Search

Background Many patients with myelodysplastic syndromes (MDS) receive red cell transfusions to relieve symptoms associated with anemia, with transfusions triggered by hemoglobin level. It is not known if patients' quality of life (QOL) improves after transfusion, nor if peri-transfusion QOL assessment (PTQA) can guide future transfusion decisions. Study design and methods We conducted a prospective pilot study of adults with MDS at three centers. Participants, who had to have hemoglobin ≥7.5, completed an MDS-specific measure of QOL (the Quality of Life in Myelodysplasia Scale, [QUALMS]) 1 day before and 7 days after red cell transfusion. A report was sent to each patient and provider before the next transfusion opportunity, indicating whether there were clinically significant changes in QOL. We assessed the proportion of patients experiencing changes in QOL, and with a follow-up questionnaire, whether they perceived their PTQA data were used for future transfusion decisions. Results From 2018 to 2020, 62 patients enrolled (mean age 73 years) and 37 completed both pre- and post-transfusion QOL assessments. Of these, 35% experienced a clinically significant increase in QUALMS score 7 days after transfusion; 46% no change; and 19% a decrease. Among those completing the follow-up questionnaire, 23% reported that PTQA results were discussed by their provider when considering repeat transfusion. Conclusions These data suggest PTQA is feasible for patients with MDS. Moreover, while helpful for some, for many others, red cell transfusion may not achieve its intended goal of improving QOL. PTQA offers a strategy to inform shared decision-making regarding red cell transfusion.

Key Points Public demographic reporting for acute leukemia trials is inadequate, and NH-White subjects are more likely to be enrolled.Larger racial-ethnic enrollment disparities were documented after federal reporting requirements, which may be from more data transparency., Visual Abstract, Data regarding racial and ethnic enrollment diversity for acute myeloid leukemia (AML) and acute lymphoid leukemia (ALL) clinical trials in the United States are limited, and little is known about the effect of federal reporting requirements instituted in the late 2000s. We examined demographic data reporting and enrollment diversity for ALL and AML trials in the United States from 2002 to 2017, as well as changes in reporting and diversity after reporting requirements were instituted. Of 223 AML trials and 97 ALL trials with results on ClinicalTrials.gov, 68 (30.5%) and 51 (52.6%) reported enrollment by both race and ethnicity. Among trials that reported race and ethnicity (AML, n = 6554; ALL, n = 4149), non-Hispanic (NH)-Black, NH-Native American, NH-Asian, and Hispanic patients had significantly lower enrollment compared with NH-White patients after adjusting for race-ethnic disease incidence (AML odds ratio, 0.68, 0.31, 0.75, and 0.83, respectively; ALL odds ratio, 0.74, 0.27, 0.67, and 0.64; all, P ≤ .01). The proportion of trials reporting race increased significantly after implementation of the reporting requirements (44.2% to 60.2%; P = .02), but race-ethnicity reporting did not (34.8% to 38.6%; P = .57). Reporting proportions according to number of patients enrolled increased significantly after the reporting requirements were instituted (race, 51.7% to 72.7%; race-ethnicity, 39.5% to 45.4%; both, P < .001), and relative enrollment of NH-Black and Hispanic patients decreased (AML odds ratio, 0.79 and 0.77; ALL odds ratio, 0.35 and 0.25; both P ≤ .01). These data suggest that demographic enrollment reporting for acute leukemia trials is suboptimal, changes in diversity after the reporting requirements may be due to additional enrollment disparities that were previously unreported, and enrollment diversification strategies specific to acute leukemia care delivery are needed.

Importance State crisis standards of care (CSC) guidelines in the US allocate scarce health care resources among patients. Anecdotal reports suggest that guidelines may disproportionately allocate resources away from patients with cancer, but no comprehensive evaluation has been performed. Objective To examine the implications of US state CSC guidelines for patients with cancer, including allocation methods, cancer-related categorical exclusions and deprioritizations, and provisions for blood products and palliative care. Design, Setting, and Participants This cross-sectional population-based analysis examined state-endorsed CSC guidelines published before May 20, 2020, that included health care resource allocation recommendations. Main Outcomes and Measures Guideline publication before or within 120 days after the first documented US case of coronavirus disease 2019 (COVID-19), inclusion of cancer-related categorical exclusions and/or deprioritizations, provisions for blood products and/or palliative care, and associations between these outcomes and state-based cancer demographics. Results Thirty-one states had health care resource allocation guidelines that met inclusion criteria, of which 17 had been published or updated since the first US case of COVID-19. States whose available hospital bed capacity was predicted to exceed 100% at 6 months (χ2 = 3.82;P = .05) or that had a National Cancer Institute–designated Comprehensive Cancer Center (CCC; χ2 = 6.21;P = .01) were more likely to have publicly available guidelines. The most frequent primary methods of prioritization were the Sequential Organ Failure Assessment score (27 states [87%]) and deprioritizing persons with worse long-term prognoses (22 states [71%]). Seventeen states’ (55%) allocation methods included cancer-related deprioritizations, and 8 states (26%) included cancer-related categorical exclusions. The presence of an in-state CCC was associated with lower likelihood of cancer-related categorical exclusions (multivariable odds ratio, 0.06 [95% CI, 0.004-0.87]). Guidelines with disability rights statements were associated with specific provisions to allocate blood products (multivariable odds ratio, 7.44 [95% CI, 1.28-43.24). Both the presence of an in-state CCC and having an oncologist and/or palliative care specialist on the state CSC task force were associated with the inclusion of palliative care provisions. Conclusions and Relevance Among states with CSC guidelines, most deprioritized some patients with cancer during resource allocation, and one-fourth categorically excluded them. The presence of an in-state CCC was associated with guideline availability, palliative care provisions, and lower odds of cancer-related exclusions. These data suggest that equitable state-level CSC considerations for patients with cancer benefit from the input of oncology stakeholders.

BACKGROUND: Brief measures of physical function such as gait speed may be useful to optimize treatment intensity for older adults with blood cancers; however, little is known about whether such assessments are already captured within oncologists’ “gestalt” assessments. METHODS: Gait speed was assessed in 782 patients ≥75 years with blood cancers at our institution, with results reported to providers after treatment decisions were made; 408 required treatment when different intensities were available per National Comprehensive Cancer Network (NCCN) guidelines. We performed structured abstractions of treatment intensity recommendations into standard intensity, reduced intensity, or supportive care, based on NCCN guidelines. We modeled gait speed and survival using Cox regression and performed ordinal logistic regression to assess predictors of NCCN-based categorizations of oncologists’ treatment intensity recommendations, including gait speed. RESULTS: Median survival by gait speed category was 10.8 months (0.8 m/s). Univariable hazard ratios for death increased for each lower category compared to ≥0.8 m/s (0.6–0.8 m/s: HR 1.76; 0.4–0.6 m/s: HR 2.30; 0.05) and did not add to a base model predicting recommended treatment intensity. CONCLUSION: In older adults with blood cancers presenting for treatment, gait speed predicted survival but not treatment intensity recommendation. Incorporating gait speed into decision-making may improve optimal treatment selection.

PURPOSE: Medication shortages in US hospitals are ongoing, widespread, and frequently involve antineoplastic and supportive medications used in cancer care. The ways shortages are managed and the ways provider-patient communication takes place are heterogeneous, but the related preferences of oncology patients are undefined. This study sought to qualitatively evaluate patient preferences. METHODS: A cross-sectional, semi-structured interview study was conducted from January to June 2019. Participants were adult oncology inpatients who received primary cancer care at the University of Chicago, had undergone treatment within 2 years, and had 1 or more previous hospitalizations during that period. Participants (n = 54) were selected consecutively from alternating hematology and oncology services. The primary outcome was thematic saturation across the domains of awareness of medication shortages, principle preferences regarding decision makers, preferences regarding allocation of therapy drugs, and allocation-related communication. RESULTS: Thematic saturation was reached after 39 participants completed the study procedures (mean age, 59.6 years [standard deviation, 14.5 years]; men made up 61.5% of the study population [mean age, 24 years]; response rate, 72.0%). In all, 18% of participants were aware of institutional medication shortages. Patients preferred having multiple decision makers for allocating medications in the event of a shortage. A majority of patients named oncologists (100%), ethicists (92%), non-oncology physicians (77%), and pharmacists (64%) as their preferred decision makers. Participants favored allocation of drugs based on their efficacy (normalized weighted average, 1.3), and they also favored prioritizing people who were already receiving treatment (1.8), younger patients (2.0), sicker patients (3.1), and those presenting first for treatment (5.3). Most participants preferred preferred disclosure of supportive care medication shortages (74%) and antineoplastic medication shortages (79%) for equivalent substitutions. CONCLUSION: In a tertiary-care center with medication shortages, few oncologic inpatients were aware of shortages. Participants preferred having multiple decision makers involved in principle-driven allocation of scarce medications. Disclosure was preferred when their usual medications needed to be substituted with equivalent alternatives. These preliminary data suggest that preferences do not align with current management practices for medication shortages.

Minimal residual disease (MRD) testing in acute myeloid leukemia is increasingly being used to assess treatment response and stratify the risk of relapse for individual patients. Molecular methods for MRD testing began with PCR-based assays for individual recurrent mutations. To date, there is robust evidence for testing NPM1, CBFB-MYH11, and RUNX1/RUNXT1 mutations using this approach, though the best timing and threshold level for each mutation varies. More recent approaches have been with PCR-based multigene panels, occasionally combined with flow cytometric techniques, and next-generation sequencing techniques. This review outlines the various techniques used in molecular approaches to MRD, the evidence behind individual mutation testing, and the novel approaches for evaluating multigene MRD so that clinicians can understand and incorporate these evaluations into their practice.

Accurate understanding of the prognosis of an advanced cancer patient can lead to decreased aggressive care at the end of life and earlier hospice enrollment.Our goal was to determine the association between high-risk clinical events identified by a simple, rules-based algorithm and decreased overall survival, to target poor prognosis cancer patients who would urgently benefit from advanced care planning.A retrospective analysis was performed on outpatient oncology patients with an index visit from April 1, 2015, through June 30, 2015. We examined a three-month window for "high-risk events," defined as (1) change in chemotherapy, (2) emergency department (ED) visit, and (3) hospitalization. Patients were followed until January 31, 2017.A total of 219 patients receiving palliative chemotherapy at the University of Chicago Medicine with a prognosis of ≤12 months were included.The main outcome was overall survival, and each "high-risk event" was treated as a time-varying covariate in a Cox proportional hazards regression model to calculate a hazard ratio (HR) of death.A change in chemotherapy regimen, ED visit, hospitalization, and at least one high-risk event occurred in 54% (118/219), 10% (22/219), 26% (57/219), and 67% (146/219) of patients, respectively. The adjusted HR of death for patients with a high-risk event was 1.72 (95% confidence interval [CI] 1.19-2.46, p = 0.003), with hospitalization reaching significance (HR 2.74, 95% CI 1.84-4.09, p 0.001).The rules-based algorithm identified those with the greatest risk of death among a poor prognosis patient group. Implementation of this algorithm in the electronic health record can identify patients with increased urgency to address goals of care.

Introduction Race-ethnic disparities in clinical trial enrollment have the potential to bias findings, limit generalizability, misdirect drug development, and reduce equitable access to novel therapy. The degree to which such disparities exist within acute myeloid leukemia (AML) North American cooperative group trials, biobanks, and correlative studies remain unclear, as are the factors that influence biobank and correlative study participation among trial enrollees. In addition, the National Cancer Institute's (NCI) mandate for Comprehensive Cancer Centers (CCC) to designate catchment areas has not been explored as a mechanism through which AML enrollment disparities can be addressed. Methods We analyzed enrollment data from the 9 Alliance/CALGB AML treatment trials, 2 biobank protocols, and 2 correlative studies active from 1998-2013 and with published results. Trial enrollees could consent to biobank and/or correlative study participation. We compared participation rates of United States (US) enrollees for the mutually exclusive racial-ethnic groups of non-Hispanic (NH)-white, NH-Black, NH-Asian, NH-Native American, and Hispanic using X 2 testing, with NH-white as the comparator and reporting odds ratios (OR) and 95% confidence intervals (CI). Rates were adjusted by national incidence according to the Surveillance, Epidemiology, and End Results program and the US Census. Testing was repeated for the 55% of participants enrolled at 15 NCI CCCs recruiting ≥10 patients, where incidence was adjusted for catchment area size and demographics. Logistic regression models, clustered by trial, were performed to assess the following predictors for biobank and correlative study participation among trial enrollees: race-ethnicity (NH-white vs non-white), site type (CCC status), age (10-year increments), sex, neighborhood urbanity (urban vs rural) and poverty ( Results There were 3041 trial enrollees at US sites; participant characteristics and demographics by race-ethnicity are shown in Table 1. 93.9% of patients participated in a biobanking study and 60.0% in a correlative study. National incidence adjusted enrollment odds by race-ethnicity are shown in the Figure (top); NH-Black, NH-Asian, and Hispanic persons were enrolled at significantly lower rates than NH-whites; NH-Native American enrollment was significantly higher. Enrollment odds were even lower for NH-Black, NH-Asian, and Hispanic enrollees at CCC sites when adjusted by catchment area incidence (Figure; bottom). Among trial enrollees, there were no univariable predictors of biobank participation, however, male sex (OR 1.12; 95% CI 1.01, 1.37; p=0.04) and NH-white race-ethnicity (OR 1.33; 95% CI 1.12,1.57; p Conclusions Across 15 years of AML cooperative group studies, there were several enrollment disparities by race-ethnicity, which were more pronounced at CCC sites. Over 90% of trial enrollees participated in biobanking, with no race-ethnic differences seen. However, correlative study participation among trial enrollees was higher for NH-whites. Taken together, these data suggest that efforts should focus on increasing trial and correlative study participant diversity-but not biobanking-within NCI-designated CCC catchment areas. Reasonable next steps include identifying key structural, provider, and patient-based barriers to trial and correlative study participation at CCC sites and developing inclusive, multilevel interventions to address them. Figure 1 Figure 1. Disclosures Eisfeld: Karyopharm (spouse): Current Employment. Stock: Pfizer: Consultancy, Honoraria, Research Funding; amgen: Honoraria; agios: Honoraria; jazz: Honoraria; kura: Honoraria; kite: Honoraria; morphosys: Honoraria; servier: Honoraria; syndax: Consultancy, Honoraria; Pluristeem: Consultancy, Honoraria. Stone: AbbVie Inc, Actinium Pharmaceuticals Inc, Aprea Therapeutics, BerGenBio ASA, ElevateBio, Foghorn Therapeutics, GEMoaB, GlaxoSmithKline, Innate Pharma, Syndax Pharmaceuticals Inc, Syros Pharmaceuticals Inc, Takeda Oncology: Other: Advisory Committee; Agios Pharmaceuticals Inc, Novartis;: Research Funding; ACI Clinical, Syntrix Pharmaceuticals, Takeda Oncology: Other: Data Safety & Monitoring. DeAngelo: Abbvie: Research Funding; Blueprint: Research Funding; Takeda: Consultancy; Autolus: Consultancy; Forty-Seven: Consultancy; Incyte: Consultancy; Jazz: Consultancy; Novartis: Consultancy, Research Funding; Pfizer: Consultancy; Servier: Consultancy; Amgen: Consultancy; Agios: Consultancy; Glycomimetrics: Research Funding. Byrd: Newave: Membership on an entity's Board of Directors or advisory committees; Vincerx Pharmaceuticals: Current equity holder in publicly-traded company, Membership on an entity's Board of Directors or advisory committees; Novartis, Trillium, Astellas, AstraZeneca, Pharmacyclics, Syndax: Consultancy, Honoraria.

Introduction Disparities in clinical trial enrollment have the potential to bias findings, limit generalizability, misdirect drug development, and reduce equitable access to novel therapy. Given the unique care delivery and clinical trial patterns of acute leukemia (AML), barriers to trial enrollment for underrepresented groups at the Comprehensive Cancer Centers (CCCs), where patients are often treated, may be distinct. Characterization of these barriers has been limited, preventing the development of interventions to overcome this disparity for AML. Methods We conducted an exploratory qualitative project to characterize barriers to trial enrollment and facilitators to overcome them through a series of focus groups and individual interviews. Participants were persons with AML who identified from one or more underrepresented groups and AML physicians, nurse practitioners, and physician assistants at three Dana-Farber/Harvard Cancer Center hospitals. A moderator guide was developed based on literature review and iteratively revised by study investigators. The questions enabled an open-ended multi-level (patient, provider, trial, institution, societal) exploration of five main domains, including familiarity and experience with clinical trials, factors influencing participation, participant concerns and enrollment supports, and facilitators to address enrollment barriers and disparities. Focus groups and interviews occurred through video conference or phone and were audio recorded and professionally transcribed. The interdisciplinary research team conducted a multi-step thematic analysis guided by framework analysis, and included prefigured and emergent codes. The team-based analysis synthesized data within and across participant types and focused on the identification of key patterns and prevalent themes related to barriers to clinical trial enrollment. Coding and analysis were assisted by NVivo 12 software. Results presented below highlight barriers that were considered relevant to the unique care delivery patterns of AML and clinically significant. Results 37 participants (19 patients; 18 providers) were recruited to focus groups (30) or interviews (7) between 5/6-6/28/2021. Participant characteristics are shown in Table 1. Patient participants were generally familiar with clinical trials. Despite limited explicit expression of enrollment hesitancy, both patients and providers noted multiple potential barriers to clinical trial enrollment specific to AML, including the rapidity from diagnosis to treatment; the inpatient setting of therapy; a lack of knowledge regarding incidence and enrollment demographics and their differences; restrictive trial design factors (e.g., hospital stays, mutation specificity); and concerns over the appropriateness and efficacy of a given trial relative to individual disease risk. Key enrollment barriers reported by participants are shown in Figure 1. Participant-identified facilitators to address these barriers are shown in Figure 2. Facilitators to overcome barriers specific to AML that were considered potentially impactful included: peri-enrollment peer support for patients, continuance of telehealth visits during trial participation, feedback and training to increase provider awareness and understanding of disparities and diverse communities, trial design augmentation to minimize restrictive barriers (centralized laboratory draws, multiple hospitalizations, white blood cell count and organ function criteria) and increase inclusivity (multiple language and culturally competent consent), and bolstering inter-institutional AML program partnerships to increase diverse patient referral. Conclusions Stakeholder-reported barriers to diverse AML clinical trial enrollment at CCCs include multiple disease-specific characteristics that reflect the unique care pattern of AML, such as concerns about restrictive trial design characteristics, the need to be inpatient, and the requirement to make decisions quickly after diagnosis. Several barriers and potential interventions at each level were considered impactful by both providers and patients. These formative data suggest a novel multi-level approach for overcoming AML enrollment disparities at CCCs is needed. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

INTRODUCTION: Since 2015, there have been between 10-20 unique chemotherapeutic shortages per year in the United States. Published ethical frameworks consider the following principles as potentially just strategies to allocate chemotherapy during shortages: prioritizing the youngest, those whose disease is most responsive to the scarce medication by volume of drug, those with worse alternatives, or those who have waited longest (Persad, Lancet 2009). To evaluate the impact of these allocation strategies on patient survival during a chemotherapy shortage, we developed a simulation model through the use-case of vincristine, a widely prescribed blood cancer chemotherapeutic recently in shortage. METHODS: We developed a model to recapitulate scarce chemotherapy allocation at the hospital level. Simulated patients enter the model when they require treatment and are ordered in a queue to receive the scarce chemotherapy-or their best alternative regimen-according to ethically-accepted strategies that prioritize by (1) youngest age, (2) greater efficacy per volume, (3) worse alternatives, or (4) first-come, first-served (the default strategy). For the vincristine use-case, demographic, disease, and treatment data were abstracted from 1689 patients treated at Dana-Farber Cancer Institute from 2015-2019; 3-year treatment regimen survival probabilities and disease risk-adjustments were abstracted from publications cited in the National Comprehensive Cancer Network guidelines. Modeled survival outcomes were validated against Surveillance, Epidemiology, and End-Results Program (SEER) data. Based on the length of the recent vincristine shortage, mean survival rates for a 9-month scenario were modeled according to individual and combination allocation strategies across varying supply levels. RESULTS: Model functions and risk-adjusted survival probabilities demonstrated no significant differences between the cohort and SEER data, respectively. During the 9-month shortage scenario, a strategy that prioritized patients by greater efficacy per volume significantly increased the mean number of patients surviving at 3-years by >5% across 34.8% of possible vincristine supply levels (grey line and shaded region in Figure) compared to the default of first-come, first-served (red line; all p5% differences in the mean number of patients surviving across 56.7% of possible supply levels (all p CONCLUSIONS: During a simulated vincristine shortage, a strategy that allocated by both greater efficacy per volume and worse alternatives improved survival compared to either alone, younger age, or the default of first-come, first-served. Moreover, this combined strategy ameliorated reductions in survival across a larger range of drug scarcity than the default. Such a model can be adapted for use for future chemotherapy shortages when multiple ethical allocation strategies exist. Figure Disclosures No relevant conflicts of interest to declare.

Background Medication shortages prevent patients from receiving optimal care. Despite the frequency with which medical trainees care for inpatients, no assessment of their experiences in medication shortage management has been performed. Objective We evaluated trainees' experiences managing medication shortages. Methods We performed a cross-sectional survey of trainees postgraduate year 2 (PGY-2) and above in medicine, anesthesiology, and emergency medicine departments at 2 academic centers in 2018–2019. Categorical and ordinal assessments evaluated shortage awareness, substitution availability, pharmacy and therapeutics committee-based restrictions, communication, and education. Regressions were performed to determine effect of PGY, department, and institution on responses. Results A total of 168 of 273 subjects completed the survey (62% response rate). Most (95%, 159 of 168) reported managing medication shortages during training; 51% (86 of 168) described managing clinically relevant shortages daily or weekly. Seventy-seven percent (129 of 168) noted equivalent alternatives were unavailable at least one-quarter of the time, and 43% (72 of 168) reported clinically necessary medications were restricted at least weekly. Fifty-four percent (89 of 168) and 64% (106 of 167) of respondents discussed clinically relevant shortages with supervising physicians or patients “some of the time” or less, respectively. Most respondents (90%, 151 of 168) reported they would benefit from shortage management training, but few (13%, 21 of 168) reported prior training. Conclusions Although trainees reported frequent involvement in clinically impactful shortage management, medication shortage communication between trainees and supervising physicians or patients appears sporadic. Medication shortage management training is uncommon but perceived as beneficial.

6523 Background: Racial and ethnic disparities in clinical trial enrollment compound inequities in drug development and the delivery of patient-centered care. Despite significant survival disparities in acute myeloid leukemia (AML), enrollment disparities data are limited. Methods: We performed a structured search and abstraction of demographic data for all United States (US) AML clinical trials from 2002-2017 listed on clinicaltrials.gov and compared the results to the incidence and demographic distribution of AML using the Surveillance, Epidemiology, and End Results program and 2010 US Census. We calculated enrollment fractions (the number of enrollees divided by the number of incident cases) for the five mutually exclusive race/ethnicity groups of non-Hispanic White (NH-White), Black (NH-Black), Asian/Pacific Islander (NH-Asian/PI), American Indian/Native Alaskan (NH-AI/AN), and Hispanic patients. We compared these using X2 testing, with NH-White as the comparator, and reported odds ratios with 95% confidence intervals (CI). To assess trends over time, we adjusted enrollment from 2005-2008 for changes in AML incidence and NH-White enrollment for a later period (2011-2014), comparing this expected enrollment fraction to the actual enrollment fraction during that later period. Results: Of 223 eligible studies (patient N=17372) on clinicaltrials.gov, 99 (44.4%) reported racial demographics (N=8417; 48.5%) and 68 (30.5%) reported race and ethnicity (N=6554; 37.7%). Enrollment and incidence proportions by race are shown in the table. Among trials reporting race and ethnicity, all groups had lower odds of enrollment compared to NH-White patients (Table). For the 99 trials reporting race data, Black and AI/AN patient enrollment odds were lower (OR 0.60 [95% CI: 0.55, 0.65]; 0.50 [95% CI: 0.33, 0.76]), but Asian/PI enrollment was not (OR 0.91 [95% CI: 0.82, 1.01]). The relative enrollment of NH-Black, NH-Asian/PI, and Hispanic patients declined later in the study period (Table). Conclusions: In AML clinical trials performed in the US from 2002-2017, NH-White patients were enrolled at higher rates compared to other racial and ethnic groups; enrollment diversity declined over time. An important first step to reducing enrollment disparities will be to improve the reporting of demographic enrollment data.[Table: see text]

Improving therapy for relapsed/refractory AML remains a challenge. We performed a retrospective analysis of outcomes following decitabine treatment in 34 patients with relapsed/refractory AML (median age, 62; median Charlson comorbidity score, 6). Decitabine, 20 mg/m2 daily, was given in 5- (25%) or 10-day (75%) cycles. Overall response rate (OR) was 30% with 21% complete remission and 9% partial remission rate. Patients with therapy-related myeloid neoplasm (t-MN) and secondary AML had a significantly higher OR compared to those with de novo AML (70 vs. 30%; p = .02). Median overall survival of all patients was 8.5 months. Median survival in patients with t-MN or secondary AML was 12.4 months compared to 8 months in those with de novo AML (p = .20). Fifteen (44%) patients proceeded to hematopoietic stem cell transplant. These data support using 10-day treatment cycles of decitabine in patients with relapsed/refractory AML, particularly for those with secondary or therapy-related AML.

BackgroundThe molecular determinants of clinical responses to decitabine therapy in patients with acute myeloid leukemia (AML) or myelodysplastic syndromes (MDS) are unclear. MethodsWe enrolled 84 adult patients with AML or MDS in a single-institution trial of decitabine to identify somatic mutations and their relationships to clinical responses. Decitabine was administered at a dose of 20 mg per square meter of body-surface area per day for 10 consecutive days in monthly cycles. We performed enhanced exome or gene-panel sequencing in 67 of these patients and serial sequencing at multiple time points to evaluate patterns of mutation clearance in 54 patients. An extension cohort included 32 additional patients who received decitabine in different protocols. ResultsOf the 116 patients, 53 (46%) had bone marrow blast clearance (

Purpose: Terminal oncology intensive care unit (ICU) hospitalizations are associated with high costs and inferior quality of care. This study identifies and characterizes potentially avoidable terminal admissions of oncology patients to ICUs. Methods: This was a retrospective case series of patients cared for in an academic medical center’s ambulatory oncology practice who died in an ICU during July 1, 2012 to June 30, 2013. An oncologist, intensivist, and hospitalist reviewed each patient’s electronic health record from 3 months preceding terminal hospitalization until death. The primary outcome was the proportion of terminal ICU hospitalizations identified as potentially avoidable by two or more reviewers. Univariate and multivariate analysis were performed to identify characteristics associated with avoidable terminal ICU hospitalizations. Results: Seventy-two patients met inclusion criteria. The majority had solid tumor malignancies (71%), poor performance status (51%), and multiple encounters with the health care system. Despite high-intensity health care utilization, only 25% had documented advance directives. During a 4-day median ICU length of stay, 81% were intubated and 39% had cardiopulmonary resuscitation. Forty-seven percent of these hospitalizations were identified as potentially avoidable. Avoidable hospitalizations were associated with factors including: worse performance status before admission (median 2 v 1; P = .01), worse Charlson comorbidity score (median 8.5 v 7.0, P = .04), reason for hospitalization (P = .006), and number of prior hospitalizations (median 2 v 1; P = .05). Conclusion: Given the high frequency of avoidable terminal ICU hospitalizations, health care leaders should develop strategies to prospectively identify patients at high risk and formulate interventions to improve end-of-life care.

A minor correction to the legend in a figure in this paper has been published and can be accessed via a link at the top of the paper. A Correction to this paper has been published: http://doi-org-443.webvpn.bjmu.tsg211.com/10.1038/s41375-021-01138-z

BACKGROUND: In patients who lack temporary or permanent decisional capacity, substituted judgment by surrogates is necessary. However, the current prevalence of surrogate decision making in the inpatient setting is unknown. The goal of this study was to better quantify how often surrogate decision makers are relied upon in the care of Medicine and Surgery inpatients. METHODS: We performed interviews of health care providers who cared for Medicine and Surgery inpatients on the acute care units and intensive care units (ICUs) of the University of Chicago Medical Center in Chicago, Illinois and Mayo Clinic Hospital in Rochester, Minnesota on a chosen date. We obtained information regarding the number of inpatients who lacked decisional capacity and how many had a surrogate decision maker. RESULTS: At University of Chicago and Mayo Clinic, 15% of Medicine inpatients and 12% of Surgery inpatients lacked decisional capacity. Eighty-five percent of the Medicine inpatients and 75% of the Surgery inpatients that lacked decisional capacity had a surrogate decision maker. A greater proportion of inpatients in the ICU setting (40% of Medicine inpatients and 45% of Surgery inpatients at University of Chicago and 32% of Medicine inpatients and 41% of Surgery inpatients at Mayo Clinic) lacked decisional capacity. CONCLUSIONS: This is the first study to quantify the extent of surrogate decision making in the inpatient setting at two major academic medical centers in the United States. At any given time, up to 15% of inpatients and 40% of ICU patients lack decisional capacity. Healthcare providers should routinely encourage patients to name a surrogate decision maker in the event of unforeseen future circumstances preventing them from making their own medical decisions. FUNDING STATEMENT: None. DECLARATION OF INTERESTS: The authors stated: "No relevant disclosures." ETHICS APPROVAL STATEMENT: The study received Institutional Review Board exemption at both the University of Chicago and Mayo Clinic.

Novel cellular therapy techniques promise to cure many haematology patients refractory to other treatment modalities. These therapies are intensive and require referral to and care from specialised providers. In the USA, this pool of providers is not expanding at a rate necessary to meet expected demand; therefore, access scarcity appears forthcoming and is likely to be widespread. To maintain fair access to these scarce and curative therapies, we must prospectively create a just and practical system to distribute care. In this article, we first review previously implemented medical product and personnel allocation systems, examining their applicability to cellular therapy provider shortages to demonstrate that this problem requires a novel approach. We then present an innovative system for allocating cellular therapy access, which accounts for the constraints of distribution during real-world oncology practice by using a combination of the following principles: (1) maximising life-years per personnel time, (2) youngest and robust first, (3) sickest first, (4) first come/first served and (5) instrumental value. We conclude with justifications for the incorporation of these principles and the omission of others, discuss how access can be distributed using this combination, consider cost and review fundamental factors necessary for the practical implementation and maintenance of this system.

The treatment of chronic phase chronic myeloid leukemia (CP-CML) was revolutionized by the introduction of tyrosine kinase inhibitors (TKIs) against the chimeric fusion protein BCR-ABL1. The introduction of imatinib, the first BCR-ABL1 TKI, changed a usually fatal disease with long-term survival of

Background Treatment of metastatic melanoma patients with immune checkpoint inhibitors is an important standard of care. Side effects are due to immune activation, can affect virtually all organ systems, and are occasionally severe. Although hematologic toxicity has been reported, we present a case of hemophagocytic lymphohistiocytosis (HLH) due to immune checkpoint inhibitor therapy. Case presentation A patient with metastatic melanoma was treated with one course of ipilimumab + nivolumab and presented 3 weeks later with severe anemia and hyperferritinemia. A bone marrow biopsy revealed necrotic tumor cells, infiltrating T cells, and hemophagocytosis. The patient was treated with high-dose steroids; 12 months later, the patient remains off all therapy and in complete remission of both HLH and metastatic melanoma. Conclusions The hemophagocytic syndromes are attributable to dysregulated immune activation and share pathophysiologic mechanisms with immune activation from checkpoint inhibitors. Increasing use of regimens that include immune checkpoint inhibition require vigilant monitoring for immune-activating side effects as they can occasionally be life threatening, as in this case of HLH.

Improved systems for detection of measurable residual disease (MRD) in acute myeloid leukemia (AML) are urgently needed, however attempts to utilize broad-scale next-generation sequencing (NGS) panels to perform multi-gene surveillance in AML post-induction have been stymied by persistent premalignant mutation-bearing clones. We hypothesized that this technology may be more suitable for evaluation of fully engrafted patients following hematopoietic cell transplantation (HCT). To address this question, we developed a hybrid-capture NGS panel utilizing unique molecular identifiers (UMIs) to detect variants at 0.1% VAF or below across 22 genes frequently mutated in myeloid disorders and applied it to a retrospective sample set of blood and bone marrow DNA samples previously evaluated as negative for disease via standard-of-care short tandem repeat (STR)-based engraftment testing and hematopathology analysis in our laboratory. Of 30 patients who demonstrated trackable mutations in the 22 genes at eventual relapse by standard NGS analysis, we were able to definitively detect relapse-associated mutations in 18/30 (60%) at previously disease-negative timepoints collected 20–100 days prior to relapse date. MRD was detected in both bone marrow (15/28, 53.6%) and peripheral blood samples (9/18, 50%), while showing excellent technical specificity in our sample set. We also confirmed the disappearance of all MRD signal with increasing time prior to relapse (>100 days), indicating true clinical specificity, even using genes commonly associated with clonal hematopoiesis of indeterminate potential (CHIP). This study highlights the efficacy of a highly sensitive, NGS panel-based approach to early detection of relapse in AML and supports the clinical validity of extending MRD analysis across many genes in the post-transplant setting.

e18323 Background: Hospital medication shortages (HMS) are pervasive throughout the U.S. healthcare system. Current management mechanisms are heterogeneous but routinely include the use of alternatives and the rationing of medications between patients. Little is known about oncology patient preferences for decision-makers, ethical allocation systems, or thresholds for disclosure during HMS. Methods: Oncology patients previously hospitalized for inpatient care within the last 24 months underwent qualitative interviews supplemented with validated instruments measuring: trust (Trust in Oncologist Scale), therapeutic alliance (Human Connection Scale), and shared decision-making (Shared Decision-Making Questionnaire). Qualitative data underwent Framework Analysis for thematic identification. Results: To date, 16 patients have been interviewed: median age 61y (31-77); 44% female; 56% married; 56% > college education; median number of treatment regimens 2 (1-6), days in hospital 18 (3-66), number of hospitalizations 2(1-8). All patients (100%) reported extremely high levels of trust, therapeutic alliance, and shared decision-making with their oncologist. Two patients (13%) reported personal experiences with HMS, 43% reported knowledge of HMS within the U.S., and no patients reported knowledge of local hospital HMS. Framework Analysis revealed that virtually all patients preferred that their oncologist act as the primary decision-maker during allocation/rationing and favored pharmacist and ethicist involvement. Most patients preferred allocation systems that prioritized efficacy, age, and degree of illness. No patients desired the use of a lottery or reciprocity-based decisions. Virtually all patients favored disclosure of shortages if alternatives were used, independent of the level of difference in efficacy/toxicity, and in the case of both chemotherapeutics and supportive medications. Conclusions: Despite ubiquitous HMS in oncology, patients are generally unaware of local HMS and prefer: multi-disciplinary decision-makers during HMS allocation, prioritized allocation schemes, and more frequent HMS disclosure than presently occurs. Study recruitment is ongoing.

This study examines national survey results from hospital pharmacy managers to investigate current drug allocation and rationing practices of US hospitals during drug shortages.

Summary Data on outcome, prognostic factors, and treatment for very elderly nonHodgkin lymphomas (NHL) is sparse. We conducted a multicentre retrospective analysis of NHL patients ‡80 years (at diagnosis) treated between 1999 and 2009. Detailed characteristics were obtained including geriatric syndromes, activities of daily living (ADLs), and co-morbidities using the Cumulative Illness Rating Scale-Geriatrics (CIRS-G). We identified 303 patients: 170 aggressive NHL (84% B cell/16% T cell) and 133 indolent NHL (82% B cell/18% T cell). Median age was 84 years (80‐95). A geriatric syndrome was present in 26% of patients, 18% had ‡1 grade 4 CIRS-G, and 14% had loss of ADLs. At 49-month median follow-up, 4-year progressionfree (PFS) and overall survival (OS) for aggressive NHLs were 31% and 44% respectively (stage I/II: PFS 53% and OS 66%; stage III/IV: PFS 20% and OS 32%; P

e21510 Background: Advance care planning (ACP) should be initiated early and readdressed often for cancer patients. We hypothesize that a rules-based algorithm based on occurrence of high-risk events can predict decreased overall survival, and can be used to target patients who would benefit from readdressing ACP. Methods: We performed a retrospective analysis of 221 patients receiving palliative chemotherapy with a diagnosis of leukemia, cholangiocarcinoma, esophageal, gastric, pancreatic, lung or urothelial cancer at the University of Chicago Medicine. Patients were included if they had an index outpatient oncology visit from April 1, 2015 through June 30, 2015. Starting at the date of index visit, we examined a three-month window for a “high-risk event,” defined as: 1. change in chemotherapy 2. emergency department visit 3. hospitalization. Patients were followed from index visit until date of death or last clinical encounter as of January 31, 2017. Each “high-risk event” was treated as a time-varying covariate in a Cox proportional hazards regression model to calculate a hazard ratio of death compared to those without an event. Results: Sixty-six percent of patients (146/221) experienced at least one high-risk event over the 3 month time frame. A change in chemotherapy regimen, an ED visit, and a hospitalization occurred in 53% (118/221), 10% (22/221) and 26% (57/221) of patients respectively. The hazard ratio of death for patients with at least one high-risk event when compared to those without was 1.86 (95% CI: 1.26-2.74, p = 0.002), when adjusted for age, gender, and race. Inpatient admission had the highest hazard of death among the high-risk events (HR 2.52: 95% CI: 1.69-3.76, p < 0.001). Conclusions: The rules-based algorithm identified patients with a greater risk of death. Implementation of this algorithm in the electronic medical record can identify patients with increased urgency to readdress goals of care.

6548 Background: ICU admissions in the last 30 days of life are an indicator of poor care. Our prior research found nearly half of terminal oncology ICU hospitalizations are potentially avoidable. Methods: Data were derived from 72 patients consecutively cared for in an academic medical center’s oncology practice who died in an ICU between July 1, 2012 and June 30, 2013. Oncologists, intensivists, and hospitalists used a standardized assessment tool to review each patient’s electronic health record from 3 months prior to hospitalization until death; they made a clinical determination of avoidability. Two investigators, blinded to the specialty, used a grounded theory approach to extract clinical themes associated with the reviewer’s determination of avoidability. Total, direct, and indirect costs were abstracted for each avoidable hospitalization. Results: Thirty-four (47%) of the examined hospitalizations were deemed avoidable. The primary themes associated with avoidability, and the percentage by specialty, were as follows: 1) failure to initiate appropriate advance care planning in the outpatient setting (68% oncologists, 55% intensivists, 65% hospitalists), 2) failure to integrate understanding of limited prognosis (23% oncologists, 24% intensivists, 26% hospitalists), and 3) failure of clinical management (6% oncologists, 21% intensivists, 6% hospitalists). A failure to educate and integrate surrogates into timely medical decision-making was a prominent secondary theme for oncologists (22%), intensivists (18%), and hospitalists (29%). The total cost per patient averaged $44,532 with direct and indirect costs averaging $25,215 and $19,317, respectively. High cost areas were ICU level care (35%) and pharmaceuticals (16%). Conclusions: The themes identified suggest potential preventative interventions, including higher rates of outpatient advance care planning, oncology inpatient communication to promote patient’s prognostic understanding, prevention of failures in clinical management, and better education and integration of surrogates. Given 8% of oncology patients expire in the ICU and 47% were identified as avoidable, the potential national annual cost of these avoidable hospitalizations is $997MM.

6615 Background: FT is an important adverse event (AE) that should be objectively measured in clinical practice. We previously developed an evidence-based FT grading system based on differences in HRQoL, analogous to the NCI-Common Terminology Criteria for Adverse Events (grade 1, mild AE; grade 2, moderate AE; grade 3, severe AE ,de Souza et al - ASCO 2015). We aimed to validate this grading system using a new sample of cancer patients (pts) and report its association with bankruptcy. Methods: FT was assessed by the COST (COmprehensive Score for financial Toxicity) in 2 sets of cancer pts. In the previously reported Development Set (DS), gradations of FT were determined by ROC analyses based on conventions for clinically meaningful small (0.2), medium (0.5) and large (0.8) effect sizes (e.s.) for independent FACT-G differences attributable to FT in pts with Stage IV cancers on chemotherapy. In the Validation Set (VS), differences in HRQoL and the odds ratio for a pt to have declared bankruptcy after the cancer diagnosis were assessed in a larger cohort of cancer pts on chemotherapy. Results: The grading system was developed in 888 cancer pts with cancer (233 pts in the DS and 655 in the VS). In the DS, ROC analyses produced 4 FT grades (G): G0, no FT, COST ≥26 (99 pts, 42%); G1, mild FT: ≥ 14-26 (71 pts, 31%); G2, moderate FT: > 0-14 (58 pts, 25%); and G3, severe FT: COST = 0 (5 pts, 2%). Applying the FT grading to the 655 pts in VS, we had: G0, 146 pts (22%); G1, 281 (43%); G2, 215 (33%); and G3, 13 (2%). The decreases in FACT-G HRQoL measured in e.s. per FT grading in comparison with G0 were small for G1: -0.4 (95%CI: -0.6 – -0.25); large for G2: -0.9 (95%CI: -1.1 – -0.7); and even larger for G3: -1.5 (95%CI: -2.0 – -0.9), all with p < 0.001. In the VS, 23 pts (4%) had declared bankruptcy after their cancer diagnosis. Compared to FT G0, the odds of having declared bankruptcy were 8.6 (95%CI: 1.1 – 67, p = 0.04) times higher for pts with FT G2, and 29 times higher (95% CI: 2.4 – 355, p = 0.008) for those with G3 FT. Conclusions: We developed a FT grading system anchored on independent differences in HRQoL. We applied the system in a different set of cancer pts and it retained its validity. We also found an larger incidence of bankruptcy after the cancer diagnosis in higher grades of FT, adding to the grading’s meaningful use.

101 Background: ICU admissions in the last 30 days of life is a quality measure endorsed by the National Quality Forum. Our prior research has demonstrated that nearly half of terminal oncology ICU hospitalizations are potentially avoidable. Methods: This was a retrospective care series of patients cared for in an academic medical center’s ambulatory oncology practice who died in an ICU during July 1, 2012 to June 30, 2013. Using a standardized assessment tool, an oncologist, intensivist, and hospitalist reviewed each patient’s electronic health record from 3 months prior to hospitalization until death and made a clinical determination of avoidability. Two investigators, blinded to the specialty of the reviewer, used a grounded theory approach to extract clinical themes associated with avoidability from the reviewers’ assessments. Results: The primary themes for avoidability identified and percent by specialty were as follows: failure to initiate appropriate advance care planning in the outpatient setting (68% oncologists, 55% intensivists, 65% hospitalists), failure to integrate understanding of limited prognosis from underlying cancer within the context of acute hospitalization (23% oncologists, 24% intensivists, 26% hospitalists), failure of clinical management (6% oncologists, 21% intensivists, 6% hospitalists), failure to recognize futility of outside hospital transfer (3% oncologists, 0% intensivists, 0% hospitalists), and failure of care coordination (0% oncologists, 0% intensivists, 3% hospitalists). A failure to educate and integrate surrogates into timely medical decision making was a prominent secondary theme for oncologists (22%), intensivists (18%), and hospitalists (29%). Conclusions: The themes identified suggest potential interventions to prevent avoidable terminal oncology ICU hospitalizations, including improved advance care planning in the outpatient setting, inpatient multidisciplinary communication to gain a better understanding of the patient’s underlying malignancy within the context of the acute hospitalization and prevent failures in clinical management, and better education and integration of surrogates in medical decision making.

A World Health Organization advisory panel recently concluded that it is ethical to use experimental medications and vaccines that have not been formally approved or tested in humans to treat perso...

Recently a mother-daughter pair with life-long thrombocytopenia presented for consultation regarding their presumed diagnoses of autoimmune-based idiopathic thrombocytopenic purpura (ITP) and their extended family history of bleeding and myeloid malignancies (Figure 1A). The proband/mother (III:4) had been diagnosed with myelodysplastic syndrome (MDS), and the daughter (IV:3) was 32 weeks pregnant. A family history revealed 10 of 28 family members with thrombocytopenia, 4 of whom also had MDS/acute myeloid leukemia (AML). Consideration was given to known alleles associated with congenital thrombocytopenia with predisposition to MDS/AML. Among RUNX1, GATA2, and CEPBA, three genes in which germline mutations predispose to myeloid malignancies, only familial platelet disorder (FPD)/germline RUNX1 mutation is associated with thrombocytopenia and platelet dysfunction.1 Recently, however, mutations within the 5′ untranslated region (UTR) of ANKRD26 on chromosome 10p12 have been associated with Thrombocytopenia 2 (THC2), an autosomal-dominant congenital thrombocytopenia, and in one series a 30-fold increase in the frequency of MDS/AML. 2–5

56 Background: Eight percent of cancer patients expire in the ICU. There is no association between spending on aggressive end of life (EOL) care and survival outcomes. This aggressive care is also associated with reduced quality of life for patients and families. The National Quality Forum endorses the number of patients admitted to the ICU in the last 30 days of life as a quality of care measure. Our hypothesis is that a significant number of oncology patients suffer avoidable terminal ICU hospitalizations. Methods: Using data from the University of Chicago (UCM) Cancer Registry, we identified patients who died in UCM adult ICU’s in FY2013. Of the 1,388 oncology deaths, 115 were in the ICU. Of those, 72 were established patients having had at least one visit with a UCM oncologist. A physician from oncology, critical care, and hospital medicine directly reviewed the electronic medical record of each patient from 3 months prior to index hospitalization until death. Using a standardized assessment tool, these physicians then determined whether the terminal hospitalization was clinically avoidable through different medical management. The primary outcome was the proportion of terminal oncology ICU hospitalizations identified as potentially avoidable by two or more reviewers. Results: Seventy-one percent of this patient population had a solid malignancy and 53% of those had metastatic disease. The ECOG performance status was ≥ 2 prior to admission in 51% of patients. Eighty-two percent had at least one prior hospitalization. During the index hospitalization, 81% were intubated, 39% had resuscitation, and 22% had hemodialysis. Two or more physician reviewers identified 34 (47%) of terminal oncology ICU hospitalizations as clinically avoidable. All three specialty physicians agreed about the avoidability 43% of the time. Conclusions: A review of terminal oncology ICU hospitalizations demonstrates that most of these patients had an advanced malignancy and poor clinical status and underwent aggressive EOL interventions. A significant number of these hospitalizations were identified as clinically avoidable by two or more physician reviewers. By identifying avoidable terminal oncology ICU hospitalizations, we can design interventions to prevent them.

Introduction Improving therapy for rel/ref AML remains a challenge. Decitabine, a DNA methyl-transferase inhibitor, initially showed promise in AML as a 5-day, first-line induction regimen and more recently as a 10-day regimen in older and unfit patients (1). However, little is known about the activity of decitabine in the rel/ref patient population despite increased use. Therefore, we sought to analyze the outcomes of these pts treated at our institution. Methods To obtain data regarding decitabine efficacy in rel/ref AML, we performed a retrospective analysis of outcomes following decitabine treatment in 34 adult pts treated at The University of Chicago from January 2009 to June 2014. Permission to access patient charts was granted by the medical centerÕs Institutional Review Board. AML was defined by WHO criteria, genetic risk grouping and complete remission (CR) was according to ELN classification; PR was defined as >50% decrease in bone marrow blasts and normalization of blood counts. Rel/ref AML was defined as either having had a prior CR with recurrence of disease or having received a prior induction regimen (1-2 cycles) without CR. Results Median pt age was 62 yrs (range, 18-81) and 60% were male. Median Charlson comorbidity index (CCI) was 5 (range, 0-8); 29% had ECOG performance status 0-1 and 71% had >2. 21 pts (62%) had de novo AML (7 with myelodysplasia-related changes), 3 (9%) had therapy-related myeloid neoplasm (t-MN), and 10 (29%) had secondary AML after myelodysplastic syndrome. 6% were in the ELN favorable genetic group, 3% intermediate-I, 18% intermediate-II, and 67% adverse; 2 cases were unevaluable. The median number of prior treatment regimens was three. 9% had received prior azacitidine, 85% had received prior HiDAC, and 38% had a prior allogeneic stem cell transplant (SCT). 34 pts received a total of 71 cycles of decitabine, 20 mg/m2 daily, in 5 or 10-day cycles every 28 days. All patients received 10-day courses, 91% had an initial 10-day course, and 74% had only 10-day courses. The median number of cycles per pt was 2; 59% received >1 cycle. 7 (21%) achieved CR and 4 (12%) had a partial response (PR), for an overall response rate (OR) of 33%. Responses occurred in 24% of pts with de novo AML, 66% with t-MN, and 50% with secondary AML. Intermediate and adverse group pts had OR of 14% and 39%, respectively. All pts achieving CR did so after 1 cycle; PR required a median of 3 cycles. Pts who achieved CR or PR had a significantly lower pretreatment WBC count (median, 9.5 vs 49.5 x 103/µL in non-responders; p=0.015) and blast percentage (44 vs 59.4; p=0.035) than those who did not. Pts with secondary AML or t-MN had a higher probability of OR compared to those with de novo AML (54 vs 23%; p=0.042). Median overall survival (OS) of all pts was 256 days; prior SCT was associated with reduced OS (p=0.017). When comparing de novo to secondary AML & t-MN, 1-year OS was not significantly different (Figure 1). Responders had a significantly longer OS (median, 622 days vs 278 days for non-responders; p=0.012). Age, race, CCI, ECOG PS, genetic risk group, prior HiDAC, dysplasia, azacitidine, and number of prior treatments did not impact OR or OS. 16 (47%) pts proceeded to SCT. During treatment, 70% had a grade 3-4 non-hematologic toxicity (based on NCI CTACE v4.0); the most common was fatigue. The median number of hospitalizations for complications per patient was 2 (range, 0-7). Causes of hospitalization were febrile neutropenia (40%), infection (22%), cytopenias (18%), rash (6%), acute kidney injury (6%), and 8% were for other causes. Conclusion Decitabine treatment of 34 adults with rel/ref AML resulted in an OR of 33% (21% CR) and allowed nearly one-half of these pts to proceed to SCT. All pts achieving CR did so after 1 cycle. Responding pts had improved OS over those without response (p=0.012). Interestingly, secondary AML or t-MN were 7.8 times more likely to achieve a response compared to de novo AML (p=0.046); lower WBC count and marrow blast percentage also correlated with higher OR. Further delineation of molecular subsets associated with response to decitabine should be evaluated in a larger prospective trial in this high-risk AML population. Citation 1. Blum KA, et al. Phase I trial of low dose decitabine targeting DNA hypermethylation in patients with chronic lymphocytic leukaemia and non-Hodgkin lymphoma: dose-limiting myelosuppression without evidence of DNA hypomethylation. Br J of Haem. Jul 2010;150(2):189-195. Figure 1. Figure 1. Disclosures Off Label Use: Decitabine is indicated for treatment of MDS but is often used to treat newly diagnosed or relapsed/refractory AML. In this study we analyzed results of patients with AML who were treated with decitabine in the relapsed/refractory setting.. Thirman:AbbVie: Research Funding; Pharmacyclics LLC, an AbbVie Company: Research Funding; Gilead: Research Funding; Merck: Research Funding; AbbVie: Research Funding; Gilead: Research Funding; Merck: Research Funding. Odenike:Sunesis: Membership on an entity's Board of Directors or advisory committees, Research Funding. Liu:Astra Zeneca/Medimmune: Consultancy; Pfizer: Consultancy; Astra Zeneca/Medimmune: Consultancy; Pfizer: Consultancy. Stock:Gilead: Membership on an entity's Board of Directors or advisory committees.

167 Background: Twenty-five percent of Medicare cancer beneficiaries use the ICU in the last month of life, and 8% of cancer patients expire there. Cancer patients who die in the ICU have worse quality of life compared with those who die at home. Terminal ICU hospitalizations also come at high financial cost, accounting for 80.2% of all terminal hospitalization costs. Our hypothesis is that a significant number of oncology patients suffer avoidable terminal ICU hospitalizations. By better characterizing these hospitalizations, we will provide the knowledge backbone for future interventions. Methods: Using data from the University of Chicago (UC) cancer registry, we identified patients who died in UC adult ICU’s in FY2013. Of the 1,388 oncology deaths, 115 were in the ICU. Of those, 73 were established patients having had at least one visit with a UC oncologist. We performed a chart abstraction to identify patient, clinical, and hospitalization characteristics. Results: The average age of this patient population was 64 years and 71% were men. Solid and hematologic malignancies comprised 71% and 29%, respectively. Of the solid malignancy patients, 54% had metastatic disease. The average duration of the patient-oncologist relationship was 693 days, and 77% had their last oncology appointment within 2 months of index hospitalization. The ECOG performance status was ≥ 2 prior to admission in 51% of patients. These patients had an average of 3 hospitalizations and 15 outpatient visits in the year prior to death. Eighty-one percent did not have advance directives and 86% did not have an outpatient palliative care consult. The average duration in the ICU was 8 days, during which 95% did not have an inpatient palliative care consult, 79% were intubated, and 38% had resuscitation attempted. Conclusions: Most cancer patients who died in the ICU had an advanced stage malignancy and/or poor clinical status prior to admission. In addition, a significant number received intubation and resuscitation. Despite evidence of adequate interaction with the healthcare system, most patients had not had a palliative care consult or advance directives. This study will allow for further exploration of the avoidability of terminal oncology ICU hospitalizations to inform future interventions.

8042 Background: Ages > 80 represent the most rapidly growing segment of society. Incidence of NHL rises exponentially with age; however, sparse data exist among the > 80 pt group. Methods: We comp...