Method for prediction of response to cardiovascular regeneration based on detecting the amount of biomarkers

Abstract

The present invention relates to a method for prediction of response to cardiovascular regeneration comprising the use of biomarkers. Further, the present invention relates to a combination of the biomarkers for use in a method for prediction of response to cardiovascular regeneration, a computer device to perform a method according to the present invention and a device adapted for carrying out the inventive method.

Claims

1. A method for prediction of response to stem cell therapy for coronary artery disease or ischemic vascular disease, wherein the method comprises (i) determining in a sample of a subject the amount of each of the following biomarkers, wherein the biomarkers include each of a growth factor selected from the group consisting of vascular endothelial growth factor (VEGF), Erythropoietin, and FGF, lymphocyte adapter protein, a glycoprotein selected from the group consisting of Vitronectin and granulocyte-colony stimulating factor (GCSF), brain natriuretic peptide (BNP), circulating endothelial progenitor cells (EPC), circulating endothelial cells (CEC), circulating thrombocytes, and circulating mononuclear cells and subpopulations, (ii) comparing the total determined amounts of all of the biomarkers in the sample to a baseline value and/or a reference of biomarkers obtained from subjects who had undergone stem cell therapy for coronary artery disease and categorized into responders or nonresponders to the stem cell therapy, and (iii) predicting, whether a response to stem cell therapy in the subject is to be expected based on whether the total determined amounts of the biomarkers in the sample are more similar to the baseline value and/or reference of the biomarkers obtained from the responders than that of the nonresponders, or whether a response to stem cell therapy in the subject is not expected based on whether the total determined amounts of the biomarkers in the sample are more similar to the baseline value and/or reference of the biomarkers obtained from the nonresponders than that of the responders.

2. The method according to claim 1, wherein the sensitivity of prediction accuracy is more than about 80%, and the specificity of prediction accuracy is more than about 80%.

3. The method according to claim 1, wherein the growth factor is selected from the group consisting of VEGF and Erythropoietin, the glycoprotein is Vitronectin and the brain natriuretic peptide is NT-proBNP.

4. The method according to claim 1, wherein the method uses further biomarkers, wherein the further biomarkers are selected from the group consisting of Cytokine, Interleukin, Interferon, Insulin-like-growth-factor-binding protein, Insulin-like growth factor, Chemokine protein and/or Multi-protein E3 ubiquitin ligase complex.

5. The method according to claim 4, wherein the Cytokine is TNF, the Interleukin is selected from the group consisting of IL-6, IL-8, and IL-10, the Interferon is human interferon gamma-induced protein 10, the Insulin-like-growth-factor-binding protein is selected from the group consisting of Insulin-like-growth-factor-binding protein-2 and Insulin-like-growth-factor-binding protein-3, the Insulin-like growth factor is IGF2, the Chemokine protein is SDF-1 and/or the Multi-protein E3 ubiquitin ligase complex is SCF.

6. The method according to claim 1, wherein the method is also used for prediction of response to induction of angiogenesis; response to tissue repair of a cardiovascular disease including myocardial infarction, stroke, peripheral ischemic vascular disease, or heart disease; or response to ischemic preconditioning.

7. The method according to claim 6, wherein the stem cell therapy comprises transplantation of CD133 positive stem cells.

8. The method according to claim 6, wherein the stem cell therapy is accompanied by coronary artery bypass graft (CABG).

9. The method according to claim 1, wherein the sample is taken from a subject suffering from heart disease and/or arteriosclerosis.

10. The method according to claim 1, wherein the method comprises profiling of the results of the comparison of at least two, three, four, five, six or more time points.

11. The method according to claim 10, wherein the sensitivity of prediction accuracy is more than 90%, and the specificity of prediction accuracy is more than 90%.

12. The method according to claim 1, wherein the method further comprises the use of clinical diagnostic parameters.

13. The method according to claim 1, wherein the method is used for profiling of angiogenesis response.

14. The method according to claim 1, wherein the method further comprises analysing an RNA and/or mRNA sequence and/or functional RNA and/or non-coding RNA and/or single nucleotide polymorphism (SNP) comprising a diagnostic signature.

15. The method according to claim 1, wherein the method further comprises analysis of pharmacokinetic and pharmacogenetic data employing RNA and/or DNA sequencing analysis and/or network pathway analysis.

16. The method according to claim 1, wherein the method further comprises analysis of phenotyping.

17. The method according to claim 1, wherein said subject is a human.

18. The method according to claim 1, wherein said sample is a blood, a serum and/or a plasma sample, and/or a tissue biopsy sample and/or a sample of circulating stem cells.

Description

(1) FIG. 1 depicts SH2B3 expression analysis in peripheral blood of responder and non-responder. Whole blood samples were obtained from 21 patients before coronary artery bypass graft (CABG) revascularization. Relative expression of SH2B3 (a) and corresponding ΔCT values (b) were calculated using the 2.sup.−ΔΔCT method. All values are presented as mean±SEM and normalized to GAPDH and POLR2A. n=13 (responder); n=8 (non-responder). ΔCT values: p=0.073.

(2) FIG. 2 shows a three-dimensional t-distributed stochastic neighbor embedding (t-SNE) calculation that refers to unsupervised ML. This independent methodology reduces the dimensionality of a given parameter set, e.g. biomarker profile, and thus calculates the variables x, y and z that likewise refer to newly calculated features, which are used to classify the patients into distinct groups. The model was subsequently fitted by a polynomial (n3) equation to visualize the z-axis as a geographic profile. The respective colors for the responder (black dot) and non-responder (grey dot) patients have been added afterwards. The classified groups have been roughly summarized by a black and a grey dashed line. Results are obtained after 3000 iterations. The calculation of the ratio between responder and non-responder is indicated for each circle. It is more likely for the non-responder group to be located at smaller z-values (z<20, ratio<42%). The responders tend to be enriched within the light grey areas (z>20) including a ratio of greater than 69%.

(3) FIG. 3 depicts ML prediction results for pre- and postoperative time points (0 days to 180 days) of the clinical and clinical & laboratory dataset to distinguish between responder and non-responder. The graph shows the true positive prediction rate of five independent feature selected ML models (AdaBoost for feature selection and Random Forest (RF) for final prediction). The error bars indicate the respective accuracy standard deviation for the constructed models that have been obtained after 100 iterations. The 100 model iterations are significantly different according to one-way ANOVA (p<0.001).

(4) FIG. 4 shows the Outcome results of the PERFECT trial. Diagnostic discrimination of primary endpoint outcome responders in the PERFECT trial by distinctive preoperative peripheral blood parameters signature linked to bone marrow/blood CD133.sup.+, CD34.sup.+, CD117.sup.+, EPC and angiogenesis response.

EXAMPLES

(5) The following examples shall merely illustrate the invention. They shall, whatsoever, not be construed as limiting the scope thereof.

Example 1

Clinical Study Design and Evaluation

(6) Induction of cardiac regeneration in patients with heart failure after myocardial infarction and ischemic cardiomyopathy has been targeted using multiple approaches including stem cell therapy. Thereby, lack of efficacy and lack of response predictability have been the main obstacles for treatment standardization and success.

(7) A clinical trial was performed to evaluate if a patient with ischemic heart failure and failed ejection fraction is an appropriate candidate for stem cell therapy and can benefit from it with high probability of cardiac regeneration. The clinical trial depicted a striking difference in cardiac recovery between responders and non-responders, which was associated with a specific signature composition of angiogenesis factors in peripheral blood. Conductance of the clinical study is detailed below.

(8) A study accompanying a controlled, prospective, randomized, double blinded multicenter trial (“Intra-myocardial Transplantation of Bone Marrow Stem Cells in Addition to Coronary Artery Bypass Graft Surgery (PERFECT)”), which was launched in the Rostock University cardiac surgery department was carried out. The trial evaluated efficacy and safety of intra-myocardial CD133.sup.+ cell injection in patients with coronary artery disease after myocardial infarction with reduced LVEF and presence of a localized kinetic/hypokinetic/hypoperfused area of the left ventricle. Prior results revealed close relations between some of regenerative factors and response of patients to the therapy (CABG or CABG plus CD133.sup.+ stem cells injection). An increase in a minimum of 5% in LVEF, measured by MRI, was selected to designate functional improvement at 6-12 months follow-up. The patients, who increased LVEF more than 5%, are defined as the responders, enhancement of less than 5% or decreased LVEF are defining the non-responders.

(9) Biomarkers

(10) Pre-specified: Distinct hematopoetic and endothelial CD133.sup.+ EPC subpopulations and angiogenesis capacity were tested in a cohort of 39 patients in bone marrow (BM) and peripheral blood (PB) employing coexpression analysis using four-laser flow cytometric methods (LSR II, Becton Dickinson, Heidelberg, Germany) for costaining panel enumeration of EPC (co-staining panel CD133, 34, 117, 184, 309, 105, 45) and circulating endothelial cells (CEC) (co-staining panel: CD31, 146, 34, 45, 105, 184, 309) as well as in vitro CFU-EC, CFU-Hill and in vivo Matrigel plug assay. NT-proBNP as well as virus analysis were performed for Epstein-Barr-Virus (EBV), Cytomegalovirus (CMV), and Parvovirus by IgG and antigen analysis in peripheral blood serum. Post hoc analysis before final data closure was performed for serum angiogenesis factors and cytokines. Post hoc analysis: BM subpopulation analysis and SH2B3 mRNA RT-PCR in peripheral blood (PB): Methods and analysis of biomarkers studied in BM CD133+ and PBMNC samples used cytometric bead array (CBA) and enzyme-linked immunosorbent assay (ELISA) and RT-PCR.

(11) Statistical Analysis

(12) The stratification of the primary analysis by centre was neglected in the sample size calculation. Instead of the analysis of covariance (ANCOVA) used in the primary analysis, the two-sample t-test scenario with equal variances was considered. Sample size was determined with the assumption of a two-sided type I error (α) at 5% and a type II error (β) at 10% (i.e. a power at 90%). The scenario of a difference in LVEF at month 6 post-operation between the two treatment arms of 4 to 5% was considered as a clinically relevant difference. With a difference of 4.5 and a standard deviation of 7.5, at least n=60 patients per group were considered necessary and, with an additional 15% drop-out rate, a total of at least 142 patients were to be randomized. Sample size was calculated using the commercial program nQuery Advisor 5.0, section 8, table MTTO-1 (Hofmann W K, de Vos S, Elashoff D, et al., Lancet 2002; 359(9305):481-6). Computation was realized using central and non-central t-distribution where the non-centrality parameter is √n δ/√2 and δ is defined as effect size |μ1-μ2|/σ (O'Brien R G, Muller K E, Signified power analysis for t-tests through multivariate hypothesis. In L K Edward (Ed.) Applied analysis of variance in behavioral science, 1993, New York, Marcel Dekker).

(13) Statistical analyses, final data set calculation, and preparation were performed by Koehler GmbH, Data analysis with machine learning identifying key features and classification of the comprehensive patient data was obtained by employing supervised and unsupervised machine learning (ML) algorithms (Kuhn M, J Statistical Software, 2008; 28(5):1-26). The data were pre-processed while removing features with low variance and high correlation for dimension reduction following best practices recommendations. Missing measurements were filled with zeros as frequently used in standard data imputation practices. The following supervised algorithms were compared: AdaBoost, Support Vector Machines (SVM) and Random Forest (RF) (Forman G and Cohen I, 2004. “Learning from Little: Comparison of Classifiers Given Little Training” doi: 10.1007/978-3-540-30116-5_17). Small clinical datasets are often prone to overfitting. Classifiers were employed that are suitable for training on small data sets for a comparison of features given little training and chose the most appropriate algorithm according to accuracy and robustness towards overfitting (Saeb A T, Al-Naqeb D. Scientifica (Cairo). 2016; 2016:2079704. doi:10.1155/2016/2079704. Epub 2016 May 30). Supervised ML models have been 10-fold cross-validated. Feature selection was then applied from AdaBoost and RF to further reduce the number of features to less than 20. We employed t-distributed stochastic neighbour embedding (t-SNE) for unsupervised machine learning classification and nonlinear dimensionality reduction (Maaten L V D & Hinton G, Visualizing Data using t-SNE. Journal of Machine Learning Research, 2008; 9:2579-2605. Doi http://jmIr.org/papers/v9/vandermaaten08a.html).

(14) Results

(15) Analysis of Patients baseline characteristics in the group of patients for safety set (SAS) and the group of patients per-protocol set (PPS) followed the description of pre-specified cohort analyses SAS (n=77) and PPS (n=58) placebo vs. CD133+. Post hoc analysis was additionally performed to analyse factors influencing primary endpoint outcome. For this, patients were grouped as responders (increase in LVEF>5% at 180 days) or and non-responders (increase in LVEF<5% at 180 days). According to this post hoc analysis 35/58 (60.3%) patients were treatment responders and 23/58 (39.7%) did not improve in LVEF. This responder/non-responder (NR) ratio was similar in the placebo group 57/43% (R/NR: 17/13 patients (pt.)) and in the CD133.sup.+ group 64/36% (R/NR: 18/10 pt.), respectively (placebo vs. CD133.sup.+: p=0.373).

(16) Efficacy Outcome Analysis

(17) The PPS efficacy analysis group (n=58) was characterized by reduced pump function post MI (measured in MRI at rest) with baseline LVEF 33.5%, SD±6.26% [Min-Max 25-49], n=58. Pre-specified primary endpoint: Six months post treatment the left ventricular function showed a considerable increase in LVEF of +9.6%±SD 11.3% [Min-Max-13-42], p<0.001 (n=58). To discriminate early improvement of left ventricular function by CABG revascularization and late myocardial reverse remodeling, additional intermediate MRI analysis at hospital discharge was available in a subgroup of patients (n=29). This revealed mainly late (day 10-180) increase of ΔLVEF by +6.5%, SD±7.92% [Min-Max-11-23], p=0.007 (n=29). In ANCOVA analysis of the primary endpoint the placebo group improved from baseline LVEF 33.5% to 42.3% at 180 days (ΔLVEF+8.8%, SE±2.17% [CI 38.0, 46.6], p<0.001; n=30) and the CD133.sup.+ group LVEF was raised from 33.5% to 43.9% (ΔLVEF+10.4%, SE±2.33% [CI 39.0, 48.5], p<0.001; n=28). Treatment group difference CD133.sup.+ versus placebo with +2.58±SE 3.13% [CI −3.7-8.9], p=0.414 was not statistically significant in ANCOVA analysis. CD133.sup.+ stem cell group displayed ΔLVEF improvement mainly in the late phase (day 10-180 ΔLVEF) with +8.8%, SD±6.38% [Min-Max 4-10], p=0.001 (n=14) versus placebo controls (day 10-180 ΔLVEF)+4.3%, SD±8.8% [Min-Max−11-23], p=0.077 (n=15).

(18) Responder (R)/Non-Responder (NR)

(19) In post hoc primary endpoint analysis treatment responders were defined as having a ΔLVEF at 180 days versus baseline higher than 5%. The results in dissemination of 35 responders in a cohort of 58 patients were characterized by an overall increase in ΔLVEF in ANCOVA at 180d/0 of +17.1%; SE±2.08% [CI 12.9; 21.3], R vs. NR, p<0.0001 (180d/0), n=58. LVEF increase was +19.1% in CD133.sup.+ vs. +13.9% in placebo, p=0.099, n=35 (data not shown). In contrast, non-responders showed a ΔLVEF at 180d/0 by 0%, SE±5.73% [CI 22.3; 44.8] p=0.287 (placebo/NR +3.3%, CD133.sup.+/NR-2.4%).

(20) Post hoc secondary endpoint: Responders showed a significant reduction in LV-dimensions (LVEDV p=0.008, LVESV p=0.0001) and reduction in NT-pro-BNP, p=0.0002 compared to non-responders. This was not reflected by a similar improvement of 6 MWT (p=0.811). The intramyocardial tissue recovery was found in responders with improvement in scar size R vs. NR-8.19 g SE±3.5 g, p=0.0238. CD133.sup.+ treated NR also displayed reduction in scar size (CD133.sup.+ NR Δscar size 180d/0: −13.9 g, SD±20.9 g placebo NR +11.9, SD±16.7 g, p=0.008, n=20) and non-viable tissue (Δnon-viable tissue 180d/0: CD133.sup.+ NR −12.4 g, SD±19.3 g vs. placebo NR +11.5 g, SD±12.0 g, p=0.004, n=19) (data not presented). This tendency was not observed in responders: scar size (CD133.sup.+ NR vs. placebo NR −1.9, SD±16.0 g vs. placebo +2.5, SD±13.2 g, p=0.398, n=33) and non-viable tissue (CD133.sup.+ NR vs. placebo NR −1.4, SD±16.7 g vs. placebo +1.8, SD±12.3 g, p=0.544, n=32). Long term survival: The medium term survival was 76.9±3.32 months (R) vs. +72.3±5.0 months (NR), HR 0.3 [CI 0.07-1.2]; p=0.067.

(21) Circulating EPC (CD133.sup.+/CD34.sup.+/CD117.sup.+) in peripheral blood were found to be reduced by a factor of two in NR versus R before treatment. For CD34.sup.+ MNC subpopulations preoperative blood levels were (R): CD34.sup.+ 0.072%, SD±0.05% vs. (NR) 0.039%, SD±0.017, RvsNR p=0.027. Similar difference was found preoperatively for CD133.sup.+, CD133.sup.+ and CD117.sup.+ subpopulations (pre-operative. RvsNR: CD133.sup.+ 0.048%, SD±0.031% vs. 0.021%, SD±0.011%, p=0.005; CD133.sup.+CD117.sup.+ 0.019%, SD±0.016% vs. 0.007%, SD±0.008%, p=0.024, n=23) (cf. table 1). This difference was not found for the comparison of placebo and CD133.sup.+ (placebo vs. CD133.sup.+ group: CD34.sup.+ p=0.975; CD133.sup.+ p=0.995; CD133.sup.+CD117.sup.+ p=0.892; n=24) (table 1). In contrast, CD146.sup.+ CEC showed higher pre-operative levels in non-responders versus responders (p=0.053) (table 1).

(22) Postoperatively, reduction of EPC in NR remained significant until discharge: peripheral blood CD34.sup.+ (NR vs. R p=0.026 pre-operative and day 10) and CD133.sup.+ CD117.sup.+ (NR vs. R p=0.024 pre-operativeop and day 10) despite postoperative increased levels of EPO (NR: preop. 16.9 U/ml, SD±14.1 U/ml; NR day 10: 42.1 U/ml, SD±23.9 U/ml; p=0.006 pre-operative/day 10) and reduction of IP10/CXCL10 (NR pre-operative: 157.6 pg/ml, SD±94.5 pg/ml; NR day 10: 95.8 pg/ml, SD±85.2 pg/ml; p=0.01 pre-operative/day 10).

(23) Treatment responders were characterized pre-operatively by lower serum levels of pro-angiogenic factors such as VEGF (p=0.056 R/NR), EPO (p=0.023 R/NR), CXCL10/IP10 (p=0.076 R/NR), higher levels of IGFBP-3 (p=0.089 R/NR) (table 1), as well as strong induction of VEGF (+26.6 pg/ml, p=0.015 pre-operative/day 10) at day 10 after intervention versus non-responders (+1.2 pg/ml, p=0.913 pre-operative/day 10) (table 1). Isolated bone marrow CD133.sup.+ cells were all tested positive for their angiogenic potential in vitro by CFU-EC and in vivo by Matrigel plug (data not shown).

(24) Thrombocyte counts were pre-operatively reduced in NR (208×109/L, SD±51.2 109/L [CI 73-311], n=23) versus R (257×109/L, SD±81.5 109/L [CI 123-620], n=35) (NR vs R: p=0.004, n=58) before treatment. Suspecting bone marrow stem cell suppression by finding reduced PB thrombocyte and CD133.sup.+ CD34.sup.+ EPC count, we tested RT-PCR gene expression analysis of SH2B3 mRNA coding for the LNK adaptor protein SH2B3 which is associated with inhibition of hematopoietic stem cell response for EPC and megakaryocytes in immediately frozen blood samples. First analysis in 21 patients revealed a tendency of increased mRNA expression in peripheral blood with non-responders (p=0.073) (cf. table 1, FIG. 1).

(25) TABLE-US-00001 TABLE 1 Analysis of angiogenesis related biomarkers in blood. Responder versus non-responder and placebo versus CD133.sup.+ groups were analysed for change in biomarkers of peripheral blood samples between preoperative (Assessment I) and day 10 postoperative (discharge). The data are derived from the patient group (cohort) with complete analysis (per protocol clinical dataset and biomarker). In this cohort all samples were immediately processed to avoid any change of the samples due to storage or transport. Data are expressed as mean values ± Standard deviation, P-value between time point 0 and 10 days, P.sup.A-value between responder/non-responder, stem cell/control for each time point (PB—peripheral blood, EPO—erythropoietin). Responder versus Non-Responder Biomarker Non- P (peripheral Time Responder P 10 responder 10 days P.sup.A R blood, unit) point (n = 15) days vs 0 (n = 8) vs 0 vs NR SH2B3  0 −1.17 ± 0.28  . . . −1.56 ± 0.51  . . . 0.073 mRNA (ΔCT %) CD34  0. 0.072 ± 0.05  0.197 0.039 ± 0.017 0.116 0.027 (% MNC) - 10 d 0.059 ± 0.048 0.027 ± 0.01  0.026 EPC CD133  0 0.048 ± 0.031 0.245 0.021 ± 0.011 0.932 0.005 (% MNC) - 10 d 0.041 ± 0.039  0.021 ± −0.013 0.105 EPC CD133, 117  0  0.019 ± −0.016 0.421 0.007 ± 0.008 0.765 0.024 (% MNC) 10 d 0.022 ± 0.024 0.006 ± 0.004 0.024 EPC CD146  0  1.1 ± 0.57 . . . 2.2 ± 1.3 . . . 0.053 (% MNC) - 10 d 1.72 ± 1.73 1.86 ± 1.53 0.853 CEC IGFBP-3  0 2121.9 ± 487.1  0.115 1623.7 ± 651.4  0.257 0.089 (ng/ml) 10 d 1753.6 ± 830.8  1378.4 ± 518.7  0.261 VEGF  0  24.6 ± −36.6 0.015 39.6 ± 33.4 0.913 0.056 (pg/ml) 10 d 51.2 ± 55.8  40.8 ± −44.5 0.528 IP-10  0 96.7 ± 42.6 0.04  157.6 ± 94.5  0.01  0.076 (pg/ml) 10 d 63.3 ± 28.3 95.8 ± 85.2 0.324 EPO  0 5.9 ± 3.7 0.001 16.9 ± 14.1 0.006 0.023 (mlU/ml) 10 60.1 ± 27.7 42.1 ± 23.9 0.180 Placebo versus CD133+ Biomarker (peripheral Time Stem cell Control blood, unit) point (n = 11) P (n = 13) P P.sup.A SH2B3  0 −1.35 ± 0.45  . . . −1.29 ± 0.41  . . . 0.756 mRNA (ΔCT %) CD34  0. 0.062 ± 0.037 0.128 0.064 ± 0.053 0.250 0.975 (% MNC) - 10 d 0.041 ± 0.038 0.058 ± 0.047 0.363 EPC CD133  0 0.04 ± 0.03 0.338  0.04 ± 0.029 0.619 0.995 (% MNC) - 10 d 0.032 ± 0.026 0.038 ± 0.032 0.637 EPC CD133, 117  0 0.014 ± 0.013 0.902 0.016 ± 0.017 0.265 0.892 (% MNC) - 10 d 0.015 ± 0.02  0.019 ± 0.022 0.626 EPC CD146  0 1.53 ± 1.33 . . . 1.481 ± 0.67  . . . 0.919 (% MNC) - 10 d 1.64 ± 1.55 1.87 ± 1.74 0.750 CEC IGFBP-3  0 1950.6 ± 689.9  0.139 1946.8 ± 507   0.231 0.972 (ng/ml) 10 d 1561.6 ± 783.2  1679.4 ± 742.6  0.715 VEGF  0 30.2 ± 29.1 0.142 29.6 ± 39.1 0.124 0.961 (pg/ml) 10 d  55.8 ± −58.5 38.5 ± 44.7 0.293 IP-10  0 129.2 ± 96.7  0.011 102.9 ± 34.6  0.001 0.275 (pg/ml) 10 d 83.2 ± 77.9 64.5 ± 22.7 . . . 0.457 EPO  0 7.7 ± 3.1 0.001 10.3 ± 12.6 0.001 0.561 (mlU/ml) 10 d  53.5 ± −30.6 56.4 ± 25.5 0.814

(26) To identify a diagnostic response signature for R/NR we used machine learning methods as a tool for the prediction of functional improvement after cardiac stem cell therapy and CABG surgery. First analyses were performed to particularly exclude overfitting in small populations. Then, blinded patient data from the PERFECT clinical database was investigated by, unsupervised ML, which is able to cluster similar patients in close proximity and reveals distinct groups. Investigating the underlying segmentation, the firstline supervised ML analysis was made for all time points to place patient characteristics into two distinct groups. The calculation independently assigned patient characteristics according to ΔLVEF at 180 days confirming the pre-selection criteria of >5% (cf. table 2, FIG. 2).

(27) Then machine learning algorithms were used to investigate the decisive parameters to a response signature. For this the underlying PERFECT clinical dataset and biomarker laboratory measurements were combined and analysed to validate classification specificity of parameter profiles for responders and non-responders before and after the CABG procedure. In particular, we used discriminative primary and secondary endpoint parameters as well as thrombocyte and leukocyte counts. Using only the clinical parameters (n=160) classification resulted in a specificity of responders assuming mean accuracy of 63.35% (180 days) (table 2). Combination of preoperative clinical data (n=49) and biomarker laboratory parameters (n=142), however, revealed higher sensitivity of angiogenesis/EPC/CEC related parameters in peripheral blood already preoperative with respective assuming max accuracy of 81.64%±SE 0.51% [CI 80.65-82.65] (n=31) (table 2). Interestingly, 17/20 relevant parameters were related to angiogenesis parameters, bone marrow EPC/CEC responses, NTproBNP, and SH2B3 gene expression in peripheral blood (table 2). Using both clinical and biomarker parameters preoperative prediction accuracy for responders was 79.35%±SE 0.24% [CI 78.87-79.84] (n=31) and for non-responders 83.95%±SE 0.93% [CI 82.10-85.80] (n=31). Postoperative evaluation at day 10 (n=382) revealed a prediction accuracy of 82.12%±SE 0.28% [CI 81.56-82.67] (n=31) (R) and 85.89%±SE 0.67% [CI 84.56-87.22] (n=31) (NR), while day 0-180 combined clinical and biomarker analysis (n=522) allowed a prediction accuracy of 94.77%±SE 0.43% [CI 93.92-95.63] (n=31) (R) and 92.44%±SE 0.60% [CI 91.24-93.64] (n=31) (NR) (cf. FIG. 3).

(28) Feature selection based on our machine learning approach led to the identification of decisive factors for lack of response and the induction of cardiac regeneration, which can be used for diagnostic R/NR selection before and monitoring of during treatment. The core factors for laboratory diagnosis in peripheral blood were NT-proBNP, VEGF, erythropoietin, vitronectin, circulating EPC/CEC/Thrombocytes, SH2B3 mRNA expression, the CFU-Hill assay/Matrigel plug for peripheral blood, as well as weight and LVESV index. We found a statistical correlation of the identified factors and calculated their diagnostic use for the selection of responder and non-responder patients using repeated cross-validation (cf. FIG. 4).

(29) A laboratory biomarker subset was selected together with specific features of the clinical trial by ML. The computationally selected features and biomarkers are depicted in table 2 below. Accuracy of prediction was determined above 80%.

(30) TABLE-US-00002 TABLE 2 Machine learning selected parameters for diagnostic discrimination of responders and non-responders. Computationally selected features for the clinical trial data and laboratory biomarker Weights for subset of the Rostock group the selected (day 0 - preoperative) N = 31 Features NT proBNP 9 718 VEGF 7 810 Erythropoietin 4 262 Vitronectin 3 898 CFU_Hill 2 871 CD45Neg_EPC 2 186 CD117_184_PB_EPC_IHG 2 146 CD45_117_184_EPC 2 118 CD45_133_146_PB_CEC 1 969 Thrombocytes 1 951 IGFBP-3 1 922 CD133 pro ml PB_IHG 1 910 CD146_PB_CEC 1 799 CD105_PB_CEC 1 793 CD45_133_34_105_PB_CEC 1 489 MatrigelPlug_PB_31 1 475 CD45_133_34_117_309_EPC 1 420 Delta_CT_SH2B3 1 393 Weight 1 363 LVESV 1 352 Accuracy: 81 64%.sup. 

(31) The invention illustratively described herein suitably may be practised in the absence of any element or elements, limitation or limitations which is/are not specifically disclosed herein. Thus, for example, in each instance herein any of the terms “comprising”, “consisting essentially of” and “consisting of” may be replaced with either of the other two terms. The terms and expressions which have been employed are used as terms of description and not of limitation, and there is no intention that in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention claimed. Thus, it should be understood that although the present invention has been specifically disclosed by preferred embodiments and optional features, modification and variation of the concepts herein disclosed may be resorted to by those skilled in the art, and that such modifications and variations are considered to be within the scope of this invention as defined by the appended claims.

(32) All references cited in this specification are herewith incorporated by reference with respect to their entire disclosure content and the disclosure content specifically mentioned in this specification.

ABBREVIATION CODE

(33) AE=Adverse Event

(34) AESI=Adverse Event of Special Interest

(35) AHA=American Heart Association

(36) ANCOVA=Analysis of covariance

(37) BM=Bone marrow

(38) BMSC=Bone marrow stem cells

(39) CABG=Coronary Artery Bypass Graft

(40) CAP-EPC=Concentrated Ambient Particles—Endothelial Progenitor Cells

(41) CBA=Cytometric Bead Array

(42) CCS=Canadian Cardiovascular Society

(43) CCTRN=Cardiovascular Cell Therapy Research Network

(44) CD=Cluster of Differentiation

(45) CEC=Circulating endothelial cells, CEC panel, CDs measured in PB

(46) CFU=Colony-forming unit

(47) CI=Confidence interval

(48) CMV=Cytomegalovirus

(49) EA=Early Antigen

(50) EBNA1=EBV-Nuclear Antigen 1

(51) EBV=Epstein-Barr-Virus

(52) EC=Endothelial Cells

(53) ECG=Echocardiography

(54) ELISA=Enzyme-Linked Immunosorbent Assay

(55) EPC=Endothelial Progenitor Cells, EPC panel, CDs measured in PB

(56) EPO=Erythropoietin

(57) GMP=Good Manufacturing Practice

(58) HR=Hazard ratio

(59) HIF=Hypoxia-Inducible Factor, transcription factor

(60) ICH GCP=Tripartite Guidelines Guideline for Good Clinical Practice

(61) IGF-1=Insulin-like Growth Factor 1

(62) IGFBP2/3=Insulin-like Growth Factor-Binding Protein 2/3

(63) IHG=Analysis performed in accordance with ISHAGE guidelines

(64) IL=Interleukin

(65) IP-10=Interferon Gamma-induced Protein 10 also known as C-X-C motif chemokine 10 (CXCL10)

(66) LMCA=Left Main Coronary Artery

(67) LVEDV=Left Ventricular End Diastolic Volume

(68) LVEF=Left Ventricular Ejection Fraction

(69) LVESD=Left Ventricular End Systolic Dimension

(70) MACE=Major Adverse Cardiovascular Events

(71) ML=Machine learning

(72) MNC=Mononuclear cells

(73) MRI=Magentic Resonance Imaging

(74) 6MWT=6-Minute Walk Test

(75) NT-proBNP=B-type Brain Natruretic Peptide

(76) PB=Peripheral blood

(77) PBMNC=mononuclear cells isolated from peripheral blood

(78) PCI=Percutaneous Coronary Intervention

(79) PEI=Paul-Ehrlich Institute

(80) PPS=Group of patients for per-protocol set

(81) SAE=Serious adverse event

(82) SAS=Group of patients for safety set

(83) SDF-1=Stromal Cell-derived Factor 1

(84) SH2B3=Lnk [Src homology 2-B3 (SH2B3)] belongs to a family of SH2-containing proteins with important adaptor functions

(85) SCF=Stem Cell Factor

(86) STEMI=ST—segment Elevation Infarction

(87) SUSAR=Suspected Unexpected Serious Adverse Reaction

(88) TNF=Tumor Necrosis Factor

(89) t-SNE=t-distributed stochastic neighbour embedding

(90) VCA=Virus-Capsid-Antigen

(91) VEGF=Vascular Endothelial Growth Factor

(92) VEGF rec=Vascular Endothelial Growth Factor Receptor

(93) VEGFR2/KDR=Vascular Endothelial Growth Factor Receptor 2/Kinase Insert Domain Receptor