Unraveling gut gram-negative antibiotic-resistant colonization dynamics in hematologic cancers: Insights from bioinformatics and immune signatures.

Abstract

10031 Background: Infections account for ~60% of cancer-related deaths. Hematologic cancers have a 3x higher infection-related mortality than solid tumors, with resistant Gram-negative (GN) bacteria causing ~50% of bloodstream infections, highlighting the need to study the microbiome, antibiotic resistance, and infection risks. Methods: Stool samples from newly diagnosed hematologic cancer patients were collected at baseline, post-chemotherapy, and subsequent admission at Amir Hospital, a referral cancer center in southern Iran, to analyze microbial colonization dynamics. Patients enrolled in a 16-month observational program to investigate the correlation between clinical factors and infectious outcomes. Carbapenem-resistant and ESBL-producing were cultured on MacConkey agar with meropenem and ceftriaxone. ESBL and carbapenemase production assessed adhering to CLSI guidelines. To support our findings, we used the microbioTA database to identify highly elevated 16S rRNA expression in blood and lymph nodes of hematologic cancer patients, exploring microbiome-host interactions worldwide. We used gutMgene, GIMICA, and AMIDIS databases to explore key microbe-immune factor associations trough network centrality analysis of the immune factors. Central factors further examined in the Amir Cancer Registry datasets to assess their association with infectious events. STATA v27 used for statistical analyses. Results: Among 73 pediatric patients, GN drug-resistant bacteria was detected in 51 before hospitalization. Escherichia coli (86.6% of positive samples) and Klebsiella pneumoniae (9.5%) were the predominant pathogens. Drug-resistant E. coli persisted across samples, indicating gut colonization, consistent with microbioTA data showing E. coli detection at baseline and post-induction therapy. ESBL and carbapenemase-producing strains were 56.8% and 15.8%. Colonized patients had a 13.8% mortality rate, with bloodstream infections and typhlitis more common in K. pneumoniae and carbapenem-resistant strains. Previous antibiotic exposure, malignancy relapse, and colonization status were risk factors for mortality and infection. Investigation of the microbioTA database identified 10 datasets from Asia, Europe, and America revealed the detection of Bacillus cereus in 9 datasets, followed by E.coli (8) and Enterobacterales like Salmonella enterica (5) and K.pneumoniae (3), approving the high impact of E.coli worldwide. IL-4, IL-6, and TNF-α showed high centrality, with retrospective analysis linking their upregulated serum baseline levels to infection outcomes in Amir hospital datasets. Conclusions: Our study links GN microbial colonization, traced by elevated immune markers, to infectious complications, highlighting the need for microbiota-specific diagnostic and treatment protocols.