Search

4520 Background: In the phase 3 JAVELIN Bladder 100 trial, avelumab 1L maintenance + best supportive care (BSC) significantly prolonged overall survival (OS) vs BSC alone in patients (pts) with advanced UC that had not progressed on 1L platinum-based chemotherapy (HR, 0.69 [95% CI: 0.56, 0.86; 1-sided P= 0.0005]). We report post hoc analyses in previously unreported clinical and genomic subgroups. Methods: In JAVELIN Bladder 100 (NCT02603432), eligible pts had unresectable locally advanced or metastatic UC without progression after 4-6 cycles of 1L gemcitabine + cisplatin or carboplatin, and were randomized to receive avelumab + BSC (n = 350) or BSC alone (n = 350). The primary endpoint was OS, in all randomized pts and pts with PD-L1+ tumors (Ventana SP263 assay). In this exploratory analysis, we analyzed OS in disease stage and site subgroups, in pts with PD-L1+ tumors who received 1L gemcitabine + carboplatin, and in genomic subtypes (RNAseq whole-transcriptome profiling of tumor tissue) defined using data from The Cancer Genome Atlas (TCGA 2017). Interaction tests were not performed. Results: Prolonged OS was observed in the avelumab + BSC arm vs the BSC alone arm in pts with upper or lower tract tumors, metastatic or locally advanced (LA) and unresectable disease (prior to chemotherapy), and lymph node-only disease post-chemotherapy (Table). OS was also prolonged with avelumab + BSC in pts in PD-L1+ tumors who had received 1L gemcitabine + carboplatin, consistent with findings in the overall population. In genomic subtypes, the OS benefit for avelumab + BSC was apparent across TCGA subtypes except luminal. Conclusions: An OS benefit was seen for avelumab 1L maintenance + BSC vs BSC alone across subgroups of interest. Results are consistent with previously reported findings, further supporting avelumab 1L maintenance as a standard of care for pts with advanced UC that has not progressed with 1L platinum-containing chemotherapy. Clinical trial information: NCT02603432. [Table: see text]

425 Background: A randomized phase III trial (JAVELIN Bladder 100; NCT02603432) to investigate avelumab (anti–PD-L1) as 1L maintenance therapy in patients with advanced UC met its primary objective, demonstrating significantly prolonged overall survival (OS) with Ave + BSC vs BSC alone in all randomized patients and in patients with PD-L1+ tumors. We report efficacy and safety in Japanese patients enrolled in this study. Methods: Eligible patients with unresectable locally advanced or metastatic UC that had not progressed with 4-6 cycles of gemcitabine with either cisplatin or carboplatin were randomized 1:1 to receive maintenance Ave (10 mg/kg IV every 2 weeks) + BSC or BSC alone, stratified by best response to 1L chemotherapy (complete/partial response vs stable disease) and by visceral vs nonvisceral disease when initiating 1L chemotherapy. The primary endpoint was OS, assessed from randomization in all randomized patients and in patients with PD-L1+ tumors (Ventana SP263 assay). Secondary endpoints included progression-free survival (PFS) per blinded independent central review and safety. Results: Japanese patients (n=73) were randomized to receive Ave + BSC (n=36) or BSC alone (n=37); 52.8% vs 62.2% had PD-L1+ tumors, respectively. Median OS (95% CI) was 24.7 months (18.2-not estimable [NE]) with Ave + BSC vs 18.7 months (12.8-33.0) with BSC alone (HR, 0.81 [95% CI; 0.409-1.585]) in all randomized patients and 18.6 months (9.4-NE) with Ave + BSC vs 19.4 months (11.7-33.0) with BSC alone (HR, 1.00 [95% CI, 0.413-2.412]) in patients with PD-L1+ tumors. Median PFS (95% CI) was 5.6 months (1.9-9.4) with Ave + BSC vs 1.9 months (1.9-3.8) with BSC alone (HR, 0.63 [95% CI, 0.358-1.113]) in all randomized patients and 5.6 months (1.8-11.2) with Ave + BSC vs 1.9 months (1.9-3.8) with BSC alone (HR, 0.62 [95% CI, 0.298-1.301]) in patients with PD-L1+ tumors. The most common treatment-emergent adverse events (all grade; grade ≥3) in the Ave + BSC arm were pyrexia (10 [27.8%]; 0), nasopharyngitis (7 [19.4%]; 0), and anemia (7 [19.4%]; 4 [11.1%]). Conclusions: Ave 1L maintenance + BSC was efficacious and tolerable in Japanese patients with advanced UC, and results were generally consistent with those in the overall population. Clinical trial information: NCT02603432.

431 Background: Bacillus Calmette-Guerin (BCG) intravesical instillation is the most effective immunotherapy for non-muscle-invasive bladder cancer (NMIBC), however there are few reliable markers to elucidate the efficacy of BCG therapy. Germline copy number polymorphisms (CNPs) are expected to affect various diseases including human malignancies, but the significance of CNPs in NMIBC treated with BCG therapy remains unclear. FAM81A located on 15q22.2 was reported as one of tumor-associated ETS shared target genes in prostate cancer. PCSK6 located on 15q26.3 was reported to regulate proliferation and tumor progression in several cancers. The purpose of this study is to determine the prognostic value of CNPs for NMIBC treated with BCG therapy. To our knowledge, this is the first report to confirm CNPs as a potential biomarker for assessing the efficacy of immunotherapy. Methods: Array comparative genomic hybridization (CGH) was performed to search for candidate whole genome-wide CNPs related to NMIBC susceptibility. Next, quantitative real-time polymerase chain reaction was carried out to evaluate the effect of BCG therapy for 57 Japanese patients with NMIBC treated with BCG intravesical instillation. Results: Eleven CNPs were associated with NMIBC risk in array CGH. FAM81A and PCSK6 copy number according to those CNPs examined showed significant relationship with disease progression in NMIBC treated with BCG. The means of the relative copy numbers of patients with CNP and those without it were 1.58 and 2.10 for FAM81A ( P < 0.0001), and 1.06 and 1.80 for PCSK6 ( P < 0.0001), respectively. Univariate Cox proportional hazards regression analysis showed that FAM81A ( P = 0.0022), and PCSK6 ( P = 0.0147) copy number had a significant effect on progression-free survival. In multivariate analyses, FAM81A copy number was an independent prognostic factor for progression-free survival ( P = 0.0419, RR = 7.59, 95% CI, 1.07–153.42). The combination of FAM81A or PCSK6 CNP was the most significant prognostic biomarker to predict the efficacy of BCG therapy for NMIBC ( P = 0.0002). Conclusions: Germline DNA CNPs may be a potential new biomarker for estimating the efficacy of BCG therapy in Japanese patients with NMIBC.

724 Background: Nivolumab is a standard treatment for previously treated advanced renal-cell carcinoma (RCC). However, nivolumab is effective in only a limited number of patients; therefore, we evaluated the prognostic value of several biomarkers, including inflammation-based prognostic scores and changes in these scores following nivolumab treatment in Japanese patients with metastatic RCC. Methods: We retrospectively reviewed the medical records of 65 patients with previously treated metastatic RCC and who received nivolumab. MSKCC and IMDC risk, inflammation-based prognostic scores, including neutrophil/lymphocyte ratio (NLR), platelet/lymphocyte ratio (PLR), lymphocyte/monocyte ratio (LMR), and Glasgow prognostic score before and 6 weeks after the treatment were recorded. Categorical variables influencing disease-specific and overall survival were compared using Cox proportional-hazards regression models. Results: Univariate analysis showed that MSKCC risk score ( P = 0.0052), lactate dehydrogenase (LDH) ( P = 0.0266), LMR ( P = 0.0113), and PLR ( P = 0.0017) had a significant effect on disease-specific survival. Multivariate analyses showed that PLR and LDH were found to be independent prognostic factors for disease-specific survival ( P = 0.0008, RR = 7.95, 95% CI, 2.16–51.64 and P = 0.0123, RR = 3.92, 95% CI, 1.37–10.80, respectively). The combination of PLR and LDH was the most significant prognostic biomarker in metastatic RCC for disease-specific ( P < 0.0001) and overall ( P < 0.0001) survival. Changes in LMR and PLR in response to nivolumab were significant prognostic factors for disease-specific survival ( P < 0.0001 and P = 0.0477, respectively). Conclusions: The combination of PLR and LDH may be a potential biomarker for estimating disease-specific and overall survival in Japanese patients with metastatic RCC treated by nivolumab. If changes of inflammation-based prognostic scores in response to nivolumab treatment might be improved, nivolumab treatment should be continued.

516 Background: Radical cystectomy is still the standard treatment for muscle-invasive bladder cancer (MIBC), while the patients with MIBC are not always appropriate candidates due to multiple comorbidities. We establish novel treatment strategy by trimodal treatment. Methods: The regimen was gemcitabine 300 mg/m2, and cisplatin 30mg/m2 in day 1 and concomitant irradiation 2Gy/Fr, 5 fraction per week. Irradiation was administered to whole pelvis up to 30Gy, then boost to true pelvis until total 48Gy to 54Gy. Extensive transurethral resection (TURBT) was performed and we confirmed pathological stage ≥T2. TURBT was also performed after chemoradiotherapy to evaluate the pathological response to treatment. This study was approved in our institutional review board (ID: H23-89) and the information was opened on UMIN (ID: UMIN000006363). We analyzed their treatment efficacy and survival. Results: The patients were 29 men and 9 women, median age was 76.5 y.o. and median follow up was 23 months (1 - 112). Clinical stage T2, T3, T4, N1 and N2 were 23, 10, 5, 4, 2 cases, respectively. The 2- and 5-year metastatic-free survival (MFS), bladder-recurrence free survival (bRFS), cancer-specific survival (CSS), and overall survival (OS) rates after treatment were 91.7 and 84.0%, 59.7 and 42.6%, 87.3 and 87.3%, and 87.3 and 81.8%, respectively. Salvage cystectomy was performed 3 patients and they were still alive. CR rate was 78.9% and overall response rate was 92.1%. cT stage and valiant histology was not associated with treatment response. The patients achieved CR had significant good prognosis in CSS (p=0.0149) and OS (p=0.0149) compared with non-responders. In cox hazard model, CR achievement was significant prognostic factors for OS (p =0.0015, HR 6.804e+26, 95% CI 56.94-1.631e+86). Patients were able to receive 3 to 5 cycle GC radiation and any grade 3 or more adverse event was 7 (18.4%) cases. no treatment related death was recorded. Conclusions: In selected patients, GC radiation for MIBC may provide good oncological outcomes as bladder preservation strategy.

625 Background: Although titrated-dose axitinib offers a good objective response rate (ORR) for metastatic renal cell carcinoma (RCC), its optimal initial dose is unclear because its dosing regimen is based on a hypertensive reaction. We investigated whether axitinib pharmacogenetics were related to clinical efficacy/adverse events, and calculated a model to predict clinical efficacy and adverse events, using pharmacokinetic data based on gene polymorphisms in a phase 2 trial. Methods: We prospectively evaluated objective response and adverse events to establish a prediction model in 44 consecutive patients with advanced RCC, treated with axitinib between October 2013 and March 2017. Gene polymorphisms, including ABC transporters ( BCRP and MDR1) and UGT1A were analyzed by whole-exome sequencing, Sanger sequencing, and DNA chips. To construct the prediction model for area under curve (AUC), we used an exponential regression model with gene polymorphisms and dosage as covariates. To further validate this prediction model, we prospectively compared the calculated AUC in this model with actual AUCs in 13 additional consecutive patients. Results: Actual AUC was significantly correlated with the best ORR ( P = 0.0002), and adverse events (grade 2–3 hand–foot syndrome [ P = 0.0055]; grade 2 hypothyroidism [ P = 0.0381]); it was also significantly correlated with calculated AUC ( P < 0.0001). Calculated AUC was also significantly associated with best ORR ( P = 0.0044), grade 2–3 hand–foot syndrome ( P = 0.0191) and grade 2 hypothyroidism ( P = 0.0082). Surprisingly, hypertension was associated with neither ORR nor AUC. In the validation study, calculated AUC before axitinib treatment precisely predicted actual AUC ( P = 0.0079) in 13 additional consecutive patients. Conclusions: Our pharmacogenetics-based AUC model may offer a more benign method to determine optimal initial axitinib doses than hypertension, and could contribute to more precise treatment of individuals with advanced RCC. Clinical trial information: UMIN000011147.

TPS4590 Background: Only 5%-15% of patients (pts) with advanced bladder cancer attain long-term survival with standard first-line cisplatin-based chemotherapy. Programmed death 1 (PD-1)/PD-L1 inhibitors have proven effective in recurrent, advanced urothelial cancer. Emerging data suggest these agents may also be useful in the first-line setting. In KEYNOTE-052, first-line pembro, an anti–PD-1 antibody, demonstrated antitumor activity and acceptable safety in cisplatin-ineligible pts with advanced urothelial cancer. KEYNOTE-361 (NCT02853305) is a randomized, open-label, phase 3 study of pembro with or without chemotherapy versus chemotherapy alone in pts with advanced urothelial carcinoma. Methods: Key eligibility criteria include age ≥18 years; histologically or cytologically confirmed unresectable/metastatic urothelial carcinoma of the renal pelvis, ureter, bladder, or urethra; measurable disease (RECIST v1.1, investigator review); no prior systemic chemotherapy ([neo]adjuvant platinum-based chemotherapy with recurrence > 12 months after completion is allowed); ECOG PS 0-2; and provision of a tumor sample for biomarker analyses. Pts will be randomly assigned 1:1:1 to receive pembro 200 mg every 3 weeks (Q3W), pembro + investigator’s choice of chemotherapy (gemcitabine [1000 mg/m2 on day 1 and 8 Q3W] + cisplatin [70 mg/m2 Q3W]), or chemotherapy alone. Cisplatin-ineligible pts randomly assigned to chemotherapy will receive gemcitabine + carboplatin [AUC 5 Q3W]. Chemotherapy choice must be selected before randomization. Treatment will continue until progressive disease, unacceptable adverse events (AEs), or 35 cycles of pembro (pembro arms only). Response will be assessed Q9W for the first year and Q12W thereafter. AEs will be evaluated throughout and graded per NCI CTCAE v4.0. Primary end points are progression-free survival (RECIST v1.1 per central review) and overall survival; secondary end points include objective response rate and safety and tolerability. Efficacy outcomes will be compared for pembro vs chemotherapy and pembro + chemotherapy vs chemotherapy. Enrollment is ongoing; ~990 pts will be enrolled. Clinical trial information: NCT02853305.

542 Background: Chronic kidney disease (CKD) is a risk factor for the cardiovascular disease, which affect to the patients’ survival, while an EORTC randomized control trial did not show superiority of partial nephrectomy (PN) against localized renal cell carcinoma (RCC) in overall survival compared to radical nephrectomy (RN). We aimed to evaluate the role of operative methods affecting the survival and tried to estimate a predictive model for high risk CKD after surgery. Methods: We reviewed the data of 357 cases with clinical T1 RCC treated by RN (292 cases, RN group) or PN (65 cases, PN group) at Yamaguchi University Hospital or its related hospitals. We supposed the cases with CKD stage 3b or higher after surgery as high risk CKD. And we set the primary endpoints as the ratio of the cases with high risk CKD and overall survival after surgery. Results: Median follow-up period after surgery was 70 months (3-161). Statistically significant difference in performance status and clinical T stage were observed between the groups, but not in other patients’ characteristics. Mean values of preoperative eGFR were 69.2 and 65.5 ml/min/1.73m2 in RN and PN group, which decreased to 46.0 and 57.9 at 5 years after surgery, respectively. There was a significant difference in the incidence of high risk CKD between RN (39.3%) and PN group (2.2%) at 5 years after surgery. During follow up period, 17 cases (4.4%) were inducted to dialysis, there was no difference in the incidence between the RN and PN group. Multivariate analysis showed that eGFR ( < 72 ml/min/1.73m2; Odds ratio 15.3), proteinuria (Odds ratio 3.84), smoking (Odds ratio 2.76), BMI ( > 23; Odds ratio 2.66) and age ( > 67 years old; Odds ratio 2.47) could be significant predictive factors for high risk CKD at 5years after surgery. Our predicting model for high risk CKD showed 86.8% of sensitivity and 74.8% of specificity. But there was no significant difference in overall survival between the RN and PN group. Conclusions: Although there was a significant difference in the incidence of high risk CKD between RN and PN group, operative methods did not affect to the survival. Postoperative high risk CKD could be predictable by preoperative clinical factors.

328 Background: The patients with advanced upper tract urothelial carcinoma (UTUC) have poor prognosis. Previous reports have not shown that neoadjuvant chemotherapy for UTUC improves patient survival and it is difficult to select appropriate candidates for neoadjuvant chemotherapy. We aimed to predict appropriate candidates for neoadjuvant chemotherapy. Methods: We reviewed the data of 536 patients with UTUC undergone total nephroureterctomy in the cohort of Yamaguchi Uro-Oncology Group (Group 1) and 150 cases in the cohort of Osaka Medical College (Group 2). We performed a multivariate analysis to the Group 1 to decide independent predictive factors and validated them in another cohort (Group 2). We evaluated the clinical significance of the predictive factors in cancer-specific survival (CSS) and overall survival (OS) between the cohort with and without perioperataive chemotherapy. Results: Multivariate analysis showed ≥cT3, cN+ and elevated CRP (≥0.31mg/dl) could become independent prognostic factors in Group 1, which showed 0.70 of AUC in predicting cancer death. We validated these factors in Group 2 and confirmed that they also showed 0.78 of AUC in predicting cancer death. Then, we divided the patients into 4 risk groups based on the number of the predictive factors: score 0 (202 cases), score 1 (152 cases), scores 2 (51 cases) and scores 3 (30 cases). Kaplan-Meier Curve showed that 5 year CSS in score 0, 1, 2, and 3 were 84.7%, 67.7%, 32.2% and 22.0%, respectively. Finally, we analyzed the difference in CSS and OS between the 138 cases with and 296 cases without perioperative chemotherapy. A significant difference in CSS and OS between the cases with and without perioperative chemotherapy was shown only in the cases with all the predictive factors: score 3. Additionally, fraction of patients with lymphovascular invasion of the surgical specimen were significantly increased with score advance. Conclusions: Survival in the patients with UTUC was predictable by preoperative factors. The patients with all the predictive factors, ≥cT3, cN+ and elevated CRP, were supposed to be the best candidates for neoadjuvant chemotherapy against UTUC.

256 Background: Clusterin (CLU) is induced by androgen receptor (AR) pathway inhibition and its overexpression confers treatment resistance. Lipid nanoparticle (LNP) facilitates tumor uptake and intracellular processing through an enhanced permeation and retention effect (EPR), currently with multiple products undergoing clinical evaluation. Gene silencing using small interfering RNA (siRNA) is a promising approach but in vivo delivery remains a major barrier. In our study, we investigated the efficacy siRNA tumor delivery using LNP systems in enzalutamide-resistant (ENZ-R) castration-resistant prostate cancer (CRPC) model. Methods: To validate the effect of LNP siRNA tumor delivery in vivo, gene silencing of a reporter gene, luciferase (LUC), in PC3-M-luc stable cell line was treated with LNP LUC-siRNA and examined for Luc activity by the IVIS imaging system. Next, we investigated the efficacy of LNP CLU-siRNA tumor delivery and LNP CLU-siRNA sensitized AR knockdown activity in ENZ-R CRPC LNCaP in vitro and in vivo models. Results: LNP LUC-siRNA exhibited LUC silencing effects in PC-3M-luc xenograft and metastatic models. LNP CLU-siRNA suppressed PSA and decreased AKT and ERK phosphorylation in ENZ-R LNCaP cells in vitro. LNP CLU-siRNA sensitized AR antisense oligonucleotides (ASO) activity, more potently inhibiting ENZ-R cell growth rates and increased apoptosis when compared to AR-ASO monotherapy. In vivo,combinatory treatment of LNP CLU-siRNA and AR-ASO significantly suppressed tumor growth and serum PSA levels compared to LNP LUC -siRNA (control) and AR-ASO in ENZ-R LNCaP xenografts. Conclusions: The LNP CLU-siRNA systems sensitized AR knockdown resulting in inhibition of ENZ-R LNCaP cell growth, providing pre-clinical proof-of-principle as a promising AR co-targeting approach in ENZ-R CRPC.

94 Background: MDV3100 is a potent androgen receptor (AR) antagonist with activity in castration resistant prostate cancer (CRPC); however, progression to MDV3100-resistant (MDV-R) CRPC frequently occurs with rising serum PSA levels, implicating AR full length or variants in disease progression. We studied the activity of Generation 2.5 antisense oligonucleotide (ASO) targeting the AR full length (ARfl) and splice variants in MDV-R CRPC models. Methods: and Results: ThreeASOs targeting exon 1, intron 1, or exon 8 were designed to suppress ARfl and known AR splice variants. We generated by selection MDV-R LNCaP-derived sub-lines that uniformly expressed high levels of both ARfl and AR-V7 compared to CRPC LNCaP cell lines. MDV-3100 induced time- and dose-dependent increases in ARfl and AR-V7 protein levels; ARfl levels were ~20-fold higher than AR-V7. All 3 AR-ASO decreased ARfl and PSA expression. Exon 1 ASO decreased expression of both ARfl and AR-V7 in MDV-R-LNCaP cells; in contrast, exon 8 ASO decreased ARfl without reducing AR-V7 levels. Exon 1 ASO also most potently suppressed ARfl and splice variants in M12 cells stably overexpressing AR splice variants AR-V7 and AR-V567es. Despite these differential effects on ARfl and splice variant knockdown, the AR ASO similarly inhibited cell growth and induced apoptosis and G1 cell cycle arrest in LNCaP-derived CRPC and MDV-R cell lines. In 22RV-1 cells (which express endogenous ARfl and AR-V7), exon 1 ASO more potently suppressed ARfl and AR-V7 levels, AR transcriptional activity and AR-regulated gene expression compared to exon 8 ASO, but inhibition of cell growth did not differ significantly. Exon 1 ASO was evaluated in vivo in MDV-R49F CRPC LNCaP xenografts; mean tumor volume and serum PSA levels decreased significantly by 40% and 50%, respectively, compared to controls. Conclusions: While MDV-3100 induces both ARfl and AR-V7 levels, the biologic consequences appear cell line dependent and mainly driven by ARfl. AR-ASO knockdown of ARfl and its splice variants suppresses MDV-R LNCaP tumor growth, providing pre-clinical proof of principle to support clinical evaluation in post-AR pathway inhibitor CRPC.