Utility of donor-derived cell-free DNA testing after lung transplantation in the precision medicine era

Abstract

Donor-derived cell-free DNA (dd-cfDNA) is a validated, highly sensitive, plasma molecular biomarker of allograft injury after solid organ transplantation. Robust experiences with dd-cfDNA testing after kidney and heart transplantation have generated interest in this biomarker within the lung transplantation (LTx) community. A growing body of evidence now provides increased insight into dd-cfDNA utility for molecular monitoring of lung allograft health after transplantation. The expanding understanding of lung allograft injury to appropriately frame the advancing role of dd-cfDNA in the evolution of the diagnostic approach after LTx is described. Performance characteristics of both laboratory-based shotgun-sequenced testing from the Genome Transplant Dynamics (GTD) and Genomic Research Alliance for Transplantation (GRAfT) consortia, as well as commercially available central lab-based algorithmic next-generation sequenced dd-cfDNA tests for lung transplant recipients (LTR) (AlloSure, CareDx and Prospera, Natera) are described. Kinetics of dd-cfDNA in LTRs over time, in multiple different clinical scenarios, from several investigator groups are aggregated. Phenotypes of lung allograft injury, such as acute lung allograft dysfunction and associated dd-cfDNA patterns and performance, are identified in alignment with established definitions and evolving molecular injury insights. Certain patterns of molecular injury that may predict long-term outcomes including chronic lung allograft dysfunction and mortality are examined. Lastly, clinical approaches to testing and interpretation of dd-cfDNA results in LTRs, a practical approach to using dd-cfDNA, and a rational framework for interpreting dd-cfDNA results in LTRs are presented.

Introduction

Donor-derived cell-free DNA (dd-cfDNA) is a validated, highly sensitive, plasma molecular biomarker of allograft injury after solid organ transplantation.^1–38 Robust experiences with dd-cfDNA testing after kidney and heart transplantation have generated interest in this biomarker within the lung transplantation (LTx) community.^35–38 In kidney and heart transplant recipients, dd-cfDNA testing serves as a reasonable rule-out test for rejection events, as few non-alloimmune sources of injury occur in these allografts. In contrast to kidney and heart transplantation, the lung allograft is susceptible to various alloimmune and non-alloimmune sources of injury; as such, uncertainty regarding the role of dd-cfDNA in the overall diagnostic framework after LTx has persisted.^39,40 In the decade since feasibility of dd-cfDNA for the detection of lung allograft injury was first established,¹ growing evidence (Figure 1) have provided deeper insight into the utility of dd-cfDNA molecular monitoring after LTx.^1–33 We describe the contemporary role of dd-cfDNA as an important element of the evolving diagnostic strategy for detecting allograft injury after LTx.

Figure 1.

Accumulating evidence for dd-cfDNA testing in LTRs.^1–33 dd-cfDNA, donor-derived cell-free DNA. rd-cfDNA, recipient-derived cell-free DNA. LTR, lung transplant recipients. GTD, Genome Transplant Dynamics. GRAfT, Genomic Research Alliance for Transplantation. Seq., sequencing. BLTR, bilateral lung transplant recipient. SLTR, single lung transplant recipient. CLAD, chronic lung allograft dysfunction. ALAD, acute lung allograft dysfunction. ASL, AlloSure-Lung. ACR, acute cellular rejection. AMR, antibody-mediated rejection. RefCV, reference change value. PL, Prospera-Lung. DQS, donor quantity score. UTSA, University of Texas San Antonio. BLAD, baseline lung allograft dysfunction. EMI, extreme molecular injury. TBBx, transbronchial biopsy.

Lung allograft injury diagnostic approach evolution

Our understanding of the complexity of lung allograft injury has improved and continues to evolve.^41–75 Figure 2 describes the evolution of the allograft injury framework after LTx and Figure 3 describes the phases of lung allograft injury and associated median LTR survival over time.^{1,3,12,14,41–75} Median 50% survival of 6.7 years post-LTx (Figure 3) was assessed in the International Society for Heart and Lung Transplantation (ISHLT) registry cohort transplanted between 2010–2017; during which time approximately 12–18% of LTx were performed for recipients with cystic fibrosis.⁷² This median 50% survival remains the shortest of all major solid organs and may be decreasing due to the minimization of LTx for cystic fibrosis and the expansion of aging LTRs with interstitial lung diseases. This underlying demographic shift highlights the need for less invasive injury markers to help manage complex LTRs today.

Figure 2.

Layered evolution of lung allograft injury diagnostic framework.^{1,3,12,14,41–69,76–84} dd-cfDNA, donor-derived cell-free DNA. LTx, lung transplantation. LTRs, lung transplant recipients. ISHLT, International Society for Heart and Lung Transplantation. CMV, cytomegalovirus. BOS, bronchiolitis obliterans syndrome. AMR, antibody-mediated rejection. PGD, primary graft dysfunction. SAB, single antigen bead. DSA, donor-specific antibody. CARV, community acquired respiratory virus. ATS, American Thoracic Society. ERS, European Respiratory Society. RAS, restrictive allograft syndrome. FVC, forced vital capacity. CLAD, chronic lung allograft dysfunction. BLAD, baseline lung allograft dysfunction. BAL, bronchoalveolar lavage. ALAD, acute lung allograft dysfunction. STS, short telomere syndrome.

Figure 3.

Lung allograft injury phases and median LTR survival over time.^41–72 a, ISHLT registry median 50% survival, 2010–17 cohort.⁷⁰ LTR, lung transplant recipients. PGD, primary graft dysfunction. BLAD, baseline lung allograft dysfunction. ACR, acute cellular rejection. AMR, antibody mediated rejection.

The introduction of the first pathologic grading scheme for transbronchial biopsy (TBBx) histologic interpretation in 1990 and routine post-LTx spirometry in 1993 established the early clinical, pathologic, and functional diagnostic approach to allograft injury in an era defined by watchful waiting for harmful events requiring intervention.^{41,43,44,46,50,60} Subsequently, multiple groups identified several additional drivers of allograft injury including gastroesophageal reflux, certain infections, air pollution, and airway neutrophilia with and without responsiveness to chronic azithromycin therapy.^{30,42,45–58} The impact of alloantibody development on allograft injury was identified in 2004 with further evidence accumulating over the succeeding decade.^48,62 The first ISHLT chronic lung allograft dysfunction (CLAD) consensus statement was published in 2019, and the first ISHLT Bronchoscopy and Bronchoalveolar Lavage (BAL) consensus statement was published in 2020.^{59,61,64–69} An evolving recognition that multiple processes drive allograft injury led to a more nuanced refinement of the injury framework from the earlier clinical, pathologic, and functional injury approach. The multiple-injury phenotypes framework involves evaluating diverse sources of injury which may or may not produce acutely observable pathologic changes or be associated with acute decline in lung function.

The adoption of ex vivo lung perfusion (EVLP) has enabled real-time physiologic and molecular assessment of donor lungs prior to implantation allowing objective characterization of injury that was previously inferred indirectly or evaluated only after transplantation. As donor lungs are isolated during EVLP, emerging data suggest that the quantity of cfDNA released during perfusion can provide a molecular readout of donor lung injury, ischemia reperfusion stress, and organ-level cellular turnover before recipient exposure.⁷³ Therefore, incorporation of EVLP cfDNA measurements represents an early extension of the multiple injury phenotype framework enabling molecular risk stratification and organ selection prior to transplantation.

The Genome Transplant Dynamics (GTD) study group from Stanford first explored dd-cfDNA in LTRs in 2015.¹ The GTD group focused their analyses on the concordance of dd-cfDNA measurements, quantitated using shotgun sequencing after genotyping the donor and recipient, with clinical and pathologic features of acute cellular rejection (ACR). Area under the receiver operating characteristics curve (AUROC) for the detection of moderate-high grade ACR (≥A3) using dd-cfDNA was found to be 0.9, and mild or higher grade ACR (≥A2) AUROC was found to be 0.76. These findings established the viability of dd-cfDNA for the detection of ACR in LTRs. Subsequently, the GTD and Genomic Research Alliance for Transplantation (GRAfT) consortia validated dd-cfDNA, again using shotgun sequencing, for the detection of ACR and further evaluated dd-cfDNA for the detection, severity, and prognostic ability in the setting of multiple different injury phenotypes described in Section 5 and delineated in Figure 1.^{2,3,6,7,9–13,18,19,23–25,27,30,34}

Publication of the VALID and ALARM studies in 2022 offered the first clinical validation and interpretive data for the two Clinical Laboratory Improvement Amendments/College of American Pathologists (CLIA/CAP) certified laboratory-developed dd-cfDNA tests (Prospera and AlloSure, respectively) for use in lung transplant recipients (LTR).^12,14 The findings from these studies informed the United States Centers for Medicare and Medicaid Services decision to incorporate dd-cfDNA molecular monitoring after LTx into its local coverage determination in 2023.⁹⁶ Since that time, multiple groups have published analyses highlighting the utility of dd-cfDNA molecular monitoring in various clinical scenarios. In aggregate, dd-cfDNA represents the first clinically available, minimally invasive, plasma-based, lung allograft injury biomarker that identifies molecular quiescence, enables detection of subclinical injury, and may inform when targeted intervention in specific clinical contexts might be appropriate. While additional emerging biomarkers may offer greater molecular endotype specificity, they are beyond the scope of this paper.^76–83

Commercially available lab-developed dd-cfDNA tests

Three CLIA/CAP certified central lab-developed dd-cfDNA tests using next-generation sequencing (NGS) platforms are available in the United States: AlloSure (CareDx, Inc., Brisbane CA), Prospera (Natera, Inc., Austin TX), and TRAC (Eurofins Transplant Genomics, Lenexa KS). Additionally, three decentralized lab-developed kit-based dd-cfDNA tests are available worldwide: AlloSeq (CareDx, Inc., Brisbane CA) using NGS, Accept cfDNA (Devyser/One Lambda, Thermo Fisher Scientific, Inc., Waltham MA) also using NGS, and GraftAssureDx (iMDx: Insight Molecular Diagnostics, Inc., Nashville TN) using droplet digital polymerase chain reaction (PCR). Table 1 delineates the characteristics of the two commercially available central lab-based tests for LTRs: AlloSure and Prospera. There are several similarities and some notable differences between these two tests.^12,14,85,86 The AlloSure algorithm assesses 405 equal distribution biallelic single nucleotide polymorphisms (SNPs) with sufficient separation and low linkage across all 22 somatic chromosomes^88–90 where Prospera targets > 13000 SNPs across 4 somatic chromosomes^90,91; both of which are agnostic to race/ethnicity. The implications of these differences are minimal when considering the fraction of dd-cfDNA to total cfDNA in most clinical situations.

Table 1.

Commercially Available Central Lab-based dd-cfDNA Tests for LTRs

Parameter	AlloSure (CareDx)	Prospera (Natera)
Operations
CLIA/CAP certified test	Yes	Yes
CLIA-certified central-lab location	Brisbane, California USA	San Carlos, California USA
EMR integration	Yes	Yes
Test characteristics
Testing platform	NGS	NGS
Donor genotyping required	No	No
Chromosomes and SNPs assessed	22 chromosomes 405 SNPs	4 chromosomes > 13000 SNPs
Routine reporting parameter	Percent donor-derived to total cfDNA fraction	Percent donor-derived to total cfDNA fraction
Lower limit of detection (LoD)	0.04%	0.08%
Upper limit of quantitation (LoQ)	16%	25%
dd-cfDNA copies/mL available	Donor DNA Count (DDC)	Donor Quantity Score (DQS)
Interpretation
Population-based threshold for elevated result in BLTRs	0.9%	1%
Approach to SLTRs	Use change from baseline	Mass-correction at central-lab; use BLTR threshold
BLTR reference change value (RefCV)	0–2 years post-LT: 70% > 2 years post-LT: 73%	< 3 years post-LT: 72.7%
Science
Seminal paper(s)	ALARM12 J Heart Lung Transplant 2022;41:458–66.	VALID14 Transplant Direct 2022;8:e1317. SLAM22 JHLT Open 2024;8:100155.
Ongoing studies	ALAMO Registry85 NCT05050955 Recruitment complete n=550, 21 sites Follow-up period: 3 years Follow-up complete: 2028 Planned extension: 5 years Extension complete: 2030	LAMBDA-002 Registry86 NCT05170425 Not actively recruiting n=154, 11 sites Follow-up period: 3 years Estimated completion: 2027 Trifecta-Lung (ATAGC)87 University of Alberta NCT05837663 Actively recruiting n=120/600, 17 sites Follow-up period: 1.5 years Estimated completion: 2027

dd-cfDNA, donor-derived cell-free DNA. CLIA, Clinical Laboratory Improvement Amendments. CAP, College of American Pathologists. EMR, electronic medical record. NGS, next-generation sequencing. SNPs, single nucleotide polymorphisms. BLTR, bilateral lung transplant recipients. SLTR, single lung transplant recipients. RefCV, reference change value. ATAGC, Alberta Transplant Applied Genomics Centre.

Differences in AlloSure and Prospera test characteristics become relevant when interpreting dd-cfDNA copies/mL, reported as AlloSure Donor DNA Count (DDC) and Prospera Donor Quantity Score (DQS). Because neither platform performs direct molecule counting or unique molecular identifier-based correction, and each assay employs distinct SNP selection strategies, chromosomal sampling, and proprietary approaches for total cfDNA calculation, dd-cfDNA copies/mL results are inherently platform-dependent. As such, while dd-cfDNA fraction is likely comparable across assays, these technical differences limit cross-platform comparability and necessitate assay-specific approaches for clinical interpretation of dd-cfDNA copies/mL results.^31,32

Both tests use similar population-based thresholds in bilateral LTRs (BLTR); however, their approaches diverge for single LTRs (SLTR). AlloSure reports dd-cfDNA results unaltered irrespective of allograft type; where Prospera applies a mass-correction method by doubling the SLTR results at the central laboratory before reporting the result to the center.²² The mass-correction strategy was first described by the GTD/GRAfT consortia investigators in 2022 who identified that this approach approximates bilateral values reasonably well in SLTRs with both a stable allograft and native lung.¹⁰ However, native lung instability due to continued deterioration, infections, malignancies, or other injuries may change circulating recipient-derived cfDNA (rd-cfDNA); thus, altering SLTR dd-cfDNA fraction, which is unaccounted for by mass-correction.^31,32 The GTD/GRAfT consortia data further demonstrated that mass-correction of SLTRs resulted in significantly lower dd-cfDNA values compared to true BLTRs in the setting of antibody-mediated rejection (AMR) (mass-corrected SLTR=2.17% (1.55–6.5) vs. BLTR=2.66% (1.55–6.93), p < 0.01) and significantly higher dd-cfDNA values in the setting of ACR (mass-corrected SLTR=1.74% (1.44–2.29) vs BLTR=1.18% (0.52–1.52), p=0.03).¹⁰ Longitudinal assessment from the GTD/GRAfT consortia suggests that the relationship between median dd-cfDNA release in SLTRs vs BLTRs may not be proportional over time (Figure 4) where SLTRs and BLTRs nadir at approximately 3 and 6 months, respectively, post-LTx and diverge to a greater degree over the subsequent 12–18 months. These differences suggest that longitudinal assessment may lead to more informed patient management decisions rather than single time point measurements.

Figure 4.

dd-cfDNA kinetic patterns in BLTRs vs SLTRs^a,10+. a, GTD/GRAfT consortia. dd-cfDNA, donor-derived cell-free DNA. BLTR, bilateral lung transplant recipients. SLTR, single lung transplant recipients.

dd-cfDNA kinetics in LTRs

Kinetics of dd-cfDNA after LTx have been described by multiple groups (Table 2) in aggregate and stratified into BLTR and SLTR subgroups.^1–34 Table 3 describes the reference change value (RefCV), the upper limit of the variability assigned to the combined analytic and biologic variability, parameters for both AlloSure and Prospera in multiple distinct cohorts. AlloSure parameters were described in two cohorts of BLTRs: 0–2 years post-LTx from the ALARM cohort (n=35 LTR with mean of 3.7 AlloSure results per patient) and > 2 years post-LTx from the Vanderbilt cohort (n=51 LTR with mean of 3.5 AlloSure results per patient).^13,15 Within-subject variation (CV_I) remains low over time which suggests that trending values in the same patient may be useful to identify injurious events. Additionally, there is significant between-subject variation (CV_G) and index of individuality (II) less than 0.6 suggesting that population thresholds may be less valuable than within-patient changes in results over time. RefCV were calculated for both cohorts of BLTRs: RefCV=70% in 0–2 years post-LTx and RefCV=73% > 2 years post-LTx. Prospera parameters were described in the Michigan State cohort of BLTRs (n=18 with median of 3 Prospera results per patient) within 3 years of LTx. Similar to AlloSure, Prospera CV_I is low and CV_G is high yielding a low II (0.36). Prospera RefCV was then calculated at 72.7%.¹⁷

Table 2.

LTx dd-cfDNA Cohort Characteristics

Parameter	GTD/GRAfT1–3,6,7,9–13,18,19,23–25,27,30	VALID14	ALARM12
LTR enrollment	305	104	198
dd-cfDNA results	2013	204	410
Centers	4 GTD: Stanford GRAfT: Johns Hopkins University of Maryland Inova Fairfax	1 Ohio State University	4 Johns Hopkins University of Maryland Inova Fairfax University of Texas San Antonio
Sponsor	NIH, NHLBI	Natera, Inc.	CareDx, Inc.
dd-cfDNA test evaluated	Shotgun sequencing	Prospera	AlloSure
Design	Longitudinal cohort	Cohort	Longitudinal cohort with per protocol management
Enrollment period	GTD: 12/2010–12/2012 GRAfT: 7/2015–10/2020	9/2020–6/2021	3/2020–9/2020
Follow-up period	GTD: 5 years; complete GRAfT: 5 years; ongoing	93% of samples collected within 550 days of LTx	Acute: 3 months; complete Long-term: ongoing
BLTRs	72.7%	78.6%	82%
Age at transplant (years)	Mean: 55.4 SD: 13.9	Median: 62 IQR: 55.5–68	Mean: 58.5 SD: 12.7
Female sex	47.2%	40%	46%
Race
White	80.7%	93.2%	64%
African American	14.7%	5.8%	18%
Other	4.6%	1%	18%
Native lung disease
Interstitial lung diseases	47.5%	48.2%	62.6%
Obstructive lung diseases	17.7%	41.4%	15.7%
Cystic fibrosis	14.2%	2.6%	7%
Pulmonary hypertension	2.9%	6.2%	3%
Others	17.7%	1.6%	11.6%
Endpoint(s)	ALAD, CLAD, death	AR, allograft injuries	ALAD
NLM identifier	GTD: NCT01985412 GRAfT: NCT0243070

dd-cfDNA, donor-derived cell-free DNA. LTx, lung transplantation. LTR, lung transplant recipients. BLTR, bilateral lung transplant recipients. GTD, Genome Transplant Dynamics. GRAfT, Genomic Research Alliance for Transplantation. NIH, National Institutes of Health. NHLBI, National Heart Lung and Blood Institute. ALAD, acute lung allograft dysfunction. AR, acute rejection.

NLM, National Library of Medicine.

Table 3.

dd-cfDNA Reference Change Parameters in Quiescent BLTRs

	AlloSure	AlloSure	Prospera
Parameter	0–2 years post-LTx	> 2 years post-LTx	< 3 years post-LTx
Within-subject coefficient of variation (CVI)	25%	26%	26.2%
Between-subject robust coefficient of variation (CVG)	53%	47%	73%
Index of individuality (II), CVI / CVG	0.47	0.56	0.36
Reference Change Value (RefCV)a	70%	73%	72.7%
N LTRs	35	51	18
Cohort	ALARM13	Vanderbilt15	Michigan State17

^a,clinically relevant difference between results that exceeds analytic and biologic variability alone.

dd-cfDNA, donor-derived cell-free DNA. BLTRs, bilateral lung transplant recipients. LTx, lung transplantation. LTR, lung transplant recipients.

Lung allograft injury phenotypes

Allograft injuries after LTx are multifactorial and evolve over time.^{1–34,39–75} Primary graft dysfunction (PGD)^49,63 occurs in the early hours to days following LTx. Increasing PGD severity is independently associated with allograft failure and death after LTx.^39,70 The GTD/GRAfT consortia graded PGD severity and obtained dd-cfDNA on day 3 post-LTx in 99 LTRs.⁷ They showed a significantly higher day 3 dd-cfDNA in LTRs with PGD vs those without PGD (median (IQR) 12.2% (8.2–22%) vs 8.5% (5.6–13.2%) respectively, p=0.01). They further showed significantly higher day 3 dd-cfDNA in LTRs with PGD who progressed to CLAD vs LTRs with PGD who did not progress to CLAD (median (IQR) 22.4% (11.8–27.6%) vs 9.9% (6.7–14.9%) respectively, p=0.007) with median follow-up time of 24 ± 10.5 months post-LTx. Multivariable logistic regression was performed and showed that higher day 3 dd-cfDNA was significantly associated with increased risk of developing CLAD (logOR (SE) 1.66 (0.7), p=0.02) when accounting for infections and number of ACR episodes. AUROC was 0.86 for the multivariable model including PGD, race, infections, ACR episodes, and dd-cfDNA.⁷ Importantly, these results are clinically meaningful as they identify early molecular injury (MI) severity and facilitate classification of LTRs with PGD into the higher and lower CLAD risk groups at day 3. Median plus IQR results for the lower CLAD risk PGD group fall within quantitation limits for both tests.^7,12,14

Baseline lung allograft dysfunction (BLAD), defined as post-LTx peak lung function which fails to reach FEV₁ and forced vital capacity (FVC) > 80% predicted in BLTRs (ISHLT consensus definition pending), is associated with significantly increased risk of death after LTx (HR 1.97, 95%CI 1.05–3.69, p=0.03) compared to LTRs without BLAD in multiple studies.^29,34,64,65 The GTD/GRAfT consortia assessed the prognostic impact of dd-cfDNA fraction at various early post-LTx timepoints, dd-cfDNA decay over 90-days post-LTx, and at the time of BLAD diagnosis.³⁴ At no timepoint was dd-cfDNA fraction associated with either the development of BLAD or the composite of subsequent CLAD and death. Subsequently, the Prague group assessed dd-cfDNA copies/mL using Prospera in BLTRs with BLAD vs non-BLAD controls and showed significantly higher dd-cfDNA copies/mL in patients with BLAD (39 vs 26 copies/mL respectively, p=0.01%).³¹ Dd-cfDNA copies/mL was associated with significantly increased risk of BLAD on univariate logistic regression (OR 3.1, 95%CI 1.4–7, p=0.004), and remained significantly associated in the multivariable model. Apart from other extrapulmonary etiologies of BLAD, these data suggest that early higher allograft-specific injury burden may be masked by only evaluating dd-cfDNA fraction as a result of high circulating rd-cfDNA associated with recipient injury burden from the transplant procedure and prolonged recovery.^{31,34,64,65,75}

Acute lung allograft dysfunction (ALAD) has been defined in multiple ways by various groups in both BLTRs and SLTRs (ISHLT consensus definition pending).^{12,14,16–18,22–24} The VALID investigators assessed the utility of Prospera to monitor LTR for acute rejection (AR) defined as the composite of ACR and AMR, and allograft injuries defined as the composite of acute infection, ACR, AMR, and CLAD/neutrophilic-responsive airways disease (NRAD), in a single-center cohort study (Table 2).¹⁴ Bronchoscopy and TBBx were performed per center protocol without regard to dd-cfDNA results. 104 LTR with 204 dd-cfDNA results were enrolled and 103 LTR (81 BLTR, 22 SLTR) with 195 dd-cfDNA results were analyzed. Levels of dd-cfDNA were higher at the time of AR diagnosis compared to stable LTRs (median (IQR) for ACR 1.43% (0.67–2.32%), p < 0.001 and AMR 2.5% (2.06–3.79%) p < 0.001 vs stable 0.46% (0.21–0.78%)). AR subgroup analyses were performed which showed significantly higher dd-cfDNA fraction during AMR when compared to ACR (p=0.03). Levels of dd-cfDNA were also higher at the time of infection and CLAD/NRAD diagnoses compared to stable LTRs (median (IQR) for infection 0.74% (0.46–1.38), p=0.02 and CLAD/NRAD 1.6% (0.57–2.6%), p < 0.001 vs stable). Further subgroup analyses were performed which showed lower dd-cfDNA fraction during infections when compared to ACR (p=0.07). Prospera test performance characteristics are summarized in Table 4. Both AUROC and negative predictive value (NPV) are high in the AR only cohort and the allograft injuries cohort suggesting Prospera serves as a reasonable rule-out test for AR or allograft injuries in LTRs.¹⁴ Similar results were shown in the single-center Michigan State cohort of 109 LTRs with 276 dd-cfDNA results that identified Prospera AUROC of 0.78 for the detection of ALAD, defined as the composite of acute infection, ACR, AMR, and for-cause procedure results without pathologic findings.¹⁷ Subsequently, the SLAM investigators assessed the performance of Prospera in SLTRs in an AR only cohort.²² Mass-correction was performed by doubling the SLTR dd-cfDNA results. Similar Prospera performance was shown using 1% threshold in mass-corrected SLTRs compared to the VALID cohort suggesting that Prospera serves as a reasonable rule-out test for AR in SLTRs.

Table 4.

Performance Characteristics of dd-cfDNA for Lung Allograft Dysfunction with Commercially Available Central Lab-based Tests

Parameter	AlloSure 0.9% in ALADa cohort	AlloSure 1% in ALADa cohort	AlloSure 73% RefCV in ALADb cohort of BLTR
AUROC	0.89	0.82	0.87
Sensitivity	90.0%	73.9%	87%
Specificity	82.5%	87.7%	78%
Positive predictive value	63.2%	43.4%	74%
Negative predictive value	96.1%	96.5%	89%
N LTRs	175	175	71
Cohort	ALARM12	ALARM12	Vanderbilt16
Parameter	Prospera 1% in allograft injuriesc cohort	Prospera 1% in ARd only cohort	Prospera 1% in ARe only cohort of mass-corrected SLTR
AUROC	0.76	0.91	0.85
Sensitivity	59.9%	89.1%	77.8%
Specificity	83.9%	82.9%	84.6%
Positive predictive value	43.6%	51.9%	38.3%
Negative predictive value	91.0%	97.3%	96.8%
N LTRs	103g	95g	85
Cohort	VALID14	VALID14	SLAM22

^a,acute infection, ACR, AMR, with or without decline in FEV₁.

^b,acute infection, ACR, suspected AMR, or decline in FEV₁ > 10%.

^c,acute infection, ACR, AMR, CLAD/NRAD.

^d,ACR, AMR.

dd-cfDNA, donor-derived cell-free DNA. ALAD, acute lung allograft dysfunction. BLTR, bilateral lung transplant recipients. SLTR, single lung transplant recipients. RefCV, reference change value. LTR, lung transplant recipient. ACR, acute cellular rejection.

AMR, antibody-mediated rejection. FEV₁, forced expiratory volume in 1 s. OP, organizing pneumonia. ALI, acute lung injury. AR, acute rejection. CLAD, chronic lung allograft dysfunction.

The ALARM investigators assessed the utility of AlloSure to monitor LTRs during routine clinical care in a multicenter longitudinal cohort study with the incidence of ALAD, defined as the composite of acute infection, ACR, AMR; with or without decline in FEV₁, as the primary endpoint (Table 2).¹² Bronchoscopy with BAL and TBBx were performed per study protocols guided by both elevated dd-cfDNA results and clinical events (see ALARM study supplemental Figures E1a and E1b¹²). Additionally, post-interventional management, intensified dd-cfDNA testing, and return to molecular dd-cfDNA surveillance were performed per study protocols. 198 LTR with 410 dd-cfDNA results were enrolled and 175 LTR (144 BLTR, 31 SLTR) with 380 dd-cfDNA results were analyzed. Levels of dd-cfDNA were higher at the time of ALAD diagnosis compared to stable LTRs (median (IQR) 1.7% (0.63–3.1%) vs 0.35% (0.22–0.75%) respectively, p < 0.001). Subgroup analyses were performed and did not show a difference in dd-cfDNA between ACR and AMR (p=0.26) or between acute rejection (AR, ACR and AMR) and infection (median (IQR) 1.6% (0.38–3.4%) vs 1.8% (0.84–2.74%) respectively, p=0.82). AlloSure test performance characteristics are summarized in Table 4. Both AUROC and NPV are high suggesting AlloSure < 0.9% or < 1% serves as a reasonable rule-out test for ALAD in LTRs.¹² Subsequently, the Vanderbilt group assessed the performance of AlloSure RefCV in BLTRs.^15,16 71 BLTRs > 2 years post-LTx were enrolled in the Vanderbilt cohort. ALAD was defined as the composite of acute infection, ACR, suspected AMR, or decline in FEV₁ > 10%. Similar AlloSure performance was shown using RefCV=73% in BLTR compared to the ALARM cohort thresholds of 0.9% and 1%. These data further suggest that AlloSure RefCV < 73% serves as a reasonable rule-out test for ALAD in BLTRs.^15,16

The GTD/GRAfT consortia investigators assessed molecular injury in the setting of pathologic ACR (ACR +MI), defined as dd-cfDNA ≥1% plus positive TBBx pathology for ACR, using the ISHLT Revised 2007 LTx Pathology grading scheme,⁵¹ occurring after day 30 post-LTx (Table 2).¹⁹ 205 LTRs were enrolled and 188 LTRs were analyzed; of whom, 80 LTRs with 153 positive TBBx results representing 112 ACR episodes were included in the ACR group. Importantly, dd-cfDNA was not used as a screening method to proceed with TBBx; rather, TBBx was performed per protocol and paired with dd-cfDNA results at the time of analysis. Median (IQR) time to ACR diagnosis was 115 (33−359) days. Maximal ACR grade experienced by the 80 LTRs with ACR events were: 28 (35%) grade A1, 37 (46%) grade A2, 12 (15%) grade A3, and 3 (4%) grade A4. After excluding dd-cfDNA results within 30 days of LTx, 62 LTRs had 110 dd-cfDNA results paired with ACR diagnoses. dd-cfDNA median (IQR) level was significantly higher for moderate-to-severe ACR events: 1.98% (0.92–2.81) for grades A3/A4 vs. 0.68% (0.3–1.48%) for grade A1 and 0.61% (0.25–1.13%) for grade A2, p=0.032, using a linear fixed effects model. As Table 5 denotes, LTRs with ACR+MI experienced significantly increased hazard (HR 3.32) of CLAD or death. Moreover, LTRs with high-risk ACR (dd-cfDNA ≥1% at ACR diagnosis) were at significantly increased hazard (HR 3.13) of CLAD or death relative to LTRs with low-risk ACR (dd-cfDNA < 1% at ACR diagnosis).¹⁹ These data represent important findings for therapeutic decision-making for LTRs with ACR. Elevated dd-cfDNA may indicate significant injury burden that increases the risk of long-term adverse events which may heighten awareness of the need for follow-up dd-cfDNA testing to confirm therapeutic responsiveness.

Table 5.

Association of AR (ACR/AMR) Plus Molecular Injury (MI) with CLAD, Death, and the Composite of CLAD or Death^a

ACR+MIb	CLAD		Death		Composite of CLAD or death
Multivariable models19	HR (95% CI)	p	HR (95% CI)	p	HR (95% CI)	p
ACR+MIc	3.97 (1.27–12.4)	0.018	3.01 (1.18–7.66)	0.021	3.32 (1.31–8.40)	0.012
ACR events stratified by dd-cfDNA leveld
High risk ACR (dd-cfDNA ≥1%)	4.65 (1.98–10.9)	0.001	2.37 (1.003–5.6)	0.049	3.13 (1.41–6.93)	0.005
Low risk ACR (dd-cfDNA < 1%)	1.15 (0.43–3.10)	0.78	0.77 (0.28–2.08)	0.61	1.15 (0.51–2.6)	0.73
AMR+MIe	CLAD		Death		Composite of CLAD or death
Multivariable modelsf,24	HR (95% CI)	p	HR (95% CI)	p	HR (95% CI)	p
AMR+MI	2.46 (1.34–4.52)	0.004	2.49 (1.43–4.35)	0.005	2.00 (1.18 3.38)	0.01
Clinical AMR60					2.36 (1.42 3.87)	0.001
AMR+MI without Clinical AMR					2.14 (1.10 4.19)	0.026
AMR+MI without Clinical AMR, ACR, or Pathogens					2.45 (1.01 5.94)	0.047

^a,GTD/GRAfT consortia.

^b,ACR+MI defined as dd-cfDNA ≥1% plus positive TBBx for ACR.⁵¹

^c,adjusted for recipient age, sex, transplant type, PGD grade, native lung disease, center, ACR grade.

^d,model c adjustments without ACR grade.

^e,AMR+MI defined as dd-cfDNA ≥1% plus positive serum anti-HLA DSA.

^f,adjusted for recipient age, sex, PGD3, lung allocation score, transplant type, center.

dd-cfDNA, donor-derived cell-free DNA. ACR, acute cellular rejection. TBBx, transbronchial biopsy. AMR, antibody-mediated rejection. CLAD, chronic lung allograft dysfunction. HLA, human leukocyte antigen. DSA, donor specific antibody.

The GTD/GRAfT consortia investigators assessed molecular injury in the setting of donor specific antibodies (DSA), defined as dd-cfDNA ≥1% plus positive anti-human leukocyte antigen (HLA) DSA testing and identified this construct as molecular AMR (AMR+MI), for the risk of the composite endpoint of CLAD or death (Table 2).²⁴ 273 LTRs who underwent adjudication were enrolled and 209 LTR were analyzed; of whom, 61 (29.2%) LTRs met criteria for AMR+MI. Clinical AMR, as defined by the ISHLT 2016 lung AMR consensus statement,⁵⁹ occurred in 42 (20.1%) LTRs. Importantly, TBBx was not required to meet the criteria for clinical AMR or AMR+MI. The majority of LTRs with AMR+MI (n=58, 95%) expressed class II or mixed class I and II DSA, 48 (79%) of whom expressed DQ DSA specifically; the small remainder (n=6, 5%) expressing only class I DSA. As Table 5 denotes, LTRs with AMR +MI, irrespective of the diagnosis of clinical AMR, experienced significantly increased hazard (HR 2) of CLAD or death. Moreover, LTRs with AMR+MI without clinical AMR, ACR, or pathogens, experienced significantly increased hazard (HR 2.45) of CLAD or death.²⁴ These data may influence the decision to proceed with acute AMR therapy as elevated dd-cfDNA may indicate significant injury burden that increases the risk of long-term adverse events.

The GTD/GRAfT consortia investigators assessed extreme molecular injury (EMI), defined a priori as dd-cfDNA > 5% occurring after day 45 post-LTx irrespective of TBBx performance or result (Table 2).¹⁸ 302 LTRs were enrolled and 238 LTR with 1564 dd-cfDNA results were analyzed. Median (IQR) follow-up time was 4.86 (1.64–4.96) years. Median (IQR) dd-cfDNA at EMI diagnosis was 7.93% (5.83–12.73%) in the 38 (16%) LTRs that experienced EMI during the study period. Figure 5 shows the dd-cfDNA kinetics of LTRs with EMI who did and did not experience the composite of severe CLAD or death. LTRs positive for the composite endpoint experienced persistently higher dd-cfDNA values over the first 4.5 years post-LTx compared to LTRs who did not experience the composite endpoint. LTRs with EMI were at significantly increased hazard (HR 2.78) of the composite endpoint of severe CLAD or death (Table 6). Additionally, post hoc analyses of > 4% and > 3% EMI definitions showed similar significantly increased hazard (HR 3.66 and HR 3.75, respectively) of the composite endpoint of severe CLAD or death (Table 6).¹⁸ Subsequently, the Temple Program retrospectively assessed EMI, defined as dd-cfDNA > 5% using Prospera, in 84 LTRs with 124 dd-cfDNA results paired with TBBx and BAL data collected within 4 weeks of dd-cfDNA testing at their center.²⁶ 24 (29%) LTRs experienced EMI with median dd-cfDNA of 6.89% at diagnosis. Similar to the GTG/GRAfT findings, LTRs with EMI in the Temple cohort were more likely to develop CLAD (aOR=12.14, 95% CI 10.86–18.3, p=0.04) on multivariable logistic regression after adjusting for ACR, AMR, DSA, infection, and other comorbidities. Taken together, these data show that LTRs with EMI are more likely to develop CLAD or death compared to LTRs who have never had extreme elevations in dd-cfDNA.^18,26

Figure 5.

Extreme Molecular Injury (EMI) dd-cfDNA patterns in LTRs^a,18. a, GTD/GRAfT consortia.dd-cfDNA, donor-derived cell-free DNA. CLAD, chronic lung allograft dysfunction. LT, lung transplantation.

Table 6.

Association of Extreme Molecular Injury (EMI) with Severe CLAD, Death, and the Composite of CLAD or Death^a,18

	Severe CLAD		Death		Composite of CLAD or death
EMI dd-cfDNA Definitions Multivariable Modelsb	HR (95% CI)	p	HR (95% CI)	p	HR (95% CI)	p
> 5% (a priori definition)	3.9 (1.42–10.73)	0.008	3.88 (1.96–7.7)	< 0.001	2.78 (1.26–6.22)	0.012
> 4% (post hoc analysis)					3.66 (2.00–6.70)	< 0.001
> 3% (post hoc analysis)					3.65 (1.91–6.98)	< 0.001

^a,GTD/GRAfT consortia.

^b,multivariable model adjusted for recipient age, race, sex, native lung disease, PGD3, transplant type, center, prior ACR episodes.

dd-cfDNA, donor-derived cell-free DNA. CLAD, chronic lung allograft dysfunction. PGD3, primary graft dysfunction grade 3. ACR, acute cellular rejection.

The GTD/GRAfT consortia investigators assessed sustained molecular injury (SMI) burden, defined by the mean dd-cfDNA over the 3 post-AR/OP months, after LTx (Table 2).²⁵ 93 LTRs with 339 dd-cfDNA results after AR/OP were analyzed. 41 (44.1%) LTRs experienced the composite of CLAD or death during the median (IQR) follow-up period of 140 (48−354) days following AR/OP. During the 3-month post AR/OP follow-up period, a median (IQR) of 3 (2−4) dd-cfDNA results were available per LTR. Table 7 delineates the multivariable Cox proportional hazards models which show that mean 3-month dd-cfDNA levels were significantly associated with increased hazard of CLAD (HR 2.28), death (HR 2.71), and the composite of CLAD or death (HR 2.62). LTRs were then grouped into tertiles of lowest < 0.48% n=31, medium 0.48–0.84% n=30, and highest > 0.84% n=32 dd-cfDNA values. Risk of CLAD (p=0.042), death (p=0.038), and the composite of CLAD or death (p=0.008) increased in a proportionally escalating manner from the low to high tertiles. Figure 6 shows 3 distinct LTR examples, one per dd-cfDNA tertile, in the post AR/OP follow-up period and associated predicted risk profiles for CLAD, death, and the composite of CLAD and death.²⁵ While these data require validation, they highlight the prognostic utility of dd-cfDNA levels on long-term adverse events in LTRs following AR/OP.

Table 7.

Association of Sustained Molecular Injury (SMI)^a burden after AR with CLAD, death, and the composite of CLAD or death^b,25.

	CLAD		Death		Composite of CLAD or death
SMI dd-cfDNA multivariable models	HR (95% CI)	p	HR (95% CI)	p	HR (95% CI)	p
Model 1c	2.5 (1.31–4.75)	0.005	3.76 (1.92–7.36)	< 0.001	2.84 (1.67–4.83)	< 0.001
Model 2d	2.28 (1.20–4.32)	0.012	2.71 (1.44–5.08)	0.002	2.62 (1.53–4.47)	< 0.001

^a,mean dd-cfDNA over 3-months.

^b,GTD/GRAfT consortia.

^c,adjusted for race, transplant type, transplant center.

^d,model 1 adjustments plus mean 3-month FEV₁.

dd-cfDNA, donor-derived cell-free DNA. CLAD, chronic lung allograft dysfunction. FEV₁, forced expiratory volume in 1 s

Figure 6.

Sustained Molecular Injury (SMI) burden dd-cfDNA patterns in LTRs^a,25. a, GTD/GRAfT consortia. dd-cfDNA, donor-derived cell-free DNA. AR, acute rejection. OP, organizing pneumonia. CLAD, chronic lung allograft dysfunction.

Clinical approaches to dd-cfDNA testing in LTRs

In aggregate, published data support the utility of dd-cfDNA surveillance and for-cause testing to identify allograft injury in LTRs, assess injury burden and persistence, and identify LTRs at higher risk for CLAD and death. Currently, substantial variability exists in frequency and duration of dd-cfDNA testing in the LTx literature^1–33 and at LTx centers worldwide. Ongoing registries^85,86 and planned extensions⁸⁵ lengthen the surveillance dd-cfDNA testing duration, at deescalating frequency, to 5 years post-LTx. Importantly, surveillance bronchoscopic evaluation of the lung allograft remains critical for multiple purposes including infection identification, visualization of the anastomoses and large airways, therapeutic clearance, and airway interventions.⁶⁸ Based on existing evidence, it may be reasonable for lung transplant programs to defer surveillance TBBx for LTRs in whom %dd-cfDNA remains persistently low, in the absence of other clinical concerns, given high test NPV.^{12,14,16,22,33}

Figure 7 describes a real-world clinical approach to dd-cfDNA testing in LTRs. The dd-cfDNA surveillance testing frequency and duration described in Figure 7 were chosen clinically for multiple reasons. Unlike heart and kidney transplants, current median 5-year LTR survival rate is approximately 60% and 5-year freedom from CLAD approaches 50%; neither of which have meaningfully improved since the introduction of calcineurin inhibitors in 1983 suggesting conventional approaches are unlikely to improve these outcomes.⁷² Moreover, accumulating data support the potential prognostic role of dd-cfDNA testing for the prediction of CLAD onset and progression which may enable earlier intervention; and may identify an at-risk cohort that could be targeted for CLAD interventional trials.^{7–12,14,18,25,26,30,69–71} For-cause dd-cfDNA testing is performed when the lung allograft shows signs or symptoms of acute or chronic dysfunction. These for-cause parameters were chosen to assess the severity and burden of molecular injury concomitant with clinical suspicion for, or diagnoses of, allograft dysfunction phenotypes; also consistent with ongoing studies.^85–87 Additionally, dd-cfDNA testing is performed monthly for 3 months after injury diagnosis, with or without significant immunomodulation, to assess for sustained molecular injury which may increase the risk for early CLAD or death.^25,30

Figure 7.

Clinical dd-cfDNA testing approaches in LTRs^a. dd-cfDNA, donor-derived cell-free DNA. LTx, lung transplantation. AR, acute rejection. ACR, acute cellular rejection. AMR, antibody mediated rejection. ALAD, acute lung allograft dysfunction. CLAD, chronic lung allograft dysfunction. TBBx, transbronchial biopsy. FEV1, forced expiratory volume in 1 s. FVC, forced vital capacity.

Figure 8 describes the current overarching surveillance and interventional approach after LTx. To provide further insight into the utility of dd-cfDNA as a component of the overarching surveillance and interventional approach, an example of a differential interpretation pathway is shown in Figure 9. Walking through the differential interpretation of dd-cfDNA results: interpretation begins by evaluating SLTR and BLTR differently. If dd-cfDNA is not elevated in the absence of other concerns (green path), it is reasonable to reconsider whether routine surveillance TBBx is necessary against the associated risks to the LTR, as was performed in the ALARM study.^{12,31,92–94} Additionally, the Ohio State program assessed elimination of their 9-month post-LTx protocolized TBBx in 78 LTRs undergoing dd-cfDNA surveillance monitoring.³³ LTRs were stratified by their 9-month ( ± 30 days) Prospera result into high (≥1%) or low (< 1%) risk groups. No differences were found in 12-month post-LTx episodes of treated AR, DSA incidence, or FEV₁, in the 49 LTRs (8 high-risk; 41 low-risk) who did not undergo 9-month protocol TBBx compared to the 29 LTRs who were biopsied at that time.

Figure 8.

Surveillance and interventional approach in LTRs. dd-cfDNA, donor-derived cell-free DNA. CMV, cytomegalovirus. CARV, community acquired respiratory virus. PCR, polymerase chain reaction. BAL, bronchoalveolar lavage. BW: bronchial wash. TBBx, transbronchial biopsy.

Figure 9.

Differential interpretation of dd-cfDNA results in LTRs^a. dd-cfDNA, donor-derived cell-free DNA. DSA, donor specific antibody. MI, molecular injury. AMR, antibody-mediated rejection. ACR, acute cellular rejection. FVC, forced vital capacity. FEV₁, forced expiratory volume in 1 s. ALAD, acute lung allograft dysfunction. CLAD, chronic lung allograft dysfunction. CT, computed tomography scan. CMV, cytomegalovirus. CARV, community acquired respiratory virus. PCR, polymerase chain reaction. GERD, gastroesophageal reflux disease. BAL, bronchoalveolar lavage. BW, bronchial wash. TBBx, transbronchial biopsy.

Continuing through the differential pathway (Figure 9): if dd-cfDNA is elevated (yellow and red paths), multiple clinical and laboratory assessments are performed and analyzed prior to proceeding with bronchoscopy and TBBx. These intermediate assessments are intended to further refine the likelihood of finding actionable pathologic injury by identifying AMR+MI, functional allograft decline, radiographically observable injuries, infections, and/or reflux-related injury prior to bronchoscopy and TBBx. It is possible to find either isolated or, albeit less likely, repeated elevated dd-cfDNA values in the absence of identifiable allograft injury physiology or pathology.^12,14 While pulmonary reserve after LTx is necessarily different than in native non-diseased lungs, substantial reserve exists in well-size matched, well-recruited, maximally functional lung allografts. It is therefore logical that longitudinal subclinical allograft injury is likely to have occurred, impacting reserve, prior to LTRs presenting with functional allograft decline by spirometry or overt signs of failure, such as hypoxia. As previously described, SMI burden characterized by chronic persistent elevation in dd-cfDNA is associated with increasing risk of CLAD or death.²⁵ Thus, repeated or persistently elevated dd-cfDNA results in the absence of identifiable injury physiology or pathology should heighten awareness and evaluation for the risk of early CLAD or death in LTRs. While the differential pathway shown in Figure 9 represents an approach for dd-cfDNA result interpretation today, it should be noted that the LTx community is poised to benefit from the contemporaneous expansion of molecular diagnostics that are presently emerging which will, in addition to dd-cfDNA, further refine precision medicine approaches for LTRs.^76–83

Limitations and unanswered questions

Limitations and unanswered questions (Table 8) exist within the current literature and in dd-cfDNA testing after LTx. First, the available data are predominantly observational in nature^{1–11,13–34} and, at present, do not establish whether serial testing and clinical intervention in response to dd-cfDNA elevation alone will improve short and long-term outcomes in LTRs. Additionally, field consensus regarding dd-cfDNA scenario-based interpretation and respective interventional strategies remains to be assessed using an expert-opinion Delphi survey-based approach. Second, data addressing dd-cfDNA monitoring in pediatric LTRs is not presently available. Pediatric LTRs; particularly very young children, are a uniquely challenging population and adult dd-cfDNA data should not be generalized to this group. Third, while a single-center retrospective study exists describing dd-cfDNA utility in a dominantly Latinx population,²⁷ data regarding implications for other ethnicities is not available. Fourth, dd-cfDNA testing has not been validated in the setting of simultaneous multi-organ transplantation, such as simultaneous lung-kidney or lung-liver or heart-lung. However, in the cases of sequential multi-organ transplants from different donors, such as kidney after lung, the relative contribution of each organ to total dd-cfDNA is clinically calculatable through donor and recipient genotyping. Fifth, clinical dd-cfDNA testing outside North America relies on individual laboratories using the kit-based assays described in Section 3. As the 2025 Sensitization in Transplantation Assessment of Risk (STAR) Working Group on dd-cfDNA report noted: analytical consistency across labs and platforms using kit-based assays is essential⁹⁵; thus, each lab must validate reproducibility and quantitation limits using their specific instruments. The STAR report also noted that laboratory guidelines, quality assurance and quality control standards have not yet been developed for decentralized kit-based dd-cfDNA testing.⁹⁵ Moreover, LTRs typically follow-up with their transplant center worldwide and are likely to be longitudinally tested by the same lab which may minimize the impact of between-lab variance on the individual LTR. Sixth, patients show variable rd-cfDNA levels at various stages after LTx which may impact dd-cfDNA fractional results.^31,32 Perioperative injury and systemic inflammation elevate rd-cfDNA which may mask higher-level allograft molecular injury by depressing dd-cfDNA fraction.^31,32,74 Certain malignancies and timing of specific chemotherapeutic agents, or other disease states which result in substantial self-cell death or loss of chromosomal SNP heterogeneity, may impact dd-cfDNA results. Both commercially available dd-cfDNA assays report dd-cfDNA copies/mL (AlloSure DDC and Prospera DQS) to provide additional context in these situations. Caution should be employed when interpreting dd-cfDNA copies/mL alone or in the context of fractional results as relatively few studies have assessed this relationship.^31,32

Table 8.

Key Unanswered Questions Amenable to Future Investigations in LTRs

Clinical and Operational Questions
1. Assessments of long-term outcomes of serially tested LTRs with dd-cfDNA
2. Interventional studies after dd-cfDNA elevation on long-term outcomes
3. Cost-effectiveness of dd-cfDNA serial monitoring vs standard of care approaches
4. Implications of dd-cfDNA elevation in the absence of clinically identifiable source(s) of injury (subclinical molecular injury)
5. Utility of dd-cfDNA in combination with other functional assessments of allograft health such as remote microspirometry
6. Implications of dd-cfDNA elevation in the setting of non-HLA DSA and anti-HLA DSA phenotyping on therapeutic approaches
7. Reflux-associated molecular injury burden before and after surgical gastroesophageal intervention(s)
8. Clinical scenario-specific expert opinion guided approaches for use, interpretation, and action on dd-cfDNA results (Delphi method)
9. Reproducibility and generalizability of kit-based dd-cfDNA testing approaches
Translational Questions
1. Implications of methylome assessments on the cellular sources of dd-cfDNA
2. Utility of combinations of dd-cfDNA and multi-omic assessments such as transcriptomics from the bronchial brush or tissue, proteomics from the airway or plasma, microbiome assessments, and/or others
3. Implications of donor-derived and/or recipient-derived inflammatory derangements assessed by copies/mL vs. fractional results
Questions in Special Populations
1. Utility of dd-cfDNA testing in pediatric LTRs
2. Implications of differences in LTR ethnicity on dd-cfDNA utility
3. Differences in molecular injury burden in LTRs with short telomere syndrome (STS)84
4. Utility of cfDNA approaches in LTRs with malignancy

Conclusions and future directions

Dd-cfDNA is a validated, highly sensitive, nonspecific plasma molecular biomarker of lung allograft injury.^1–34 The accumulating body of evidence (Figure 1) supporting the utility of dd-cfDNA for molecular injury monitoring after LTx in the precision medicine era is summarized and clinical approaches for testing and interpretation are described.^1–33 At present, ISHLT definitions of injuries in LTRs have not incorporated molecular diagnostic approaches. It may be reasonable for our colleagues who are leading these consensus efforts to consider whether the data are mature enough and if the time is ripe for potential inclusion. Future directions for molecular diagnostics trials and assay clinical development should prioritize multimodality approaches that combine and/or sequence injury markers such as dd-cfDNA with multiomics approaches of decreasing invasiveness to enhance specificity for distinct molecular injury endotypes in LTRs. These molecular endotyping approaches may facilitate earlier and targeted interventions to extend LTR quality and quantity of life.^76–83

Abbreviations:

ACR

acute cellular rejection

ALAD

acute lung allograft dysfunction

ALAMO

AlloSure Lung Assessment and Metagenomics Outcomes Study

AMR

antibody mediated rejection

AR

acute rejection

AUROC

area under the receiver operating characteristic curve

BAL

bronchoalveolar lavage

BLTx

bilateral lung transplant

BLTR

bilateral lung transplant recipient

BLAD

baseline lung allograft dysfunction

CAP

College of American Pathologists

CARV

community acquired respiratory virus

cfDNA

cell-free DNA

CLAD

chronic lung allograft dysfunction

CLIA

Clinical Laboratory Improvement Amendments

CV_I

within-subject coefficient of variation

CV_G

between-subject coefficient of variation

dd-cfDNA

donor-derived cell-free DNA

DSA

donor specific antibody

EMI

extreme molecular injury

EVLP

ex vivo lung perfusion

FEV₁

forced expiratory volume in 1 s

FVC

forced vital capacity

GRAfT

Genomic Research Alliance for Transplantation

GTD

Genome Transplant Dynamics

II

index of individuality

LoD

limit of detection

LoQ

limit of quantitation

LSG

latent subgroup

LTx

lung transplantation

LTR

lung transplant recipient

MI

molecular injury

NGS

next-generation sequencing

NHLBI

National Heart Lung and Blood Institute

NIH

National Institutes of Health

NLM

National Library of Medicine

NRAD

neutrophilic responsive airways disease

OP

organizing pneumonia

PCR

polymerase chain reaction

PGD

primary graft dysfunction

RAS

restrictive allograft syndrome

rd-cfDNA

recipient-derived cell-free DNA

RefCV

reference change value

SLTx

single lung transplant

SLTR

single lung transplant recipient

SMI

sustained molecular injury

STAR

Sensitization in Transplantation Assessment of Risk

TBBx

transbronchial biopsy

US

United States

UTSA

University of Texas San Antonio