The Lung Allocation Score and Other Available Models Lack Predictive Accuracy for Post-Lung Transplant Survival

Clinical Deployment of Machine Learning Tools in Transplant Medicine: What Does the Future Hold?

Medical applications of machine learning (ML) have shown promise in analyzing patient data to support clinical decision-making and provide patient-specific outcomes. In transplantation, several applications of ML exist which include pretransplant: patient prioritization, donor-recipient matching, organ allocation, and posttransplant outcomes. Numerous studies have shown the development and utility of ML models, which have the potential to augment transplant medicine. Despite increasing efforts to develop robust ML models for clinical use, very few of these tools are deployed in the healthcare setting. Here, we summarize the current applications of ML in transplant and discuss a potential clinical deployment framework using examples in organ transplantation. We identified that creating an interdisciplinary team, curating a reliable dataset, addressing the barriers to implementation, and understanding current clinical evaluation models could help in deploying ML models into the transplant clinic setting.

Cell-Free DNA Maps Tissue Injury and Correlates with Disease Severity in Lung Transplant Candidates

Rationale: Plasma cell-free DNA levels correlate with disease severity in many conditions. Pretransplant cell-free DNA may risk stratify lung transplant candidates for post-transplant complications. Objectives: To evaluate if pretransplant cell-free DNA levels and tissue sources identify patients at high risk of primary graft dysfunction and other pre- and post-transplant outcomes. Methods: This multicenter, prospective cohort study recruited 186 lung transplant candidates. Pretransplant plasma samples were collected to measure cell-free DNA. Bisulfite sequencing was performed to identify the tissue sources of cell-free DNA. Multivariable regression models determined the association between cell-free DNA levels and the primary outcome of primary graft dysfunction and other transplant outcomes, including Lung Allocation Score, chronic lung allograft dysfunction, and death. Measurements and Main Results: Transplant candidates had twofold greater cell-free DNA levels than healthy control patients (median [interquartile range], 23.7 ng/ml [15.1-35.6] vs. 12.9 ng/ml [9.9-18.4]; P < 0.0001), primarily originating from inflammatory innate immune cells. Cell-free DNA levels and tissue sources differed by native lung disease category and correlated with the Lung Allocation Score (P < 0.001). High pretransplant cell-free DNA increased the risk of primary graft dysfunction (odds ratio, 1.60; 95% confidence interval [CI], 1.09-2.46; P = 0.0220), and death (hazard ratio, 1.43; 95% CI, 1.07-1.92; P = 0.0171) but not chronic lung allograft dysfunction (hazard ratio, 1.37; 95% CI, 0.97-1.94; P = 0.0767). Conclusions: Lung transplant candidates demonstrate a heightened degree of tissue injury with elevated cell-free DNA, primarily originating from innate immune cells. Pretransplant plasma cell-free DNA levels predict post-transplant complications.

The transformative potential of artificial intelligence in solid organ transplantation

Solid organ transplantation confronts numerous challenges ranging from donor organ shortage to post-transplant complications. Here, we provide an overview of the latest attempts to address some of these challenges using artificial intelligence (AI). We delve into the application of machine learning in pretransplant evaluation, predicting transplant rejection, and post-operative patient outcomes. By providing a comprehensive overview of AI's current impact, this review aims to inform clinicians, researchers, and policy-makers about the transformative power of AI in enhancing solid organ transplantation and facilitating personalized medicine in transplant care.

Lung Transplantation in a New Era in the Field of Cystic Fibrosis

Lung transplantation for people with cystic fibrosis (PwCF) is a critical therapeutic option, in a disease without a cure to this day, and its overall success in this population is evident. The medical advancements in knowledge, treatment, and clinical care in the field of cystic fibrosis (CF) rapidly expanded and improved over the last several decades, starting from early pathology reports of CF organ involvement in 1938, to the identification of the CF gene in 1989. Lung transplantation for CF has been performed since 1983, and CF now accounts for about 17% of pre-transplantation diagnoses in lung transplantation recipients. Cystic fibrosis transmembrane conductance regulator (CFTR) modulators have been the latest new therapeutic modality addressing the underlying CF protein defect with the first modulator, ivacaftor, approved in 2012. Fast forward to today, and we now have a growing CF population. More than half of PwCF are now adults, and younger patients face a better life expectancy than they ever did before. Unfortunately, CFTR modulator therapy is not effective in all patients, and efficacy varies among patients; it is not a cure, and CF remains a progressive disease that leads predominantly to respiratory failure. Lung transplantation remains a lifesaving treatment for this disease. Here, we reviewed the current knowledge of lung transplantation in PwCF, the challenges associated with its implementation, and the ongoing changes to the field as we enter a new era in the care of PwCF. Improved life expectancy in PwCF will surely influence the role of transplantation in patient care and may even lead to a change in the demographics of which people benefit most from transplantation.

Machine Learning-Based Prognostic Model for Patients After Lung Transplantation

Importance

Although numerous prognostic factors have been found for patients after lung transplantation (LTx) over the years, an accurate prognostic tool for LTx recipients remains unavailable.

Objective

To develop and validate a prognostic model for predicting overall survival in patients after LTx using random survival forests (RSF), a machine learning algorithm.

Design, Setting, and Participants

This retrospective prognostic study included patients who underwent LTx between January 2017 and December 2020. The LTx recipients were randomly assigned to training and test sets in accordance with a ratio of 7:3. Feature selection was performed using variable importance with bootstrapping resampling. The prognostic model was fitted using the RSF algorithm, and a Cox regression model was set as a benchmark. The integrated area under the curve (iAUC) and integrated Brier score (iBS) were applied to assess model performance in the test set. Data were analyzed from January 2017 to December 2019.

Main Outcomes And Measures

Overall survival in patients after LTx.

Results

A total of 504 patients were eligible for this study, consisting of 353 patients in the training set (mean [SD] age, 55.03 [12.78] years; 235 [66.6%] male patients) and 151 patients in the test set (mean [SD] age, 56.79 [10.95] years; 99 [65.6%] male patients). According to the variable importance of each factor, 16 were selected for the final RSF model, and postoperative extracorporeal membrane oxygenation time was identified as the most valuable factor. The RSF model had excellent performance with an iAUC of 0.879 (95% CI, 0.832-0.921) and an iBS of 0.130 (95% CI, 0.106-0.154). The Cox regression model fitted by the same modeling factors to the RSF model was significantly inferior to the RSF model with an iAUC of 0.658 (95% CI, 0.572-0.747; P < .001) and an iBS of 0.205 (95% CI, 0.176-0.233; P < .001). According to the RSF model predictions, the patients after LTx were stratified into 2 prognostic groups displaying significant difference, with mean overall survival of 52.91 months (95% CI, 48.51-57.32) and 14.83 months (95% CI, 9.44-20.22; log-rank P < .001), respectively.

Conclusions and relevance

In this prognostic study, the findings first demonstrated that RSF could provide more accurate overall survival prediction and remarkable prognostic stratification than the Cox regression model for patients after LTx.

Systems prediction of chronic lung allograft dysfunction: Results and perspectives from the Cohort of Lung Transplantation and Systems prediction of Chronic Lung Allograft Dysfunction cohorts

Background

Chronic lung allograft dysfunction (CLAD) is the leading cause of poor long-term survival after lung transplantation (LT). Systems prediction of Chronic Lung Allograft Dysfunction (SysCLAD) aimed to predict CLAD.

Methods

To predict CLAD, we investigated the clinicome of patients with LT; the exposome through assessment of airway microbiota in bronchoalveolar lavage cells and air pollution studies; the immunome with works on activation of dendritic cells, the role of T cells to promote the secretion of matrix metalloproteinase-9, and subpopulations of T and B cells; genome polymorphisms; blood transcriptome; plasma proteome studies and assessment of MSK1 expression.

Results

Clinicome: the best multivariate logistic regression analysis model for early-onset CLAD in 422 LT eligible patients generated a ROC curve with an area under the curve of 0.77. Exposome: chronic exposure to air pollutants appears deleterious on lung function levels in LT recipients (LTRs), might be modified by macrolides, and increases mortality. Our findings established a link between the lung microbial ecosystem, human lung function, and clinical stability post-transplant. Immunome: a decreased expression of CLEC1A in human lung transplants is predictive of the development of chronic rejection and associated with a higher level of interleukin 17A; Immune cells support airway remodeling through the production of plasma MMP-9 levels, a potential predictive biomarker of CLAD. Blood CD9-expressing B cells appear to favor the maintenance of long-term stable graft function and are a potential new predictive biomarker of BOS-free survival. An early increase of blood CD4 + CD57 + ILT2+ T cells after LT may be associated with CLAD onset. Genome: Donor Club cell secretory protein G38A polymorphism is associated with a decreased risk of severe primary graft dysfunction after LT. Transcriptome: blood POU class 2 associating factor 1, T-cell leukemia/lymphoma domain, and B cell lymphocytes, were validated as predictive biomarkers of CLAD phenotypes more than 6 months before diagnosis. Proteome: blood A2MG is an independent predictor of CLAD, and MSK1 kinase overexpression is either a marker or a potential therapeutic target in CLAD.

Conclusion

Systems prediction of Chronic Lung Allograft Dysfunction generated multiple fingerprints that enabled the development of predictors of CLAD. These results open the way to the integration of these fingerprints into a predictive handprint.