The management of recurrent urinary tract infection (rUTI) is challenging and its natural history is not fully understood. Moreover, the correct approach to the management of symptomatic recurrence is an essential step in the antimicrobial stewardship principles achievement in everyday clinical practice. Artificial intelligence proffers the ability of computer systems to perform human brain tasks across various topics in all aspects of everyday life. Artificial intelligence (AI) seems a good tool to use to guide antimicrobial choice in recurrent UTI. Here, we aim to define a neural network for predicting the clinical and microbiological efficacy of the antimicrobial treatment in a large cohort of women affected by recurrent UTI, to use in everyday clinical practice.
Among all patients who had undergone antimicrobial treatment for uncomplicated lower urinary tract infections, between January 2012 and December 2020, 1,043 patients were finally selected and enrolled. All microbiological and clinical data at the enrollment and at the first follow-up after the symptomatic episode were collected. All data about the previous use of antibiotics and antibiograms were included in the analysis. The data were analyzed by using the commercially available software program NeuralWorks Predict. These data were compared with univariate and multivariate analysis results.
After a few runs of AI learning and prediction processes, the use of artificial neural networks in women with recurrent cystitis showed a sensitivity of 87.8% and specificity of 97.3% in predicting clinical and microbiological efficacy of the prescribed antimicrobial drug at the follow-up evaluation. Statistical and AI analyses allowed selection of the previous use of fluoroquinolones (yes, HR=4.23, p=0.008) and cephalosporins (yes, HR=2.81, p=0.003) in the last three months and the presence of E.coli with resistance against cotrimoxazole (yes, HR=3.54, p=0.001) as the most influential variables affecting the output decision in predicting the fluoroquinolones-based therapy failure. Moreover, the previous E.coli with resistance against fosfomycin (yes, HR=2.67, p=0.001) and amoxicillin-clavulanic acid (yes, HR=1.94, p=0.001) in at least one antibiogram seems the most influential variable affecting the output decision in predicting the cephalosporins and cotrimoxazole-based therapy failure.
We demonstrated the feasibility and reliability of AI applications to guide antimicrobial choice in the management of recurrent cystitis in everyday clinical practice, reporting a good recurrence predicting performance. This tool seems interesting also in the achievement of antimicrobial stewardship principles.
