Background
Evidence suggests that ATB and gut microbiota composition are associated to anti-PD-1 outcomes in RCC patients. Network analysis of the gut microbiota represents a novel tool to determine “guilds” of bacterial communities.
Methods
Here, we investigated guilds’ relationships with prior exposure to ATB, TKI [axitinib (axi), sunitinib (suni) or other] and clinical outcomes of RCC patients treated with nivolumab (nivo).
Results
Considering prior exposure to ATB (n = 11, 22%), axi (n = 13, 19%), suni (n = 49, 71%) or other TKI (n = 20, 29%), n = 36 (52%) patients were NR and 33 (48%) R to nivo. Cross-validation of overall fecal microbiota composition stratified RCC patients with different predictive power (ATB=84%; axi=81%; suni=69%; outcome=49%). Network analysis revealed six guilds (G1 to G6). G1-G2 behaved in an opposite way and topologically separated by negative correlations. G1-G2 were both related to NR, while G1 was dominated by species related to ATB. Conversely G2 was mainly represented by species related to no ATB (X2=8.98, P = 0.0027) and more susceptible to prior TKI exposure (where axi and suni behaved in an opposite way) compared to the other guilds (X2=10.68, P = 0.0011). G4 was mainly inhabited by species related to other TKI (no axi and no suni). Random forest analysis found definite bacterial species able to drive the stratification into guilds of the global RCC network, such as Akkermansia muciniphila for R and Dorea formicigenerans for no ATB.
Conclusions
We show that ATB and suni are the most powerful external features able to drive fecal microbiota compositional shifts in RCC patients treated with nivo. Analysis of the gut microbiota using bacterial communities guilds provides a novel theranostic approach to stratify RCC patients.