Search

Aim: Germline BRCA1/2 mutation screening is critical in ovarian cancer patients to identify a subset of women who might benefit from PARP inhibitors. A recent clinical study suggests that women with either germline or somatic BRCA1/2 mutations may benefit from PARP inhibitors1. Tumor mutation screening could expand the number of ovarian cancer patients that might consider PARP inhibitor treatment2. Methods: Previously untreated high grade epithelial ovarian cancer patients were prospectively enrolled between January 2011 and December 2013 in accordance with IRB protocol. Patients were offered germline and tumor BRCA1/2 mutation screening. Germline mutation screening was performed on DNA from blood via custom amplicon assay and next generation sequencing. DNA from FFPE tumor samples was sequenced using custom hybridization enrichment and next generation sequencing. Variants and large rearrangements were identified and classified according to established criteria. Results: 263 patients were enrolled and germline testing was completed. Germline mutation analysis revealed 39 deleterious mutations (15% mutation rate, 25 BRCA1 and 14 BRCA2 mutations). Somatic tumor analysis has been completed for 88 patients receiving upfront surgical treatment (out of 123 total upfront surgery patients). 14 (16%) had germline deleterious mutations and all 14 mutations were independently identified in the tumor, giving 100% concordance between the assays. An additional 8 (9%) deleterious somatic mutations were identified, representing an overall BRCA1/2 mutation rate of 25%. One somatic BRCA2 mutation was observed in a patient with a previously identified germline BRCA2 mutation. Conclusions: Data from this prospective cohort support previously reported frequencies of BRCA1/2 deleterious mutations in high grade epithelial ovarian tumors2. Next generation sequencing assays on blood and tumor DNA were 100% concordant. Tumor sequencing identified additional BRCA1/2 defects. Further molecular profiling and clinical follow-up is ongoing to evaluate outcomes in these subgroups of ovarian cancer patients with BRCA1/2 defects. References 1. Ledermann et al, 2013 ASCO Annual Meeting 2. Hennessy et al, 2010 JCO 28:3570-6. Disclosure: K. Timms, C. Neff, J. Potter, S. Chau, J. Chen, D. Williams, M. Perry, B. Morris, A. Gutin and J. Lanchbury: is an employee and stockholder of Myriad Genetics, Inc.; K. Lu: Participating MD Anderson researchers do not receive funds from Myriad Genetics, Inc.All other authors have declared no conflicts of interest.

morbidity. The goal of this study was to evaluate the outcomes of women with early-stage cervical cancer treated with a nonradical surgical approach. Methods: Between March 1991 and July 2013, 51 women with earlystage cervical cancer were treated with simple hysterectomy or cone biopsy and concomitant bilateral pelvic lymphadenectomy or sentinel lymph node biopsy. Data on patient demographics, clinical stage, perioperative complications, pathology findings, and diseasefree interval were collected prospectively. Results: The median age of the woman was 34 years (range, 19– 77 years). Twenty-five women had squamous cell carcinoma (SCC), 22 had adenocarcinoma (AC), and 3 had adenosquamous (AS) carcinoma. Thirty women had FIGO stage 1A1 disease, 8 women had stage IA2 disease, and 13 women had stage 1B1 disease. Twentytwo (43%) and 29 (57%) women underwent simple hysterectomy with lymphadenectomy and cone biopsy with lymphadenectomy, respectively. Median tumor size was 10 mm (range, 2–11 mm). Lymphovascular space invasion (LVSI) was present in 18 women (35%). The median depth of invasion was 2.0 mm (range, 0.6– 12 mm), 2.0 mm (range, 0.1–4.5 mm), and 2.0 mm (range, 1.7– 4.0 mm) for women with SCC, AC, and AS, respectively. The margins were clear in all women. Two women received adjuvant chemoradiation (one had deep stromal invasion on cone biopsy with LVSI and one had two micrometastases to pelvic nodes). More than 95% (49 of 51 women) of patients had their Foley catheter removed on the day of surgery or postoperative day 1. There were no intraoperative or postoperative complications, and the average blood loss was estimated at 143 mL. Median follow-up was 21 months (range, 1– 112 months). None of the 51 women with early-stage cervical cancer developed a recurrence during their follow-up time (95% CI: 0–6%). Conclusions: Nonradical surgery in appropriately selected early-stage cervical cancer patients appears to be associated with a very low perioperative complication rate and excellent oncologic outcomes. While large phase II and III studies are underway, this approach seems to be a safe and reasonable option in well-selected patients.

5512 Background: Screening for Lynch Syndrome (LS) in endometrial cancer (EC) has traditionally been based on early age at diagnosis and family history. Universal tumor testing has been proposed gi...

5510 Background: Metformin reduces cancer incidence and improves overall survival in diabetic patients. Preclinical studies have shown that metformin decreases endometrial cancer (EC) cell growth b...

Background Cord blood transplantation (CBT) is an increasingly used alternative to bone marrow or peripheral blood stem cell transplantation, but is associated with slower neutrophil and platelet recovery. One strategy to overcome the delayed hematopoietic recovery of CB is to correct the decreased fucosylation of CB cell surface molecules which is thought to impair homing to the bone marrow of the limited numbers of stem and progenitor cells in the CB graft. Our pre-clinical murine xenograft models have demonstrated that human CB-derived progenitor cells treated with recombinant human fucosyltransferase-VI (ASC-101:America Stem Cell Inc.) prior to infusion resulted in more rapid and higher levels of human engraftment as compared to untreated CB progenitors (Robinson et al Exp Hematol, 2012). We therefore sought to study this novel strategy in a clinical trial where recipients with hematologic malignancies receive a double CBT where one CB unit is fucosylated prior to infusion. Objective In an effort to improve engraftment following cord transplant, we tested the ability of a 30-minute ex vivo fucosylation treatment to shorten time to neutrophil and platelet recovery. Methods Cell Processing: The unmanipulated CB unit with the highest total nucleated cell (TNC) dose was thawed, washed on the Sepax device (Biosafe) and infused first. The second CB unit which had the smaller TNC dose was thawed and washed using 10% Dextran-40/5% human serum albumin and the cells treated at 106cells/ml for 30 minutes at room temperature with recombinant human fucosyltransferase VI (ASC-101) and substrate GDP-fucose (America Stem Cell Inc). The fucosylated cells were then washed on the Sepax and infused. After the procedure fucosylated (CD34+ CLA+) CD34+ cells increased from a median of 33% to 99%. Clinical 10 patients with AML (5), MDS (2), NHL (2), or HL (1) have been enrolled to date. Three patients are too early to evaluate engraftment; therefore, first 7 evaluable patients are reported here. Median age was 55 (range 26 -62) years and median weight 87 (range 61-97) kg. All patients were conditioned with fludarabine 160mg/m2, melphalan 140mg/m2, and ATG 3mg/kg. Tacrolimus and MMF were used for GVHD prophylaxis. Results Median time to absolute neutrophil count ≥ 0.5 X 109 /L was 14 (range 12-28) days. Median time to platelet count ≥ 20 X 109/L was 33 (range 18-100) days. One patient had secondary graft failure and was rescued with backup autologous stem cells. Four patients had engraftment of the fucosylated unit and 2 of the unfucosylated unit. Two patients developed grade 2 acute graft versus host disease. No infusion related toxicities were seen. Conclusion Ex vivo fucosylation appears to be a safe, simple and rapid approach to enhancing neutrophil and platelet engraftment in the setting of double cord transplantation. Accrual to the trial continues and updated results will be presented. Disclosures: Miller: America Stem Cell Inc: Equity Ownership. Paradiso:America Stem Cell Inc: Equity Ownership. Koh:America Stem Cell Inc: Equity Ownership.

Materials/Methods: Nine hundred fifty-two patients with prostate cancer were treated with permanent interstitial implantation as monotherapy using a transrectal ultrasound-guided approach. Real-time intraoperative treatment planning software which incorporates inverse planning optimization was used for all patients. Nine hundred twenty-one patients were treated with I-125 and 31 patients were treated with Pd-103. Two hundred fiftynine patients (27.2%) received neoadjuvant androgen deprivation in most cases to achieve pre-implantation volume reduction for prostate volumes >50 cc. The median prostate volume was 35.9 cc (range, 7.6-99.7 cc). The breakdown between favorable and intermediate risk disease was 79% and 20%, respectively. Dose-volume constraints for this inverse-planning system included: prostate V100 95%, maximal urethral dose 140, the 7 year PSA-RFS was 98% compared to 93% for patients with lower D90 values (p Z 0.01). Multivariate analysis demonstrated that the D90 dose level was the only significant predictor of PSA relapse-free survival (p Z 0.04 value HR 0.32). The median maximum urethral dose was 134%, and the median urethra D20% was 127%. Maximum urethral doses exceeding 150% of the prescription dose were noted in 28% of patients. The median maximum rectal dose was 64% of the prescription dose (range, 29% -193%) and the median volume of the rectum receiving the prescription dose rectum V100 cc was 0.3 cc (range, 0-7.4 cc). No dosimetric parameter was identified linked to late urinary or rectal toxicities. Conclusions: Dosimetric analysis performed on the day of the brachytherapy procedure of a large cohort of patients confirms excellent coverage of the prostate with the prescription dose. Our data demonstrate that the D90 was the single most important variable associated with improved biochemical tumor control in these patients. These data further highlight the critical need for intraoperative real time dose modifications and corrections to achieve optimal dose delivery to the prostate to insure maximal tumor control probability. Author Disclosure: M. Ashamalla: None. M. Kollmeier: None. G.N. Cohen: None. X. Pei: None. M.J. Zelefsky: None.

4528 Background: The role of cisplatin-based chemotherapy for the treatment of locally advanced or metastatic urothelial carcinoma is well established. For patients (pts) who cannot receive cisplatin owing to renal insufficiency, substitution with carboplatin was associated with inferior response rate and overall survival (OS). To address this unmet need, we conducted a phase II study of gemcitabine (Gem), paclitaxel (Tax), and doxorubicin (Adria) in this group. Methods: The primary endpoint was overall response rate (ORR); secondary endpoints were toxicity, OS, and the safety and efficacy of pegfilgrastim 6 mg (G-CSF) given immediately after chemotherapy on Day 1. A Simon 2-stage design was chosen to detect ORR of 40% and to reject ORR of 25%. Eligible pts had metastatic or unresectable urothelial carcinoma, no prior chemotherapy, performance status ≤ 2, glomerular filtration < 60 ml/min, and no need for dialysis; all gave informed consent. Brain metastases were excluded, as were clinically significant heart disease, peripheral neuropathy, and liver or bone marrow dysfunction. Treatment consisted of 900 mg/m2 Gem (fixed rate of 10 mg/m2/min), 135 mg/m2 Tax, and 40 mg/m2 Adria administered with same-day GCSF every 14 d to a maximum of 9 cycles. Tumors were evaluated after every 3 cycles. Results: Forty pts were enrolled January 2008 through November 2011, and 39 could be assessed for response. Median age was 72 (range, 51–89) and 11 pts (28.2%) were women. There were 7 complete and 15 partial responses, for an ORR of 56.4% (95% CI 39.6-72.2). Notable grade 3 and 4 nonhematologic toxicities in the first 2 cycles were dyspnea and mucositis (1 pt each). There were no treatment-related deaths and no toxicity attributed to same-day G-CSF. Median OS was 14.4 mo with median follow-up of 12.6 mo for all pts, 15.6 mo for 10 who were alive. Conclusions: Gem-Tax-Adria is effective as first-line treatment for metastatic or locally advanced urothelial carcinoma, and it can safely be given to pts with renal insufficiency. Same-day G-CSF also appears to be safe and effective in this setting. Phase III study is warranted. Clinical trial information: NCT00478361.

5596 Background: Advanced-stage (III/IV) uterine papillary serous carcinoma (UPSC) has a median overall survival (OS) of ~ 3 yrs. The study objective was to determine factors associated with long-term survival in advanced stage UPSC. Methods: We performed a retrospective review of pts diagnosed with stage III or IV UPSC between 1993 and 2012. Summary statistics were used to describe demographic and clinical characteristics. OS was estimated with the Kaplan-Meier estimator. Fisher’s exact test and the Wilcoxon rank sum test were used to compare pts surviving >3 yrs with those surviving 3 yrs was 0.462, >4 yrs was 0.310, and >5 yrs was 0.228. Thirty-six (14%) pts survived >3 yrs and 37 (14%) survived 3 yrs compared to pts surviving

TPS6100 Background: BRAF mutations are found in 40% of pts with newly diagnosed PTC but much more prevalent in recurrent PTC, ranging from 78-86%, suggesting they play an active role in tumor progression. Activating mutations in BRAF result in constitutive activation of MEK and subsequently ERK. ERK mediates activation of nuclear transcription factors coordinating expression of genes involved in proliferation and survival of malignant cells. PTC usually has an excellent prognosis, especially in younger pts and in pts who respond to the only effective treatment: surgery followed by radioactive iodine (RAI). However, more advanced stage disease at diagnosis, older age, and lack of RAI avidity are associated with a high rate of recurrence and distant metastasis, imparting a worse overall survival. Long-term outcomes depend highly on initial surgery outcome. Pts at highest risk of recurrence and death are those with gross residual disease after surgery and macroscopic tumor invasion. VEM is an inhibitor of the activated form of the BRAF serine-threonine kinase enzyme. The drug is FDA approved for the treatment of advanced melanoma and is currently being studied in a phase 2 trial in metastatic BRAF-mutated PTC. This is a neoadjuvant trial to determine if pharmacodynamic changes in the tumor correlate with response to drug. Methods: Pts with primary or recurrent BRAF mutated PTC who are planned for surgical resection are eligible. Pts will undergo baseline core biopsy prior to starting VEM 960mg bid. After 56 days of treatment they will undergo surgery and post-treatment specimen will be collected. Pts with widely metastatic PTC may continue VEM. The primary endpoint is to determine whether changes in ERK phosphorylation responses after treatment with VEM correlate with clinical response at day 56 in pts with locally advanced, BRAF mutated PTC. Secondary endpoints include assessments of safety of neoadjuvant VEM, response by RECIST, rate of surgical complications, persistent disease at the surgical site at 1 year, and mechanisms of resistance to VEM. The trial began enrolling pts in December 2012, with a total of 22 pts planned. This trial has no sponsor. Clinical trial information: NCT01709292.

Conclusions: Moderate hypofractionation with SIB and RA technique showed to be safe, with a good acute toxicity profile. Longer follow-up is needed to assess definitively late toxicity and clinical outcome. Author Disclosure: F. Alongi: None. A. Fogliata: None. S. Arcangeli: None. S. Castiglioni: None. G. Taverna: None. P. Navarria: None. L. Cozzi: None. P. Grazziotti: None. A. Tozzi: None. M. Scorsetti: None.

9061 Background: Although great progress has been made in the control of chemotherapy-induced vomiting (CIV), prevention of chemotherapy-induced nausea (CIN) has been less successful. Preliminary data suggests that some families of lesser used antiemetic agents, such as the cannabinoids, may have greater efficacy against nausea than against vomiting. Methods: Adult solid tumor patients (pts) receiving cyclophosphamide ≤ 1500 mg/m2 (C) and/or doxorubicin ≥ 40 mg/m2 (A) were eligible. Pts could have received prior mildly emetogenic chemotherapy (EC). Pts were not eligible who were receiving other moderately or highly EC, were receiving cranial, abdominal or pelvic radiotherapy, had CIV/CIN with previous chemotherapy, had other causes for nausea/vomiting besides chemotherapy, or were scheduled to receive other antiemetics. Pts with habitual cannabinoid use were not eligible. All pts received palonosetron 0.25 mg (PALO) and dexamethasone 10 mg (DXM) IV before chemotherapy. Patients were randomized double-blind to receive dronabinol 5 mg (D) or matching placebo (P) 3 times a day for 5 days. Nausea, vomiting and toxicity data was collected daily for 5 days. Results: 62 pts were entered on study – female/male 61/1, White/Black/Hispanic/Other 45/14/2/1, median age (range) 58 (29-76). No significant difference was noted in CIV-dependent endpoints (including No Vomiting, Complete Response, or Complete Protection) or in rescue medication use. However pts receiving D had a shorter duration of nausea – Mean number of days of nausea (D vs P) 1.86 days vs 3.10 days (p=0.027) – and a trend toward greater frequency of No Nausea (D vs P) 37% vs 17% (p=0.143). Common toxicities included fatigue (D/P 17/11), headache (D/P 16/16), dizziness (D/P 14/7), constipation (D/P 14/11), and diarrhea (D/P 13/6) No pt discontinued therapy due to mood changes. Conclusions: Low-dose D decreased the duration of CIN and increased the frequency of No Nausea in pts receiving C and/or A. Agents to prevent CIV (such as PALO and DXM) and to prevent CIN (such as D) have complementary activity and result in improved overall control when used together.

330 Background: Primary carcinomas of the urethra (PCU) are rare and often advanced when diagnosed. Given the low incidence of this disease, there is a paucity of data regarding treatment standards. The purpose of this report was to study the treatment response and survival in a relatively large cohort of patients with PCU, with emphasis on modern platinum-containing chemotherapy regimens, in combination with surgery for (locally) advanced PCU. Methods: We performed a retrospective chart review of consecutive patients with PCU seen at the MD Anderson Cancer Center over a recent 5-year period (2005-2010). Clinical outcome was measured as best response to chemotherapy (either neo-adjuvant or for metastatic disease). Kaplan-Meier estimates were generated for survival and Cox proportional hazard was used for prognostic factors for survival. Results: We identified 44 consecutive pts with PCU. The majority (64%) were female, and the median age at diagnosis was 66.5 yrs. Squamous cell carcinoma and adenocarcinoma were the most common histologic subtypes of PCU in our cohort. 43% and 16% presented with lymph node–positive (LN+) and distant metastases, respectively. For all evaluable pts, the overall survival (OS) was 31.7 mo. Platinum-containing neoadjuvant chemotherapy achieved a response rate of 72%. 21/44 pts had locally advanced or LN+ PCU and were treated with chemotherapy followed by surgery. In this cohort, the median OS from chemotherapy initiation was 25.6 months. Among pts with LN+ PCU at diagnosis, 4/9 (44%) were alive as of July 2011, with a minimum follow-up of >3 years. Conclusions: We report the clinical outcomes in the to the best of our knowledge largest cohort of consecutive pts with PCU. Our data suggests that modern platinum-containing regimens are effective in a subset of advanced PCU. Neoadjuvant treatment leads to prolonged disease-free survival in a substantial proportion of lymph node positive cases.

297 Background: PUC are a rare subset of urothelial cancers with a paucity of available literature regarding their clinical behavior. Methods: Retrospective analysis of outcomes among all patients (pts) with predominant PUC (pPUC) seen at MDACC from 1999-2010. Results: 31 pts with pPUC histology were identified (median age:63.5yrs; 83.3% male; TNM stage:cT1N0,n=4;cT2N0,n=8;cT3b-4aN0,n=6; cT4b, N+ or M+ n = 13). Median survival for all pts was 17.7mo (Stage I-III vs IV: 38.1 vs 13.3mo). Of 18 pts with potentially surgically resectable PUC (18 mo after surgery. The most common site of recurrence was in the peritoneum (13/18pts), with relapses occurring even in those with pCR at surgery. At least 4 pts developed CNS involvement with lepto-meningeal disease. In pts presenting with metastatic disease who were treated with chemotherapy, the median survival was 12.6 months. Conclusions: Plasmacytoid bladder tumors are a very aggressive subset with overall poor outcomes. Though chemotherapy sensitive with downstaging seen with neoadjuvant chemotherapy, there are few long-term survivors. There is a strong predilection for recurrences along the peritoneal lining, and peritoneal carcinomatosis should remain high on a clinician’s differential when patients develop obstructive bowel symptoms.

254 Background: The addition of bevacizumab to capecitabine-based chemoradiotherapy (CRT) for LAPC has been shown to be safe. The aim of this study was to determine the safety, tolerability and maximum tolerated dose (MTD) of the addition of erlotinib to this treatment regimen. Methods: Seventeen patients with CT-staged biopsy-proven non-metastatic unresectable LAPC were enrolled between March 2008 and October 2010. Prior chemotherapy was permitted. All patients received 50.4 Gy (GTV only) in 28 fractions with concurrent capecitabine, bevacizumab and erlotinib. Dose was escalated using a continual reassessment method. Two patients each were enrolled at dose levels (DLs) 1-4 and 9 patients at DL 5. Bevacizumab was escalated from 5mg/Kg every two weeks (DLs 1-4) to 10mg/Kg (DL 5); erlotinib from 100 mg/day (DLs 1-2) to 150 mg/day (DLs 3-5); and capecitabine from 400mg/m2 twice daily on days of radiation (DL 1) to 600mg/m2 (DLs 2-3) to 825 mg/m2 (DLs 4-5). Reassessment for potential resection was performed 6-8 weeks later. Results: With a median follow-up of 10 months (range 3-23), no grade 3 toxicities were observed in DLs 1-4. Three (33%) patients at DL 5 developed a grade 3 acute toxicity (2 diarrheas and 1 rash). No grade 4 or 5 toxicities were seen. DL 4, with a posterior probability of 0.122 of dose limiting toxicity, was selected as the MTD. Median survival was 19.4 months and time to distant progression was 9.8 months. Patients treated at DLs 4 and 5 had a median survival of 24 months. Of 5 patients who underwent margin-negative resections, 4 were originally deemed unresectable and 1 was borderline; 4 were treated at DLs 4 or 5 (36% of patients treated at these DLs); 3 patients had excellent pathological responses (complete response, 5% viable tumor, and 20% viable tumor) at pancreatectomy and are alive at 13, 21 and 22 months respectively with no local or distant failures. Conclusions: The combination of erlotinib, bevacizumab and capecitabine with radiotherapy for LAPC is safe and tolerable. Both the promising survival and the high rate of resectability at the higher dose levels suggest that this strategy of dual inhibition of growth factor receptor pathways during CRT warrants continued evaluation.

544 Background: The goal of this phase I trial was to determine the maximal tolerated dose (MTD) of concurrent capecitabine, bevacizumab and erlotinib with preoperative radiation therapy (RT) for rectal cancer, with a secondary objective of evaluating pathologic complete response (pCR). Methods: Eligibility criteria included patients with clinical stage II-III rectal adenocarcinoma within 12 cm from the anal verge. Exclusion criteria included uncontrolled hypertension, recent myocardial infarction or angina, coagulopathy, impaired renal function, and history of gastrointestinal perforation. Patients were treated in cohorts of 2, in 4 escalating dose levels (DL), using the continual reassessment method, as shown in the table. Patients underwent surgery at least 9 weeks after the last dose of bevacizumab. Results: Nineteen patients were enrolled in the study: 1 withdrew early because of insurance issues and was non-evaluable. The clinical stage was T3N0 in 3, T3N1 in 13, T3N2 in 1 and T4N0 in 1 patient. No patient had grade 4-5 acute toxicity; 1 patient had grade 3 acute toxicity (hypertension, DL 2) during or within 4 weeks after RT. The MTD was not reached; the probability of dose limiting toxicity at DL 4 was 0.15. All 18 evaluable patients underwent surgery, with low anterior resection in 7 (39%), proctectomy with coloanal anastomosis in 5 (28%), and abdominoperineal resection in 6 (33%) patients. Eight patients (44%) had pCR, and an additional 8 (44%) had ≤ 10% viable tumor in the surgical specimen. Fifteen patients (83%) had T downstaging. Three patients developed grade 3 post-operative complications (ileus, small bowel obstruction and infection). After a median follow-up of 21 months (range 9-39 months), no patient had local recurrence, but 1 patient developed distant metastasis. Conclusions: The combination of preoperative RT with concurrent capecitabine, bevacizumab and erlotinib was well tolerated. The pCR rate of 44% is promising and warrants further investigation. [Table: see text]

5016 Background: Everolimus is an oral rapamycin analog that selectively inhibits mTOR. PTEN mutations occur in 40-60% of endometrial cancer (EC). Although PTEN loss leads to constitutive activation of AKT and up-regulation of mTOR, it is not clear which EC patients will benefit from this targeted therapy. PS6K is a downstream marker of mTOR activation. Prior studies suggest KRAS mutations are associated with resistance to everolimus. We evaluated primary tumor specimens to determine if expression of biomarkers in the mTOR pathway or KRAS mutations would predict response to therapy. Methods: Correlative studies evaluating PTEN and PS6K expression by immunohistochemistry (IHC) and KRAS mutational analysis were performed from tissue blocks of primary tumor. Response to therapy was estimated with Fisher’s exact test. Positive predictive value (PPV) for each variable was calculated. Tumor from patients enrolled in a single institution, open-labeled, single arm, Phase II study in pretreated patients with recur...

246 Background: The role of cisplatin-based chemotherapy for the treatment of locally advanced or metastatic UC is well established. Pts with impaired renal function cannot receive cisplatin, however, and substitution with carboplatin yielded inferior results. We initiated a phase II study to assess the efficacy of gemcitabine, paclitaxel, and doxorubicin (GTA) treatment for UC in pts with renal insufficiency. Methods: Eligible pts had metastatic or unresectable UC of bladder, urethra, or upper tract, no prior chemotherapy, glomerular filtration rate < 60 ml/min, ECOG performance status ≤ 2, left ventricular ejection fraction > 40%, and adequate hematopoietic and hepatic function. Pts were excluded if they had brain metastasis, peripheral neuropathy ≥ grade 2, significant heart disease within 6 months of enrollment, or required hemodialysis. Outpatient treatment consisted of 900 mg/m2 gemcitabine, 135 mg/m2 paclitaxel, and 40 mg/m2 doxorubicin on day 1 every 2 weeks. Pegfilgrastim 6 mg sc was given immediately after GTA on day 1, or on day 2 if requested by the pt. Tumor evaluation was repeated every 3 cycles (6 weeks); treatment duration was limited to 9 cycles. A Simon 2-stage design was chosen to detect a target overall response rate of 40% and to reject a response rate of 25% or less. Results: Twenty-five pts enrolled and 21 could be assessed for response. Median (range) age was 72.8 years (53.4, 89.3) and 8 pts (32.0%) were female. RECIST responses occurred in 12 pts (4 complete; 8 partial), for an overall response rate of 57.1% (95% CI 34.0-78.2). Notable grade 3 and 4 toxicities were anemia (9 pts), thrombocytopenia (3 pts), neutropenia (2 pts), dyspnea (1 pt), mucositis (1 pt), and sepsis (1 pt). No adverse events were attributed to same-day pegfilgrastim and there were no treatment-related deaths. Eleven pts (44.0%) died of progressive disease; the median (range) follow-up time for 14 surviving pts was 5.5 months (2.3, 17.5). Conclusions: GTA has been well-tolerated in the setting of renal insufficiency, with an observed response rate in advanced UC exceeding the targeted response. The study will continue to a maximum of 72 pts. [Table: see text]

Abstract 362 Delayed engraftment and low rates of platelet transfusion independence are frequently observed after CB transplantation (CBT). We conducted a study of ex-vivo co-culture of CB mononuclear cells with either third party haploidentical family member marrow derived MSCs (N=8) or off-the-shelf mesenchymal progenitor cells (MPCs) from Angioblast (N=24). Patients received a double cord blood transplant, with one of the 2 units undergoing ex vivo expansion using this system. MSCs create a microenvironment that promotes expansion and fosters the differentiation of hematopoietic cells. Patients must have had two CB units matched in at least 4/6 HLA antigens, with a minimum of 1×107 TNC/Kg per unit. Method: Diagnoses were AML/MDS (n=21), ALL (n=6), NHL (n=2), CLL (n=2), and HD (n=1). Fourteen patients (44%) were in CR (CR1, n=3, CR2 or more, n=11) and 18 (56%) had active disease at CBT. Preparative regimen: myeloablative fludarabine, melphalan, thiotepa and ATG (n=32), with rituximab in the 4 NHL/CLL cases. GVHD prophylaxis: tacrolimus and MMF. Median weight was 75.2 Kg (range, 15–118) and median age was 35.3 years (2.8-62 years). Donor-recipient HLA matching was 6 of 6 in 5%, 5 of 6 in 28% and 4 of 6 in 67% of the cases, respectively. Ex-vivo EXP: 100 ml of marrow was aspirated from the family donor and MSCs generated in ten T175 flasks, which took ∼21 days (n=8) or one vial of Angioblast MPCs was thawed and expanded to confluence in 10 flasks within 4 days (n=24). The CB unit with the lowest TNC dose was then thawed, divided into 10 fractions, and each placed into 1 flask containing the confluent layers of MSCs in expansion media with SCF, FLT3-ligand, G-CSF and TPO. After 7 days at 37°C, the non-adherent cells were removed from each flask, placed into each of ten one-liter Teflon-coated culture bags (American Fluoroseal) and cultured for an additional 7 days (14 days total), while 50 ml of media/growth factors was added to the flasks to culture the remaining adherent layer during that time period. On day 14 the cells from the bags and the flasks were combined, washed and infused along with a second unmanipulated CB unit. Result: The median number of total nucleated cell (TNC) and CD34+ cells infused/Kg in unmanipulated CB was 2.35 × 107 (range 0.2–8.2) and 0.95 × 105 (range 0–4). The median number of TNC and CD34+ cells infused/Kg after EXP was 5.8 × 107 (range, 0.3–14.4) and 8.7 × 105 (range, 0–93.4). This represented a median expansion of 14-fold (range 1–30) for TNC and 40-fold (range 4–140) for the CD34+ cells. Median time to neutrophil and platelet engraftment was 15 days (range 9–42) and 40 days (range 13–62). There were no toxicities attributable to the EXP cells. Thirty-one (97%) and 26 (81%) of all patients engrafted neutrophils and platelets, respectively. One patient died before engraftment. Thirty and one-hundred day non-relapse mortality is respectively 6% and 19%. Median donor(s) chimerism was 100% in the mononuclear, T lymphocyte and myeloid cell populations. On transplant day+21, EXP unit contributed with a mean of 19% of mononuclear cell, 16% of T cell, and 14% of myeloid chimerism. Subsequently, hematopoiesis was increasingly derived from the unexpanded unit with long-term engraftment provided by the unexpanded unit by six months posttransplant. Acute grade II-IV and III-IV GVHD rate was 50% and 16%; 25% of the grade II-IV GVHDs occurred beyond 100 days, and two patients developed chronic GVHD. With a median follow-up of 9 months, 11 patients are alive; actuarial one-year survival is 40%. Mortality was due to relapse in 26% and non-relapse causes in 74% of patients. Conclusion: MSC-CB Exp is feasible and leads to fast engraftment of neutrophils and platelets, and high-rates of platelet transfusion independence. Disclosures: No relevant conflicts of interest to declare.

10060 Background: Sarcoma metastasizing to the heart is a difficult and complicated clinical problem, yet local control is feasible and effective for select patients with cardiac metastasis. Diligence is required to detect these lesions while still amenable to treatment, and echocardiogram remains the best tool for detecting cardiac disease. Methods: We reviewed all echocardiograms (1330) performed on 307 pediatric sarcoma patients treated at UT M. D. Anderson Cancer Center between 1997 and 2008. Measures of cardiac function and pathology (including size of ventricles and atria, function of ventricles and valves, presence of metastasis, and presence of pericardial effusion) were assigned numerical values. Fisher's exact test and the Wilcoxon rank-sum test compared clinical characteristics of patients with and without cardiac metastases. Results: The prevalence of cardiac metastases was 1.6% (5/307) with 95% confidence interval 0.5% to 3.8%. The presence of cardiac metastasis positively correlated with pericardial effusion (p = 0.001) and tricuspid valve insufficiency (p = 0.014). The probability of a patient with pericardial effusion having a cardiac metastasis was 28.6% (95% CI: 3.7%-71%). Of the 5 patients with documented sarcoma metastasis to the heart, 3 had widespread refractory disease, were given no cardiac-specific therapy, and rapidly died from disease. One patient who had widespread disease controlled with chemotherapy and radiation had open resection of 2 cardiac metastases which resulted in site-specific disease control for 6 months before succumbing to progressive extra-cardiac disease. Another patient had isolated cardiac metastasis, treated with open resection of 2 metastases followed by adjuvant chemotherapy, and has been without recurrent cardiac metastases for 5 years and 9 months. Conclusions: Durable local control for sarcoma metastasizing to the heart is possible and effective for select patients. Pericardial effusion merits evaluation for potential associated cardiac metastasis. Echocardiogram monitoring of sarcoma patients remains important, and should include screening for possible cardiac metastasis. No significant financial relationships to disclose.

20756 Background: Predicting inpatient mortality has clinical and financial implications and helps improve the care of advanced cancer patients and their families. Models are available for mortalit...

5502 Background: The finding of PTEN mutations in 40–60% of EC suggests this pathway is important in the pathogenesis of this disease. Loss of PTEN leads to constitutive activation of AKT, which le...

Both G-CSF and GM-CSF (alone or in combination) may be used for mobilization of hematopoietic progenitor cells in patients undergoing autologous stem cell transplantation (ASCT). It has been suggested that GM-CSF use during mobilization may impact graft composition and therefore clinical outcomes. METHODS: We prospectively evaluated patients ≤ 70 years old with relapsed CD20+ NHL who were candidates for ASCT. Additional eligibility criteria included adequate marrow and organ function. Patients with history of pelvic radiation, > 3 prior chemoregimens or > 6 cycles of fludarabine chemotherapy were excluded. Patients recieved chemotherapy with ifosfamide 3.33 g/m2 daily × 3 days, etoposide 150 mg/m2 × 6 doses and rituximab (375 mg/m2 on day 1 and 1 g/m2 on day 8). Using a Bayesian adaptive randomization based on treatment outcomes, patient’s were randomized to receive G-CSF 12 μg/kg/d (Group G) or G-CSF 12 μg/kg/d plus GM-CSF 500 μg/d (Group G/GM). We assumed that the success rate for each treatment arm had a β prior distribution with mean 0.90 and variance 0.03. Cytokines started 24 hours after completion of chemotherapy and continued until completion of apheresis. RESULTS: Forty-three patients were randomized to Group G and 41 patients to Group G/GM. In each arm 1 patient withdrew consent after randomization. Baseline characteristics were similar in the 2 groups (Table 1). Both regimens were equally well tolerated. Data are presented as intent to treat analysis. Thirty-nine patients (90.7%) in Group G and 35 patients (85.4%) in Group G/GM collected ≥ 4 × 106 CD34+ cells/kg. The probability that Group G has a higher success rate than Group G/GM is 0.778. The median CD34+ cell dose collected was 10.3 × 106/kg (range, 0.1–59) and 7.5 × 106/kg (range, 0.7–73) in Groups G and G/GM respectively (P=NS). A median of 2 apheresis procedures were required in both arms. Seventy-three patients have undergone ASCT. After a median follow up time of 14.5 months (range, 0.6–38.5) in Group G and 14.0 months (range, 1.1–39.9) in Group G/GM, the 3 year PFS is 75% (95% CI 57.9–99.4) and 77% (95% CI 65–91.5) respectively (P=0.41). CONCLUSION: Our study does not support the hypothesis that using G-CSF plus GM-CSF versus G-CSF alone for progenitor cell mobilization alters graft composition in a way that impacts clinical outcomes after ASCT for NHLs. Baseline Patient Characteristics *Missing data 1 patient G-CSF, N (%) G-CSF + GM-CSF, N (%) 43 (51) 41 (49) AGE 59 10 (23.2) 12 (29.3) GENDER (Male/Female) 29/14 (67.4/32.6) 24/17 (58.5/41.5) HISTOLOGY Low grade 4 (9.3) 7 (17.1) Intermediate grade 39 (90.7) 34 (82.9) ANN ARBOR STAGE >I 18 (41.9) 18 (43.9) LDH>Normal* 15 (34.9) 11 (27.5) Figure Figure