Search

TPS8606 Background: Current standard of care immunotherapy plus chemotherapy options for first-line extensive-stage small-cell lung cancer (ES-SCLC) are associated with modest improvements in median OS and PFS. In the KEYNOTE-604 study, first-line pembrolizumab plus etoposide/platinum (EP) significantly improved PFS (HR 0.75; 95% CI, 0.61‒0.91; P = 0.0023) compared with placebo plus EP in ES-SCLC; OS was also longer with pembrolizumab plus EP vs placebo plus EP but did not reach statistical significance (HR 0.80; 95% CI, 0.64‒0.98; P = 0.0164). Preclinical and clinical data suggest that blocking the interaction between the T-cell immunoreceptor with immunoglobulin and ITIM domains (TIGIT) and its ligands CD112 and CD155 with the anti-TIGIT humanized monoclonal antibody vibostolimab (MK-7684) yields promising antitumor activity when combined with pembrolizumab, with or without chemotherapy, including in patients with lung cancer. The current phase 3 study, KeyVibe-008 (NCT05224141), is comparing the efficacy and safety of first-line treatment with MK-7684A, a co-formulation of vibostolimab plus pembrolizumab, in combination with EP vs atezolizumab plus EP in patients with ES-SCLC. Methods: This multicenter, randomized, double-blind, phase 3 study is enrolling patients aged ≥18 years with histologically or cytologically confirmed, previously untreated ES-SCLC. Patients must have measurable disease per RECIST v1.1; ECOG PS of 0 or 1; no active CNS metastases/carcinomatous meningitis, autoimmune disease, neurologic paraneoplastic syndromes, pneumonitis, or interstitial lung disease; and must provide a pretreatment tumor sample. Patients are randomized 1:1 to receive up to 4 cycles of EP (cisplatin or carboplatin) in combination with MK-7684A (vibostolimab 200 mg + pembrolizumab 200 mg) Q3W or atezolizumab (1200 mg) Q3W, followed by MK-7684A or atezolizumab, respectively, until disease progression, unacceptable AEs, intercurrent illness, protocol violation, or investigator/patient decision. Randomization is stratified by ECOG PS (0 vs 1), LDH (≤ULN vs > ULN), liver metastases (yes vs no), and brain metastases (yes vs no). The primary endpoint is OS. Secondary endpoints include PFS, ORR, and duration of response per RECIST v1.1 by blinded independent central review; safety; and patient-reported outcomes (PROs). Tumor imaging occurs at baseline, every 6 weeks until 48 weeks, and every 9 weeks thereafter until disease progression, start of new anticancer treatment, withdrawal of consent, or death. PROs are assessed using validated instruments including the EORTC quality of life and EuroQol questionnaires. AEs are graded according to NCI CTCAE v5.0. Enrollment is ongoing worldwide. Clinical trial information: NCT05224141.

2578 Background: Combination strategies to improve the efficacy of single agent immune checkpoint inhibitors (ICIs) are increasingly being explored, with one strategy being the addition of vascular endothelial growth factor (VEGF) inhibition. Having shown promise in the treatment of hepatocellular carcinoma and renal cell carcinoma, NCT03583086 is a multi-institutional, phase I/II study of combination vorolanib and nivolumab in both naïve and refractory thoracic tumors that progressed on at least one prior line of platinum-based chemotherapy. Though structurally similar to the tyrosine kinase inhibitor, sunitinib, vorolanib was designed to have a more favorable safety profile with comparable efficacy. Here we present safety data across all cohorts and interim efficacy analyses of the SCLC and NSCLC with primary resistance to ICI-based therapy cohorts, both of which have now completed enrolment. Methods: The maximum tolerated dose determined in phase I was vorolanib 200mg daily and nivolumab 240mg q2 weeks. Phase II uses a two-stage MinMax design across 5 cohorts with objective response rate (ORR) as the primary endpoint: NSCLC (ICI naïve, primary refractory, and acquired resistance), SCLC, and thymic carcinoma. Primary refractory is defined as radiographic progression of disease within 12 weeks of ICI initiation. Results: As of January 2021, 75 patients have been enrolled across all cohorts. Stage 1 of the SCLC and primary refractory NSCLC cohorts have completed accrual at 18 and 15 patients, respectively. In the SCLC cohort, disease-control rate (DCR) was 7% and no objective responses were achieved among 14 evaluable patients. In the primary refractory NSCLC cohort, DCR was 57% and ORR 7% (1 partial response) among 14 evaluable patients. A total of 140 treatment-related adverse events (TRAEs) have been reported, 13 (9%) were grade 3 and there were no grade 4/5 events. Fatigue (9%), nausea (6%), diarrhea (6%), ALT elevation (5%), and AST elevation (5%) were the most common all grade TRAEs. The most common grade 3 TRAEs were ALT elevation and hypertension. Conclusions: This therapeutic strategy of nivolumab plus vorolanib appears to be a well-tolerated combination with a manageable safety profile. Adding VEGF inhibition may offer additional disease control in the setting of NSCLC with primary resistance to ICIs, but neither the SCLC or primary refractory NSCLC cohorts achieved the pre-determined target number of objective responses for progression to stage 2 of the study. Enrolment in the other 3 cohorts as well as exploratory correlatives are ongoing. Clinical trial information: NCT03583086.

6551 Background: Immune cell (IC) infiltrates in primary tumors (PT) have been identified as prognostic markers in head and neck squamous cell carcinoma (HNSCC). IC densities may differ among PT, lymph node metastasis (LNM) and recurrent disease (RD) and by primary disease site (oral cavity- OC, oropharynx- OP, hypopharynx- HP, larynx- L). Here, we compare CD3 and CD20 IC densities in PT, LNM and RD in paired samples from different disease sites and determine the prognostic impact of IC infiltrates. Methods: Tissue microarrays with 425 PT, 198 LNM and 46 RD samples--each in triplicate--were stained immunohistochemically for CD3 and CD20 in the same slide. Immune cell densities per mm2 were determined using a digital image analysis software (QuPath). Individual means were calculated from triplicates of each sample. IC infiltrates from different sample types (PT, LNM, RD) and primary tumor sites were compared using Kruskal-Wallis and Mann-Whitney-U tests. Paired samples were compared using Wilcoxon signed rank test. IC densities were classified as CD3 high/low and CD20 high/low for each primary tumor site using the individual median as a cut-off. Overall survival (OS) was calculated using the Kaplan-Meier method. P-values for each hypothesis were corrected using a false discovery rate of 5%. Results: CD3 and CD20 IC densities differed significantly by sample type (both p0.05). CD3 densities were significantly higher than CD20 densities in all sample types. CD3high patients had the best prognosis in all sites except for OC (qlow/CD20high patients had the worst OS compared to CD3low/CD20low and CD3high/CD20high or even CD3high/CD20low patients (p=0.018) who had the best prognosis. Conclusions: IC densities of CD3 and CD20 vary by sample type and primary site. Except for OC, in all sites the prognostic impact is determined by CD3high, whereas in OSCC only the combination of CD3 and CD20 IC densities achieves a good prognostic value. Interestingly, CD3low/CD20high patients have the worst overall survival in OC patients. Further work is needed to understand the interaction of B- and T cell infiltrates in the tumor, especially in OC.

e20595 Background: Immune checkpoint inhibitors (ICIs) have changed the landscape of NSCLC therapy. In previous study, gender, smoking history, and PD-L1 status were found to influence the efficacy of ICI monotherapy in NSCLC. This meta-analysis evaluated the clinical and molecular factors predictive of benefit from the addition of ICIs to first-line (1L) chemotherapy in metastatic NSCLC. Methods: Using the random effect model, we computed and compared the pooled hazard ratio (HR) of progression-free survival (PFS) and overall survival (OS) among the selected subgroups. The correlation between PD-L1 expression level and outcome was analyzed by meta-regression. Results: 7 phase III randomized trials comparing CIT vs. chemo in the 1L treatment of stage 4 NSCLC were included. CIT evenly improved the PFS irrespective of age, gender, histology, smoking history and ECOG performance status (PS). While patients (pts) with liver metastasis or ALK/EGFR aberration had greater PFS with the addition of ICI to Bevacizumab (BEV) based chemo regimen, the derived benefit was no longer statistically significant among pts treated with non-BEV-based regimens. Whereas the PFS superiority conferred by CIT was noticed across all PD-L1 expression subgroups, the benefit correlated with PD-L1 level and was more pronounced in the “PD-L1 high” subset. Except for pts with EGFR/ALK abnormalities, squamous histology or liver metastasis, CIT yielded a consistent amelioration of OS across the other selected subgroups. Conclusions: Survival advantage associated with 1L CIT in metastatic NSCLC was observed in different pts population including those for which single-agent ICI has limited therapeutic benefit. Our findings support the role of chemo +/- VEGF blockade as enhancer of ICIs activity even in “less immunogenic” context. [Table: see text]

e21121Background: Monoclonal antibodies targeting the PD-1/PD-L1 pathway have demonstrated significant activity in the treatment of metastatic non–small cell lung cancer (NSCLC). However, the major...

Purpose To evaluate the effectiveness of donepezil compared with placebo in cancer patients with fatigue as measured by the Functional Assessment for Chronic Illness Therapy–Fatigue (FACIT-F). Patients and Methods Patients with fatigue score ≥ 4 on a scale of 0 to 10 (0 = no fatigue, 10 = worst possible fatigue) for more than 1 week were included. Patients were randomly assigned to receive donepezil 5 mg or placebo orally every morning for 7 days. A research nurse contacted the patients by telephone daily to assess toxicity and fatigue level. All patients were offered open-label donepezil during the second week. FACIT-F and/or the Edmonton Symptom Assessment System (ESAS) were assessed at baseline, and days 8, 11, and 15. The FACIT-F fatigue subscale score on day 8 was considered the primary end point. Results Of 142 patients randomly assigned to treatment, 47 patients in the donepezil group and 56 in the placebo group were assessable for final analysis. Fatigue intensity improved significantly on day 8 in both donepezil and placebo groups. However, there was no significant difference in fatigue improvement by FACIT-F (P = .57) or ESAS (P = .18) between groups. In the open-label phase, fatigue intensity continued to be low as compared with baseline. No significant toxicities were observed. Conclusion Donepezil was not significantly superior to placebo in the treatment of cancer-related fatigue.

9021 Background: To better understand outcome heterogeneity in pts with KRAS mutant lung cancers, we analyzed the largest multi-institutional database of pts with metastatic KRAS mutant lung adenocarcinomas. Methods: We reviewed data of all pts who consented to LCMC between 2009-2015. Pts with known KRAS status were included in analyses. Mutation data along with co-mutations were obtained along with clinical outcomes. We evaluated baseline characteristics and association of KRAS data with overall survival (OS), calculated from date of distant metastasis to death, in a univariate and multivariable analyses. The median follow-up was 2.15 years (95% CI: 2.01-2.27). Results: 1655 (86%) of 1918 pts’ data were analyzed. Comparative characteristics are summarized in the Table. Among 450 (23%) pts with KRAS mutations: 58% female, 93% ever smokers, and median age 65 years. Main KRAS subtypes: G12C 39%, G12D and G12V were 18% each. Never smokers with KRAS mutation were more likely to have G12D subtypes (18%; p < 0.001). Pts with KRAS mutation, the median OS was 1.96 years, with 2-year OS rate of 49%. Co-mutations (1-16) were checked in all KRAS mutant pts. Co-mutations were rare (14 pts; 3%). KRAS co-mutations were associated with improved OS in multivariable analyses (HR 0.35; 95% CI: 0.13-0.97; p = 0.04), but not KRAS main subtypes / codons. Co-mutation STK-11 in particular (17 of 92 pts; 18%) was associated with poor OS in univariate (HR 2.16; 95% CI: 1.03-4.54; p = 0.04) and multivariable analyses (HR 2.31; 95% CI: 1.18-5.50; p = 0.02). Pts with KRAS mutations had a trend towards a shorter survival (median OS 1.96 vs. 2.22 years; p = 0.08) and a decreased 2-year OS when compared to KRAS wildtype [49% (95% CI: 44-54%) vs. 55% (95% CI: 52-58%)], respectively. Conclusions: KRAS mutation is a significant predictor of worse survival outcomes in pts with metastatic lung adenocarcinomas. The presence of STK-11 co-mutation was associated with especially poor OS. [Table: see text]

1575 Background: Lung cancer is the leading cause of cancer-related mortality linked with smoking, though only 6-18% of heavy smokers die of lung cancer. We hypothesized that major stressful life events are a risk factor for developing lung cancer. Methods: In our matched case-control study, cases (CA) were lung cancer patients diagnosed within past 12 months. Controls (CO) were patients without a prior history of malignancy. CA and CO were matched for age, gender and smoking status. Smokers had at least 10 packs/years history of smoking. Data was collected using standardized research questionnaire on 11 major stressful life events using Holmes and Rahe stress scale. The primary endpoint was odds of having a major stressful life event. A sample of 360 patients (120 CA and 240 CO), was needed to achieve 80% power to detect an odds ratio (OR) of 2.00 using Chi-Square test with a P = 0.05 significance. The study was IRB approved at each institution. Results: Between May 2015 and December 2016, 324 patients were enrolled (23 were excluded due to prior cancer history or incomplete information). 301 (CA = 102; CO = 199) were included in the final analysis. The two groups were well matched in median age (CA = 64.4 years; CO = 63.9years), gender (CA-Male = 48%; CO-Male = 49.2%) and smoking status (ever smoker, CA = 86%; CO = 85%). There was no difference in lifetime stressful life event between CA and CO (95% vs 93.9% P = 0.68%). However, CA were significantly more likely to have had a major stressful life event within the past 5 years than controls (CA = 77.4% vs CO = 65.8%, P = 0.03, (OR = 1.78). Serious life-threatening illness of an immediate family member (P = 0.04) and retirement (P = 0.07) within the past 5 years were noticeably more common among CA. Holmes-Rahe stress score in the last 5 years was higher in men (86.3 vs 63.3, P = 0.07) and those > 65 years old (82.4 vs 57.2,P = 0.04) as compared with CO and in those with squamous histology than with adenocarcinoma (115.6 vs 63.4, P = 0.005). Conclusions: Patients with lung cancer (CA) were significantly more likely to have had a major stressful life event within the past 5 years than the matched controls (CO), especially in older men with squamous histology. Major stressful life events should be considered a risk factor for developing lung cancer.

e14562Background: irAE occur as a result of autoimmunity triggered by ICI during cancer treatment. We hypothesize that more patients (pts) will develop any grade irAE if treated with ICI targeting ...

3511 Background: BRAF V600 mutations, present in 5-10% of patients (pts) with metastatic colorectal cancer (mCRC), are poor prognostic markers and are associated with a low response to the combinat...

114 Background: Although dealing with end of life (EOL) issues is common and can be stressful in cancer care, very little is taught and written in regards to how healthcare professionals should communicate their final good-bye with their terminally ill patients. There is a wealth of literature available about communicating “bad news”, but a literature review elicited scant data about this topic. Methods: We conducted a pilot survey on 88 house staff physicians to evaluate the frequency, their level of comfort, whether they had prior training and their knowledge about communicating their final good-bye at the EOL and their interest to learn more about this communication skill. Results: 42% were first year trainees across all specialties. 61% encountered at least 1 patient monthly at the EOL. 45% felt comfortable saying good-bye to a patient they were unlikely to see alive again. 54% have expressed interest in learning how to communicate their final good-bye to their patient at the EOL. However, only 5.7% have reported receiving a thorough training and 2.2% endorsed being very knowledgeable in this communication skill. 81% were interested to learn these skills from experienced mentors. Conclusions: Communicating the final good-bye with the terminally ill patient is an important skill that is not routinely taught to our physician-trainees. The majority of them expressed their interest to learn this skill from experienced mentors. Palliative care healthcare professionals can play a pivotal role in our physician training programs by mentoring our trainees on communicating their final good-bye to their terminally ill patient effectively and compassionately. Once implemented, further studies will be needed to assess the impact of experienced mentors on the comfort and knowledge of our physician-trainees in this communication skill.

8524 Background: There is limited information about the characteristics and outcomes of inpatient palliative care consults in cancer centers. Two mobile teams (MT) each with a physician, fellow, and a nurse, provide consultation to hospitalized patients (pts) with complex symptoms. Methods: We analyzed the pts characteristics and outcomes during a two-month period. The charts were reviewed for demographics, cancer data, reason for consultation, symptoms, interventions, and outcomes. Results: Sixty-one pts were analyzed. Pain was the main reason for a consult request in 46 pts (75%), delirium in 10 (16%), anxiety in 4 (7%) and constipation in 5 (8 %). Some pts had more than one reason. 56(92%) pts had metastatic disease, diagnosis for ≤1 year in 26 (42%) and ≤2 years in 44 (72%) (Median =17 months). The Edmonton Symptom Assessment Scale (ESAS), 0–10, is summarized in the table. We uncovered new issues at consultation, such as 20 pts (38%) screened positive for alcoholism with CAGE questionnaire and the 34 pts (56%) had clinical delirium. Features of opioids toxicity such as constipation (N=43;70%), confusion (N=35;57%) (Mean MMSE = 23 ± 5), hallucinations (N=21;34%), myoclonus (N=16;26%) and miosis (N=18;29%) were frequently identified. The MT interventions included imaging studies (23%), enema (43%), laxatives (49%), neuroleptics (54%), metoclopramide (39%), corticosteroids (25%). Half (N=30;50%) of the pts had opioid rotation and/or had counseling (N=27;46%). One out of two pts (N=30;49%) required transfer to the palliative care unit. Conclusions: Most pts had previously undiagnosed opioid toxicity, delirium, and other symptoms. Opioid toxicity occurred secondary to rapid opioid escalation, possibly linked to chemical coping, and psychosocial distress. The outcome of these pts improved by opioid rotation, adding laxatives, metoclopramide, neuroleptics, and steroids. [Table: see text] No significant financial relationships to disclose.

235 Background: The MD Anderson Clinical and Translational Research Center (CTRC) focuses on Phase I and early Phase II clinical trials and is managing approximately 156 studies annually. This 18-room combined treatment and laboratory unit manages an average of 12,500 research pharmacologic time points and over 3,000 individual shipments yearly. In FY12, 74% of shipments were done on the day of collection (DOC), whereas 26% of the shipments were shipped periodically as required by the corresponding protocol. All studies are monitored at each step for errors - minor (simple DOC label corrections) or major (improper or missed specimen). Activity is analyzed and reported monthly. Methods: In 2010, to ensure accuracy of timing, collection method, proper storage, specimen condition, and destination accuracy, we developed a robust proprietary Oracle-based laboratory tracking system (LabTracker). In addition, we designed an Advanced Research Management and Data Analysis interface (ARMADA) which provides daily step by step protocol guidance using an innovative color-icon touch-screen display of all required protocol activities for patients in the CTRC. LabTracker contains all auditable items required in the clinical trial and markedly reduces work duplication and handwritten data transcription. We hypothesized that this input system would help decrease errors and improve research quality. Results: In FY10 there were 2,506 DOC shipments with an overall error rate of 3.07%. In FY12 errors decreased to 1.31% out of 3,128 DOC shipments. In FY10, there were 14,507 timepoints with 28 major errors (0.19%). In 2012, with 13,490 timepoints, major errors decreased to 19 (0.076%). LabTracker database also supports billing and has catalogued over 275,000 samples. This system provides real-time accounting and has revolutionized the quality control system of the CTRC as well as the research finance component with 98% charge capture efficiency within 30 days. Conclusions: The LabTracker/ARMADA systems are high volume phase I protocol management tools which have improved quality, decreased redundant work, and support a real time financial management system for complex phase I clinical trials.

e20534 Background: The role of Physical Performance Test (PPT) as a screening tool for patients in geriatric oncology requiring a Comprehensive Geriatric Assessment (CGA) has not been studied so far. We undergo this study to assess PPT as a screening tool in comparison with Karnofsky Performance Status (KPS) and CGA. Methods: One hundred patients, aged ≥ 70 and diagnosed with cancer participated in our study. Inclusion criteria were knowledge of Arabic, French or English and absence of significant cognitive impairment. Exclusion criteria were: KPS

3103 Background: Although PK evaluation is a prevalent and critical component of cancer drug development, optimization of the cost-value relationship of PK studies has not been evaluated. Methods: We reviewed 26 phase I protocols that were activated and performed in the Clinical & Translational Research Center at U.T. MD Anderson Cancer Center from 2010 to 2012. 23 protocols met the residual analysis criteria of a normal and independent distribution. We collected protocol-specific data including medication, route, frequency, mechanism of action, number of enrolled patients, number of PK time points, and PK charges. Results: All 23 protocols were funded by industry sponsors. During 2010-2012: the total number of patients was 560; the median number of patients enrolled per protocol at our site was 18 (range 3-84); the median number of total PK studies per patient performed on each protocol was 19.46 (4.11-36.92); the median charge of PK time points was $42.33 (35.02-80.00); the median charge of all PK time points per patient on a protocol was $813.00 (329.00-1676.00); the median number of total PK studies per patient on a protocol during cycle 1 was 15.20 (2.40-27.47); the median charge of PK studies during cycle 1 was $43.24 (36.93-80). The number of PK time points and the amount of PK charges did not significantly correlate with monotherapy vs. combination regimens, oral vs. parenteral route, or small molecules vs. other (cytotoxic, vaccine, monoclonal antibodies). However, protocols (all cycles combined) with small molecules had a trend of lower PK time point median charges (41.97 vs. 52.67; p=0.09). Analysis of PKs in cycle 1 demonstrated that protocols with small molecules had a trend of higher PK time point median charges (61.33 vs. 42.38; p=0.08). Conclusions: The number and charges of PK studies was independent of the study drug or route. Future studies comparing budgeted vs. performed PKs and industry vs. investigator-initiated trials are warranted.

3106 Background: NRAS mutations in cancers increase MAPK signaling. It is plausible that NRAS mutations may be associated with sensitivity to drugs targeting the MAPK pathway. Methods: Tumors from 689 patients with advanced cancers referred to the Clinical Center for Targeted Therapy (phase I program) were screened in a CLIA-certified laboratory for NRAS mutations and if feasible, for PIK3CA, KRAS, BRAF mutations and PTEN aberrations. Whenever possible, patients with NRAS mutations were treated with agents targeting the RAF-MEK pathway. Results: Of the 689 patients, 58 (8%) had NRAS mutations. NRAS mutations were most prevalent in melanoma (26/170, 25%), thyroid (5/26, 19%), endometrial (2/27, 7%), non-small cell lung (1/31, 3%), ovarian (1/60, 2%) and colorectal cancers (1/74, 1%). NRAS mutations were not seen in any of the tested head and neck squamous cell cancers (n = 44); sarcomas (n = 41) and breast cancers (n = 25). None (0/62, 0%) of the patients tested for KRAS had simultaneous NRAS and KRAS mutations, and only 2/42 (5%) tested for BRAF had simultaneous NRAS and BRAF mutations. Of 58 patients with NRAS mutations, 14 (24%) were enrolled in trials that included MEK axis inhibitors. Of these 14 patients, 13 had melanoma and 1 had ovarian cancer and 4/14 (29%) demonstrated tumor shrinkage >10% from baseline. In comparison, only 22/187 (12%) patients with wild-type/unknown NRAS, KRAS and BRAF treated on the same trials demonstrated tumor shrinkage >10% (p=0.086). Conclusions: NRAS mutations are diagnosed in variety of advanced cancers. Patients with NRAS mutations demonstrated a trend to better antitumor activity with MEK axis inhibitors compared to patients without NRAS, KRAS or BRAF mutations. Further studies are warranted to delineate the role of NRAS mutations in patients treated with MEK axis inhibitors.

10535 Background: BRAF mutations are found in diverse malignancies. The current study assesses the clinical characteristics and outcomes of patients with mutant (mut) BRAF advanced malignancies. Me...

9543 Background: Alcoholism is frequently underdiagnosed in cancer patients. Preliminary studies suggest that it is associated with increased symptom expression and opioid dose escalation. The CAGE...

9503 Background: We reported cytokine gene polymorphisms as a significant factor for severe pain at time of registration in 446 White patients newly diagnosed with non-small cell lung cancer. This ...

9587 Background: Methadone is an effective and very inexpensive opioid for cancer pain treatment. It has been reported as difficult to use in the outpatient setting due to its variable relative pot...

9064 Background: Methadone (ME) has been used increasingly for pain control and for maintenance in drug addiction programs. Its use is increasing in cancer patients (pts), mainly as part of opioid rotation. Some recent reports suggest that ME may prolong QTc interval and cause torsade de pointes in pts on high dose ME. The purpose of our study was to determine the effect of initiation of ME on QTc interval in pts with cancer pain. Methods: We enrolled 101 pts in this prospective study. These pts had never been on ME before. Pts were followed clinically and electrocardiographically for QTc changes from baseline. EKG was obtained at baseline (QTc0), 2 (QTc2), 4 (QTc4), and 8 (QTc8) weeks. We recorded other contributing factors for QTc prolongation such as medication interactions with ME and electrolyte disturbances. QTc is prolongation is defined as > 430 ms in males and > 450 ms in females. In our study, significant QTc prolongation was defined as = 10% increase from baseline or QTc = 500 msec. Results: 74 pts (73%) had normal QTc (group A) and 27 pts (27%) had prolonged QTc (group B) at baseline. Significant increase in QTc in group A males occured in 1 of 16 (6%) at week 2, 2 of 10 (20%) at week 4, and 0 of 8 (0%) at week 8. Significant increase in QTc in group A females occured in 3 of 32 (10%) at week 2, 1 of 21 (5%) at week 4, and 0 of 12 (0%) at week 8. These pts had multiple contributing factors for QTc prolongation. QTc > reference for males occured in 6 of 16 (37%) at week 2, 4 of 10 (40%) at week 4, and 1 of 8 (12%) at week 8; in females, 1 of 32 (3%) at week 2, 3 of 21 (14%) at week 4, and 1 of 12 (8%) at week 8 ( Table 1 ). 2 of 27 pts (7%) from group B had a significant prolongation at week 2: one of them had an increase from 498 to 509 ms, then to 512 at week 4 and 486 at week 8. Conclusions: Baseline prolonged QTc is a common finding. QTc prolongation = 500 ms is rare in pts receiving ME for cancer pain. No data exists for other opioids. There was one case of temporary increase in QTc > 500 ms. There was no evidence of severe arrhythmias or torsade de pointes clinically or on EKG. ME dose was = 50 mg/day in the majority of these pts. [Table: see text] No significant financial relationships to disclose.

9003 Background: Fatigue is the most frequent symptom in advanced cancer. No standard treatment is available. We previously found that open-label donepezil significantly improved fatigue by day 3 and 7 in patients (pts) on opioids for cancer pain (Fisch et al, ASCO 2003). The purpose of this study was to compare donepezil (D) with placebo (P) for fatigue in pts with advanced cancer. Methods: In this randomized, double-blind, placebo-controlled trial, pts with fatigue score = 4 on a 0 to 10 scale (10 = worst fatigue) for > 1 week, hemoglobin = 10g/dl for = 4 weeks, and no major contraindication to D were randomized to receive D 5 mg or P orally every morning for 7 days. All pts were offered open-label D during week 2. Assessment included: research nurse daily phone call for fatigue and toxicity evaluation, Edmonton Symptom Assessment System (ESAS), Functional Assessment for Chronic Illness Therapy-Fatigue (FACIT-F), Sleeping Pattern Assessment, and overall effectiveness of the treatment. The FACIT-F fatigue subscale score on day 8 was considered the primary endpoint. Results: 103 pts were evaluable for final analysis. Mean difference in scores for symptoms intensity between baseline and day 8 are shown in Table 1 . FACIT-F fatigue subscale score at day 8 decreased a mean of 6 (10.6 SD) in the D arm (p < 0.001) and 7.2 (9.5 SD) in the P arm (p < 0.001). There was no significant difference in fatigue improvement between both arms according to the FACIT-F subscale (p = 0.57) and ESAS fatigue (p = 0.18) scores, and no significant difference in sleep quality score between D and P. On day 15 of the open-label phase, mean fatigue intensity remained significantly improved as compared to baseline on FACIT-F fatigue subscale (p < 0.001) and ESAS fatigue (p < 0.001) scores. No significant toxicities were observed. Conclusions: Both donepezil and placebo resulted in significant fatigue improvement. Donepezil was not significantly superior to placebo after one week. Our pilot findings are probably due to placebo effect. No significant financial relationships to disclose. [Table: see text]

9028 Background: Most referrals to acute palliative care (PC) services occur late in the trajectory of the disease, although an earlier intervention can decrease patients’ (pts) symptoms distress. The purpose of this study was to determine the time interval between first palliative care consultation (PC1) and death (D) in pts diagnosed with advanced cancer (aCA) at our comprehensive cancer center and whether such interval has increased over time. Methods: The study group was 2,868 consecutive pts who had their PC1 during a 30-month period. We reviewed the charts for information about demographics, cancer type, date of cancer diagnosis, aCA diagnosis, PC1, and D. aCA was defined as locally recurrent or metastatic. Results: 1,404 pts (49%) were female, 1,791 (62%) were < 65 years old, 2,563 (89%) had solid cancer, and 2,004 (70%) were white. The median PC1-D, aCA- PC1, and aCA-D intervals were 40, 114, and 243 days respectively. The median PC1-D interval (days) was: 47 for pts with solid cancer vs 14 for pts with hematological malignancy (p < 0.0001); 44 for pts < 65 years old vs 36 for pts = 65 years old (p = 0.002); 45 for females vs 37 for males (p = 0.004); 40 for white pts vs 41 for pts from other ethnicities (p = 0.42). The median PC1-D interval in 5 consecutive half-years was 46, 56, 42, 41, and 34 days respectively (p = 0.02). The total number of pts referred for PC1 in this period increased 20%, from 544 to 654. The ratio of PC involvement period in the aCA-D interval (PC1-D/aCA-D) decreased from 0.30 to 0.26 over the 5 half-year periods (p = 0.0004) ( Table ). Conclusions: Patients with solid cancers, younger pts, and females pts were referred earlier to acute PC. Referral timing was not affected by ethnicity. The interval between first palliative care consult and death has decreased over time. Education is needed among referring physicians to increase this interval. Further research on increasing acute PC access and its impact on PC1-D interval is needed. [Table: see text] No significant financial relationships to disclose.