ALBERT

Description: Introduction: We have previously reported that clinically relevant, dramatic reductions occur in intestinal bacterial diversity during allogeneic hematopoietic stem cell transplant (allo-HSCT). These are likely attributable to antibiotic exposure, nutritional alterations, and intestinal mucosa injury from high-dose chemotherapy. Patients undergoing autologous hematopoietic stem cell transplantation (AHCT) also receive antibiotics and experience nutritional alterations due to mucositis and other gastrointestinal toxicities. We hypothesized that the pattern of dysbiosis seen in AHCT patients would reflect the changes in allo-HSCT patients. Here, we present a novel analysis of microbiota diversity in AHCT patients from two independent institutions. Methods: We retrospectively identified a cohort of 365 patients (median age 60 years) who underwent AHCT for treatment of hematologic malignancy between May 2009 to February 2018 at two large-volume transplant centers in the US. The population was diverse in terms of histology, conditioning regimens and remission status prior to transplant, with 179 (49%) patients diagnosed with multiple myeloma, 153 (42%) patients diagnosed with lymphoma, and 33 (9%) patients with other diseases. Stool samples from the selected patients were collected approximately weekly during inpatient hospitalization. Sequencing of the V4-V5 region of the bacterial 16S rRNA genes from all samples was performed on the Illumina platform at a central site. Microbial diversity was measured by the Simpson reciprocal a-diversity index (S). We defined the pre-AHCT period as days -10 to 0, and computed median values for patients with multiple samples within that period. We additionally defined monodomination of the microbiota as a single operational taxonomic unit comprising 〉30% of bacterial abundance. For comparison, we sequenced samples from 17 healthy volunteers and used a public dataset of sequences from 313 healthy volunteers from the NIH Human Microbiome Project (HMP). Median pre-transplant microbial diversity in the healthy patient and AHCT groups was compared by a pairwise Wilcox test to a retrospective cohort of allo-HSCT patients. Results: We evaluated 857 samples from 365 adult patients undergoing AHCT, with 316 patients from Memorial Sloan Kettering Cancer Center (MSKCC) and 49 patients from Duke University Medical Center (DUMC). Median pre-transplant diversity in AHCT patients from both centers was significantly lower than in normal controls (Fig 1A) (HMP vs MSKCC AHCT, S=12.05 vs. 9.19, p75% by day +14. Conclusion: Microbial diversity is reduced prior to transplant in both AHCT and allo-HSCT patients. Loss of diversity after AHCT occurs across centers and the degree of injury is comparable to the dysbiosis in allo-HSCT patients. Preliminary analysis suggests that lower diversity may correlate with worse progression-free survival (PFS) in myeloma patients in our diverse AHCT cohort. Given the known associations of alterations in microbiota composition with toxicities and overall survival in allo-HSCT patients, further evaluation of microbiota injury and its associations with toxicities, PFS, and overall survival (OS) in AHCT patients is warranted. Figure 1: A: The median Simpson reciprocal a-diversity index (S) of pre-transplant (days -10 to 0) samples of AHCT and allo-HSCT patients from two centers, as well as two cohorts of healthy volunteers, was plotted and a pairwise Wilcox test was performed, with p-values as indicated. B: (S) was plotted against time relative to allo-HSCT (on L) and AHCT (on R), for samples collected from day -10 to day +30. Larger values indicate greater diversity. C: Microbiota composition and changes in bacterial monodominance after transplant (days -14 to +28); the most common genus post-transplant is Streptococcus. Figure 1. Figure 1. Disclosures Peled: Seres Therapeutics: Research Funding. Sauter:Juno Therapeutics: Consultancy, Research Funding; Sanofi-Genzyme: Consultancy, Research Funding; Spectrum Pharmaceuticals: Consultancy; Novartis: Consultancy; Precision Biosciences: Consultancy; Kite: Consultancy. Shah:Amgen: Research Funding; Janssen: Research Funding. Perales:Novartis: Other: Personal fees; Incyte: Membership on an entity's Board of Directors or advisory committees, Other: Personal fees and Clinical trial support; Merck: Other: Personal fees; Takeda: Other: Personal fees; Abbvie: Other: Personal fees. Jenq:Ziopharm Oncology: Consultancy; Seres Therapeutics, Inc.: Membership on an entity's Board of Directors or advisory committees; MicrobiomeDx: Consultancy; Seres Therapeutics, Inc.: Patents & Royalties.

Description: Abstract 817 Strategies to suppress GVHD are often associated with broader suppression of the immune system leading to a compromised GVT effect. Using experimental models, we have demonstrated a novel strategy to enhance GVT effects and explicitly suppress GVHD using genetically engineered T lineage cells over-expressing TNF-Related Apoptosis Inducing Ligand (TRAIL). TRAIL can induce apoptotic signals through death receptor (DR) 4 and 5 molecules (only DR5 in mice) expressed on target cells. Expression of DR5 is higher on certain tumors and can be enhanced on others using small molecules rendering them susceptible to TRAIL mediated killing. TRAIL is therefore an attractive candidate for genetic engineering of donor T cells to enhance their GVT potential. We evaluated the effect of TRAIL over-expression (TRAIL+) in donor T cells (mature and precursor) on GVHD and GVT. Mature T cells derived from donor B6 splenocytes were transduced with a lentiviral TRAIL expression vector. The transduced TRAIL+ T cells were adoptively transferred on day 0 into lethally irradiated CBF1 recipients of T cell depleted allografts and LB27.4 tumor (B6 ^ CBF1+LB27.4) to assess their GVHD and GVT activity. TRAIL+ T cells displayed significantly enhanced antitumor immunity compared to T cells transduced with a control vector against LB27.4 tumor cell lines in vitro and upon transfer into tumor bearing allo-BMT recipients (p

Description: Introduction: Although the role of Type I IFNs in initiating host defense against pathogens is well established, recent work highlights the regenerative function of this cytokine family. Yet, despite its involvement in tissue repair, the cellular targets and mechanisms of action by which Type I IFNs act during tissue regeneration are poorly understood. Results: Here, we describe a hitherto unrecognized regenerative function for the RIG-I/MAVS/IFN-I pathway through direct effects on epithelial regeneration. Mice deficient in the RIG-I adaptor MAVS were more sensitive to intestinal barrier damage after total body irradiation (TBI) and, like RIG-I deficient mice, developed worse graft-versus-host disease (GVHD) in a preclinical model for allogeneic hematopoietic stem cell transplantation (allo-HSCT). This phenotype was not associated with changes in the intestinal microbiota, but with a defect in epithelial regeneration. Moreover, in contrast to previous reports in steady-state conditions and after viral challenge, we found that interferon-α/β receptor (IFNAR) signaling in non-hematopoietic epithelial cells was crucial for tissue regeneration after acute damage. Importantly, we could demonstrate that this pathway could be therapeutically targeted with RIG-I agonists, which promoted barrier integrity and prevented GVHD. Mechanistically, Type I IFNs (either RIG-I-induced or recombinant) could promote intestinal stem cell (ISC) growth in crypt organoid cultures, suggesting that stimulation of the ISC compartment led to epithelial regeneration. Conclusion: Our findings suggest that activation of RIG-I and IFN-I can promote regeneration of intestinal epithelial cells and thus offers an innovative therapeutic strategy for the management of acute intestinal injury. Disclosures van den Brink: Merck: Honoraria; Boehringer Ingelheim: Consultancy, Other: Advisory board attendee; Regeneron: Honoraria; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees; Tobira Therapeutics: Other: Advisory board attendee.

Description: Abstract 1335 Poster Board I-357 Alloreactive T cells are crucial for graft-versus-host-disease (GVHD) pathophysiology, and we hypothesized that controlling their trafficking can ameliorate GVHD. P-selectin is a dimeric glycoprotein found on most inflamed endothelium, which interacts with multiple lectin-type molecules on leukocytes, including T cells. We used murine allogenienc BMT models to study GVHD and found that P-selectin−/− recipients exhibited significantly less GVHD mortality and morbidity, as well as decreased GVHD of the skin, liver and small bowels. However, WT and P-selectin−/− allo-BMT recipients had comparable large bowel GVHD. This decrease in target organ and systemic GVHD was associated with diminished infiltration of alloactivated T cells into the Peyer's Patches and small bowels, coupled with increased numbers of donor T cells in the spleen and secondary lymphoid organs (SLO) on day 14 and day 35 post-transplant. However, donor alloreactive T cells in WT and P-selectin−/− allo-BMT recipients had similar alloactivation and apoptosis, and donor alloactivated T cells from WT and P-selectin−/− allo-BMT recipients with GVHD showed similar proliferation in vitro in a mixed leukocyte reaction, suggesting that the inflammatory environment in WT and P-selectin−/− recipients was comparable. Finally, non-transplanted P-selectin−/− mice, and P-selectin−/− mice which had received the allo-BMT conditioning regimen but not a donor graft, had similar cellularity in the majority of tissues examined as corresponding WT controls. This suggests that the differential cellularity of donor alloactivated T cells in WT and P-selectin−/− allo-BMT recipients with GVHD is probably largely dependent on trafficking and tissue infiltration during inflammation. Since P-selectin glycoprotein ligand 1 (PSGL1) is the best-described P-selectin ligand, and all leukocytes constitutively bear high levels of membrane PSGL1, we next hypothesized that PSGL1−/− donor alloreactive T cells would be defective in trafficking into GVHD target organs, and that PSGL1−/− donor T cells would cause decreased target organ damage, systemic GVHD, and mortality. However, allo-BMT recipients of WT and PSGL1−/− donor T cells had comparable survival and clinical GVHD scores, and further analyses on day 14 post-transplant revealed that recipients of WT and PSGL1−/− donor T cells also had similar numbers of donor alloactivated T cells in the spleen, liver, mesenteric and peripheral lymph nodes, and Peyer's Patches. Additionally, WT and PSGL1−/− donor T cells had comparable proliferation as measured by CFSE dilution, and comparable alloactivation in vivo as determined by levels of CD25, CD44, and CD62L, suggesting similar T cell function. As PSGL1−/− and WT donor T cells appeared to have equal functionality and accumulated in GVHD target tissues and lymphoid tissues in a similar fashion, we asked whether PSGL1−/− T cells might display other P-selectin ligands. Flow cytometric analyses of T cells from non-transplanted PSGL1−/− mice, and analyses of PSGL1−/− alloactivated T cells on day 14 after allo-BMT, revealed that these cells displayed substantial levels of cell-surface P-selectin ligands as defined by positive staining with recombinant P-selectin-IgG-Fc fusion protein at levels similar to those found on WT T cells, suggesting that although absence of P-selectin on host tissues may ameliorate GVHD, multiple donor leukocyte P-selectin ligands interact meaningfully with P-selectin. Our studies suggest that P-selectin may be required for trafficking into inflamed tissues but not SLO, and that donor T cells may utilize multiple P-selectin ligands apart from PSGL1 to interact with P-selectin and traffic into inflamed tissues during GVHD. We conclude that targeting P-selectin may be a viable target for GVHD prophylaxis or treatment. Disclosures No relevant conflicts of interest to declare.

Description: Background: Hepatic SOS is a serious complication after allogeneic HSCT. It has been reported elsewhere that in the period between 1979-2007, the overall mean incidence of hepatic SOS was 13.7% (95%CI:13-15). While prophylactic procedures are widely used, there is no uniform consensus on the optimal strategy for the prevention of this disease. Low dose UFH have been used but its benefit in the prevention of SOS and the potential associated risk of serious bleeding complications have not been fully established. Methods: We evaluated 730 adult allograft recipients transplanted 01/2003-12/2013 for hematologic malignancies [401 (55%) acute leukemia/ MDS, 235 (32%) lymphoma, 64 (9%) multiple myeloma, 30 (4%) CML/myeloproliferative disorder]. The majority received myeloablative conditioning (n = 466, 65%) with either total body irradiation-based (n = 243) or busulfan-based (n = 233), followed by non-myeloablative conditioning (n = 143, 19%) and reduced intensity (RIC) (n = 111, 15%). Approximately half of the patients had T-cell depleted (TCD) grafts (402, 55%). All patients had SOS prophylaxis with low dose heparin 100 units/kg given as continuous IV over 24 hours from admission to day +21 or engraftment. Prior allogeneic HSCT, non-malignant or refractory disease, dual prophylaxis with UFH/ursodiol or therapeutic anticoagulation at the time of allograft admission were all excluded. Results: The median age was 51 (range 21-65) years, and 415 (56%) were male. During the 10 year study period only 15 patients developed hepatic SOS with a day 100 incidence of 2% (95%CI:1-3). When evaluated by conditioning intensity, myeloablative and reduced intensity conditioning recipients had a day 100 incidence of 3% (95%CI:2-4) (Figure) whereas none of the NMA recipients developed SOS. Univariate analysis showed that age, HLA-match, previous hepatitis B or C exposure did not influence SOS development (Table). There was a significantly higher incidence of hepatic SOS in recipients of unmodified (3%) versus TCD grafts (1%), p= 0.025. The median time to SOS onset was 29 days (range 5-57). Six patients received supportive care only whereas 9 had defibrotide therapy. Of those, 7 patients developed severe SOS resulting in multi-organ failure (5 supportive care, 2 defibrotide). Four of 15 patients diagnosed with hepatic SOS are alive at last follow up (7 deaths attributed to SOS). Bleeding complications occurred in 26/730 patients (4%) of which only 4 (0.5%) patients had significant grade 3-4 bleeding. The day 100 grade II-IV and III-IV acute liver GVHD is 2% (95%CI:1-3) and 0.5% (95%CI:0.2-1), respectively. The incidence of day 100 TRM was 9% (95%CI:7-11) for the entire cohort of patients. Conclusion: To our knowledge, this is the largest analysis evaluating the use of low dose UFH for the prevention of hepatic SOS. We demonstrated that UFH prophylaxis is associated with a low incidence of hepatic SOS, is well tolerated, and severe bleeding complications are uncommon. Low dose UFH may be consider as a strategy for the prevention of hepatic SOS in selected patients. Further prospective studies comparing UFH to other hepatic SOS prophylaxis methods are warranted. Table. Univariate analysis of variables potentially associated with day 100 SOS. Variable (95% CI) p-value Age (years) 0.745 〈 40 2% (1-6) ≥ 40 2% (1-3) HLA-match 0.857 8/8 2% (1-4) ≤ 7/8 2% (1-5) Graft 0.025 TCD 1% (0.3-2) Unmodified 3% (2-6) Conditioning regimen 0.054 Myeloablative/RIC 3% (2-4) Non-Myeloablative 0 (NA) Myeloblative conditioning type 0.353 Busulfan-based 2% (1-5) TBI-based 3% (2-6) Hepatitis B & C Serologies 0.089 Negative 2% (1-3) Positive 8% (0.4-30) Figure 1. Figure 1. Disclosures Bhatt: Spectrum: Consultancy. Off Label Use: Use of low dose unfractionated heparin for the prevention of hepatic SOS. van den Brink:Boehringer Ingelheim: Consultancy, Other: Advisory board attendee; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees; Regeneron: Honoraria; Merck: Honoraria; Tobira Therapeutics: Other: Advisory board attendee. Giralt:TAKEDA: Consultancy, Honoraria, Research Funding; AMGEN: Consultancy, Research Funding; JAZZ: Consultancy, Honoraria, Research Funding, Speakers Bureau; SANOFI: Consultancy, Honoraria, Research Funding; CELGENE: Consultancy, Honoraria, Research Funding.

Description: Abstract 1043 Thymopoiesis is a highly complex process involving cross-talk interactions between developing thymocytes and the supporting non-hematopoietic stromal microenvironment, which includes highly specialized thymic epithelial cells (TECs). Paradoxical to its importance for continually generating a diverse repertoire for effective adaptive immunity, the thymus undergoes profound atrophy with age. Age-related thymic involution is characterized by severe structural dysregulation of the supporting epithelial microenvironment (and in humans linked to a buildup of fatty tissue), reduced thymopoiesis, and subsequently reduced export of na•ve lymphocytes into the periphery. Together this degeneration in thymic function significantly narrows the T cell receptor repertoire and can causally linked to increased infection, autoimmunity and malignancy. Moreover, progressive thymic involution can also be a considerable hindrance to the regeneration of adaptive immunity following cytoreductive treatments such as chemotherapy or the conditioning required for successful hematopoietic stem cell transplant. Despite considerable work, little is understood about the underlying causes of age-related thymic involution. We have recently demonstrated a novel role for interleukin-22 (IL-22), a recently identified cytokine predominantly associated with maintenance of barrier function at mucosal surfaces, in endogenous thymic regeneration from acute immune injury. Our studies suggested that 1) the depletion of DP thymocytes triggers, 2) upregulation of IL-23 by dendritic cells (DCs), which induces 3) the production of IL-22 by intrathymic innate lymphoid cells (ILCs). IL-22 promotes the proliferation and survival of TECs, therefore this cascade of events leads to regeneration of the supporting microenvironment and, ultimately, to rejuvenation of thymopoiesis. Given these recent findings demonstrating a role for IL-22 in endogenous thymic regeneration following acute immune injury, one hypothesis would be that a breakdown in the IL-22 pathway contributes towards chronic age-related thymus involution. However, in contrast to this initial hypothesis, our studies revealed that rather than being depleted with age, there was actually a significant increase in the level of intrathymic IL-22 in aged (18+ months old) compared to young (2 months old) mice (Figure 1a). These findings highlighted that, in addition to being triggered by the depletion of CD4+CD8+ double positive thymocytes during acute immune injury, the IL-22 regenerative pathway can also be activated by the chronic atrophy that is a hallmark of age-related thymic involution. Similar to our findings in models of thymic injury in young mice, we found that these increased levels of IL-22 with age were predicated on the increased production of IL-22 by thymic innate lymphoid cells (Figure 1b). Moreover, in keeping with our findings in young mice with acute thymic injury, intrathymic levels of IL-22 in aged mice correlated with those of IL-23 - production of which by dendritic cells was significantly increased with age (Figure 1c). As predicted by this increase in the production of IL-22 with age, TECs from aged mice displayed all the hallmarks of increased IL-22 signaling including increased expression of the IL-22 receptor (Figure 1d) as well as increased phosphorylation of STAT-3 (Y705) (Figure 1e). However, although in vitro incubation of aged TECs with IL-22 led to increased proliferation, consistent with our findings in young mice, in vivo analysis revealed significantly reduced proliferation among TECs in aged mice (Figure 1f), as has been previously reported. Given the role for inflammasome components in mediating thymic involution, it is possible that although endogenous regenerative pathways are triggered with age (in the case of IL-22 likely due to the depletion of DP thymocytes), these regular processes fail in the face of an overwhelming inflammatory milieu in the thymus with age. Although further studies need to elucidate the specific inhibitory interactions constraining thymic regeneration, it is clear that strategies harnessing these endogenous pathways for enhancing immunity in the aging thymus first need to overcome these negative stimuli for effective regeneration. Disclosures: No relevant conflicts of interest to declare.

Description: Reduced oral intake (ROI) is common in patients that have received cytotoxic therapy for hematologic malignancies. Here we examine the interplay between nutrition, intestinal bacteria and intestinal barrier function, focusing on preclinical mouse models. By 3 days following 9 Gy of total body irradiation (TBI), individually-housed C57BL/6 mice were observed to have an approximately 40-60% reduction in oral intake (Fig A), and 3 days later displayed marked changes in the microbiome by 16S deep sequencing, including an expansion of Bacteroides thetaiotaomicron (Fig B). These mice also had lost roughly 20% of their body weight (Fig C), and showed compromise of their intestinal barrier assayed by oral challenge with FITC-dextran (Fig M). We hypothesized that the ROI could be a driving factor behind the microbiome changes. To assess this, we limited the access of unirradiated mice to 2 grams of normal chow a day for 7 days (an approximately 50% reduction in oral intake) with unlimited water. Microbiome profiling by 16S sequencing showed many changes that were similar to those seen in irradiated mice, including an expansion of Bacteroides thetaiotaomicron (Fig D). Because Bacteroides thetaiotaomicron has been shown to utilize glycans derived from colonic mucin, we evaluated the consumption of porcine gastric mucin by intestinal bacteria in vitro with a colorimetric assay that quantifies polysaccharides. We found that following ROI, mice harbored intestinal bacteria that were significantly more efficient at utilizing mucin glycans (Fig E). Corroborating this, a histological characterization of the colonic mucus thickness following fixation with Carnoy's solution showed that after a week of ROI, unirradiated mice developed significant thinning of the colonic mucus layer (Fig F), though intestinal barrier function otherwise remained intact as assayed by FITC-dextran gavage (data not shown). We asked how ROI could favor mucolytic bacteria and hypothesized that ROI could reduce bacterial fermentation of dietary carbohydrates which largely results in production of short-chain fatty acids. Evaluating the pH of the colonic lumen showed that ROI led to a significant elevation in pH (Fig G). Ion-chromatography mass spectrometry of colonic stool samples showed a reduction in acetate, propionate, butyrate, and lactate following ROI, and interestingly an increase in succinate (Fig I). Corroborating this, RNA sequencing of intestinal bacteria following ROI demonstrated a downregulation of phosphoenolpyruvate carboxykinase, a metabolic enzyme that in Bacteroides species has been previously been found to be associated with conversion of succinate to propionate (Fig H). Finally, we evaluated in our ROI model the impact of administration of oral sugars by supplementing a well-absorbed sugar (glucose), and a poorly-absorbed prebiotic sugar in the drinking water of mice. We found that compared with glucose, supplementation with the prebiotic more effectively acidified the colonic lumen of normal mice (Fig J), and completely prevented thinning of the mucus layer in mice undergoing ROI (Fig K). Additionally, in mice following 9 Gy of total body irradiation, daily administration of the prebiotic was highly effective in preventing weight loss (Fig L) and nearly completely abrogated compromise of intestinal barrier function (Fig M). Conclusions: Reduced oral intake following cytotoxic therapy contributes to thinning of the colonic mucus layer, an effect that appears to be independent of the cytotoxic effects on the host and is rather mediated by changes in the intestinal microbiome, including reduced metabolism of organic acids. Strategies to restore colonic acidity with prebiotic sugars may successfully target this phenomenon with potential clinical benefits, particularly in preventing disruption to the intestinal barrier. Figure Figure. Disclosures Jenq: MicrobiomeDx: Consultancy; Seres Therapeutics, Inc.: Patents & Royalties; Ziopharm Oncology: Consultancy; Seres Therapeutics, Inc.: Membership on an entity's Board of Directors or advisory committees.

Description: Abstract 2296 Poster Board II-273 CB transplantation (CB-T) may be curative for patients with high-risk or advanced hematologic malignancies. However, given there are no randomized trials comparing survival after CBT with the more traditional approach of matched related donor transplantation (MRD-T) or unrelated donor transplantation (URD-T), how CB-T compares to MRD-T and URD-T transplantation is not established. Therefore, we conducted a retrospective study comparing survival after CB-T (n=67) with MRD-T (n=96) and URD-T (n=163) performed 10/05-3/09 for the treatment of hematologic malignancies. Our hypothesis was that 1 year survival is comparable between hematopoietic stem cell (HSC) sources. Consecutive adult and pediatric recipients of first allograft for the treatment of acute leukemia in remission (CR1-3), myelodysplasia (MDS, ≤5% blasts at work-up), or non-Hodgkins or Hodgkin lymphoma were eligible for analysis. The median age of CB-T recipients (37 years, range

Description: Introduction: Dysbiosis of the gut microbiome during hematopoietic stem cell transplantation (HCT) is associated with adverse post-transplant outcomes such as graft-versus-host disease, bloodstream infections, and mortality. In order to learn more about the role of the microbiome in HCT in adverse clinical outcomes, researchers collect stool samples from patients at various time points throughout HCT. However, unlike blood samples or skin swabs, stool collection requires active subject participation, particularly in the outpatient setting, and may be limited by patient aversion to handling stool. By providing study participants with compensation for their stool samples, we hypothesize that we can significantly increase stool collection rates. Methods: We performed a prospective cohort study on the impact of financial incentives on stool collection rates for microbiome studies. The intervention group consisted of allogeneic (allo)-HCT patients from 05/2017-05/2018 who were compensated with a $10 gas gift card for each stool sample. The intervention group was compared to a historical control group consisting of allo-HCT patients from 11/2016-05/2017 who provided stool samples before the incentive was implemented. To control for potential changes in collections over time, we also compared a contemporaneous control group of autologous (auto)-HCT patients from 05/2017-05/2018 with a historical control group of auto-HCT patients from 11/2016-05/2017; neither auto-HCT groups were compensated. Allo-HCT patients were required to give samples at pre-HCT, day 0 (the day of HCT), and days 7, 14, 21, 30, 60, and 90 post-HCT. Auto-HCT patients were required to give samples at pre-HCT and days 7, 14, and 90 post-HCT. Collection rates were defined as the number of samples provided divided by the number of time points for which we attempted to obtain samples. Patient characteristics were summarized by proportions for categorical variables and median with interquartile ranges for continuous variables. Chi-square tests or Fisher's exact tests were used to compare categorical variables, as appropriate, and Wilcoxon Rank Sum tests or t-tests were used to compare continuous variables, as appropriate. This study was approved by the Duke Institutional Review Board, and informed consent was obtained from all patients. Results: There were 35 allo-HCT patients in the intervention group, 19 allo-HCT patients in the historical control group, 142 auto-HCT patients in the contemporaneous control group, and 75 auto-HCT patients in the historical control group. Groups were similar with regard to baseline demographics such as age, race, and gender. While allo-HCT patients were more likely to have leukemia and auto-HCT patients were more likely to have lymphoma and multiple myeloma, there were no differences in disease rates across the study periods. Allo-HCT patients in the intervention group had significantly higher average overall collection rates when compared to the historical control group allo-HCT patients (80% vs 37%, p