Abstract Rising rates of antibiotic treatment failure highlight the complexity of resistance mechanisms. While conventional, genetically encoded resistance is well established, recent studies have uncovered widespread non-canonical mechanisms driven by phenotypically insensitive subpopulations hiding in seemingly susceptible populations, such as heteroresistance, persistence or adaptive resistance. These variants are not only clinically significant but also act as evolutionary precursors to fully resistant populations, rendering infections increasingly difficult to control. Yet standard antibiotic susceptibility tests lack the resolution and dynamic range needed to detect this wide spectrum of resistance mechanisms and their evolutionary progression, often misguiding clinical decisions, obscuring mechanisms of failure, and underestimating the true epidemiology of antibiotic resistance. To address these limitations, we introduce the Dilution-and-Delay (DnD) assay—a practical, high-resolution approach that implements two basic principles of bacterial growth in the antibiotic susceptibility testing format. Using well-defined synthetic communities and strains, we demonstrate that the DnD assay quantifies viable cells over more than eight orders of magnitude and detects rare antibiotic-insensitive cells at frequencies as low as 1 in 100 million. It additionally reports the standardized MIC as a secondary output, thereby capturing the average antibiotic response of the majority population and heterogeneity of minority subpopulations within a single assay. We scaled this approach for high-throughput application, classifying ∼120 previously uncharacterized clinical isolates of Klebsiella pneumoniae, Enterobacter cloacae , Escherichia coli , Pseudomonas aeruginosa and Acinetobacter baumannii —across five antibiotics. This robust, quantitative, and scalable platform opens the door to next-generation antibiotic susceptibility testing, with broad utility in basic research, clinical diagnostics, and epidemiological surveillance. This improved testing will guide evidence-based clinical decisions and prevent the rapid evolution of antibiotic resistance, thereby counteracting the rising rate of treatment failure. Significance Antibiotic treatment failure due to inappropriate prescription remains a critical threat to global health. It also accelerates the evolution of resistance, making infection increasingly difficult to control. A key underlying issue is that current antibiotic susceptibility testing lacks the resolution and dynamic range needed to detect a wide range of resistance mechanisms and their evolutionary progression, leading to misinformed clinical decisions. While population analysis profiling (PAP) offers sufficient accuracy, it is too labor-intensive for routine use. Single-cell approaches are technically demanding and resource-intensive. Here, we present a susceptibility testing strategy that preserves the operational efficiency and compatibility of conventional workflows, while dramatically extending their resolution and dynamic range. Our approach is scalable, quantitative, and well-suited for high-throughput applications in both research and clinical settings—closing a critical diagnostic gap in the fight against antibiotic treatment failure.
Ma et al. (Mon,) studied this question.