducing parenchymal atrophy-fibrosis. Function (left ventricular ejection fraction) and short-term risk of failure are strongly determined by parenchymal injury. These discoveries can guide molecular diagnostic applications, either as a central MMDx system or adapted to other platforms. MMDx can also help calibrate noninvasive blood-based biomarkers to avoid unnecessary biopsies and monitor response to therapy.

Revisiting transplant immunology through the lens of single-cell technologies

Solid organ transplantation (SOT) is the standard of care for end-stage organ disease. The most frequent complication of SOT involves allograft rejection, which may occur via T cell- and/or antibody-mediated mechanisms. Diagnosis of rejection in the clinical setting requires an invasive biopsy as there are currently no reliable biomarkers to detect rejection episodes. Likewise, it is virtually impossible to identify patients who exhibit operational tolerance and may be candidates for reduced or complete withdrawal of immunosuppression. Emerging single-cell technologies, including cytometry by time-of-flight (CyTOF), imaging mass cytometry, and single-cell RNA sequencing, represent a new opportunity for deep characterization of pathogenic immune populations involved in both allograft rejection and tolerance in clinical samples. These techniques enable examination of both individual cellular phenotypes and cell-to-cell interactions, ultimately providing new insights into the complex pathophysiology of allograft rejection. However, working with these large, highly dimensional datasets requires expertise in advanced data processing and analysis using computational biology techniques. Machine learning algorithms represent an optimal strategy to analyze and create predictive models using these complex datasets and will likely be essential for future clinical application of patient level results based on single-cell data. Herein, we review the existing literature on single-cell techniques in the context of SOT.

The Histological Spectrum and Clinical Significance of T Cell-mediated Rejection of Kidney Allografts

T cell-mediated rejection (TCMR) remains a significant cause of long-term kidney allograft loss, either indirectly through induction of donor-specific anti-HLA alloantibodies or directly through chronic active TCMR. Whether found by indication or protocol biopsy, Banff defined acute TCMR should be treated with antirejection therapy and maximized maintenance immunosuppression. Neither isolated interstitial inflammation in the absence of tubulitis nor isolated tubulitis in the absence of interstitial inflammation results in adverse outcomes, and neither requires antirejection treatment. RNA gene expression analysis of biopsy material may supplement conventional histology, especially in ambiguous cases. Lesser degrees of tubular and interstitial inflammation (Banff borderline) may portend adverse outcomes and should be treated when found on an indication biopsy. Borderline lesions on protocol biopsies may resolve spontaneously but require close follow-up if untreated. Following antirejection therapy of acute TCMR, surveillance protocol biopsies should be considered. Minimally invasive blood-borne assays (donor-derived cell-free DNA and gene expression profiling) are being increasingly studied as a means of following stable patients in lieu of biopsy. The clinical benefit and cost-effectiveness require confirmation in randomized controlled trials. Treatment of acute TCMR is not standardized but involves bolus corticosteroids with lymphocyte depleting antibodies for severe, refractory, or relapsing cases. Arteritis may be found with acute TCMR, active antibody-mediated rejection, or mixed rejections and should be treated accordingly. The optimal treatment ofchronic active TCMR is uncertain. Randomized controlled trials are necessary to optimally define therapy.

Combining Donor-derived Cell-free DNA Fraction and Quantity to Detect Kidney Transplant Rejection Using Molecular Diagnoses and Histology as Confirmation

BACKGROUND

Donor-derived cell-free DNA (dd-cfDNA) fraction and quantity have both been shown to be associated with allograft rejection. The present study compared the relative predictive power of each of these variables to the combination of the two, and developed an algorithm incorporating both variables to detect active rejection in renal allograft biopsies.

METHODS

The first 426 sequential indication biopsy samples collected from the Trifecta study ( ClinicalTrials.gov # NCT04239703) with microarray-derived gene expression and dd-cfDNA results were included. After exclusions to simulate intended clinical use, 367 samples were analyzed. Biopsies were assessed using the molecular microscope diagnostic system and histology (Banff 2019). Logistic regression analysis examined whether combining dd-cfDNA fraction and quantity adds predictive value to either alone. The first 149 sequential samples were used to develop a two-threshold algorithm and the next 218 to validate the algorithm.

RESULTS

In regression, the combination of dd-cfDNA fraction and quantity was found to be significantly more predictive than either variable alone ( P = 0.009 and P < 0.0001). In the test set, the area under the receiver operating characteristic curve of the two-variable system was 0.88, and performance of the two-threshold algorithm showed a sensitivity of 83.1% and specificity of 81.0% for molecular diagnoses and a sensitivity of 73.5% and specificity of 80.8% for histology diagnoses.

CONCLUSIONS

This prospective, biopsy-matched, multisite dd-cfDNA study in kidney transplant patients found that the combination of dd-cfDNA fraction and quantity was more powerful than either dd-cfDNA fraction or quantity alone and validated a novel two-threshold algorithm incorporating both variables.

N6-methyladenosine regulators-related immune genes enable predict graft loss and discriminate T-cell mediate rejection in kidney transplantation biopsies for cause

Objective

The role of m6A modification in kidney transplant-associated immunity, especially in alloimmunity, still remains unknown. This study aims to explore the potential value of m6A-related immune genes in predicting graft loss and diagnosing T cell mediated rejection (TCMR), as well as the possible role they play in renal graft dysfunction.

Methods

Renal transplant-related cohorts and transcript expression data were obtained from the GEO database. First, we conducted correlation analysis in the discovery cohort to identify the m6A-related immune genes. Then, lasso regression and random forest were used respectively to build prediction models in the prognosis and diagnosis cohort, to predict graft loss and discriminate TCMR in dysfunctional renal grafts. Connectivity map (CMap) analysis was applied to identify potential therapeutic compounds for TCMR.

Results

The prognostic prediction model effectively predicts the prognosis and survival of renal grafts with clinical indications (P< 0.001) and applies to both rejection and non-rejection situations. The diagnostic prediction model discriminates TCMR in dysfunctional renal grafts with high accuracy (area under curve = 0.891). Meanwhile, the classifier score of the diagnostic model, as a continuity index, is positively correlated with the severity of main pathological injuries of TCMR. Furthermore, it is found that METTL3, FTO, WATP, and RBM15 are likely to play a pivotal part in the regulation of immune response in TCMR. By CMap analysis, several small molecular compounds are found to be able to reverse TCMR including fenoldopam, dextromethorphan, and so on.

Conclusions

Together, our findings explore the value of m6A-related immune genes in predicting the prognosis of renal grafts and diagnosis of TCMR.

CARARIME: Interactive web server for comprehensive analysis of renal allograft rejection in immune microenvironment

Background

Renal transplantation is a very effective treatment for renal failure patients following kidney transplant. However, the clinical benefit is restricted by the high incidence of organ rejection. Therefore, there exists a wealth of literature regarding the mechanism of renal transplant rejection, including a large library of expression data. In recent years, research has shown the immune microenvironment to play an important role in renal transplant rejection. Nephrology web analysis tools currently exist to address chronic nephropathy, renal tumors and children’s kidneys, but no such tool exists that analyses the impact of immune microenvironment in renal transplantation rejection.

Methods

To fill this gap, we have developed a web page analysis tool called Comprehensive Analysis of Renal Allograft Rerejction in Immune Microenvironment (CARARIME).

Results

CARARIME analyzes the gene expression and immune microenvironment of published renal transplant rejection cohorts, including differential analysis (gene expression and immune cells), prognosis analysis (logistics regression, Univariable Cox Regression and Kaplan Meier), correlation analysis, enrichment analysis (GSEA and ssGSEA), and ROC analysis.

Conclusions

Using this tool, researchers can easily analyze the immune microenvironment in the context of renal transplant rejection by clicking on the available options, helping to further the development of approaches to renal transplant rejection in the immune microenvironment field. CARARIME can be found in http://www.cararime.com.

The MMDx® diagnostic system: A critical re-appraisal of its knowledge gaps and a call for rigorous validation studies

Transcriptomics generates pathogenetic insights not obtainable by histology, but translation of these insights into diagnostic tests is not a trivial task. This opinion-piece critically appraises declarative MMDx statements, such as the infallibility of machine learning algorithms, measurements of gene expression with >99% precision, and "unambiguous reclassifications" of contentious biopsies such as those with borderline change, polyomavirus nephropathy, chronic active T-cell or mixed rejection, isolated intimal arteritis, and renal medullary pathology. It is shown that molecular diagnoses that do not agree with histology cannot be attributed primarily to pathology reading errors. Neither can all molecular calls derived from arbitrary binary thresholds be automatically accepted as the ground truth. Important other sources of discrepancies between clinico-pathologic and molecular calls include: (a) organ being studied, (b) disease definition, (c) clinical histologic, and gene expression heterogeneity within the same diagnostic label, (d) size and composition of comparator groups, (e) molecular noise, (f) variability in output of different machine learning algorithms, and (g) the nonavailability of a molecular classifier for chronic active TCMR. Carefully designed clinical trials are needed to determine which of the proposed indications of MMDx provide incremental value over existing standard of care protocols.

Artificial Intelligence: Present and Future Potential for Solid Organ Transplantation

Artificial intelligence (AI) refers to computer algorithms used to complete tasks that usually require human intelligence. Typical examples include complex decision-making and- image or speech analysis. AI application in healthcare is rapidly evolving and it undoubtedly holds an enormous potential for the field of solid organ transplantation. In this review, we provide an overview of AI-based approaches in solid organ transplantation. Particularly, we identified four key areas of transplantation which could be facilitated by AI: organ allocation and donor-recipient pairing, transplant oncology, real-time immunosuppression regimes, and precision transplant pathology. The potential implementations are vast—from improved allocation algorithms, smart donor-recipient matching and dynamic adaptation of immunosuppression to automated analysis of transplant pathology. We are convinced that we are at the beginning of a new digital era in transplantation, and that AI has the potential to improve graft and patient survival. This manuscript provides a glimpse into how AI innovations could shape an exciting future for the transplantation community.

Use of a donor-derived cell-free DNA assay to monitor treatment response in pediatric renal transplant recipients with allograft rejection

BACKGROUND

Detection of donor-derived cell-free DNA (dd-cfDNA) reliably identifies allograft rejection in pediatric and adult kidney transplant (KT) recipients. Here, we evaluate the utility of dd-cfDNA for monitoring response to treatment among pediatric renal transplant recipients suffering graft rejection.

METHODS

58 pediatric transplant recipients were enrolled between April 2018 and March 2020 and underwent initial dd-cfDNA testing to monitor for rejection. Allograft biopsy was performed for dd-cfDNA scores >1.0%. Patients with histologically proven rejection formed the study cohort and underwent appropriate treatment. Results of dd-cfDNA, serum creatinine (SCr), biopsy findings, and treatment outcomes were evaluated. Standard statistical analyses were applied.

RESULTS

Nineteen of 58 (31%) patients had dd-cfDNA score >1.0%, of which 18 (94.7%) had biopsy-proven rejection. Median dd-cfDNA value was 1.90% (interquartile range 1.43%-3.23%), and biopsy results showed 11 patients (61.1%) with antibody-mediated rejection (AMR), 2 patients (11.1%) with T-cell mediated rejection (TCMR), and 5 patients (27.7%) with mixed AMR/TCMR. SCr at time of biopsy was 1.28 ± 1.09 mg/dl. Following treatment, dd-cfDNA scores decreased for all types of rejection but still remained >1.0% in both AMR (1.50% [0.90%-3.10%]) and mixed (1.40% [0.95%-4.15%]) groups. Repeat dd-cfDNA values were <1.0% for patients with TCMR (0.20%-0.28%). SCr showed minimal change from pre-treatment levels regardless of rejection subtype.

CONCLUSIONS

Patients with TCMR may be reliably followed by dd-cfDNA; however, it remains unclear whether persistently elevated dd-cfDNA levels in AMR is a reflection of ongoing subclinical rejection or an inherent limitation of the assay's utility.

Archetypal Analysis of Injury in Kidney Transplant Biopsies Identifies Two Classes of Early AKI

All transplanted kidneys are subjected to some degree of injury as a result of the donation-implantation process and various post-transplant stresses such as rejection. Because transplants are frequently biopsied, they present an opportunity to explore the full spectrum of kidney response-to-wounding from all causes. Defining parenchymal damage in transplanted organs is important for clinical management because it determines function and survival. In this study, we classified the scenarios associated with parenchymal injury in genome-wide microarray results from 1,526 kidney transplant indication biopsies collected during the INTERCOMEX study. We defined injury groups by using archetypal analysis (AA) of scores for gene sets and classifiers previously identified in various injury states. Six groups and their characteristics were defined in this population: No injury, minor injury, two classes of acute kidney injury ("AKI," AKI1, and AKI2), chronic kidney disease (CKD), and CKD combined with AKI. We compared the two classes of AKI, namely, AKI1 and AKI2. AKI1 had a poor function and increased parenchymal dedifferentiation but minimal response-to-injury and inflammation, instead having increased expression of PARD3, a gene previously characterized as being related to epithelial polarity and adherens junctions. In contrast, AKI2 had a poor function and increased response-to-injury, significant inflammation, and increased macrophage activity. In random forest analysis, the most important predictors of function (estimated glomerular filtration rate) and graft loss were injury-based molecular scores, not rejection scores. AKI1 and AKI2 differed in 3-year graft survival, with better survival in the AKI2 group. Thus, injury archetype analysis of injury-induced gene expression shows new heterogeneity in kidney response-to-wounding, revealing AKI1, a class of early transplants with a poor function but minimal inflammation or response to injury, a deviant response characterized as PC3, and an increased risk of failure. Given the relationship between parenchymal injury and kidney survival, further characterization of the injury phenotypes in kidney transplants will be important for an improved understanding that could have implications for understanding native kidney diseases (ClinicalTrials.gov #NCT01299168).

Transcripts associated with chronic lung allograft dysfunction in transbronchial biopsies of lung transplants

Transplanted lungs suffer worse outcomes than other organ transplants with many developing chronic lung allograft dysfunction (CLAD), diagnosed by physiologic changes. Histology of transbronchial biopsies (TBB) yields little insight, and the molecular basis of CLAD is not defined. We hypothesized that gene expression in TBBs would reveal the nature of CLAD and distinguish CLAD from changes due simply to time posttransplant. Whole-genome mRNA profiling was performed with microarrays in 498 prospectively collected TBBs from the INTERLUNG study, 90 diagnosed as CLAD. Time was associated with increased expression of inflammation genes, for example, CD1E and immunoglobulins. After correcting for time, CLAD manifested not as inflammation but as parenchymal response-to-wounding, with increased expression of genes such as HIF1A, SERPINE2, and IGF1 that are increased in many injury and disease states and cancers, associated with development, angiogenesis, and epithelial response-to-wounding in pathway analysis. Fibrillar collagen genes were increased in CLAD, indicating matrix changes, and normal transcripts were decreased-dedifferentiation. Gene-based classifiers predicted CLAD with AUC 0.70 (no time-correction) and 0.87 (time-corrected). CLAD related gene sets and classifiers were strongly prognostic for graft failure and correlated with CLAD stage. Thus, in TBBs, molecular changes indicate that CLAD primarily reflects severe parenchymal injury-induced changes and dedifferentiation.

Diagnostic Biomarkers and Immune Infiltration in Patients With T Cell-Mediated Rejection After Kidney Transplantation

T cell-mediated rejection (TCMR) is an important rejection type in kidney transplantation, characterized by T cells and macrophages infiltration. The application of bioinformatic analysis in genomic research has been widely used. In the present study, Microarray data was analyzed to identify the potential diagnostic markers of TCMR in kidney transplantation. Cell-type identification by estimating relative subsets of RNA transcript (CIBERSORT) was performed to determine the distribution of immune cell infiltration in the pathology. Totally 129 upregulated differently expressed genes (DEGs) and 378 downregulated DEGs were identified. The GO and KEGG results demonstrated that DEGs were mainly associated with pathways and diseases involved in immune response. The intersection of the two algorithms (PPI network and LASSO) contains three overlapping genes (CXCR6, CXCL13 and FCGR1A). After verification in GSE69677, only CXCR6 and CXCL13 were selected. Immune cells Infiltration analysis demonstrated that CXCR6 and CXCL13 were positively correlated with gamma delta T cells (p < 0.001), CD4⁺ memory activated T cells (p < 0.001), CD8⁺ T cells (p < 0.001) and M1 macrophages (p = 0.006), and negatively correlated with M2 macrophages (p < 0.001) and regulatory T cells (p < 0.001). Immunohistochemical staining and image analysis confirmed the overexpression of CXCR6 and CXCL13 in human allograft TCMR samples. CXCR6 and CXCL13 could be diagnostic biomarkers of TCMR and potential targets for immunotherapy in patients with TCMR.

The molecular phenotypes of injury, steatohepatitis, and fibrosis in liver transplant biopsies in the INTERLIVER study

To extend previous molecular analyses of rejection in liver transplant biopsies in the INTERLIVER study (ClinicalTrials.gov #NCT03193151), the present study aimed to define the gene expression selective for parenchymal injury, fibrosis, and steatohepatitis. We analyzed genome-wide microarray measurements from 337 liver transplant biopsies from 13 centers. We examined expression of genes previously annotated as increased in injury and fibrosis using principal component analysis (PCA). PC1 reflected parenchymal injury and related inflammation in the early posttransplant period, slowly regressing over many months. PC2 separated early injury from late fibrosis. Positive PC3 identified a distinct mildly inflamed state correlating with histologic steatohepatitis. Injury PCs correlated with liver function and histologic abnormalities. A classifier trained on histologic steatohepatitis predicted histologic steatohepatitis with cross-validated AUC = 0.83, and was associated with pathways reflecting metabolic abnormalities distinct from fibrosis. PC2 predicted histologic fibrosis (AUC = 0.80), as did a molecular fibrosis classifier (AUC = 0.74). The fibrosis classifier correlated with matrix remodeling pathways with minimal overlap with those selective for steatohepatitis, although some biopsies had both. Genome-wide assessment of liver transplant biopsies can not only detect molecular changes induced by rejection but also those correlating with parenchymal injury, steatohepatitis, and fibrosis, offering potential insights into disease mechanisms for primary diseases.

The Trifecta Study: Comparing Plasma Levels of Donor-derived Cell-Free DNA with the Molecular Phenotype of Kidney Transplant Biopsies

BACKGROUND

The relationship between the donor-derived cell-free DNA fraction (dd-cfDNA[%]) in plasma in kidney transplant recipients at time of indication biopsy and gene expression in the biopsied allograft has not been defined.

METHODS

In the prospective, multicenter Trifecta study, we collected tissue from 300 biopsies from 289 kidney transplant recipients to compare genome-wide gene expression in biopsies with dd-cfDNA(%) in corresponding plasma samples drawn just before biopsy. Rejection was assessed with the microarray-based Molecular Microscope Diagnostic System using automatically assigned rejection archetypes and molecular report sign-outs, and histology assessments that followed Banff guidelines.

RESULTS

The median time of biopsy post-transplantation was 455 days (5 days to 32 years), with a case mix similar to that of previous studies: 180 (60%) no rejection, 89 (30%) antibody-mediated rejection (ABMR), and 31 (10%) T cell-mediated rejection (TCMR) and mixed. In genome-wide mRNA measurements, all 20 top probe sets correlating with dd-cfDNA(%) were previously annotated for association with ABMR and all types of rejection, either natural killer (NK) cell-expressed (e.g., GNLY, CCL4, TRDC, and S1PR5) or IFN-γ-inducible (e.g., PLA1A, IDO1, CXCL11, and WARS). Among gene set and classifier scores, dd-cfDNA(%) correlated very strongly with ABMR and all types of rejection, reasonably strongly with active TCMR, and weakly with inactive TCMR, kidney injury, and atrophy fibrosis. Active ABMR, mixed, and active TCMR had the highest dd-cfDNA(%), whereas dd-cfDNA(%) was lower in late-stage ABMR and less-active TCMR. By multivariate random forests and logistic regression, molecular rejection variables predicted dd-cfDNA(%) better than histologic variables.

CONCLUSIONS

The dd-cfDNA(%) at time of indication biopsy strongly correlates with active molecular rejection and has the potential to reduce unnecessary biopsies.

CLINICAL TRIAL REGISTRATION NUMBER

NCT04239703.

US payer budget impact of a microarray assay with machine learning to evaluate kidney transplant rejection in for-cause biopsies

AIM

This study evaluates the economic impact to US commercial payers of MMDx-Kidney used in conjunction with histologic evaluation of for-cause kidney transplant biopsies.

MATERIALS AND METHODS

An Excel-based model was developed to assess the cost impact of histology plus MMDx-Kidney versus histology alone for the evaluation of potential rejection in kidney transplant patients who receive a for-cause biopsy. Different model time periods were assessed, ranging from 1 to 5 years post-biopsy. A targeted literature review was used to identify parameter estimates, validated by two external clinicians with expertise in managing kidney transplant rejection. A sensitivity analysis was conducted to evaluate the relative impact of key clinical and cost parameters. In particular, the model identified the magnitude of MMDx-Kidney's impact on graft failure from rejection that would be required for MMDx-Kidney to be cost-neutral.

RESULTS

By more accurately characterizing rejection, MMDx-Kidney is estimated to increase antirejection treatment costs by $1,126 per test. Nevertheless, a break-even analysis shows that the costs of MMDx-Kidney and anti-rejection medication, as well as the costs associated with an increase in the number of patients with functioning transplants, may be offset by reductions in costs associated with graft failure (i.e. costs of hospitalizations, dialysis, and repeat transplants) over 5 years, assuming MMDx-Kidney reduces annual graft failure from rejection by at least 5%. For the base case, with a 25% relative reduction in annual rate of graft failures from rejection, MMDx-Kidney increases overall costs incurred in the first year of the model but starts generating savings by the second year of the model.

CONCLUSIONS

Compared with histologic evaluation of for-cause kidney transplant biopsies alone, the use of MMDx-Kidney in conjunction with histologic evaluation improves the diagnoses of graft dysfunction and may have the potential to generate overall savings from reductions in rejection-related graft failure.

Donor-Specific Antibody Is Associated with Increased Expression of Rejection Transcripts in Renal Transplant Biopsies Classified as No Rejection

BACKGROUND

Donor -specific HLA antibody (DSA) is present in many kidney transplant patients whose biopsies are classified as no rejection (NR). We explored whether in some NR kidneys DSA has subtle effects not currently being recognized.

METHODS

We used microarrays to examine the relationship between standard-of-care DSA and rejection-related transcript increases in 1679 kidney transplant indication biopsies in the INTERCOMEX study (ClinicalTrials.gov NCT01299168), focusing on biopsies classified as NR by automatically assigned archetypal clustering. DSA testing results were available for 835 NR biopsies and were positive in 271 (32%).

RESULTS

DSA positivity in NR biopsies was associated with mildly increased expression of antibody-mediated rejection (ABMR)-related transcripts, particularly IFNG-inducible and NK cell transcripts. We developed a machine learning DSA probability (DSA_Prob) classifier based on transcript expression in biopsies from DSA-positive versus DSA-negative patients, assigning scores using 10-fold cross-validation. This DSA_Prob classifier was very similar to a previously described "ABMR probability" classifier trained on histologic ABMR in transcript associations and prediction of molecular or histologic ABMR. Plotting the biopsies using Uniform Manifold Approximation and Projection revealed a gradient of increasing molecular ABMR-like transcript expression in NR biopsies, associated with increased DSA (P<2 × 10^-16). In biopsies with no molecular or histologic rejection, increased DSA_Prob or ABMR probability scores were associated with increased risk of kidney failure over 3 years.

CONCLUSIONS

Many biopsies currently considered to have no molecular or histologic rejection have mild increases in expression of ABMR-related transcripts, associated with increasing frequency of DSA. Thus, mild molecular ABMR-related pathology is more common than previously realized.

A 2-fold Approach to Polyoma Virus (BK) Nephropathy in Kidney Transplants: Distinguishing Direct Virus Effects From Cognate T Cell-mediated Inflammation

BACKGROUND

BK nephropathy (BKN) in kidney transplants diagnosed by histology is challenging because it involves damage from both virus activity and cognate T cell-mediated inflammation, directed against alloantigens (rejection) or viral antigens. The present study of indication biopsies from the Integrated Diagnostic System in the International Collaborative Microarray Study Extension study measured major capsid viral protein 2 (VP2) mRNA to assess virus activity and a T cell-mediated rejection (TCMR) classifier to assess cognate T cell-mediated inflammation.

METHODS

Biopsies were assessed by local standard-of-care histology and by genome-wide microarrays and Molecular Microscope Diagnostic System (MMDx) algorithms to detect rejection and injury. In a subset of 102 biopsies (50 BKN and 52 BKN-negative biopsies with various abnormalities), we measured VP2 transcripts by real-time polymerase chain reaction.

RESULTS

BKN was diagnosed in 55 of 1679 biopsies; 30 had cognate T cell-mediated activity assessed by by MMDx and TCMR lesions, but only 3 of 30 were histologically diagnosed as TCMR. We developed a BKN probability classifier that predicted histologic BKN (area under the curve = 0.82). Virus activity (VP2 expression) was highly selective for BKN (area under the curve = 0.94) and correlated with acute injury, atrophy-fibrosis, macrophage activation, and the BKN classifier, but not with the TCMR classifier. BKN with molecular TCMR had more tubulitis and inflammation than BKN without molecular TCMR. In 5 BKN cases with second biopsies, VP2 mRNA decreased in second biopsies, whereas in 4 of 5 TCMR classifiers, scores increased. Genes and pathways associated with BKN and VP2 mRNA were similar, reflecting injury, inflammation, and macrophage activation but none was selective for BKN.

CONCLUSIONS

Risk-benefit decisions in BKN may be assisted by quantitative assessment of the 2 major pathologic processes, virus activity and cognate T cell-mediated inflammation.

Machine Learning Applications in Solid Organ Transplantation and Related Complications

The complexity of transplant medicine pushes the boundaries of innate, human reasoning. From networks of immune modulators to dynamic pharmacokinetics to variable postoperative graft survival to equitable allocation of scarce organs, machine learning promises to inform clinical decision making by deciphering prodigious amounts of available data. This paper reviews current research describing how algorithms have the potential to augment clinical practice in solid organ transplantation. We provide a general introduction to different machine learning techniques, describing their strengths, limitations, and barriers to clinical implementation. We summarize emerging evidence that recent advances that allow machine learning algorithms to predict acute post-surgical and long-term outcomes, classify biopsy and radiographic data, augment pharmacologic decision making, and accurately represent the complexity of host immune response. Yet, many of these applications exist in pre-clinical form only, supported primarily by evidence of single-center, retrospective studies. Prospective investigation of these technologies has the potential to unlock the potential of machine learning to augment solid organ transplantation clinical care and health care delivery systems.

Technology-Enabled Care and Artificial Intelligence in Kidney Transplantation

Purpose of Review

Artificial intelligence (AI), machine learning, and technology-enabled remote patient care have evolved rapidly and have now been incorporated into many aspects of medical care. Transplantation is fortunate to have large data sets upon which machine learning algorithms can be constructed. AI are now available to improve pretransplant management, donor selection, and post-operative management of transplant patients.

Recent Findings

Changes in patient and donor characteristics warrant new approaches to listing and organ acceptance practices. Machine learning has been employed to optimize donor selection to identify patients likely to benefit from transplantation of higher risk organs, increasing organ discard and reducing waitlist mortality. These models have greater precisions and predictive ability than currently employed metrics including the Kidney Donor Profile Index and the expected posttransplant survival models. After transplant, AI tools have been developed to optimize immunosuppression management, track patients adherence, and assess graft survival.

Summary

AI and technology-enabled management tools are now available throughout the transplant journey. Unfortunately, those are frequently not available at the point of decision (patient listing, organ acceptance, posttransplant clinic), limiting utilization. Incorporation of these tools into the EMR, the Donor Net® organ offer system, and mobile devices is vital to ensure widespread adoption.

Artificial intelligence and kidney transplantation

Artificial intelligence and its primary subfield, machine learning, have started to gain widespread use in medicine, including the field of kidney transplantation. We made a review of the literature that used artificial intelligence techniques in kidney transplantation. We located six main areas of kidney transplantation that artificial intelligence studies are focused on: Radiological evaluation of the allograft, pathological evaluation including molecular evaluation of the tissue, prediction of graft survival, optimizing the dose of immunosuppression, diagnosis of rejection, and prediction of early graft function. Machine learning techniques provide increased automation leading to faster evaluation and standardization, and show better performance compared to traditional statistical analysis. Artificial intelligence leads to improved computer-aided diagnostics and quantifiable personalized predictions that will improve personalized patient care.

Renal allograft DARCness in subclinical acute and chronic active ABMR

Gene expression profiling of renal allograft biopsies revealed the Duffy antigen receptor for chemokines (DARC) as being strikingly upregulated in antibody‐mediated rejection (ABMR). DARC has previously been shown to be associated with endothelial injury. This study aimed at assessing the value of DARC immunohistochemistry as diagnostic marker in ABMR. The study was performed on 82 prospectively collected biopsies of a clinically well‐defined population (BORTEJECT trial, NCT01873157) of DSA‐positive patients with gene expression data available for all biopsies. Diagnostic histologic assessment of biopsies was performed according to the Banff diagnostic scheme. DARC expression was focally accentuated, on peritubular capillaries (PTC) mostly in areas of interstitial fibrosis and/or inflammation. DARC positivity was associated with diagnosis of ABMR and correlated with DARC gene expression levels detected by microarray analysis. Still, as previously described, a substantial number of biopsies without signs of rejection showed DARC‐positive PTC. We did not observe significantly reduced graft survival in cases showing histologic signs of ABMR and being DARC‐positive, as compared to DARC‐negative ABMR. In summary, the upregulation of DARC, detected by immunohistochemistry, is associated with but not specific for ABMR. We did not observe reduced graft survival in DARC‐positive patients.

Lupus nephritis and ANCA-associated vasculitis: towards precision medicine?

Historically the treatment of lupus nephritis (LN) and anti-neutrophil cytoplasmic antibody (ANCA) vasculitis was 'one size fits all'; however, with the emergence of precision medicine initiatives, the field is moving towards more personalized treatment approaches. The recent development of a more accurate and reproducible histopathological classification system for LN could lead to better disease categorization and therefore more targeted therapies. A better understanding of the pathophysiology of LN has provided evidence that not only T but also B cells play an important role, opening new opportunities for individualized treatment approaches. Recent trials have shown calcineurin inhibitors and the anti-CD20 antibodies rituximab and ofatumumab to be effective in the treatment of LN, adding new treatment options. State-of-the-art targeted therapy in ANCA-associated vasculitis (AAV) takes interindividual heterogeneity in disease severity, type of ANCA antibody [myeloperoxidase versus proteinase 3 (PR3)] and the risk for side effects of therapy into consideration. In addition, within an individual, induction therapy differs from maintenance therapy, the same holding true in incident and relapsing disease. Rituximab is now widely used in AAV and it has become clear that prolonged B cell depletion, as in LN, must be achieved to obtain a long-lasting clinical response, especially in anti-PR3-associated disease. Still, despite these advances, molecular and genetic markers are rarely incorporated into diagnostic and treatment algorithms and true precision medicine remains an aspiration that hopefully can be achieved.

Augmenting the Transplant Team With Artificial Intelligence: Toward Meaningful AI Use in Solid Organ Transplant

Advances in systems immunology, such as new biomarkers, offer the potential for highly personalized immunosuppression regimens that could improve patient outcomes. In the future, integrating all of this information with other patient history data will likely have to rely on artificial intelligence (AI). AI agents can help augment transplant decision making by discovering patterns and making predictions for specific patients that are not covered in the literature or in ways that are impossible for humans to anticipate by integrating vast amounts of data (e.g. trending across numerous biomarkers). Similar to other clinical decision support systems, AI may help overcome human biases or judgment errors. However, AI is not widely utilized in transplant to date. In this rapid review, we survey the methods employed in recent research in transplant-related AI applications and identify concerns related to implementing these tools. We identify three key challenges (bias/accuracy, clinical decision process/AI explainability, AI acceptability criteria) holding back AI in transplant. We also identify steps that can be taken in the near term to help advance meaningful use of AI in transplant (forming a Transplant AI Team at each center, establishing clinical and ethical acceptability criteria, and incorporating AI into the Shared Decision Making Model).

Correlation of Donor-Derived Cell-free DNA with Histology and Molecular Diagnoses of Kidney Transplant Biopsies

BACKGROUND

Circulating donor-derived cell free DNA (cfDNA), a minimally invasive diagnostic tool for kidney transplant rejection, was validated using traditional histology. The Molecular Microscope (MMDx) tissue gene expression platform may provide increased precision to traditional histology.

METHODS

In this single-center prospective study of 208 biopsies (median=5.8 months) post-transplant, we report on the calibration of cfDNA with simultaneous biopsy assessments using MMDx and histology by Area under the curve (AUC) analyses for optimal criterion, as well as for, previously published cfDNA cut-offs ≤0.21% to 'rule-out' rejection and ≥1% to 'rule-in' rejection.

RESULTS

There were significant discrepancies between histology and MMDx, with MMDx identifying more antibody-mediated rejection (65; 31%) than histology (43; 21%); the opposite was true for T-cell mediated rejection [TCMR; histology: 27 (13%) vs MMDx: 13 (6%)]. Most of the TCMR discrepancies were seen for histologic borderline/1A TCMR. AUC Curves for cfDNA and prediction of rejection were slightly better with MMDx (AUC=0.80; 95%CI: 0.74-0.86) vs. histology (AUC=0.75; 95%CI: 0.69-0.81). A cfDNA≤0.21% had similar sensitivity (~91%) to 'rule-out' rejection by histology and MMDx. Specificity was slightly higher with MMDx (92%) compared with histology (85%) to 'rule-in' rejection using cfDNA criterion≥1%. Strong positive quantitative correlations were observed between cfDNA scores and molecular acute kidney injury (AKI) for both 'rejection' and 'nonrejection' biopsies.

CONCLUSIONS

Molecular diagnostics using tissue gene expression and blood-based donor-derived cell-free DNA may add precision to some cases of traditional histology. The positive correlation of cfDNA with molecular AKI suggests a dose-dependent association with tissue injury irrespective of rejection characteristics.

Discovering novel injury features in kidney transplant biopsies associated with TCMR and donor aging

We previously characterized the molecular changes in acute kidney injury (AKI) and chronic kidney disease (CKD) in kidney transplant biopsies, but parenchymal changes selective for specific types of injury could be missed by such analyses. The present study searched for injury changes beyond AKI and CKD related to specific scenarios, including correlations with donor age. We defined injury using previously defined gene sets and classifiers and used principal component analysis to discover new injury dimensions. As expected, Dimension 1 distinguished normal vs. injury, and Dimension 2 separated early AKI from late CKD, correlating with time posttransplant. However, Dimension 3 was novel, distinguishing a set of genes related to epithelial polarity (e.g., PARD3) that were increased in early AKI and decreased in T cell-mediated rejection (TCMR) but not in antibody-mediated rejection. Dimension 3 was increased in kidneys from older donors and was particularly important in survival of early kidneys. Thus high Dimension 3 scores emerge as a previously unknown element in the kidney response-to-injury that affects epithelial polarity genes and is increased in AKI but depressed in TCMR, indicating that in addition to general injury elements, certain injury elements are selective for specific pathologic mechanisms. (ClinicalTrials.gov NCT01299168).

How will artificial intelligence and bioinformatics change our understanding of IgA Nephropathy in the next decade?

IgA nephropathy (IgAN) is the most common glomerulonephritis. It is characterized by the deposition of immune complexes containing immunoglobulin A (IgA) in the kidney’s glomeruli, triggering an inflammatory process. In many patients, the disease has a progressive course, eventually leading to end-stage kidney disease. The current understanding of IgAN’s pathophysiology is incomplete, with the involvement of several potential players, including the mucosal immune system, the complement system, and the microbiome. Dissecting this complex pathophysiology requires an integrated analysis across molecular, cellular, and organ scales. Such data can be obtained by employing emerging technologies, including single-cell sequencing, next-generation sequencing, proteomics, and complex imaging approaches. These techniques generate complex “big data,” requiring advanced computational methods for their analyses and interpretation. Here, we introduce such methods, focusing on the broad areas of bioinformatics and artificial intelligence and discuss how they can advance our understanding of IgAN and ultimately improve patient care. The close integration of advanced experimental and computational technologies with medical and clinical expertise is essential to improve our understanding of human diseases. We argue that IgAN is a paradigmatic disease to demonstrate the value of such a multidisciplinary approach.

Factors associated with kidney graft survival in pure antibody-mediated rejection at the time of indication biopsy: Importance of parenchymal injury but not disease activity

We studied the relative association of clinical, histologic, and molecular variables with risk of kidney transplant failure after an indication biopsy, both in all kidneys and in kidneys with pure antibody-mediated rejection (ABMR). From a prospective study of 1679 biopsies with histologic and molecular testing, we selected one random biopsy per patient (N = 1120), including 321 with pure molecular ABMR. Diagnoses were associated with actuarial survival differences but not good predictions. Therefore we concentrated on clinical (estimated GFR [eGFR], proteinuria, time posttransplant, donor-specific antibody [DSA]) and molecular and histologic features reflecting injury (acute kidney injury [AKI] and atrophy-fibrosis [chronic kidney disease (CKD)] and rejection. For all biopsies, univariate analysis found that failure was strongly associated with low eGFR, AKI, CKD, and glomerular deterioration, but not with rejection activity. In molecular ABMR, the findings were similar: Molecular and histologic activity and DSA were not important compared with injury. Survival in DSA-negative and DSA-positive molecular ABMR was similar. Multivariate survival analysis confirmed the dominance of molecular AKI, CKD, and eGFR. Thus, at indication biopsy, the dominant predictors of failure, both in all kidneys and in ABMR, were related to molecular AKI and CKD and to eGFR, not rejection activity, presumably because rejection confers risk via injury.