Plain Language Summary of the PLUTO Study Prediction of Lung Transplant Outcomes (PLUTO) What is this study about? This study aims to improve how doctors predict the health of lung transplant recipients over time. Many people with severe lung disease need a transplant, but even after receiving a new lung, some still face serious health issues. One of the biggest problems is chronic lung transplant dysfunction (CLAD), which can slowly damage the new lung and is currently irreversible. What is the goal of the study? Researchers want to better understand why some people do worse after a lung transplant. Researchers hope to identify early warning signs and improve diagnosis. The main goal is to build a model that can predict how well a lung transplant will function over time, using routine health data and test results from transplant patients. Who can take part in this study? People aged 15 and older who had a lung transplant between 2009 and 2027 and are being followed at one of the study centers. People who speak French and have national health insurance. People who gave written consent (or whose guardians did, if under 18). The study may also use past data from deceased patients who did not object to research use. How will the study work? The study will follow about 4,200 lung transplant recipients across many centers in France. Researchers will collect clinical data, lung function tests, biopsy results, and blood samples. Researchers will also study new biomarkers (signals in the body that may show how well a transplant is doing) found in blood or lung samples. Using these data, the investigators will build and test tools to predict transplant outcomes. Why is this research important? By understanding early signs of transplant problems, doctors can act sooner and tailor treatment for each person. This may improve long-term survival after a lung transplant and help guide future research. How long is the study? Each participant will be followed for about 3 years, and the full study will last 6 years, including data analysis.
Lung transplantation remains the sole treatment option for thousands of patients with end-stage lung diseases. However, it retains the poorest prognosis of all solid organ transplants. Current challenges include precise, timely identification of different types of rejection episodes, and early identification of risk factors for poor outcomes, particularly chronic lung allograft dysfunction (CLAD), which can take years to develop and is at present irreversible so that preventive strategies are needed. Survival after lung transplantation is limited largely by CLAD, a process by which, through a variety of mechanisms, the patient progressively loses lung function. Early diagnosis of CLAD, and identification of clinical events (such as rejection) which predispose the patient to CLAD, is a key issue in lung transplant research. The investigators aim to assess the trajectories of lung function after transplantation, to identify patterns associated with worse prognosis, and to identify the risk factors of belonging to these trajectories. Then the investigators will propose a dynamic prediction model based on repeated graft evaluation. In addition, the investigators will assess a variety of novel biomarkers, which can be measured in blood, biopsy specimens or bronchoalveolar lavage. These biomarkers will be evaluated in conjunction with the lung function trajectory, baseline known risk factors, and repeated graft evaluation parameters to identify patterns which can define more accurately graft dysfunction, predict poor outcomes at any stage following transplantation, assist clinicians in prognostication for the individual patient, and support the development of future research studies.
Study Type
INTERVENTIONAL
Allocation
NA
Purpose
DIAGNOSTIC
Masking
NONE
Enrollment
4,200
Blood sampling and biocollection for research purposes, in addition to routine clinical care. Samples include plasma (EDTA tubes), peripheral blood mononuclear cells (CPT tubes), cell-free DNA (Streck tubes), PaxGene tubes and other predefined blood samples collected at clinically indicated time points according to the PLUTO study schedule. When transbronchial biopsies (TBB) are performed for clinical indications, additional tissue fragments are collected for the biocollection. Induced sputum and stool samples may also be collected. These samples are used for biomarker analyses and for the development and validation of multidimensional prognostic models of lung graft function and outcomes.
APHP - site de Cochin
Paris, Paris 14, France
APHP - site de Bichat
Paris, Paris 18, France
CHU Strasbourg
Strasbourg, Strasbourg, France
CHU Bordeaux & INSERM U1045
Talence, Talence, France
Chu Grenoble
Grenoble, France
CML La Fondation Hôpital Saint-Joseph
Le Plessis-Robinson, France
CHU LYON
Lyon, France
CHU Nantes Hôpital Maison Blanche
Nantes, France
Hôpital Foch
Suresnes, France
Discrimination Performance of the Multidimensional Dynamic Prognostic Model for Lung Graft Outcomes, Measured by Area Under the Curve (AUC)
To develop a multidimensional and dynamic prediction system for lung graft outcomes, joint multivariate Bayesian models with shared parameters and optimal artificial intelligence approaches will be used to assess associations between longitudinal biomarkers and survival data. This approach combines a Cox model linking time-to-event data with joint mixed models estimating trajectories of repeated measures. Clinical, biological, functional, histological and immunological variables before, during and after transplantation will be included in the Cox model, while repeated post-transplantation measures will be integrated into the mixed models. The effect of repeated measures will be evaluated using the last value, slope, and cumulative effect (area under the curve). The best model will be selected based on its discrimination performance, measured by Area Under the Curve (AUC), and evaluated in internal and external cohorts.
Time frame: 72 months
Number of Distinct Graft Dysfunction Phenotypes Identified by Hierarchical Clustering Analysis
To identify novel rejection and graft dysfunction phenotypes, an unsupervised hierarchical ascendant clustering (CAH) based on principal components of histological data from transbronchial biopsies (TBB), as well as clinical and immunological characteristics, will be applied. This unsupervised method groups similar data points (TBB) so that points within the same cluster are more similar to each other than to those in other clusters. Initially, each data point is a distinct cluster. At each step, distances between clusters are calculated using Euclidean distance and Ward's method, then the closest clusters are merged until only one remains. To determine the optimal number of clusters, inertia gain will be used by calculating intra-cluster variability and selecting clusters with minimal variability. Finally, k-means consolidation will refine clusters to minimize inertia and ensure internal coherence.
Time frame: 72 months
Differential Expression Levels of Candidate Biomarkers Between Patients With and Without Graft Dysfunction (Cross-Sectional Analysis)
Candidate biomarkers have been identified based on a literature review and scientific expertise from each participating center. They will first be evaluated by the team responsible for each project or decided within pilot studies comparing diseased and healthy subjects (mainly CLAD cases and non-CLAD cases), using discrimination measures such as area under the Receiver Operating Characteristic (ROC) curve, sensitivity, specificity, positive and negative predictive values, likelihood ratios, and calibration.
Time frame: 72 months
Validation of Differential Expression of Candidate Biomarkers in an Independent External Cohort (Cross-Sectional Analysis)
Based on preliminary results, the biomarker will then be evaluated in a validation study comparing biomarker expression in diseased and non-diseased subjects (mainly CLAD cases and non-CLAD cases), using the same discrimination and calibration measures.
Time frame: 72 months
Improvement in Discrimination Performance (AUC) of the Prognostic Model After Integration of Validated Biomarkers
To identify Forced Expiratory Volume in 1 second (FEV1) trajectories, latent class mixed models will be applied, and associations between FEV1 trajectory groups and rejection phenotypes will be explored using appropriate statistical methods. The validated biomarkers will then be integrated into the multidimensional clinical models to evaluate whether they improve the prediction of lung graft outcomes. The best performing biomarker or biomarker combination, measured by gain in discrimination (AUC), calibration, or clinical utility, will be identified through comparison of multivariable models.
Time frame: 72 months
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.