Chris Wallis: Hello, and thank you for joining us for this UroToday Journal Club discussion.

Today, we're discussing a recent publication entitled nivolumab plus ipilimumab with or without live bacterial supplementation in metastatic renal cell carcinoma, a randomized Phase I trial. I'm Chris Wallis, an Assistant Professor in the Division of Urology at the University of Toronto. Joining me today is Zach Klaassen, an Assistant Professor in the Division of Urology at the Medical College of Georgia.

You can see here, the citation for this recent work published in Nature Medicine led by the team of Dr. Monte Pal.

The interaction between the microbiome and immunotherapy is both interesting and an area of active study across many tumor types. The gut microbiome has been associated with response to immunotherapy. You can see a whole proposed causal pathway here to the right. But in the context of advanced kidney cancer, specific species have been associated with a particular benefit of immunotherapy, and bifidobacterium in particular has been shown to modulate immunotherapy response.

So as most will know, the treatment landscape of advanced kidney cancer has rapidly evolved. And since about 2018, we've been in this immuno-oncology combo therapy era. And the first of these approaches that was approved and to clinical use is the dual IO approach using nivolumab and ipilimumab. And so this is the original standard combination IO approach, the one that these authors have also built as a backbone for their study.

So nivo and ipi, as we know, was first to demonstrate to have benefit in this setting in the CheckMate 214 trial. There's improvements in overall survival, progression-free survival, and objective response rate compared to the prior standard of care of sunitinib. However, the vast majority of patients still don't have a response, and 20% have immediate progression on first-line therapy. And so as the authors of this paper point out, there is a dramatic need for improvement to the current therapeutic paradigm. And one potential approach is modulation of the gut microbiome.

So the goal of their paper was to assess whether CBM588, which is a live bacterial product containing Clostridium butyricum could improve outcomes among patients with advanced kidney cancer who are being treated with nivolumab and ipilimumab. CBM588 has shown benefit in a retrospective study in improving both progression-free survival and overall survival in patients with non-small cell lung cancer who are receiving IO therapy.

So this is a single-center, open-label, investigator-initiated trial that enrolled patients aged 18 years and older with histologically confirmed clear cell RCC or sarcomatoid histology. They had to have measurable metastatic disease according to resist version 1.1 criteria and have IMDC intermediate or poor-risk classification. The performance status had to be at least 70 patients who had received prior systemic therapy for renal cell carcinoma were excluded, although prior non-checkpoint inhibitor-based adjuvant therapy was allowed.

So as mentioned, this is a randomized open-label trial. Patients were randomized in a two-to-one fashion to receive nivolumab and ipilimumab with or without CBM588. No stratification was used in this Phase I study design. Nivolumab and ipilimumab were given in standard doses for an initial 12-week period, followed by a nivolumab monotherapy maintenance. CBM588 was given 80 milligrams orally, twice daily. And this was a two 40 milligram sachets that were mixed with water and consumed each one containing at least two times 10 to the eighth colony-forming units.

And in keeping with the FDA label and approval, no dose reductions of nivolumab or ipilimumab were allowed. And treatment was continued until progression, unacceptable toxicity, or patient withdrawal of consent.

Patients underwent a baseline chest, abdomen, and pelvis CT scan, and bone scan, and central nervous system imaging were allowed, but not mandated as clinically indicated. Repeat imaging was performed at the 12-week interval until the end of protocol defined therapy or death, whichever came first. Safety assessments were made every three weeks. A radiographic response of these scans was assessed using resist version 1.1 criteria.

Additionally, stool was collected at baseline and at 12 weeks using an ONMIgene Gut kit. Whole metagenome sequencing was performed with DNA extracted and bacterial fungal loads quantified. The whole metagenome library was constructed in taxonomically profiled with species-level assignments retained.

Additionally, the authors assess the functional potential microbial communities, comparing metabolic pathways between baseline and week 12 using Wilcoxon signed-rank test to assess for differences over time between the two randomized arms.

Additionally, peripheral blood was collected at baseline and weeks seven, 12, 17, and 25 of therapy. Following collection, plasma was initially extracted for circulating cytokine analysis. And then from the remaining portion of the blood, peripheral blood monocular cells were isolated and stained for a variety of combination of antibodies to and CD3, CD4, CD8, intracellular FoxP3, CD33, HLA-DR, and CD15.

The primary outcome of this trial was the change in Bifidobacterium composition in the stool from baseline to week 12. Secondarily, the authors assess the Shannon index, which is a measure of microbial diversity over time. The best overall response rate, progression-free survival, the proportion of circulating Tregs at baseline to follow up, proportion of circulating MDSCs, cytokine levels, as well as toxicity with a particular focus on diarrhea and nausea.

The authors calculated the sample size of 30 patients would allow 80% power to detect a one standard deviation change in specific Bifidobacterium species using a t-test and a one-sided alpha of 0.05. They compared gut microbiome, composition, cytokines, regularly T cell, and myeloid-derived suppressor cell populations over time using the Wilcoxon signed-rank test, and between study arms using the Mann-Whitney U test. Survival data, as is typical, was analyzed using the Kaplan-Meier method and the log-rank test.

Now going to hand it over to Zach to walk us through the results of this interesting Phase I trial.

Zach Klaassen: Chris, thanks for that great introduction. As we can see here, this is the CONSORT diagram for the study. And there was 36 patients that were assessed for eligibility. Ultimately 30 patients under what randomization. As Chris mentioned, this was a two-to-one randomization so there was 10 allocated to the nivolumab plus ipilimumab arm, and then there was 20 randomized to the Nivo plus CBM588 arm. Ultimately 19 of these 20 underwent analysis as one patient was deemed to be likely a sarcoma and a removed from the analysis data set.

This is the patient characteristics for this study. You can see the Nivo/Ipi plus CBM588 on the right part of this table. And to the left of this is the Nivo/Ipi patients. And they were quite comparable. Amongst these groups, you can see the median ages, roughly mid-sixties. The majority of patients were male. And the majority of patients were white. You can see with regards to histological subtype, there was 70% of the patients the Nivo/Ipi arm were clear cell compared to 63% in the CB588 arm, with several other histologies, including sarcomatoid, and papillary with sarcomatoid features.

Given that this is a Nivo/Ipi-based strategy, this was IMDC prognostic intermediate and poor, with the majority of patients being intermediate risk, and roughly half of patients having a prior nephrectomy. Ultimately, all patients in this study had metastatic sites greater than or equal to two, with the most common sites being the lung and the lymph nodes.

This is the Kaplan-Meier curve for progression-free survival. And over a median follow up of 12.2 months, we can see that there was a significant improvement in progression-free survival with the CBM588 patients with the median PFS of 12.7 months compared to 2.5 months for the Nivo/Ipi patients with a hazard ratio favoring the CBM588 patients at 0.15 and a 95% confidence interval of 0.05 to 0.47.

So the Kaplan-Meier curve for overall survival, certainly with a median follow up just over one year, the median was not reached in neither of these groups, but you can see that there were, at least in the first several months, there was an early splitting of the curves between these patients.

This figure looks at best response by treatment arm. On the right we can see that patients with partial response was 20% for Nivo/Ipi compared to 58% partial response for CBM588. There was no patients with complete response in either of the arms.

This is the waterfall plot looking at best change in target lesions from baseline. And to sort of summarize this figure, the reduction in tumor volume for the CBM588 patients was 74% of the patients compared to 50% for Nivo/Ipi. And we can see here on the figure, the patients in gray are the patients that received a Nivo/Ipi plus CBM588. And on the right side of this figure, a decent number of patients receiving even up to 50% tumor reduction in volume.

This is a swimmers plot looking at response and survival characteristics. Again, to summarize this figure, the disease control rate was 79% for the CBM588 patients compared to 40% for the Nivo/Ipi patients. And focusing on the top half of this figure, the gray patients again in the CBM588, a number of patients with excellent disease control rate and continuing on treatment. And with the patients doing less or poorly, was patients with just Nivo/Ipi alone at the bottom part of this figure.

This looks like toxicities in the Phase I trial. Certainly, toxicities is important to assess in these patients. And this table looks at greater than or equal to grade 2 toxicities. On the right, we can see with CBM588, there was one grade 4 toxicity, which was neutrophil count decrease with no grade 4 toxicities in the Nivo/Ipi group. In terms of grade 3, 47% for the CBM588 patients compared to 50% for grade 5. And with regards to grade 2, the most common in the CBM patients was fatigue at 19%, adrenal insufficiency at 19%, transaminitis at 26%, and arthralgia or myalgia and 22%.

Next several sets of slides will look at the microbiome assessment. This is the change in Bifidobacterium species. And looking at these two figures, to summarize these two, there was no significant change in relative abundance of the Bifidobacterium species from baseline to week 12 with the Bifidobacterium +/- CBM588. However, there was a significant increase in Bifidobacterium species in patients receiving CBM588 and responding to treatment as we can see here on the far right of this slide with a P-value of 0.024.

This slide looks at the microbiome assessment, specifically looking at species that where there was a decrease in abundance and response to treatment. And we can see that, again, on the far right of each of these four figures, this is the patients that received CBM588 and also had a response to treatment. And we see that reduction in the desulfovibrio species, specifically desulfuricans, marinus, vulgaris, and sulfodismutans, all of these species with a decrease in abundance in the gut, is specifically in the responders, were prognostic.

Again, looking at the increases in abundance of gut microbiome species, and this is looking at Bifidobacterium longum. Again, the responders having a significant P-value of 0.03. And the Enterococcus faecalis, again in the CBM responders, a P-value of 0.010.

So several discussion points from this exciting Phase I trial. The results of this RCT suggest that supplementation with live bacterial products may augment the activity of checkpoint inhibitors. Although these results must be taken in context of a small sample size, there's several key findings from this study.

First, is that there was a significant improvement in PFS with the addition of CBM588 to Nivo/Ipi. Secondly, this was consistent results obtained favoring CBM588 for response rate and overall survival. And finally, there was no significant change in Bifidobacterium species with CBM588 therapy. However, there was an increase in these organisms amongst those who responded to therapy.

Ultimately these findings reinforce recent observations documenting synergy between immunotherapy and microbiome modulation. The authors also note that there's several challenges of studies in the microbiome space, including this study, with regards to difficulty monitoring patient diets, as well as probiotic restriction in the control arm patients.

So in conclusion, this study suggests that a live bacterial product may augment the activity of checkpoint inhibitors. And that given what appears to be an acceptable safety profile, it is important to validate these findings in larger series and across different tumor types.

Thank you very much for your attention today, and we hope you enjoyed UroToday Journal Club discussion.