Investigators Examine Biomarkers of Response to Avelumab in Urothelial Carcinoma

September 20, 2020

Biomarkers of immune activity were linked with overall survival benefit to frontline maintenance avelumab in patients with advanced urothelial cancer, according to data presented at the ESMO Virtual Congress 2020 from an exploratory analysis of the phase 3 JAVELIN Bladder 100 trial.

Biomarkers of immune activity were linked with overall survival (OS) benefit to frontline maintenance avelumab (Bavencio) in patients with advanced urothelial cancer, according to data presented at the ESMO Virtual Congress 2020 from an exploratory analysis of the phase 3 JAVELIN Bladder 100 trial (NCT02603432).1

Results showed that none of the established biomarkers that were evaluated, either alone or in combination, were capable of optimally predicting OS benefit with maintenance avelumab; this included elevated PD-L1 on either tumor cells (TCs) or immune cells (ICs), and were irrespective of tumor mutational burden (TMB) status or elevated TMB and mutation signatures, regardless of PD-L1 status.

Moreover, additional biomarker candidates were identified that beg for further investigation. For example, investigators identified gene expression signatures linked with various immune cell types, including innate and adaptive effector cells. They discovered several alleles encoding high-infinity FcγRIIA and FcγRIIA variants, which could potentially indicate FcR-mediated antitumor mechanisms. Gene signatures of tumor growth–promoting pathways, such as angiogenesis, hedgehog, Notch, and TGFβ, were also recognized; these signatures could potentially reduce antitumor immune responses.

“The JAVELIN Bladder 100 study enabled an exploratory analysis of predictive biomarkers that may help to identify patients who would obtain an OS benefit with avelumab first-line maintenance plus best supportive care (BSC) versus best supportive care alone,” Srikala S. Sridhar, MD, MSc, FRCPC, co-investigator and clinician investigator in the Cancer Clinical Research Unit at Princess Margaret Cancer Centre, said in a presentation during the congress. “Multiple other biomarkers were identified and potentially these could be explored in future research.”

The PD-L1 inhibitor has a native IgG1 isotype backbone that binds Fc receptors. Results from the phase 3 JAVELIN Bladder 100 trial showed that first-line maintenance treatment with avelumab plus BSC resulted in a significant OS benefit in patients with advanced urothelial carcinoma that had not progressed on first-line platinum-based chemotherapy.2

Biomarker analyses are needed to adequately characterize disease and identify opportunities for novel treatments to improve outcomes for patients, said Sridhar. Although PD-L1 protein expression, TMB, and gene expression signatures have been considered to be candidate biomarkers, they have shown variable predictive utility, which has limited their application to patient selection.

Because the JAVELIN Bladder 100 trial was the first to compare an immune checkpoint inhibitor against BSC and demonstrate a survival advantage, investigators had a unique opportunity to examine the predictive utility of several emerging biomarkers in urothelial carcinoma, added Sridhar.

In the trial, 700 patients with unresectable locally advanced or metastatic urothelial cancer who had achieved a complete response, partial response, or stable disease with standard frontline chemotherapy comprised of either cisplatin plus gemcitabine or carboplatin plus gemcitabine were randomized 1:1 to avelumab plus BSC (n = 350) or BSC alone (n = 350). Participants received treatment until disease progression, intolerable toxicity, or withdrawal.

The primary end point of the trial was OS and the primary analysis populations included all patients who underwent randomization as well as a subgroup of patients with PD-L1 positivity. Results presented during the ASCO 2020 Virtual Scientific Program showed that the median OS in the randomized patient population was 21.4 months in the avelumab arm (95% CI, 18.9-26.1) versus 14.3 months in the control arm (95% CI, 12.9-17.9; HR, 0.69; 95% CI, 0.56-0.86; P <.001).

During the ESMO Virtual Congress 2020, Sridhar presented the initial data from a preplanned exploratory analysis that examined the association between biomarkers, treatment, and outcomes in the entire intent-to-treat (ITT) population.

To generate hypotheses pertaining to the predictive utility of tumor biomarkers, investigators collected tumor tissue from primary or metastatic lesions before patients received first-line chemotherapy. These tissue samples underwent biomarker analysis per IHC or next-generation sequencing. Then, investigators conducted ad hoc exploratory analyses of the associations between biomarkers found in the tumors and OS.

Prespecified thresholds were utilized to examine associations between PD-L1 expression by TC or IC, and median biomarker values were used to set exploratory thresholds to decrease imbalance between the subgroups analyzed.

“As is common with many biomarker studies, not all samples were evaluable for all biomarkers,” explained Sridhar. “These analyses were all retrospective and exploratory and P values were not corrected for multiple comparisons. As such, they do not represent prospective demonstration of statistical significance or clinical utility.”

PD-L1 Expression

First, the investigative team examined the association between PD-L1 expression and OS benefit with avelumab. Results showed that the prespecified PD-L1–positive population (25% or greater PD-L1 positivity on TC and/or IC; n = 358) met its primary end point for OS, with a HR of 0.56 (95% CI, 0.40-0.79). “I’ll remind you that the HR in the ITT population was 0.69,” said Sridhar.

When looking at the TC and IC subgroups, investigators noted that they both enhance for outcomes. In the TC subgroup (n = 113), the HR was 0.35 (95% CI, 0.186-0.653) and the HR was 0.61 (95% CI, 0.429-0.867) in the IC subgroup (n = 326).

“In the TC subgroup, it may appear that the HR is lower, and the median OS is greater, but know that the BSC arm here doesn’t perform as well and there were fewer patients in this subgroup,” explained Sridhar. “In the IC subgroup, the HR is similar to the prespecified population. What we take away from this is that neither PD-L1–positive TC nor IC alone fully predicted OS or could identify all responders.”

The Role of TMB

When examining the OS benefit with maintenance avelumab in subgroups of patients defined by TMB and PD-L1 status, they found that neither markers alone could fully predict for OS benefit.

In the exploratory analysis, investigators used a median of 7.66 nonsynonymous SNVs/Mb to define subgroups compared with the established threshold of 10, noted Sridhar. Patients who had TMB above the median experienced an improvement in OS (HR, 0.48; 95% CI, 0.332-0.707); this was observed even in patients overall who were PD-L1 negative (HR, 0.85; 95% CI, 0.616-1.181).

“Similarly, in patients who were TMB low, we also saw a benefit in those who were PD-L1 positive,” said Sridhar. “This suggests that neither TMB nor PD-L1 status alone could fully predict OS benefit.”

Examining Mutation Profiles

Investigators then scored tumors according to mutation signatures from the COSMIC database and found that signatures that resulted from a C > T were linked with lower HRs. When examining DNA damage repair genes implicated by the Memorial Sloan Kettering Cancer Center group, the genes that were mutated also had lower HRs, although this was not corrected for TMB.

“Not only is TMB important, but also the type and location of mutations plays a critical role,” said Sridhar. “Both [of these factors] may influence utility of TMB assessment.”

Tumor Gene Expression and Immune Cell Gene Expression

When examining the association between gene expression and OS benefit with avelumab, investigators found that immune-related genes were associated with OS benefit.

“In our volcano plots, we saw a striking number of immune-related genes that crossed the prespecified threshold of 0.001 [in the avelumab plus BSC samples] compared with BSC [samples],” said Sridhar. “What is interesting is that a number of these genes are components of both the innate and adaptive immune response and some of the genes have previously been shown in other biomarker studies of immune checkpoint inhibitors.” The top 5 genes expressed included CD8, CXCL9, IFNG, LAG3, and TIGIT.

When looking more closely at the association between immune cell gene expression signatures and OS with avelumab, investigators noted that many different components of the immune system appeared to play a role in benefit. “This can potentially explain why TMB and PD-L1 alone doesn’t explain the whole story,” noted Sridhar.

Several immune cells expressing the Fc receptor were revealed during the analysis. It has previously been observed that FCGR2A and FCGR3A are able to bind with variable affinity to the IgG antibody, according to Sridhar. Other data have demonstrated that with other agents like cetuximab (Erbitux) the presence of high affinity FcγR alleles could be linked with OS benefit.

To evaluate this further, investigators assigned patients into subgroups based on whether they had 2 or more high affinity alleles; patients who did, experienced a more favorable OS benefit with avelumab (HR, 0.53; 95% CI, 0.384-0.732) versus those who had less than 2 (HR, 1.03; 95% CI, 0.682-1.548).

“This raises a number of questions about whether FcγR association plays a key role and if this is specific to certain immune checkpoint inhibitors,” said Sridhar. “I think this certainly warrants further investigation.”

Evaluating Established Immune Active Gene Signatures

Lastly, investigators explored the relationships between 2 established immune active gene signatures and OS outcomes. Specifically, they looked at JAVELIN-Immuno, which is a 26-gene signature that had been evaluated in a trial with avelumab plus axitinib (Inlyta) in renal cell carcinoma, and the T-cell inflamed 18-gene signature that had been defined in a total of 200 patients with 9 tumor types.

Patients were assigned to subgroups based on whether they fell above or below a median score. Both of the gene signatures appeared to predict for outcome in those who fell above the median, according to Sridhar. The HR was 0.55 with the JAVELIN-Immuno signature (95% CI, 0.383-0.796) and 0.49 with the T-cell inflamed signature (95% CI, 0.332-0.719) in the subgroups of patients above the median.

“The JAVELIN Bladder 100 trial led to the approval of first-line maintenance avelumab in patients with advanced urothelial carcinoma independent of PD-L1 expression,” concluded Sridhar. “Combinations of biomarkers may help to further characterize patients with advanced urothelial carcinoma who will benefit from avelumab first-line maintenance.”


  1. Powles T, Loriot Y, Bellmunt J, et al. Avelumab first-line (1L) maintenance + best supportive care (BSC) vs BSC alone for advanced urothelial carcinoma (UC): association between clinical outcomes and exploratory biomarkers. Presented at: 2020 ESMO Virtual Congress; September 19-21, 2020; Virtual. Abstract 699O.
  2. Powles T, Park SH, Voog E, et al. Maintenance avelumab + best supportive care (BSC) versus BSC alone after platinum-based first-line (1L) chemotherapy in advanced urothelial carcinoma (UC): JAVELIN Bladder 100 phase III interim analysis. J Clin Oncol. 2020;38(suppl 18):LBA1. doi:10.1200/JCO.2020.38.18_suppl.LBA1