The MOTR study was an international, prospective, cross-sectional, multicenter study that evaluated dd-cfDNA baseline levels in patients with MOT. The studies involving human participants were reviewed and approved by local or central institutional review boards (IRBs, Supplementary Table S1). All PK, HK and LK transplant recipients who met the eligibility criteria were consented according to the IRB-approved protocol. The study has been performed in full adherence to the Declaration of Helsinki.

Patients were considered eligible for enrollment into the study if they met the following inclusion criteria: (1) age ≥18 years at the time of signing informed consent, (2) history of pancreas-kidney, kidney-heart, or liver-kidney transplant, (3) either considered clinically stable (defined as no history of rejection or no clinical indication of rejection within 3 months prior to enrollment in the study), or scheduled to undergo a biopsy of at least one of the transplanted organs, (4) able to read, understand, and provide written informed consent, and (5) willing and able to comply with the study-related procedures. Patients were excluded from this study if, at the time of enrollment they (1) were pregnant, (2) were on dialysis, (3) had received an allograft from an identical twin, and/or (4) had more than 2 unique organ transplant types (i.e.: pancreas, liver, heart).

Clinical and biological data related to recipient characteristics, donor characteristics, transplant procedure, and contemporaneous laboratory tests including SCr, lipase, amylase, ALP, AST, ALT, DSA, and viral loads, were collected for patients in both the stable arm and the biopsy-matched arm (Table 1).

dd-cfDNA samples were analyzed using the methodology previously described (the Prospera™ test; Natera, Inc., Austin, TX) [29]. Blood samples for dd-cfDNA testing were collected in two 10 mL Streck Cell-Free DNA BCT tubes and shipped at room temperature to Natera’s Clinical Laboratory Improvement Amendments-certified and College of American Pathologists-accredited laboratory. Samples were processed within 8 days of collection, or according to previously validated standard operating protocol [30]. The algorithm was designed to provide a single composite value by combining non-recipient cfDNA for both dd-cfDNA % and DQS, which is an estimate of the concentration of dd-cfDNA in plasma reported as genomic copies per mL (cp/mL). Blood draws for dd-cfDNA testing were performed within 0–7 days before biopsy in the biopsy-matched arm, or according to clinical standard in the stable arm. Results of the investigational test (dd-cfDNA) were not shared with the patients or the attending physicians, and thus not used in patient management.

Allograft biopsies of kidney, pancreas, heart, and liver transplants were graded as previously reported [31–34]. Biopsy reports were centrally reviewed by pathologists for data accuracy.

dd-cfDNA levels in PKT, HKT, LKT recipients were compared to real-world cohorts of KT and HT recipients matched for time post-transplant.

Statistical analyses were performed in Python (3.8.10). Normally distributed variables were reported as mean and standard deviation; non-normally distributed variables were reported as median and interquartile range (IQR). Categorical variables were represented as counts and percentages and evaluated using Chi-square test as appropriate. Kruskal-Wallis tests were performed for groupwise comparisons. For significant p-values, post-hoc tests of chi-square or Mann Whitney U tests were performed, as appropriate, with Bonferroni corrections for multiple testing.

Random forest classifiers were applied to identify the relationship between known clinical variables and positivity of dd-cfDNA (i.e., high risk by dd-cfDNA) testing. dd-cfDNA positivity was defined using the clinical thresholds established for kidney-alone transplant rejection (dd-cfDNA%≥1.0% OR DQS≥78.0 cp/mL) for the PKT and HKT cohorts, and an investigational, unvalidated threshold for liver-alone transplant rejection (dd-cfDNA%≥10.0%) based on prior studies [35–37] for the LKT cohort. Modeling was not feasible in the HKT cohort due to the limited number of dd-cfDNA-positive cases. Modeling included PKT cases with complete known matched eGFR (estimated glomerular filtration rate), amylase and lipase values, and LKT cases with complete known matched eGFR, ALP, AST, and ALT values, and was trained on these values along with time from transplant, Kidney Donor Profile Index (KDPI), and body mass index (BMI) and optimized for F1 score. The Shapley values of these features were assessed on a holdout random set of 20% of the complete cases.

Spearman correlation analyses were performed between dd-cfDNA (%, cp/mL) and the clinical biomarkers relevant to each transplanted organ, as described above. For each variable pair, correlation was calculated using cases with complete data.

This study used data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donor, wait-listed candidates, and transplant recipients in the US, submitted by the members of the Organ Procurement and Transplantation Network (OPTN). The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN and SRTR contractors. Contemporaneous data from the Scientific Registry of Transplant Recipients (SRTR) was assessed for comparison.

The data reported here have been supplied by the Hennepin Healthcare Research Institute (HHRI) as the contractor for SRTR. The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the SRTR or the U.S. Government.

Contemporaneous data from SRTR were queried to compare the MOTR study cohorts (PKT, HKT, and LKT) to the corresponding national cohorts for each MOT type, to assess representativeness and generalizability. Variables included MOT recipient age, sex, BMI, race/Ethnicity, and repeat transplant status as indicated in Supplementary Table S2.

Between October 2020 and June 2022, 385 subjects from 18 sites were enrolled in the MOTR Study (Supplementary Table S1). Of these, 38 subjects were excluded due to ineligibility or lack of valid dd-cfDNA testing (e.g., insufficient plasma volume). Of the remaining 347 transplant recipients, 183 had PKT, including 169 with a simultaneous pancreas-kidney transplant (SPKT; 92.4%), 7 with a pancreas-after-kidney transplant (PAKT; 3.8%). 7 patients (3.8%) underwent repeat transplant following a kidney-pancreas transplant: Four received kidney re-transplant after SPK, and three received a pancreas retransplant (one after PAK, two after SPK). 57 subjects had HKT, including 54 with a simultaneous heart-kidney transplant (SHKT; 94.7%) and 3 with a kidney-after-heart transplant (KAHT; 5.3%). 107 subjects had LKT including 80 with a simultaneous kidney-liver transplant (SLKT; 74.8%), 24 with a kidney-after-liver transplant (KALT; 22.4%), two with a liver-after-kidney transplant (LAKT; 1.9%), and one (0.9%) with a liver re-transplant following a kidney-after-liver transplant (Figures 1, 2).

The MOTR cohort was 64.6% male (n = 224), had a median age of 57.1 y (IQR: 48.7–65.9), and a median BMI of 28.2 kg/m² (IQR: 24.9–32.7). Patients were 51.6% White, 27.7% African American, and 8.1% Hispanic and/or Latino. In the PKT, HKT, and LKT cohorts, 3.8% (7/183), 5.3% (3/57), and 0.9% (1/107) of the transplanted kidneys were donated by living donors, respectively. The median KDPI across all deceased donor’s kidney transplants was 23.0 (IQR: 6.0–41.0). The majority (72%) of MOTR participants were considered clinically stable at enrollment, and the remaining 28% underwent biopsy; (for-cause: n = 49, 50.5%; for-protocol: n = 48, 49.5%) (Table 1). For-cause and for-protocol biopsies were observed in 18.6% and 3.3% of PKT, 12.3% and 47.4% of HKT, and 7.5% and 14.0% of LKT recipients, respectively; 70% of all biopsies were performed on the transplanted kidney.

The covariates of the PKT, HKT, and LKT subcohorts, such as MOT recipient demographics including age, sex, and ethnicity, were also compared to a contemporaneous national cohort’s covariates as represented in data requested from the SRTR (https://www.srtr.org/) (Supplementary Table S2).

In this analysis, a single dd-cfDNA test from each patient was used, along with a matched biopsy, when available. The median time from most recent transplant to dd-cfDNA testing was 23.5 months (IQR: 8.4–52.7) and was significantly longer in the PKT cohort (33.8 months) compared to the HKT (13.1 months) and LKT (19.0 months) cohorts (p < 0.05).

Among the PKT recipients, there were no significant differences in the median (IQR) dd-cfDNA % or DQS between the SPKT (dd-cfDNA%: 0.23% (0.13–0.47); DQS: 9 (4–17) cp/mL), PAKT (dd-cfDNA%: 0.13% (0.11–0.23); DQS: 4 (3–7) cp/mL) or those who underwent kidney or pancreas retransplant after PKT (dd-cfDNA%: 0.29% (0.21–0.44); DQS: 16 (6–22) cp/mL) cohorts and the KT cohort (dd-cfDNA%: 0.23% (0.12–0.48); DQS: 9 (5–20) cp/mL) (p = 0.38, 0.31, and 0.55, respectively) (Figure 3A). The distribution of total cfDNA paralleled that of both dd-cfDNA (%) and DQS (cp/mL) among these groups (Supplementary Figure S1A).

Similarly, among the HKT recipients, there were no significant differences in the median (IQR) dd-cfDNA % or DQS (cp/mL) between the SHKT (dd-cfDNA%: 0.21% (0.12–0.42); DQS: 10 (7–18) cp/mL) or the KAHT cohorts (dd-cfDNA%: 0.32% (0.30–0.39; DQS: 22 (15–23) cp/mL) and the KT cohort (dd-cfDNA%: 0.24% (0.12–0.51); DQS: 10 (5–22) cp/mL (p = 0.36, 0.47, respectively) (Figure 3B). Both the median dd-cfDNA % and DQS (cp/mL) were significantly higher among the SHKT cohort when compared to the HT cohort (dd-cfDNA%: 0.21% (0.12–0.42); DQS: 10 (7–18) cp/mL) vs. (dd-cfDNA%: 0.06 (0.02–0.21), DQS: 2 (1–11) cp/mL, p < 0.001) (Figure 3B). There were no differences in total cfDNA when comparing SHKT and KAHT to KT and SHKT and KAHT to HT cohorts (Supplementary Figure S1B).

Among the LKT recipients, both the median (IQR) dd-cfDNA % and DQS for the SLKT (dd-cfDNA%: 3.92% (1.92–6.99); DQS: 147 (75–258) cp/mL) and KALT cohorts (dd-cfDNA%: 3.35% (1.59–5.14); DQS: 35 (73–284) cp/mL) were significantly higher than the KT (dd-cfDNA%: 0.23% (0.12–0.49); DQS: 9 (5–22) cp/mL; p < 0.001 for both) (Figure 3C). Lower total cfDNA levels in the SLKT compared to the KALT cohort may contribute to differences between dd-cfDNA% and DQS (cp/mL) observed between them (Supplementary Figure S1C).

While this study did not contain sufficient patients with BPAR to assess appropriate dd-cfDNA cutoffs for rejection risk in MOT, we felt it was still illustrative to interpret the data in the context of the validated kidney-only threshold, as we would expect the dd-cfDNA values for kidney-only transplant patients to approximate a lower-bound for the appropriate MOT rejection risk threshold. A total of three patients in the SPKT cohort had kidney BPAR, two of which had dd-cfDNA above the 1.0% threshold used to define increased rejection risk in KT recipients (Supplementary Figure S2A): one recipient with T-cell-mediated rejection (TCMR) had dd-cfDNA levels of 3.11% and 63 cp/mL, while two patients with chronic active antibody-mediated rejection (AMBR) had dd-cfDNA% of 0.51% and 1.59% and DQS of 21 and 55 cp/mL. Three patients among the LKT cohort had BPAR in the kidneys (Supplementary Figure S2B), including TCMR 1A with dd-cfDNA of 1.32%, 86 cp/mL, chronic active ABMR (2.31%, 114 cp/mL), and ABMR (6.98%, 698 cp/mL). No patients in the HKT cohort showed acute rejection on biopsy (Supplementary Figure S2B). Of the remaining 91 biopsies, 68 showed no rejection, 18 showed other findings not representing rejection, and 5 consisted of limited sample tissue and were thus unable to be diagnosed.

Of the 240 non-rejecting PKT and HKT recipients, 21 (19 PKT and 2 HKT) exceeded KT dd-cfDNA thresholds (≥1.0% and/or ≥78 copies/mL). All but one of these transplants were simultaneous; the remaining one was a kidney re-transplant after SPK. Seven patients (33.%) underwent biopsies (6 kidney, 1 heart), all without significant findings. The remaining 14 patients (13 PKT, 1 HKT) were considered clinically stable and did not undergo biopsy; one of these had substantially elevated amylase levels, while the other 13 were clinically stable by all parameters apart from dd-cfDNA. Of the 219 patients with dd-cfDNA below KT threshold, only one experienced BPAR (chronic-active ABMR).

Of the 107 subjects in the LKT cohort, 12 (11.2%; 11 SLKT, 1 KALT) had dd-cfDNA above 10%, which prior studies have posited as an appropriate threshold for rejection in liver-only transplants [35–37]. Two of these subjects received surveillance biopsies of the kidney, both demonstrating no rejection; no patients received biopsies of the liver.

We sought to identify the clinical variables in MOT recipients that were associated with elevated dd-cfDNA levels using a random forest classifier with the PKT (Figure 4A) and LKT (Figure 4B) cohorts. The traditional markers for respective organ function were chosen as variables. In the PKT cohort, the most influential variables included amylase, time since transplant, and lipase. Shorter time post-transplant, higher amylase values, and higher lipase values were associated with higher likelihood of elevated dd-cfDNA (Figure 4A). Looking at the continuous relationship between dd-cfDNA and the laboratory markers in the PKT cohort, significant positive correlations were observed in the PKT cohort between dd-cfDNA% and amylase and lipase (p < 0.001, 0.006, respectively), as well as between DQS and amylase and lipase (p = 0.006, 0.027, respectively). eGFR did not show a significant association with dd-cfDNA in either the random forest model or using Spearman correlations (Supplementary Figure S3).

In the LKT cohort, the most influential variables in the random forest classifier were ALT, AST, and ALP. Higher liver enzyme profiles were associated with higher likelihood of elevated dd-cfDNA (Figure 4B). Across all dd-cfDNA levels in this cohort, significant positive correlations were observed between dd-cfDNA% and ALT and AST (p < 0.001 for both), and between DQS and ALT and AST (p < 0.001 for both). No significant correlations were found between dd-cfDNA levels and eGFR (Supplementary Figure S4).

As the model indicated a significant association between time post-transplant and the likelihood of elevated dd-cfDNA levels, we further assessed the relationship between dd-cfDNA and time post-transplant in the different MOT cohorts (Supplementary Figure S5). In the PKT and HKT cohorts, dd-cfDNA% and DQS (cp/mL) both decreased over time. Total cfDNA remained steady in the PKT cohort while the HKT cohort showed a dramatic decrease in the first 40 months post-transplant before stabilizing. Despite these fluctuations, dd-cfDNA levels in both cohorts remained below the validated thresholds for KT rejection. In the LKT cohort, dd-cfDNA% and DQS increased slightly over time, whereas the total cfDNA decreased slightly, before stabilizing.