Morales‐Barrera, Rafael

Abstract Purpose: The three-arm, phase III KEYNOTE-361 study did not meet its dual primary endpoints of progression-free survival (PFS) or overall survival (OS) with first-line pembrolizumab plus chemotherapy versus chemotherapy in advanced urothelial carcinoma. This prespecified exploratory analysis assessed the association of tumor mutational burden (TMB) and PD-L1 combined positive score (CPS) with clinical outcomes. Patients and Methods: TMB and PD-L1 CPS were determined via whole-exome sequencing and PD-L1 IHC 22C3 pharmDx, respectively. The association was evaluated in each treatment arm using logistic regression [objective response rate (ORR)] and Cox proportional hazards regression models (PFS and OS); one-sided (pembrolizumab monotherapy; pembrolizumab plus chemotherapy) and two-sided (chemotherapy) nominal P values were calculated. Significance was prespecified at α = 0.05 without multiplicity adjustment. Efficacy was evaluated by prespecified cutoffs of 175 mutations/exome (TMB) and CPS 10 (PD-L1). Results: Of the 993 treated patients, 820 (82.6%) and 993 (100%) had evaluable TMB and CPS data, respectively. Continuous TMB was positively associated with ORR, PFS, and OS for pembrolizumab monotherapy (one-sided P < 0.001, P < 0.001, and P = 0.007, respectively); PFS and OS for pembrolizumab plus chemotherapy (one-sided P = 0.007 and P = 0.010, respectively); and OS for chemotherapy alone (two-sided P = 0.040). Continuous PD-L1 CPS showed evidence of anticipated association with ORR and PFS for pembrolizumab monotherapy. The subgroup with TMB ≥175 mutations/exome and PD-L1 CPS ≥10 had the highest PFS and OS improvements with pembrolizumab alone or with chemotherapy versus chemotherapy alone. Conclusions: These data suggest that TMB may be predictive of the response to pembrolizumab alone or with chemotherapy in advanced urothelial carcinoma.

654 Background: MSI and TMB correlate with ICI efficacy. We assessed the association between TMB and MSI with outcomes in pts with aUC treated with anti-PD1/L1 in a ‘real-world’ cohort. We hypothesize that pts with high TMB and MSI-high (H) have better outcomes with ICI. Methods: In this retrospective study we included pts from 13 sites treated with ICI for aUC. We calculated overall response rate (ORR), median overall and progression-free survival (mOS, mPFS) using KM method from ICI start. TMB and MSI were assessed by NGS. TMB was analyzed as dichotomous (≥10 vs <10 mut/Mb) and continuous variable. The analysis was stratified by therapy line (1L/upfront, 2+L). Multivariable models were adjusted by Khaki factors for 1L/upfront and Bellmunt factors for 2+L. We separately analyzed pts on maintenance avelumab (mAV). MSI is reported with descriptive statistics only. Results: 342 pts were treated with ICI 1L/upfront or 2+L (69% pure UC, median age at ICI start 69, 77% men, 90% White, 16% upper tract primary, 12% liver metastases [mets], 78% ECOG PS 0-1). Median f/u from ICI start was 34 months (mo). ORR was 36% (95%CI 26-46%) in pts with TMB≥10 vs 30% (95%CI 23-36%) with TMB<10 (OR=1.4 [95%CI 0.8-2.5] p=0.3). ORR was 75% (6/8) in pts with MSI-H and 30% (60/200) with MSI stable (MSI-S). mPFS was similar between TMB≥10 and <10 groups (6 vs 4 mo; HR=0.88 [95%CI 0.61-1.26] p=0.48) and between MSI-H and MSI-S (5 mo). mOS was numerically longer (not significant) in pts with TMB≥10 vs TMB<10: 25 vs 21 mo; (HR=0.7 [95%CI 0.50-1.09] p=0.12) and with MSI-H and MSI-S (NR and 20 mo). A significant association was found between TMB (continuous variable) and OS (HR=0.97 [95%CI 0.95-0.99] p=0.01). TMB analyses by therapy line are shown in Table. In the 1L/upfront setting (N=144), ORR 71% (5/7) vs 33% (45/135), mPFS 5 and 4 mo, mOS NR and 22 mo for MSI-H and MSI-S, respectively. Pts on mAV (n=77) were 77% men, 92% White, 16% upper primary, 13% liver mets, 84% ECOG PS 0-1 with median f/u 15 mo from avelumab start; mOS was NR for MSI-H and 24 mo for MSI-S; mPFS was 4 mo for both groups. Of 3 pts on mAV with MSI-H: 1 CR, 2 SD; of 58 pts with MSI-S: 4CR, 4PR, 24 SD, 21 PD, 5 unknown. Conclusions: MSI-H had numerically higher ORR and higher TMB was associated with longer OS with ICI, but further validation is needed. Limitations: retrospective design, small sample size for MSI-H cases, lack of randomization and central scan review, selection/confounding. [Table: see text]

540 Background: The 3-arm, open-label, phase 3 KEYNOTE-361 study (NCT02853305) evaluated first-line pembro ± chemo vs chemo in advanced UC regardless of PD-L1 status. The trial did not meet its primary end points of superior PFS and OS with pembro + chemo vs chemo and thus analysis of pembro monotherapy (mono) vs chemo was exploratory. We explored the association of TMB status and PD-L1 combined positive score (CPS) with clinical outcomes in KEYNOTE-361. Methods: In pts with TMB and/or PD-L1 data, the association between TMB (via whole exome sequencing) and PD-L1 (via PD-L1 IHC 22C3 pharmDx) and clinical outcomes (ORR, PFS, and OS) was evaluated. In each treatment arm, the hypotheses regarding the associations were evaluated using logistic regression (ORR) and Cox proportional hazards regression (PFS; OS), and 1-sided (pembro; pembro + chemo) and 2-sided (chemo) P values were calculated; significance was prespecified at α = 0.05 without multiplicity adjustment. Clinical utility was assessed using prespecified cutoffs of 175 mut/exome (TMB) and CPS 10 (PD-L1). Clinical data cutoff was April 29, 2020. Results: 820/993 pts (82.6%) had evaluable TMB data (pembro, 252; pembro + chemo, 282; chemo, 286). TMB (log10) was significantly positively associated with ORR, PFS, and OS for pembro ( P < 0.001, < 0.001, and 0.007, respectively) and PFS and OS for pembro + chemo ( P= 0.007 and 0.010, respectively). The area under the receiver operating characteristics (AUROC) curve (95% CI) for discriminating response was 0.64 (0.56-0.71) for pembro, 0.53 (0.46-0.60) for pembro + chemo, and 0.52 (0.45-0.59) for chemo. Efficacy by TMB cutoff is reported in the Table. All 993 pts had PD-L1 data (pembro, 302; pembro + chemo, 349; chemo, 342). PD-L1 was significantly positively associated with PFS for pembro ( P= 0.006) and ORR for pembro + chemo ( P= 0.042) but not chemo. Efficacy by PD-L1 CPS is reported in the Table. Conclusions: Strong associations were observed between TMB and all 3 clinical outcomes (ORR, PFS, and OS) with pembro mono in the first-line setting and a reduced association was observed between TMB and clinical outcomes with pembro + chemo. No consistent associations were observed between PD-L1 and clinical outcomes with pembro mono or pembro + chemo. Clinical trial information: NCT02853305. [Table: see text]

Abstract Background: PD-L1 blockade with durvalumab (D) is approved as second-line therapy in platinum-refractory metastatic urothelial cancer (mUC). Adding CTLA-4 blockade with tremelimumab (T) may improve responses, particularly in pts with low PD-L1 expression. Here we report safety and efficacy in the mUC cohort of the dose-expansion phase of a phase I, multicenter, open-label study of D+T in pts with advanced solid tumors (NCT02261220). Methods: Pts with platinum-refractory mUC received D 20 mg/kg + T 1 mg/kg Q4W for 4 months, followed by D 10 mg/kg Q2W for a total of 12 months of treatment. The primary objectives were safety in the mUC cohort and antitumor activity in the subgroup with tumor cell and immune cell PD-L1 expression &lt;25% (assessed by SP263 IHC assay). Secondary objectives included antitumor activity per RECIST v1.1 in other subgroups. Results: As of Oct 20 2017, 168 pts received treatment and had ≥24 weeks' follow-up. Median age was 65.5 yr (range, 35-85), 78.0% were male, 81.0% had visceral metastases (32.1% liver metastases) and 32.1% had received &gt;1 prior line of chemotherapy. Median duration of follow-up was 11.6 mo. Treatment-related AEs occurred in 75.6% of pts and were Grade 3-4 in 28.6%. One pt died due to a treatment-related AE (pulmonary hemorrhage). Treatment-related AEs led to discontinuation of therapy in 11.9% of pts. Confirmed objective response rate (ORR) was 35/168 (20.8%) including 4 CRs. Responses occurred early (median 1.8 mo) and are durable (median DOR not reached, range 1.9-24.9 mo). At 6 months, the PFS rate was 25.4% and the OS rate was 60.9%. Median PFS was 1.9 mo and median OS was 9.5 mo. Clinical activity was seen regardless of PD-L1 status, but patients with either tumor or immune cell expression ≥25% had numerically higher response rates (29.4% vs 15.1%) and 6-month OS rates than those with &lt;25% expression (Table). Conclusions: D+T had a manageable safety profile and encouraging antitumor activity and survival rate in previously treated mUC regardless of PD-L1 status. Response and survivalPD-L1 ≥25% (n=68)PD-L1 &lt;25% (n=86)PD-L1 unknown (n=14)Total (N=168)Confirmed ORR (CR+PR)(95% CI), %29.4 (19.0-41.7)15.1 (8.3-24.5)14.3 (1.8-42.8)20.8 (15.0-27.8)Ongoing ORR, %60.092.310074.3Disease control rate (CR+PR+SD≥24 weeks) (95% CI), %32.4 (21.5-44.8)24.4 (15.8-34.9)42.9 (17.7-71.1)29.2 (22.4-36.7)Median PFS, months (95% CI)3.5 (1.9-3.7)1.8 (1.8-1.9)4.9 (1.8-NE)1.9 (1.8-3.4)PFS 6-month rate, %26.122.638.525.4Median OS, months (95% CI)*18.9 (8.1-NE)8.0 (4.8-10.0)16.4 (7.3-16.4)9.5 (8.1-18.9)OS 6-month rate, %66.451.991.760.9NE, not estimable *Should be interpreted with caution due to limited follow-up at data cutoff Citation Format: Arjun V. Balar, Amit Mahipal, Enrique Grande, Victor M. Villalobos, Sebastien Salas, Taek Won Kang, Se Hyun Kim, Thomas Powles, Frank Tsai, Aung Naing, Albiruni Razak, Yohann Loriot, Ji Youl Lee, Sang Joon Shin, Rafael Morales-Barrera, Natasha Angra, Feng Xiao, Shaad Abdullah, Michiel S. Van der Heijden. Durvalumab + tremelimumab in patients with metastatic urothelial cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr CT112.