Microbial dysbiosis as predictor of benefit from CBM588 as an adjunct to immune checkpoint blockade (ICB)–based first line therapies in metastatic renal cell carcinoma (mRCC).

4519 Background: The Clostridium butyricum -based live biotherapeutic, CBM588, demonstrated signals for enhanced clinical activity with first-line ICB combinations in two randomized phase I trials for mRCC. However, impact of microbial dysbiosis on the benefit of CBM588 supplementation is unclear. Herein, we examined prognostic and predictive value of microbial dysbiosis with CBM588. Methods: We analyzed clinical outcomes and stool whole-genome sequencing data from a combined cohort from two randomized phase I clinical trials. Both enrolled treatment-naïve patients with mRCC and randomized them to receive nivolumab/ipilimumab (NCT03829111) or nivolumab/cabozantinib (NCT05122546) alone (standard of care SOC), or with CBM588 (SOC+CBM588). Baseline TOPOSCORE, a stool metagenomic dysbiosis index linked to ICB outcomes, was analyzed continuously (S score positively correlated with dysbiosis) and categorically (SIG1+ dysbiosis phenotype vs SIG2+ non-dysbiosis). Associations with progression free survival (PFS) and objective response rate (ORR, per RECIST 1.1) with SOC and SOC + CBM588 were assessed. Results: Fifty-nine patients were included: 39 in the SOC + CBM588 arm and 20 in the SOC arm. Median age was 65 (range 36-90), with the majority male (69.5%), clear-cell (88.1%) and intermediate/poor risk (67.8%) mRCC. Baseline clinical characteristics were comparable across arms. ORR was 66.7% in SOC+CBM588 arm versus 20.0% in the SOC arm (p = 0.001). Median PFS was 32.1 months (95% CI 16.6-NR) with SOC+CBM588 versus 3.7 months (95% CI 2.6-17.0) with SOC (HR 0.36 95% CI 0.19, 0.67, p = 0.001). At baseline, S score was similar across arms (p = 0.16), and SIG1+ was seen in 55.6% of patients in the SOC+CBM588 arm versus 44.4% in the SOC arm (p = 0.441). While differences in PFS were not statistically significant in the SIG2+ cohort between SOC+CBM588 and SOC alone (32.0 vs 10.9 months, HR 0.49 95% CI 0.19-1.3, p = 0.149), in the SIG1+ setting, a PFS improvement was seen for SOC+CBM588 versus SOC alone (24.9 vs 2.8 months, HR 0.19 95% CI 0.07-0.53, p < 0.001). Notably, S score was associated with response to SOC+CBM588 among patients who received nivolumab/ipilimumab backbone (median S score 0.668 in response vs 0.540 in no response, p = 0.046), while no such association was observed in patients treated with nivolumab/cabozantinib backbone (median S score 0.681 vs 0.821, p = 0.554, respectively). Conclusions: Addition of CBM588 to ICB-based first-line combinations conferred a disproportionately greater benefit in patients with a dysbiotic stool phenotype, underscoring actionability of TOPOSCORE as a potential predictive biomarker in this setting. The upcoming phase III randomized placebo-controlled BIOFRONT trial will assess clinical activity of CBM588 and advance microbiome-based predictive biomarker development in mRCC.