Search

3000 Background: NY-ESO-1 is expressed in ~70% of synovial sarcomas (SS). NY-ESO-1c259T cells recognizing an NY-ESO-1 derived peptide complexed with HLA-A*02 are being studied in SS. Methods: Eligible patients (pt) are HLA-A*02:01, 02:05 or 02:06, with unresectable, metastatic or recurrent SS expressing NY-ESO-1. Primary endpoint of ORR (CR+PR) is evaluated in high (≥ 50% tumor cells express 2+/3+) and low (≥ 1+ in ≥ 1% cells, not exceeding 2+/3+ in ≥ 50% cells) NY-ESO-1 expressers with different lymphodepleting regimens. Secondary endpoints are safety, DOR, PFS, OS, and gene-marked cell persistence. Lymphocytes are obtained by leukapheresis, isolated, activated, transduced to express NY-ESO-1c259T, and expanded. Target dose is 1–6 × 109cells. Disease is assessed at wk 4, 8 and 12 post-T-cell infusion, and then every 3 months. Results: 34 pt have been enrolled with 24 treated. 50% are male; median age is 30 yr (range 15 – 73). 12/15 pt in cohort 1 were treated. ORR was 50% (1 CR; 5 PR). Time to response was 6 wk (range 4-9) and median DOR 31 wk (range 13-72). Cohort 3 was closed due to only 1 PR out of 5 pt. Evaluation is ongoing in cohorts 2 (6 enrolled; 5 treated) and 4 (8 enrolled; 2 treated) as of 1/9/17. The most common AE are leukopenia (96%), nausea and pyrexia (88%), neutropenia (88%), lymphopenia (83%), anemia (79%), and thrombocytopenia (79%). 11 events of CRS were reported (3 G3; 1 G4), with no events of seizure, cerebral edema or fatal neurotoxicity; all resolved with supportive therapy. One fatal SAE (bone marrow failure) occurred in cohort 2; investigations have not identified a mechanism by which NY-ESO-1c259T may have caused this event. Conclusions: NY-ESO-1c259T has promising efficacy and acceptable safety. CRS is not associated with severe neurotoxicity and appears manageable with appropriate supportive care. Cohort 3 data indicate that Flu may be important for efficacy. Efficacy and safety data will be further evaluated and presented. Clinical trial information: NCT01343043. [Table: see text]

TPS3096 Background: Non-small cell lung cancer (NSCLC) accounts for 84% of lung cancer. Survival has recently been impacted by molecularly targeted therapies and checkpoint inhibitors (CPI), and the promising CPI results implicate a role for the immune system in NSCLC. 10-40% of NSCLC express NY-ESO-1 or MAGE-A10 cancer/testis antigens. These studies will evaluate the safety and antitumor activity of genetically engineered affinity enhanced TCRs (NY-ESO-1c259T or MAGE-A10c796T) directed towards a NYESO-1 or MAGE-A10 derived peptides complexed with HLA-A*02. In addition, correlative studies to evaluate persistence, phenotype, functionality of engineered T cells, mechanisms of resistance and antigen spreading will be performed. Methods: Patients (pt) are screened (NCT02636855) to identify those who have the relevant HLA-A*02 alleles and NY-ESO-1 or MAGE-A10 tumor expression. For entry into either treatment protocol, pt must have Stage IIIb or IV NSCLC, have failed at least one platinum-containing regimen (may have received CPIs), have measurable disease, ECOG 0-1, adequate organ function, and be without brain metastases, history of severe autoimmune disease or current uncontrolled illness. Following apheresis, T cells are isolated and expanded with CD3/CD28 beads, transduced with a lentiviral vector containing the NY-ESO-1c259T or MAGE-A10c795 TCR, and infused into the pt following lymphodepleting chemotherapy with fludarabine and cyclophosphamide. The NY-ESO-1c259T study is a 10 pt study utilizing a dose of 1-6 x 109 transduced T cells. The MAGE-A10c796T first-in-human study is a modified 3+3 design in up to 28 pt with escalating doses of 0.1, 1.0 and 1-6 x 109 transduced T cells, with staggered treatments to allow for safety review; dose escalation will be guided by the DLT observed and by safety review committee guidance. Response to treatment will be assessed by RECIST v1.1 at weeks 4, 8, 16, 24, every 3 months (for 2 yr) and every 6 months until disease progression. Clinical trial information: NCT02588612/NCT02592577.

TPS3098 Background: MAGE-A10 is a cancer/testis antigen that has been identified in 42, 26 and 17% of urothelial, melanoma and head and neck tumors, respectively. This study will evaluate the safety and antitumor activity of genetically engineered affinity enhanced autologous MAGE-A10c796T cells directed towards a MAGE-A10 peptide expressed on tumors in the context of HLA *02:01 and/or *02:06. Methods: This first-in-human T cell dose escalation study utilizes a modified 3+3 design to evaluate safety, including dose limiting toxicities (DLT). Secondary objectives include anti-tumor activity (overall response, duration of response, time to response, PFS, OS) and translational research assessments. Patients are screened under a separate protocol (NCT02636855). Those who are HLA*02:01 and/or *02:06 positive and have inoperable or metastatic (advanced) urothelial cancer, melanoma, or squamous cell head and neck tumors with MAGE-A10 expression and meet all other entry criteria are eligible for treatment. Patients must have received standard of care therapies and have progressive disease. Following apheresis, the T cells are isolated and expanded with CD3/CD28 beads, transduced with a lentiviral vector containing the MAGE-A10c796 TCR, and infused into the subject (Day 1) after receiving lymphodepleting chemotherapy (fludarabine 30 mg/m2/day and cyclophosphamide 600 mg/m2/day, on days -7, -6 and -5). The DLT observation period will be during the first 30 days following the infusion of MAGE-A10c796T for each patient in all groups. Up to 10 patients will be enrolled at the target dose. Disease assessments will be conducted at week 6, 12, 18 and 24, and then every 3 months until confirmation of disease progression. On study tumor biopsies and blood samples will be evaluated to compare the pre- and post-T cell infusion immune profile for association with treatment outcome. Clinical trial information: NCT02989064. [Table: see text]

TPS3094 Background: Epithelial ovarian cancer comprises the majority of malignant ovarian neoplasms (~80%) and is the leading cause of death from gynecologic cancer in the US. Due to lack of effective screening strategies, the majority (63%) of patients are diagnosed with ovarian cancer at advanced stages. New therapies are needed to address the unmet medical need of patients with ovarian cancer. 11-40% of ovarian cancers express NY-ESO-1 cancer testis/antigen. This study is evaluating affinity enhanced autologous NY-ESO-1c259T cells recognizing an NY-ESO-1 derived peptide complexed with HLA-A*02 in ovarian cancer. Methods: This single arm, open label clinical trial is evaluating safety and tolerability, antitumor activity (response rate by RECIST v1.1, progression free survival, overall survival, duration of response), and translational research endpoints. The study evaluates two lymphodepleting regimens: cyclophosphamide (enrolment completed; n = 7) and cyclophosphamide plus fludarabine (at least 10 subjects to be enrolled). Subjects must be ≥ 18 years old; HLA-A*02:01, *02:05, or *02:06 positive; have recurrent epithelial ovarian, primary peritoneal or fallopian tube carcinoma with refractory or platinum-resistant disease expressing NY-ESO-1 by IHC; have measurable disease; have ECOG status 0 or 1; and have adequate organ function. Following apheresis, the T cells are isolated and expanded with CD3/CD28 beads, transduced with a lentiviral vector containing the NY-ESO-1c259 TCR, and 1 – 6 × 109 transduced T cells are infused intravenously on Day 0 after lymphodepletion with fludarabine 30 mg/m2/day and cyclophosphamide 600 mg/m2/day on days -7 to -5. Response is assessed at weeks 4, 8, 12 and 24, and then every 3 months until confirmation of disease progression. Clinical trial information: NCT01567891.

TPS3097 Background: Myxoid/round cell liposarcomas (MRCLS) account for 6-10% of soft tissue sarcomas. Although a chemosensitive tumor, metastatic MRCLS has a poor prognosis and is inevitably fatal. More effective, durable and less toxic therapies are needed. NY-ESO-1 is a cancer/testis antigen that is expressed in 80-90% of MRCLS tumors. This study will evaluate the safety and efficacy of genetically engineered affinity enhanced autologous NY-ESO-1c259T cells recognizing an NY-ESO-1 derived peptide complexed with HLA-A*02 in MRCLS. Methods: This open label phase I/II non-randomized pilot study will evaluate efficacy (overall response rate by RECIST v1.1, time to response, duration of response, progression free survival, overall survival), safety, and translational research endpoints. Patients must meet these criteria: ≥ 18 yrs old; HLA-A*02:01, *02:05 or *02:06 positive; have advanced (metastatic or inoperable) MRCLS expressing NY-ESO-1 at 2+/3+ intensity in ≥30% of tumor cells by IHC; measurable disease; prior systemic anthracycline therapy; have ECOG status 0 or 1; and adequate organ function. Initially, ten patients are planned to be enrolled, with potential to enroll an additional 5 patients. Patients who do not receive the minimum cell dose or who do not receive the T-cell infusion may be replaced. Following apheresis, the T cells are isolated and expanded with CD3/CD28 beads, transduced with a lentiviral vector containing the NY-ESO-1c259 TCR, and 1– 8 × 109 transduced T-cells are infused intravenously on Day 1 after lymphodepletion with fludarabine 30 mg/m2/day and cyclophosphamide 600 mg/m2/day on days -7 to -5. Response is assessed at 4, 8, 12 and 24 weeks, and then every 3 months until confirmation of progression of disease. On study tumor biopsies and blood samples will be evaluated to compare the pre- and post-T cell infusion immune profile for association with treatment outcome. Clinical trial information: NCT02992743.

8006 Background: In 78 BRAF V600E mut NSCLC pts, single agent D induced an overall response rate (ORR) of 32%. The combination of D and T (DT) has demonstrated significant improvements in efficacy ...

4582 Background: No biomarker has been incorporated into clinically based prog algorithms for mRCC. In previous work in pts with ECOG PS ≤1 treated in a randomized placebo-controlled phase III tria...

4503 Background: Pazopanib and sunitinib are angiogenesis inhibitors approved for treatment of advanced RCC, but there is substantial heterogeneity in response to either treatment. We hypothesized ...

416 Background: The interaction of PDL1 (B7H1) with its receptor PD-1 on activated T cells contributes to suppression of antitumor immune responses. Tumor PDL1 expression has been associated with poor outcomes in RCC but has not been investigated as a biomarker of response in RCC patients receiving standard vascular endothelial growth factor (VEGF)-targeted therapy. Methods: Formalin-fixed paraffin-embedded tumor samples were collected at baseline from consenting patients enrolled in COMPARZ—a phase lll clinical trial comparing pazopanib and sunitinib as first-line interventions for metastatic RCC (Motzer et al, NEJM 2013). PDL1 expression was evaluated using the anti-PDL1 mouseIgG1 (clone 5H1; Thompson) on the Leica automated immunohistochemistry platform. Additional dual PDL1/CD68 staining was performed on tumor associated macrophages (TAMs). Tumor PDL1 expression was quantified by an H-score (HS). PDL1+ TAMs were assessed semiquantitatively. In addition, intra-tumor CD8+ T cells were quantified morphometrically. The association between PDL1 HS, CD8+T cell counts, and survival was investigated using Kaplan-Meier analysis. Results: HS data were available from 453 of 1110 patients. 64% of patients had negative (HS = 0) PDL1 expression (HS range 0-290), but PDL1 expression was associated with tumours containing higher numbers of infiltrating macrophages. Peripheral CD8+ T cells in the invasive margin surrounding the tumor were also observed. Patients with HS >50 (n = 61, 13%) had significantly shorter overall survival (OS) in both pazopanib (19.7 vs 31.6 mo) and sunitinib (15.3 vs 27.7 mo) arms. In both arms, patients with HS >50 with intratumoral CD8+T cell counts >300 had the shortest OS. Results were similar for progression-free survival and persisted on multivariate analysis. Conclusions: This is the largest report to show that tumors’PDL1 expression and CD8+ T cell counts are associated with treatment outcome in metastatic RCC patients. Increased PDL1, or increased PDL1 plus tumor CD8+ T cell counts, were associated with shorter OS. These findings may have major implications for future trial designs that involve PD-1 inhibitors.

4519 Background: Pazopanib and sunitinib are angiogenesis inhibitors approved for treatment of advanced renal cell carcinoma (RCC). COMPARZ, a phase III randomized clinical trial comparing pazopanib vs sunitinib for RCC, demonstrated similar efficacies for the two therapies but safety profiles differed. Our genetic analyses of previous pazopanib clinical trials found that IL8 polymorphisms may be associated with progression-free survival (PFS) and overall survival (OS). We attempted to validate these associations in the COMPARZ study. Methods: Of the 1110 participants in COMPARZ, 724 (65%) provided consent and DNA for pharmacogenetic analyses (pazopanib, N = 371; sunitinib, N = 353).Associations of IL8 polymorphisms (rs1126647 and rs4073) with PFS and OS were tested using the Cox proportional hazards model with baseline factors as covariates in a combined analysis of all patients and also separately in pazopanib-treated and sunitinib-treated patients. One-tailed P values were calculated for effects in the same direction as previously observed. Results: For PFS there was no significant association in the combined analysis or in pazopanib-treated patients, but there was a significant association in sunitinib-treated patients (P = 0.017). For OS there were significant associations in the combined analysis (P = 0.010) and in sunitinib-treated patients (P = 0.0043) but not in pazopanib-treated patients (P = 0.30). Hazard ratios (HRs) for genetic effects were not significantly different between sunitinib- and pazopanib-treated patients (two-tailed P = 0.23 for genotype-by-treatment interaction). Kaplan-Meier plots suggested a recessive genetic model in the combined data set, with median OS (95% CI) 23.7 months (15.4–29.1) for rs1126647 TT genotype compared to 35.5 months (30.8–∞) for AA or AT genotypes (HR = 1.66, P = 0.0007). Similar associations were seen for rs4073. Conclusions: Germline variants in IL8 are associated with survival outcome in patients with RCC who have received angiogenesis inhibitors. These findings may provide additional scientific insights in making treatment decisions and developing alternative therapies.

3021 Background: The interaction between PDL1 (B7H1) and its receptor PD-1 on activated T cells plays an important role in the inhibition of T-cell responses and contributes to suppression of antitumor immune responses. Tumor PDL1 expression has been associated with poor outcomes in RCC. This study investigates the correlation between PDL1 tumor expression and outcomes in RCC pts treated with paz. Methods: Using IHC, we retrospectively analyzed baseline FFPE tumor samples for PDL1 from 2 paz RCC studies: a single arm phase II trial and randomized placebo (pbo)-controlled phase III study. PDL1 expression was analyzed by MedTox Laboratories using the anti-PDL1 MouseIgG1 clone 5H1 (Thompson) on the Leica automated IHC platform. Additional dual PDL1/CD68 staining was carried out to delineate tumor and macrophage PDL1 expression. Tumor PDL1 expression was quantified by H-Score (HS) and PDL1+ macrophages were assessed semi-quantitatively. Association between PDL1H scores and PFS was investigated by Kaplan-Meier analysis using optimal cutoff of PDL1tumor HS (minimum p value, log rank test). Results: The optimal cut-point of PD-L1 tumor HS, relative to PFS, was identified as HS > 3. In the phase II study (46 available samples out of 225), HS range was 0-150 and most samples had negative (HS = 0, n = 34, 74%) or low (HS 1-3, n=4, 9%) PDL1 expression. Pts with HS > 3 (n = 8, 17%) had significantly shorter PFS (2.6 mo) than those with HS ≤ 3 (12 mo; p = .0005). In the phase III study (N = 160 available samples: paz, 113 of 290; pbo, 47 of 145), HS range was 0-280. Most patients had negative (n = 122/160, 76%) or low (n = 9/160, 6%) PD-L1 expression, with 18% (29/160) having HS > 3. Pbo-arm pts with HS > 3 (n = 6/47, 13%) had shorter PFS (2.3 vs 5.5 mo p = .0207). Paz-arm pts with HS > 3 (n = 23/113, 20%) trended toward shorter PFS (7.3 vs 11 mo, p = .1405). Conclusions: PDL1 appears to be a prognostic marker with PDL1 HS > 3 associated with shorter PFS. Limitations of the study include the retrospective nature of the analysis with limited pt samples available, low or negative PDL1 expression in the vast majority of pts, and use of archival samples that may not accurately reflect PDL1 status at study entry. Additional results (tumor volume, OS) will be presented.

4569 Background: A phase III randomized clinical trial (COMPARZ) comparing pazopanib vs sunitinib for treatment of advanced renal cell carcinoma demonstrated similar efficacies but different safety profiles for the two therapies. Elevations in serum total bilirubin have been observed in patients receiving either therapy. UGT1A1 polymorphisms are associated with elevated bilirubin in the general population (Gilbert’s syndrome). This study investigated the association between functional UGT1A1 polymorphisms and on-therapy serum total bilirubin in the COMPARZ study. Methods: Patients homozygous or compound heterozygous for UGT1A1 *28, *37, and *6 alleles were predicted to have reduced UGT1A1 function. Logistic regression adjusted for ancestry principal components was used to compare patients with on-therapy hyperbilirubinemia (≥1.5 × upper limit of normal [ULN]; pazopanib, N = 62; sunitinib, N = 34) against patients exposed to treatment and with maximum on-therapy bilirubin ≤1 × ULN (pazopanib, N = 213; sunitinib, N = 215), excluding patients with maximum on-therapy bilirubin between 1 and 1.5 × ULN (pazopanib, N = 96; sunitinib, N = 104). Results: Patients with predicted reduced UGT1A1 function had higher baseline bilirubin and also were more likely to experience hyperbilirubinemia when receiving either pazopanib (P = 6.9×10–8) or sunitinib (P = 1.8×10–3). After adjusting for baseline bilirubin, patients with predicted reduced UGT1A1 function remained more likely to experience hyperbilirubinemia when receiving pazopanib (P = 0.015) or sunitinib (P = 0.026), with odds ratio (95% CI) 3.53 (1.28–9.76) and 4.41 (1.23–15.75), respectively. Conclusions: The data suggest that some instances of hyperbilirubinemia in patients treated with pazopanib or sunitinib may be benign manifestations of Gilbert’s syndrome. Bilirubin fractionation or, if not available, UGT1A1 genotyping, would enable further characterization of liver safety risk and help in making treatment decisions.