S304 The Journal of Heart and Lung Transplantation, Vol 37, No 4S, April 2018 anti-NKp46, CD3, CD163 and CD34 antibodies for NK cells, T lympho- cytes, macrophages and endothelial cells respectively. The sections were observed with Vectra imaging system and analyzed with InForm software (Perkin Elmer). Results: The density of total NK cells was very low in non-rejecting EMB (mean: 0.45/mm 2 ) and increased in AMR and ACR EMB (mean: 13.33 vs.36.32/mm 2 respectively, NS). The ratio of intravascular to perivascular/ interstitial NK cells was increased in AMR vs. ACR (mean: 3.64 vs. 0.67 respectively; p<0.001). The density of perivascular/interstitial NK cells was increased in ACR vs. AMR (mean: 32.37/mm 2 vs. 0.67/ mm 2 respectively; p<0.05). Preliminary results of multiplex immunofluorescence demonstrated clearly both intravascular and perivascular/interstitial NK cells and their close relationship with T lymphocytes and macrophages. Automatic cell popula- tions counting using InForm program is in progress. Conclusion: NK cells are recruited in both ACR and AMR. Their compart- mentalization differs between ACR and AMR. NK cells colocalize with T lymphocytes and macrophages and appear as a cell type engaged in cardiac rejection. ( 770) The Interleukin-33/ST2 Pathway is Expressed in the Failing Human Heart and Associated with Pro-Fibrotic Remodeling of the Myocardium M.M. Huibers , 1 C.C. Tseng, 2 J. van Kuik, 1 R.A. de Weger, 1 A. Vink, 1 N. de Jonge. 2 1 Pathology, UMC Utrecht, Utrecht, Netherlands; 2 Cardiology, UMC Utrecht, Utrecht, Netherlands. Purpose: The interleukin-33 (IL-33)/ suppression of tumorigenicity 2 (ST2) pathway is a potential pathophysiological mediator of cardiac fibrosis. Soluble ST2 (sST2) is one of the main isoforms of ST2 with strong prog- nostic value in cardiac disease. The exact role of sST2 in cardiac fibrosis is unknown. The aim of this study was 1) to investigate myocardial expression of the IL-33/ST2 pathway in relation to myocardial fibrosis in end-stage heart failure patients and 2) to study whether plasma sST2 is associated with histologically determined cardiac fibrosis. Methods: In 38 patients undergoing left ventricular assist device implanta- tion, mRNA expression of sST2, total ST2 and IL-33 was measured in cardiac tissue obtained during the implantation. In the same tissue histological fibro- sis was digitally quantified and mRNA expression of pro-fibrotic signaling molecules, connective tissue growth factor (CTGF) and transforming growth factor beta 1 (TGFβ1), was measured. In addition, plasma levels of sST2 were determined. Results: Expression levels of IL-33/ST2 pathway factors in myocardial tis- sue were significantly associated with cardiac fibrosis and the expression levels of CTGF and TGFβ1. Plasma levels of sST2 did not correlate with tissue expression of ST2, the amount of fibrosis or myocardial expression of pro-fibrotic signaling proteins. Conclusion: The interleukin-33/ST2 pathway is expressed in the failing human heart and its expression is associated with cardiac fibrosis and pro- fibrotic signaling proteins, suggesting a role in pro-fibrotic myocardial remodeling. Soluble ST2 levels in the circulation did not correlate with the amount of cardiac fibrosis or myocardial ST2 expression. Therefore, other pathophysiological processes such as inflammation might also substantially affect sST2 plasma levels. ( 771) Pathological Correlation Between Apical Core Biopsies at the Time of Left Ventricular Assist Device Implantation and Excised Heart at Time of Transplant or Autopsy M. Alhussein , 1 L. Battioni, 1 K. Runeckles, 1 J. Duero Posada, 1 Y. Moayedi, 1 J. Lombardi, 1 H.J. Ross, 1 F. Billia, 1 V. Rao, 1 J. Butany, 2 M. McDonald. 1 1 Division of Cardiology, University Health Network, Toronto, ON, Canada; 2 Division of Pathology, University Health Network, Toronto, ON, Canada. Purpose: The diagnostic yield of endomyocardial biopsy is low, a possible explanation is the small sample size. At the time of LVAD implantation, core tissue biopsies may prove to be useful but some reports suggest pathologi- cal findings are non-specific. In this study, we aim to compare core biopsy findings at the time of LVAD implantation with those of the explanted heart at the time of heart transplantation (HT) or autopsy ( 768) Improving the Diagnosis of Rejection by Molecular Phenotype of Endomyocardial Biopsies: Single Center Insights from the Interheart Study L. Borgese , 1 L. Potena, 1 O. Leone, 2 V. Agostini, 2 J. Reeve, 3 M. Masetti, 1 A. Russo, 1 F. Grigioni, 1 P. Halloran. 3 1 Cardiology, University of Bologna, Bologna, Italy; 2 Pathology, University of Bologna, Bologna, Italy; 3 University of Alberta, Edmonton, AB, Canada. Purpose: Endomyocardial biopsy (EMB) is the standard of practice for diagnosis of rejection in heart transplant. However, reproducibility of EMB pathology reading (pEMB) may yield to high diagnostic variability, ques- tioning its overall accuracy. In addition, correlation of current grading of pEMB with clinical phenotypes also is uncertain. Seeking to improve the precision in rejection diagnosis and to guide the interpretation of pEMB, we used a microarray-based analysis to identify molecular phenotypes (MP) associated with rejection. Methods: 95 EMBs from 57 prospectively enrolled patients were analyzed by the MMDx platform. This analysis was based on algorithms derived from kidney transplant biopsies, in which T-Cell genes were associated with cel- lular rejection (CR), while endothelial, NK and IFN-g genes were associated with antibody-mediated rejection (AMR). Patients also underwent right heart catheterization (RHC), and cardiac ultrasound. EMB were graded according to 1990 ISHLT and 2006 working formulations for CR and 2014 guidelines for AMR. Results: Abnormal MP was found in 31(32%), and abnormal pEMB in 41(43%) of the samples. The majority (73%) of abnormal MP revealed AMR, while CR was prevalent in pEMB. ISHLT grading was highly cor- related with rejection MP, found in 65% of 2R EMB, 32% of 1R and 10% of 0R (P<0.01). After reclassifying 1R with 1990 grades, we found that none of grade 2 EMB showed rejection MP, as opposed to the 50% of 1B and 35% of 1A. Similarly, in pAMR grading, 75% of pAMR1H+ and of pAMR2 samples also had rejection MP, while only 25% of pAMR1I+ showed rejection MP. Analyzing the discrepancies between MP and pEMB, we found that of the 9 patients with rejection MP and normal pEMB, 6 developed heart failure or rejection on the subsequent EMB. Rejection MP was associated with abnormal RHC and lower ejection fraction (P<0.02), while pathology grading was not. Conclusion: Despite being grossly associated with pathology findings, MP provided a more accurate and clinically relevant diagnosis than pEMB. Heterogeneity in MP of morphologies associated to 1R grade and largely normal MP in pAMR1i+ grade suggest low specificity and lack of precision of the current grading systems. The imbalance in AMR and CR etiologies between pEMB and MP support the hypothesis of endothelial gene activation during CR in heart transplant patients ( 769) NK Cells in Endomyocardial Biopsies From Cardiac Allografts: Detection, Quantification, and Precise Localization Using Multiplex Immunofluorescence M. Terada , 1 M. Rabant, 2 C. Lesaffre, 1 J. Duong Van Huyen, 1 P. Bruneval. 1 1 INSERM U970, Paris, France; 2 INSERM U1151, Paris, France. Purpose: Among the complex cellular interplay engaged in cardiac rejec- tion, recent transcriptomic analysis of endomyocardial biopsies (EMB) has detected NK cell transcripts in the setting of antibody-mediated rejection (AMR). The precise in situ detection of NK cells in EMB supporting this molecular signature remains to be done. Methods: Formalin-fixed paraffin-embedded sections from 30 EMB [10 without rejection; 8 with acute cellular rejection (ACR) (1 with 1R, 4 with 2R, and 3 with 3R); 12 with AMR (3 pAMR 1(H+) and 9 pAMR 2)] were used for immunoperoxidase with anti-NKp46 antibody. Labeled NK cells were manually counted, their probable intravascular or perivascular/ interstitial location was determined and the surface area of myocardial tissue bites was measured for calculation of NK cell densities. To deter- mine the precise intravascular or perivascular/interstitial location of NK cells and their relationship with T lymphocytes and macrophages, multi- plex immunofluorescence labeling kit Opal (Perkin Elmer) was used with