PRECISION MEDICINE IN SURGICAL INFECTIONS: INTEGRATING MOLECULAR DIAGNOSTICS, MICROBIOME PROFILING, AND ARTIFICIAL INTELLIGENCE-A SYSTEMATIC REVIEW

Abstract

Background: Surgical infections remain major causes of postoperative morbidity, prolonged hospitalization, reoperation, implant failure, antimicrobial exposure, and increased healthcare costs. Conventional culture remains essential but may have reduced sensitivity after antimicrobial administration, in biofilm-associated infection, and when fastidious, anaerobic, low-abundance, or non-cultivable microorganisms are involved. Precision medicine offers an integrated approach combining molecular diagnostics, microbiome profiling, and artificial intelligence to individualize infection prevention, diagnosis, surveillance, and treatment. Objective: To systematically evaluate the clinical applications, diagnostic performance, potential utility, and limitations of molecular diagnostics, microbiome profiling, and artificial-intelligence-based technologies in surgical infections. Methods: This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 statement. MEDLINE/PubMed, Embase, Scopus, Web of Science, and the Cochrane Library were searched for studies published from January 2000 to January 2026. Primary studies evaluating broad-range or multiplex polymerase chain reaction, targeted or metagenomic next-generation sequencing, microbial cell-free DNA, microbiome profiling, machine learning, deep learning, or multimodal artificial intelligence for the prediction, diagnosis, characterization, surveillance, or management of surgical infections were eligible. Owing to clinical and methodological heterogeneity, the findings were synthesized narratively. Results: Eleven primary studies published between 2017 and 2026 were included. Four studies primarily evaluated molecular diagnosis, three evaluated microbiome profiling, three evaluated artificial-intelligence-assisted surveillance or wound-image analysis, and one randomized trial evaluated molecularly guided antimicrobial prophylaxis. Broad-range 16S ribosomal RNA polymerase chain reaction detected bacterial DNA in 53 of 97 conventionally culture-negative surgical-site infection specimens, corresponding to an additional detection rate of 54.6%. Microbial cell-free DNA and next-generation sequencing provided additional pathogen identification in culture-negative and implant-associated infections. In a spinal-surgery cohort, 12 of 14 surgical-site infections, approximately 86%, were caused by strains carried by patients before surgery, supporting the importance of endogenous microbial reservoirs. In colorectal surgery, intraoperative microbiome alterations and increased alpha diversity were associated with subsequent infection. Artificial intelligence models using wound images, patient-reported outcomes, and perioperative data demonstrated potential for postoperative surveillance. A multicentre image-based model achieved an area under the receiver-operating-characteristic curve of 0.98 for incision detection and 0.81 for surgical-site-infection detection, while a multimodal model reduced estimated staff review time from 51.5 hours to 9.1 hours in simulated implementation. Conclusions: Molecular diagnostics, microbiome profiling, and artificial intelligence have the potential to shift surgical infection management from reactive, culture-dependent care toward individualized and anticipatory precision medicine. Molecular testing currently has the clearest clinical role as an adjunct in culture-negative, antimicrobial-pretreated, and implant-associated infections. Microbiome profiling and artificial-intelligence-assisted surveillance remain promising but require larger prospective multicentre studies, standardized methodologies, external validation, and assessment of patient level outcomes. These technologies should complement rather than replace conventional microbiology, surgical source control, antimicrobial stewardship, and multidisciplinary clinical judgment.